Calcium/calmodulin‐dependent kinase II and nitric oxide synthase 1‐dependent modulation of ryanodine receptors during β‐adrenergic stimulation is restricted to the dyadic cleft

Abstract

Key points

• The dyadic cleft, where coupled ryanodine receptors (RyRs) reside, is thought to serve as a microdomain for local signalling, as supported by distinct modulation of coupled RyRs dependent on Ca²⁺/calmodulin‐dependent kinase II (CaMKII) activation during high‐frequency stimulation.

• Sympathetic stimulation through β‐adrenergic receptors activates an integrated signalling cascade, enhancing Ca²⁺ cycling and is at least partially mediated through CaMKII.

• Here we report that CaMKII activation during β‐adrenergic signalling is restricted to the dyadic cleft, where it enhances activity of coupled RyRs thereby contributing to the increase in diastolic events. Nitric oxide synthase 1 equally participates in the local modulation of coupled RyRs.

• In contrast, the increase in the Ca²⁺ content of the sarcoplasmic reticulum and related increase in the amplitude of the Ca²⁺ transient are primarily protein kinase A‐dependent.

• The present data extend the concept of microdomain signalling in the dyadic cleft and give perspectives for selective modulation of RyR subpopulations and diastolic events.

Abstract

In cardiac myocytes, β‐adrenergic stimulation enhances Ca²⁺ cycling through an integrated signalling cascade modulating L‐type Ca²⁺ channels (LTCCs), phospholamban and ryanodine receptors (RyRs). Ca²⁺/calmodulin‐dependent kinase II (CaMKII) and nitric oxide synthase 1 (NOS1) are proposed as prime mediators for increasing RyR open probability. We investigate whether this pathway is confined to the high Ca²⁺ microdomain of the dyadic cleft and thus to coupled RyRs. Pig ventricular myocytes are studied under whole‐cell voltage‐clamp and confocal line‐scan imaging with Fluo‐4 as a [Ca²⁺]_i indicator. Following conditioning depolarizing pulses, spontaneous RyR activity is recorded as Ca²⁺ sparks, which are assigned to coupled and non‐coupled RyR clusters. Isoproterenol (ISO) (10 nm) increases Ca²⁺ spark frequency in both populations of RyRs. However, CaMKII inhibition reduces spark frequency in coupled RyRs only; NOS1 inhibition mimics the effect of CaMKII inhibition. Moreover, ISO induces the repetitive activation of coupled RyR clusters through CaMKII activation. Immunostaining shows high levels of CaMKII phosphorylation at the dyadic cleft. CaMKII inhibition reduces I _CaL and local Ca²⁺ transients during depolarizing steps but has only modest effects on amplitude or relaxation of the global Ca²⁺ transient. In contrast, protein kinase A (PKA) inhibition reduces spark frequency in all RyRs concurrently with a reduction of sarcoplasmic reticulum Ca²⁺ content, Ca²⁺ transient amplitude and relaxation. In conclusion, CaMKII activation during β‐adrenergic stimulation is restricted to the dyadic cleft microdomain, enhancing LTCC‐triggered local Ca²⁺ release as well as spontaneous diastolic Ca²⁺ release whilst PKA is the major pathway increasing global Ca²⁺ cycling. Selective CaMKII inhibition may reduce potentially arrhythmogenic release without negative inotropy.

Key points

• The dyadic cleft, where coupled ryanodine receptors (RyRs) reside, is thought to serve as a microdomain for local signalling, as supported by distinct modulation of coupled RyRs dependent on Ca²⁺/calmodulin‐dependent kinase II (CaMKII) activation during high‐frequency stimulation.

• Sympathetic stimulation through β‐adrenergic receptors activates an integrated signalling cascade, enhancing Ca²⁺ cycling and is at least partially mediated through CaMKII.

• Here we report that CaMKII activation during β‐adrenergic signalling is restricted to the dyadic cleft, where it enhances activity of coupled RyRs thereby contributing to the increase in diastolic events. Nitric oxide synthase 1 equally participates in the local modulation of coupled RyRs.

• In contrast, the increase in the Ca²⁺ content of the sarcoplasmic reticulum and related increase in the amplitude of the Ca²⁺ transient are primarily protein kinase A‐dependent.

• The present data extend the concept of microdomain signalling in the dyadic cleft and give perspectives for selective modulation of RyR subpopulations and diastolic events.

Abbreviations

AC

adenylate cyclase

AIP

autocamtide‐2‐related‐inhibitory‐peptide

β‐AR

β‐adrenergic receptor

CaMKII

Ca²⁺/calmodulin‐dependent kinase II

CPVT

catecholaminergic polymorphic ventricular tachycardia

Epac

exchange protein activated by cAMP

I_CaL

L‐Type Ca²⁺ current

ISO

isoproterenol

LTCC

L‐type Ca²⁺ channel

l‐VNIO

vinyl‐l‐NIO hydrochloride

LV

left ventricle

NAC

N‐acetylcysteine

NCX

Na⁺/Ca²⁺ exchanger

NO

nitric oxide

NOS1

nitric oxide synthase1

P_o

open probability

PDE

phosphodiesterase

PKA

protein kinase A

PKI

protein‐kinase inhibitor

PLN

phospholamban

ROS

reactive oxygen species

RyR

ryanodine receptor

SERCA

SR Ca²⁺‐ATPase

SR

sarcoplasmic reticulum

T_F50

time to reach half‐maximal [Ca²⁺]_i

TT

T‐tubule

Introduction

In the heart, inotropy, lusitropy and chronotropy are increased by catecholamine‐mediated activation of β‐adrenergic receptors (β‐ARs), for example during emotional stress or physical exercise. This results in increased cardiac output to meet the metabolic demand, and this adaptation is largely due to modulation of myocyte Ca²⁺ handling. Upon activation, β‐ARs stimulate adenylate cyclase (AC) via G proteins to produce cAMP. Activation of the cAMP‐dependent kinase protein kinase A (PKA) leads to the phosphorylation of several Ca²⁺ handling proteins: L‐type Ca²⁺ channels (LTCCs), resulting in enhanced L‐type Ca²⁺ current (I _CaL); ryanodine receptors (RyRs), facilitating Ca²⁺ release from the sarcoplasmic reticulum (SR); and phospholamban (PLN), reducing inhibition of the SR Ca²⁺‐ATPase (SERCA) leading to an enhancement of Ca²⁺ uptake into the SR. The integrated response, larger and faster contraction and relaxation at higher heart rates, is the result of both frequency‐dependent modulation of Ca²⁺ handling and direct β‐receptor‐mediated signalling.

β‐AR‐dependent RyR phosphorylation has recently received particular attention because of the association of RyR mutations with catecholaminergic polymorphic ventricular tachycardia (CPVT) and sudden death (Venetucci et al. 2012). In vitro, PKA‐mediated phosphorylation at serine 2809 in RyR2 enhances RyR open probability (P _o), with one possible mechanism being the dissociation of the immunophilin FKBP12.6 from the RyR (Marx et al. 2000). However, this mechanism has been questioned and the exact mechanisms behind PKA regulation of RyRs in the intact cell, as well as its role in the β‐AR response, remain controversial (Xiao et al., 2004, 2006; Morimoto et al. 2009; Shan et al. 2010).

In addition to PKA, Ca²⁺/calmodulin‐dependent kinase II (CaMKII) is activated downstream of β‐AR and also increases RyR P _o when measured as SR Ca²⁺ leak. This was shown in small rodents, guinea pigs and rabbits (Curran et al. 2007; Ferrero et al. 2007; Ogrodnik & Niggli, 2010). CaMKII activation was at first ascribed to PKA‐dependent increases in [Ca²⁺]_i. However, CaMKII activity can also be increased via mechanisms distinct from PKA during β‐adrenergic stimulation (Pereira et al. 2007; Gutierrez et al. 2013; Curran et al. 2014). These findings have shifted the paradigm and CaMKII is viewed as a primary integrator of both frequency‐dependent signalling and β‐adrenergic signalling (rather than a secondary player, dependent on PKA).

At the molecular level, the exchange protein activated by cAMP (Epac) is a pathway for cAMP signalling that is independent of PKA (Ruiz‐Hurtado et al. 2013). Stimulation of Epac, using the specific activator 8‐CPT, led to an increased diastolic SR Ca²⁺ leak via CaMKII‐dependent phosphorylation of serine 2814 in RyR2 (RyR‐S2814) in myocytes from rat and mice hearts (Pereira et al., 2007, 2013). Additionally, direct activation of CaMKII by nitric oxide (NO) during β‐adrenergic stimulation has been reported to occur in non‐paced cells, i.e. in the absence of increased Ca²⁺ transients (Gutierrez et al. 2013; Curran et al. 2014).

Other proteins that contribute to the overall β‐adrenergic‐mediated increase in Ca²⁺ cycling are PLN and sarcolemmal LTCCs. These proteins are phosphorylated in a PKA‐dependent manner, which can be mimicked by forskolin, a direct activator of AC (Curran et al. 2007). Nevertheless, these proteins can also be modulated by CaMKII as part of the frequency‐dependent modulation (Wu et al. 2012).

Recent studies have emphasized that β‐adrenergic signalling is not necessarily homogeneous through the cardiac myocyte. Using fluorescence resonance energy transfer‐based sensors, localized cAMP signalling in subcellular compartments was demonstrated (Zaccolo & Pozzan, 2002; Nikolaev et al. 2006; Agarwal et al. 2014). Such compartmentalized cAMP (and cGMP) signalling is associated with subcellular localization of different phosphodiesterases (PDEs) that restrict cAMP diffusion and the dimension of the cAMP microdomain (Leroy et al. 2008; Mika et al. 2014). In addition, different β‐AR isoforms have specific subcellular locations and downstream targets (Gorelik et al. 2013). Local β‐adrenergic signalling compartments are reported to be localized to T‐tubules (TTs) and caveolae, where key components of the β‐AR cascade are also found (e.g. G_s proteins, AC, PDE3/4 and AKAPs).

We recently reported on the role of CaMKII microdomains around RyRs in the response to increased stimulation frequency in the absence of β‐AR stimulation (Dries et al. 2013). CaMKII‐dependent potentiation of RyR activity was restricted to RyRs in the dyadic cleft (coupled RyRs), where there is a distinct Ca²⁺ microdomain, and absent in the RyRs not coupled to the sarcolemmal membrane (non‐coupled RyRs) despite the fact that CaMKII was present in both locations.

The primary aim of the present study is to investigate whether modulation of RyRs through activation of the β‐adrenergic receptor is restricted to the coupled RyRs in the microdomain of the dyadic cleft. Second, we examine the mechanisms underlying the regulation of RyRs during β‐adrenergic stimulation. Finally, we investigate how microdomain CaMKII activation integrates with PKA signalling and global Ca²⁺ cycling. For our studies, we use pig ventricular myocytes that have a less dense TT network than rodents, more similar to that found in human ventricular myocytes (Heinzel et al. 2002; Louch et al. 2004; Jayasinghe et al. 2012).

Methods

Animal care

Healthy pigs (40‐45 kg) were housed and treated according to the European Directive 2010/63/EU. Experimental protocols were approved by the in‐house ethical committee (Ethische Commissie Dierproeven, KU Leuven), with permit numbers P10139 and P14176.

Cell isolation

Animals were killed via an overdose of pentobarbital (100 mg kg⁻¹) under full anaesthesia (absence of deep reflexes), after which the heart was quickly excised. Single cardiomyocytes were enzymatically isolated from the mid‐myocardial layer of the left ventricle (LV) as described before (Heinzel et al. 2002). A large wedge with the perfusing artery was prepared and the coronary artery was cannulated and perfused with a constant flow (4 ml min⁻¹) at 37 °C. To washout the remaining blood, the tissue was initially perfused with normal Tyrode (in mmol l^–1: NaCl 137, KCl 5.4, MgCl₂ 0.5, CaCl₂ 1.8, Na‐HEPES 11.8 and glucose 10; pH 7.4). Next, the tissue was perfused with a Ca²⁺‐free Tyrode (in mmol l^–1: NaCl 130, KCl 5.4, KH₂PO₄ 1.2, MgSO₄ 1.2, Na‐HEPES 6, glucose 20; pH 7.2) followed by the enzymatic solution [collagenase A (Roche, Basel, Switzerland) and protease XIV (Sigma‐Aldrich, St Louis, MO, USA) added to Ca²⁺‐free Tyrode] and after digestion, perfused with low Ca²⁺ Tyrode (0.18 mM CaCl₂ added to Ca²⁺‐free Tyrode). The digested tissue was minced, after which the suspension was filtered and the isolated myocytes were resuspended in normal Tyrode. Cells were allowed to recover for 1 h after isolation before starting experiments.

Electrophysiological recordings

Cells were studied under whole‐cell voltage‐clamp to record membrane currents (Axon 200B amplifier, Axon Instruments, Union City, CA, USA). Cells were constantly perfused with normal Tyrode at 37 °C and patch pipettes (2–3 MΩ) (GB 200‐8P, Science Products, Hofheim, Germany) were filled with (in mmol l^–1): K⁺‐aspartate 120, NaCl 10, KCl 20, K‐Hepes 10, MgATP 5, and K₅Fluo‐4 0.05; pH 7.2. Ca²⁺ transients were elicited by depolarizing steps (200 ms) from –70 to + 10 mV at 0.5 Hz for 30 s. This low‐frequency stimulation protocol was chosen to exclude frequency‐dependent potentiation (Dries et al. 2013). Protocols were repeated in the absence and presence of isoproterenol (ISO; 10 nm). SR Ca²⁺ content was measured by integrating the inward Na⁺/Ca²⁺ exchanger (NCX) current during fast caffeine application (10 mmol l⁻¹ for 8 s) after a conditioning train at 0.5 Hz stimulation. I _CaL was measured during pulses from –70 to +10 mV while the fast Na⁺ current was inhibited with 5 μm TTX and 200 μm lidocaine.

The PKA inhibitor PKI (PKI myristoylated‐(14‐22)‐amide; Tocris Bioscience, Bristol, UK) was used at 20 μm with incubation for 30 min and included in the pipette solution. The PKA inhibitor H‐89 (Sigma‐Aldrich, St Louis, MO, USA) was used at 10 μm with pre‐incubation for 15 min and then included in the pipette solution. AIP (CaMKII inhibitor; autocamtide‐2‐related‐inhibitory‐peptide; Tocris Bioscience) was used at 10 μm with pre‐incubation for 1 h and included in the patch pipette. CE3F4 (kindly provided by Frank Lezoualc'h, INSERM, Toulouse, France) was used at 20 μm with pre‐incubation for 30 min and included in the patch pipette. ESI‐05 (4‐methylphenyl‐ 2, 4, 6‐ trimethylphenylsulfone, Biolog, Hayward, CA, USA) was used at 10 μm with pre‐incubation for 15 min and included in the patch pipette and wash‐on during the experiments. Vinyl‐l‐NIO hydrochloride (l‐VNIO, Santa Cruz Biotechnology, Inc., Santa Cruz, CA, USA) was used at 100 μm with pre‐incubation for 30 min and in the patch pipette. N‐acetylcysteine (NAC, Sigma‐Aldrich) was used at 10 mm with pre‐incubation for 1 h and included in the patch pipette.

Ca²⁺ measurements, confocal microscopy and image analysis

[Ca²⁺]_i was reported by Fluo‐4. Confocal line scan images were recorded using a Zeiss LSM 510 confocal system. Line scans were recorded at 650 Hz with a pixel size of 0.2–0.3 μm. Following the 30 s conditioning train with depolarizing steps at 0.5 Hz, spontaneous Ca²⁺ sparks were routinely observed (15 s recording at the holding potential of –70 mV). Sparks were detected using an automated spark analysis program based on the Cheng algorithm (Dries et al. 2013). Spark frequency was calculated and normalized to line length and time as number of sparks per 100 μm per second. When analysing spark frequency in coupled vs. non‐coupled sites, the frequency was normalized to the line length of these sites. To define coupled (dyadic) and non‐coupled (non‐dyadic) release sites, an algorithm based on the time taken to reach 50% peak amplitude (T _F50) of the Ca²⁺ transient was used as described previously (Dries et al. 2013). Briefly, five consecutive Ca²⁺ transients on the line scan images were averaged and fluorescence was normalized to the diastolic [Ca²⁺]_i level (F/F ₀). From the averaged transient, the T _F50 was calculated for each pixel along the line to establish the spatial distribution of the timing of the Ca²⁺ release. T _F50 values were used as markers for the distance to the membrane (Dries et al. 2013). In the averaged Ca²⁺ transient, pixels were categorized as a coupled release site (< 0.5 μm) when T _F50 < 18 ms and a non‐coupled release site (> 2 μm) when T _F50 > 27 ms. Release sites with a T _F50 between 18 and 27 ms were excluded so to have a clearly separated population of coupled and non‐coupled RyRs. Ca²⁺ sparks were then assigned to coupled or non‐coupled sites to define spark properties in each type of RyR site (Fig. 1 Aa).

Figure 1. After ISO the spark frequency in both coupled and non‐coupled RyRs is increased, but only coupled RyRs are modulated by CaMKII .

Aa, myocytes are stimulated in the presence and absence of ISO using depolarizing steps (–70 to +10 mV; 200 ms) for 30 s at 0.5 Hz and Ca²⁺ sparks are recorded during a 15 s period after stimulation. Example of a confocal line scan image is shown. Ca²⁺ sparks are assigned to coupled and non‐coupled using the temporal cut‐off values of the averaged Ca²⁺ transients. Coupled sites (< 0.5 μm) have a T _F50 < 18 ms (shown in red) and non‐coupled sites (> 2 μm) have a T _F50 > 27 ms (shown in green). Ab, spark frequency in coupled and non‐coupled RyRs before and after ISO (n _cells = 34; N _pigs = 11); and c, SR Ca²⁺ content before and after ISO application (n _cells = 25; N _pigs = 9). Ba, example of a confocal line scan image showing Ca²⁺ transients and Ca²⁺ sparks in the presence of ISO and ISO with AIP. Ca²⁺ sparks originating from coupled regions are shown in red and from non‐coupled regions in green. Ca²⁺ sparks from intermediate sites are not colour‐coded. The effect of AIP on Ca²⁺ spark frequency in coupled and non‐coupled RyRs after ISO (n _cells = 17; N _pigs = 5); b, effect of AIP on the SR Ca²⁺ content during ISO (n _cells = 19; N _pigs = 5). C, the relationship between SR Ca²⁺ load and Ca²⁺ sparks at baseline (n _cells = 7; N _pigs = 6), ISO (n _cells = 10; N _pigs = 12) and ISO with AIP (n _cells = 9; N _pigs = 4). ^*** P < 0.001.

Phosphorylation assays

Freshly isolated myocytes were stimulated at 0.5 Hz for 15 min using a multichannel homebuilt stimulator controlled using Labview 6.0 (National Instruments, Austin, TX, USA) in the presence and absence of ISO (10 nm). After stimulation, cells were fixed with 2% paraformaldehyde and permeabilized with 0.4% Triton X‐100 in PBS. Cells were washed three times and incubated with blocking buffer (BSA 4%, 0.1% Triton X‐100 in PBS) for 1 h at room temperature. Primary antibodies were incubated overnight at 4°C (mouse IgG anti‐RyR 1:200, MA3‐925 from Thermo Scientific, Waltham, MA, USA; mouse IgM anti‐NCX 1:200, MA3‐926 from Thermo Scientific; rabbit IgG anti‐phospho‐CaMKII Th286 1:200, PA1‐14076 from Thermo Scientific). Cells were washed three times in PBS and incubated with secondary antibodies (RyR: Alexa fluor 488 goat anti‐mouse IgG; NCX: Alexa fluor 647 goat anti‐mouse IgM; Phospho‐CaMKII Th286: Alexa fluor 568 goat anti‐rabbit IgG) diluted at 1:200 in blocking buffer for 2 h at room temperature. Cells were washed three times in PBS before imaging with a confocal microscope (Nikon A1R configured on an Eclipse Ti using a 60× 1.4 NA oil immersion objective). Fluorescence intensity was measured for phospho‐CaMKII Th286 in the whole cell and local regions (coupled vs. non‐coupled RyRs) using ImageJ. Whole‐cell fluorescence was normalized to background; local regions are shown as ratio of fluorescence in coupled/non‐coupled regions. Identification of coupled and non‐coupled regions was performed by using a mask of the skeletonized NCX signals < 0.5 or > 2 μm away, respectively, as described previously (Dries et al. 2013).

Statistics

All data are presented as means ± SEM. Data have been compared using a paired Student's t test or a two‐way ANOVA with Bonferroni post hoc testing when comparing a specific blocker in coupled versus non‐coupled RyRs. Data were considered significantly different when P < 0.05 and is represented as ^* P < 0.05, ^** P < 0.01 and ^*** P < 0.001.

Results

During β‐adrenergic stimulation sparks increase globally but only coupled RyRs are modulated by CaMKII

In the present study, spark frequency increased equivalently at both coupled and non‐coupled RyRs after application of ISO (Fig. 1 Ab). SR Ca²⁺ content was also increased following ISO stimulation (Fig. 1 Ac). These data were somewhat unexpected as the literature suggests that CaMKII is the pathway for ISO‐dependent modulation of RyR and we had hypothesized that CaMKII activation, and increase of sparks, would be restricted to the dyadic cleft and coupled RyRs. We therefore validated that the restricted modulation of coupled RyRs by CaMKII during high‐frequency stimulation alone (cells conditioned at 2 Hz, no ISO; Dries et al. 2013) was intact (Fig. 2). Furthermore, global Ca²⁺ handling during β‐adrenergic stimulation with ISO was also as previously described in many species, with increased amplitude of the Ca²⁺ transient and faster relaxation, larger ICaL and greater number of Ca²⁺ sparks (Fig. 3).

Figure 2. Specific modulation of coupled RyRs during high‐frequency stimulation in the absence of ISO .

Spark frequency in coupled and non‐coupled RyRs during high‐frequency stimulation in the absence of ISO (n _cells = 15; N _pigs = 5). ^** P < 0.01.

Figure 3. Increase in global Ca²⁺ handling during β‐adrenergic stimulation .

A, Ca²⁺ transient kinetics (n _cells = 30; N _pigs = 9); B, peak Ca²⁺ current amplitude (n _cells = 13; N _pigs = 3); and C, global spark frequency before and after ISO (n _cells = 34; N _pigs = 11). ^*** P < 0.001.

Having confirmed selective regulation of coupled RyRs by high‐frequency stimulation, we next examined whether there was a specific CaMKII component to the global ISO response. Relative to ISO‐treated cells, the specific CaMKII inhibitor AIP reduced the spark frequency in coupled RyRs by 50%, without affecting the frequency of sparks at the non‐coupled RyRs (Fig. 1 Ba) or the SR Ca²⁺ content (Fig. 1 Bb). When the relationship between SR Ca²⁺ content and spark frequency was examined, the CaMKII‐specific modulation of coupled RyRs occurred independently of changes in the SR Ca²⁺ content (Fig. 1 C).

Repetitive activation of RyRs is restricted to coupled RyRs and mediated by CaMKII

At the cellular level, Ca²⁺ waves (i.e. propagating spontaneous Ca²⁺ release) contribute to potential arrhythmogenic activity due to activation of the NCX current. The likelihood of inducing Ca²⁺ waves depends on the simultaneous activation of multiple RyR clusters in close proximity. Thus, a higher number of sparks arising from the same RyR cluster, and/or a higher number of firing RyR clusters in close proximity, can increase the likelihood of wave initiation. We examined the spark activity of repetitively firing RyRs and how such activity is altered during β‐adrenergic stimulation. RyR clusters were categorized as repetitive firing sites when at least two sparks occurred within 15 s of spark recording (Fig. 4 A).

Figure 4. The activation of repetitive firing sites during ISO is higher in coupled areas and is regulated by CaMKII .

A, left: theoretical example of a confocal line scan with non‐active sites (black), sites with 1 spark (purple) and sites with 2 or more sparks (orange). Right: formulas used to determine the percentage of repetitive sites recruited and the spark frequency from repetitive firing sites. Both parameters are separately analysed for coupled and non‐coupled RyRs. B, spark frequency and percentage of repetitive sites in coupled or non‐coupled regions before (n _cells = 27; N _pigs = 11) and after ISO (n _cells = 40; N _pigs = 16). C, the spark frequency and percentage of repetitive release sites in coupled and non‐coupled sites with ISO vs. ISO with AIP (n _cells = 16; N _pigs = 5). ^* P < 0.05; ^** P < 0.01; ^*** P < 0.001.

At baseline, sparks originating from repetitive firing sites were equally prevalent in coupled versus non‐coupled RyRs (Fig. 4 Ba). In the presence of ISO, sparks from repetitive sites significantly increased for both coupled and non‐coupled RyRs. However, the effects were more pronounced for coupled RyRs, with a tripling of the spark frequency compared to a doubling of the spark frequency at non‐coupled regions (Fig. 4 Ba). One potential cause of this increased spark frequency is the recruitment of new repetitive sites (Fig. 4 Bb). These results mirrored those of the spark frequency analysis; at baseline, there were no obvious differences between coupled and non‐coupled RyRs, whereas ISO treatment more than doubled the number of repetitive sites at coupled areas.

We further tested the role of CaMKII in control of Ca²⁺ sparks using AIP. The presence of AIP abolished the majority of the ISO‐induced increases in spark frequency at coupled sites alone, without noticeably affecting non‐coupled ones (Fig. 4 Ca). Likewise, AIP reduced the recruitment of repetitive sites at coupled areas alone (Fig. 4 Cb).

Mechanisms of CaMKII activation at coupled RyRs

Activation of CaMKII could be indirect through high local [Ca²⁺]_i or through other signalling pathways. Recent studies have described a β‐adrenergic signalling cascade where CaMKII activation occurs via an NO‐dependent pathway (Gutierrez et al. 2013; Curran et al. 2014). This was tested using the nitric oxide synthase 1 (NOS1) inhibitor l‐VNIO (Fig. 5 A). l‐VNIO significantly reduced spark frequency in coupled RyRs but had no effect on non‐coupled RyRs in the presence of ISO. Furthermore, similar to the results obtained using AIP, SR Ca²⁺ content was not affected by l‐VNIO application.

Figure 5. Activation of CaMKII is likely to occur via Ca²⁺ and NOS1 and is not dependent on Epac during β‐adrenergic stimulation .

A, the role of NOS1 in CaMKII activation after ISO stimulation. The effect of l‐VNIO on Ca²⁺ spark frequency in coupled and non‐coupled RyRs (n _cells = 13; N _pigs = 3) and on SR Ca²⁺ content in the presence of ISO (n _cells = 10; N _pigs = 3). B, the role of Epac in CaMKII activation after ISO stimulation. The effect of CE3F4 (Epac1; n _cells = 15; N _pigs = 3) and ESI‐05 (Epac2; n _cells = 12; N _pigs = 3) on Ca²⁺ spark frequency in coupled and non‐coupled RyRs and on SR Ca²⁺ content (CE3F4: n _cells = 5; N _pigs = 3; ESI‐05: n _cells = 11; N _pigs = 3) in the presence of ISO. C, the role of changes in [Ca²⁺]_i in CaMKII activation after ISO stimulation. Example of a confocal line scan image with a coupled and non‐coupled region selected and corresponding local Ca²⁺ transient plots normalized to the diastolic [Ca²⁺]_i (please note that traces have been slightly shifted along the x‐axis for clarity). The rate of upstroke and Ca²⁺ transient amplitude in coupled and non‐coupled regions in the absence and presence of ISO (n _cells = 30; N _pigs = 10). ^** P < 0.01; ^*** P < 0.001.

Epac, a direct target for cAMP, has also been reported to increase CaMKII activation independently of PKA (Pereira et al. 2013; Ruiz‐Hurtado et al. 2013). To test this, we used the recently described Epac1 inhibitor CE3F4 (Courilleau et al. 2013) and Epac2 inhibitor ESI‐05 (Domínguez‐Rodríguez et al. 2015) (Fig. 5 B). CE3F4 and ESI‐05 did not affect spark frequency in either coupled or non‐coupled RyRs under ISO stimulation. Similarly, the SR Ca²⁺ content was not altered by these agents.

Finally, we examined the role of local increases in [Ca²⁺]_i during the stimulation train (Fig. 5 C). We indirectly estimated differences in local [Ca²⁺]_i by measuring the rate of upstroke of local Ca²⁺ transients at selected coupled and non‐coupled regions. The rate of upstroke may be more indicative of local transient elevations of [Ca²⁺]_i than its amplitude in the same region. Despite a large increase in the local Ca²⁺ transient amplitude in non‐coupled RyRs, the rate of upstroke at those areas remained substantially below that of coupled regions. These data therefore indicated that increases in the local [Ca²⁺]_i are far more pronounced at coupled than non‐coupled areas, and are consistent with a much greater CaMKII activation at coupled RyR clusters.

Taken together, these data indicate that the activation of CaMKII in coupled RyRs is likely to occur via local Ca²⁺ and NOS1 microdomains, and is not dependent on Epac.

CaMKII is equally distributed but has a higher level of phosphorylation in coupled sites

We previously showed an equal distribution of CaMKII in both coupled and non‐coupled RyRs (Dries et al. 2013). However, the above results suggest a greater CaMKII activation near coupled RyRs after ISO stimulation. Therefore, we investigated whether there is an increased distribution of activated CaMKII [phospho‐CaMKII Th286 (P‐CaMKII)] in coupled RyRs when myocytes are conditioned at 0.5 Hz stimulation in the presence of ISO. Triple immunolabelling for RyR, NCX and P‐CaMKII were performed and the fluorescence intensity for P‐CaMKII was measured in the presence and absence of ISO in electrically paced myocytes (Fig. 6 A). The total fluorescence intensity of P‐CaMKII was not changed after ISO (Fig. 6 Ba). However, local analysis of the fluorescence intensity in coupled relative to non‐coupled RyRs within each myocyte revealed a higher P‐CaMKII in coupled RyRs (Fig. 6 Bb). These data confirm our findings at the functional level, further supporting the hypothesis of a local CaMKII activation near coupled RyRs.

Figure 6. Distribution of P‐CaMKII in coupled and non‐coupled RyRs .

A, example of triple immunostaining of myocytes in the presence and absence of ISO. RyRs are stained in green, NCX in blue and P‐CaMKII in red. Higher magnification images of the indicated regions are shown for coupled and non‐coupled RyRs. Ba, total fluorescence intensity for P‐CaMKII in the absence and presence of ISO; b, the ratio of fluorescence intensity for P‐CaMKII in coupled and non‐coupled RyRs in the absence and presence of ISO (n _cells = 7–10 cells per pig; N _pigs = 5). *** P < 0.001.

PKA modulates spark frequency in both coupled and non‐coupled RyRs, at least partially by modulating SR Ca²⁺ load

Our mechanistic dissection uncovered that coupled RyRs are differentially regulated by ISO, in particular via local CaMKII activation, which is dependent on high local [Ca²⁺]_i and NOS1. Furthermore, CaMKII‐dependent modulation occurred in the absence of changes in SR Ca²⁺ content. These insights leave a number of questions unanswered: how is spark frequency increased by ISO at non‐coupled clusters and what is the role of PKA?; and to what extent does SR load influence spark activity at coupled and non‐coupled areas compared to RyR phosphorylation? We therefore first examined how PKA modulates coupled and non‐coupled RyRs. Application of the peptide‐based PKA inhibitor PKI during β‐adrenergic stimulation reduced the frequency of sparks arising from both coupled and non‐coupled RyRs (Fig. 7 A) and reduced the SR Ca²⁺ content (Fig. 7 B). Analysis of the relationship between SR Ca²⁺ content and spark frequency indicated that the reduced activity of coupled and non‐coupled RyRs by PKA occurred together with a reduction in SR Ca²⁺ content (Fig. 7 C). These experiments were repeated with another PKA inhibitor, H‐89, and similar results were obtained (Fig. 8). Therefore, it is possible that the observed RyR modulation depends on both direct RyR modifications (i.e. phosphorylation) and indirect regulation via changes in SR Ca²⁺ load.

Figure 7. PKA affects coupled and non‐coupled RyRs during β‐adrenergic stimulation .

A, the effect of PKI on Ca²⁺ spark frequency in coupled and non‐coupled RyRs after ISO (n _cells = 18; N _pigs = 5). B, the effect of PKI on the SR Ca²⁺ content after ISO (n _cells = 12; N _pigs = 4). C, the relationship between SR Ca²⁺ load and Ca²⁺ spark frequency at baseline (n _cells = 7; N _pigs = 6), with ISO (n _cells = 10; N _pigs = 8) and with ISO plus PKI (n _cells = 8; N _pigs = 3). ^* P < 0.05; ^** P < 0.01.

Figure 8. H‐89 affects all RyRs and global Ca²⁺ handling during β‐adrenergic stimulation .

A, the effect of H‐89 on Ca²⁺ spark frequency in coupled and non‐coupled RyRs after ISO (n _cells = 18; N _pigs = 4). B, the effect of H‐89 on the SR Ca²⁺ content (n _cells = 14; N _pigs = 4); C, peak Ca²⁺ current (n _cells = 13; N _pigs = 3); and D, Ca²⁺ transient kinetics (n _cells = 11; N _pigs = 4). *** P < 0.001.

To distinguish between a direct effect of PKA on RyR and an effect on store load, Ca²⁺ sparks were recorded after reducing SR Ca²⁺ load in the presence of ISO (Fig. 9 Aa). To this end, we conditioned cells as described before and applied a short caffeine pulse (400 ms, 10 mm caffeine) 6 s before the last Ca²⁺ transient to induce (partial) emptying of the SR before spark recording. The timing and duration of the caffeine pulse ensured an RyR activity that was not modified by caffeine at the time of spark collection. Additionally, the time to 50% relaxation of the last Ca²⁺ transient prior to spark recording was not different from the Ca²⁺ transient prior to caffeine application (data not shown), indicating that this intervention did not affect PLN phosphorylation. The intervention did, however, reduce SR Ca²⁺ content as assessed in Fig. 9 Ab.

Figure 9. A reduced SR Ca²⁺ load affects both coupled and non‐coupled RyRs during maintained β‐adrenergic stimulation .

Aa, myocytes are stimulated in the presence of ISO as described before. To reduce the SR Ca²⁺ load, a brief caffeine pulse (400 ms) is applied before the end of the stimulation train. Ab, SR Ca²⁺ content; and B, spark frequency in the presence of ISO or ISO with a reduced SR Ca²⁺ load (n _cells = 24; N _pigs = 3). C, the relationship between SR Ca²⁺ load and Ca²⁺ spark frequency at baseline, ISO (n _cells = 10; N _pigs = 8), ISO with low SR load (n _cells = 25; N _pigs = 3), and AIP and AIP with low SR load (n _cells = 11; N _pigs = 4) in coupled and non‐coupled RyRs. ^* P < 0.05; ^*** P < 0.001.

The lower SR Ca²⁺ load was associated with a reduction in the Ca²⁺ spark frequency for both coupled and non‐coupled RyRs (Fig. 9 B and Fig. 9 C, light blue symbols, compared to ISO alone in light grey symbols). However, spark frequency remained higher when compared to the baseline spark frequency with an equal SR Ca²⁺ load (light blue vs. black in Fig. 9 C). These findings indicate that the increase in SR Ca²⁺ load only partially accounts for the response of coupled and non‐coupled RyRs to β‐adrenergic stimulation. In addition, similar experiments in which the SR Ca²⁺ load was reduced were also performed in the presence of AIP. Here lowering of SR Ca²⁺ load resulted in significant reduction of spark frequency in non‐coupled RyRs but the low frequency in coupled RyRs was not further reduced, supporting the importance of the regulation of coupled RyRs via CaMKII (Fig. 9 C, green symbols compared to dark blue symbols).

CaMKII‐ and PKA‐dependent effects on global Ca²⁺ transients triggered by LTCC

We have focused thus far on the spontaneous release events at resting membrane potential and on the SR Ca²⁺ content. We next examined the contribution of CaMKII activation in the dyadic cleft on Ca²⁺ handling during depolarizing pulses.

As illustrated in Fig. 10 Aa, the increase in I _CaL under ISO is sensitive to CaMKII inhibition. Despite this reduction of ICaL, global Ca²⁺ transients are not significantly affected in amplitude or kinetics of upstroke or relaxation (Fig. 10 Ab). A possible factor here is that the SR Ca²⁺ load was not reduced by CaMKII inhibition as shown before in Fig. 1 Bb and may be related to lack of an appreciable effect on SERCA activity; this observation is supported by the maintained high rate of relaxation. Nevertheless, the effect of CaMKII inhibition on I _CaL and coupled RyRs P _o can be detected in the rate of upstroke of I _CaL‐triggered local Ca²⁺ transients in the dyadic cleft (Fig. 10 Ac).

Figure 10. Global Ca²⁺ handling is increased during β‐adrenergic stimulation via global PKA effects and local CaMKII effects .

Aa, peak Ca²⁺ current amplitude (n _cells = 10; N _pigs = 3); and b, global and c, local Ca²⁺ transient kinetics in the presence of ISO or ISO with AIP (n _cells = 14; N _pigs = 4). Ba, peak Ca²⁺ current amplitude (n _cells = 8; N _pigs = 3); and b, global Ca²⁺ transient kinetics in the presence of ISO or ISO with PKI (n _cells = 20; N _pigs = 5). ^* P < 0.05; ^** P < 0.01; ^*** P < 0.001.

In contrast, as illustrated in Fig. 10 Ba, PKI not only reduced I _CaL but also had a major effect on the global Ca²⁺ transient, reducing the amplitude and kinetics of upstroke and rate of decay (Fig. 10 Bb). This is probably due to a major reduction in SR Ca²⁺ content (Fig. 7 B), and concordant with a reduced SERCA activity, evidenced by the slowing of relaxation with PKI.

These results indicate that β‐adrenergic activation of CaMKII has effects that are spatially restricted to the dyadic cleft. In contrast, PKA has global effects on systolic Ca²⁺ handling and Ca²⁺ transient amplitude during β‐adrenergic stimulation, at least partly indirectly, via alterations in SR load.

Discussion

The present data show that different signalling microdomains underlie the integrated response to β‐adrenergic modulation of LTCCs, PLN and RyRs. CaMKII activation is confined to the dyadic cleft, where it enhances the P _o of RyRs and LTCCs and contributes to the increase of the local Ca²⁺ release and diastolic events. Enhanced SR Ca²⁺ uptake and SR Ca²⁺ load are PKA‐dependent and major determinants of the global increase of Ca²⁺ cycling.

The dyadic cleft is a microdomain for CaMKII and NOS1‐dependent modulation of RyRs

During β‐adrenergic stimulation, CaMKII‐dependent modulation of RyRs exclusively occurs in the restricted area of the dyadic cleft. These data are in line with our previous report of CaMKII activation during high‐frequency stimulation in the absence of ISO (Dries et al. 2013). NOS1 equally participates in the local modulation of coupled RyRs. However, the current data cannot identify the relationship between CaMKII and NOS1 because inhibition of either pathway reduces spark frequency to baseline levels. This is consistent with a sequential and interdependent activation, but the current data cannot inform about the hierarchy of this dyad‐specific signalling.

NOS1 could enhance RyR activity directly through nitrosylation or indirectly via nitrosylation of CaMKII, or both. Data obtained in rabbit myocytes (Curran et al. 2014) suggest that CaMKII is activated downstream of NO. These authors show that after pre‐incubation of purified CaMKII with Ca²⁺ and CaM, application of SNAP (NO donor) resulted in a high CaMKII activity. This was not observed when pre‐incubated with EGTA, indicating that high levels of Ca²⁺ are a prerequisite for CaMKII activation by NO. Erickson et al. (2015) recently reported that NO treatment after Ca²⁺/CaM binding can result in the autonomous activation of CaMKII via S‐nitrosylation. This CaMKII activation mechanism was suggested earlier by Guttierez et al. (2013) who furthermore showed that CaMKII was activated independently of [Ca²⁺]_i when ISO was applied in quiescent cells. It should be noted, however, that previous studies used high ISO concentrations, 1 μm (Gutierrez et al. 2013) and 250 nm (Curran et al. 2014) as opposed to 10 nm used in this study.

If NOS1 is the primary step in leading to RyR phosphorylation, little is known regarding how β‐AR could activate NOS1. NOS1, RyRs and CaMKII are spatially colocalized (Barouch et al. 2002; Bendall et al. 2004; Damy et al. 2004) and this could create a local NO microdomain near coupled RyRs. Recently, Curran et al. (2014) proposed a role for β1‐AR and NOS1 activation via AKT signalling. However, the exact mechanism of NOS1 activation upon β‐adrenergic signalling remains elusive.

Although data in the present study are compatible with a NOS1 activation of CaMKII, the data also suggest that direct activation of CaMKII by Ca²⁺ is involved in RyR regulation. Indeed, CaMKII activation required the increased Ca²⁺ levels associated with the conditioning pulses (which are PKA‐dependent): when the amplitudes of the Ca²⁺ transients during the conditioning pulses were reduced with PKI (Fig. 10 B), the activity of coupled receptors also decreased. While this could imply co‐activation of PKA and CaMKII, reducing the SR Ca²⁺ load, even without reducing PKA activity (Fig. 9 C), also reduced sparks in coupled sites, suggesting again that the CaMKII activation is dependent on local Ca²⁺. The local Ca²⁺ levels in the dyadic cleft during depolarizing pulses are indeed much higher than in the global cytosol (e.g. Acsai et al. 2011 in a similar myocyte cell type) and analysis of the local Ca²⁺ transients during the conditioning pulses is consistent with a different local [Ca²⁺]_i between coupled and non‐coupled RyRs (Fig. 5 C).

Another pathway for CaMKII activation is through reactive oxygen species (ROS). So far there is no conclusive evidence indicating that ROS production during β‐adrenergic stimulation can directly activate CaMKII. NO‐dependent CaMKII modulation has been reported to occur independently of ROS, as direct ROS scavenging during β‐adrenergic stimulation did not affect SR Ca²⁺ sparks (Gutierrez et al. 2013) and NOX2 inhibition during ISO did not shift the SR load–leak relationship (Curran et al. 2014). In line with these studies, global ROS scavenging using NAC did not affect the spark frequency in coupled or non‐coupled sites, nor did it affect SR Ca²⁺ content (Fig. 11). Likewise, Bovo et al. (2012, 2015) showed an increased ROS production during β‐adrenergic stimulation but it was only the direct oxidation of RyRs that changed RyR sensitivity. On the other hand, we recently showed that enhanced RyR activity at coupled sites during high‐frequency stimulation depends on the interdependent activation of NOX2 and CaMKII (Dries et al. 2013), these data being consistent with a ROS microdomain in the dyadic cleft that may regulate CaMKII activity.

Figure 11. Global ROS scavenging does not affect RyRs and SR Ca²⁺ content during β‐adrenergic stimulation .

A, the effect of NAC on Ca²⁺ spark frequency in coupled and non‐coupled RyRs after ISO (n _cells = 11; N _pigs = 3); and B, on the SR Ca²⁺ content (n _cells = 8; N _pigs = 3).

In the present study, local CaMKII activation at coupled sites occurred independently from Epac. These data are in line with the study of Curran et al. (2014), where neither direct activation of Epac nor increases in cAMP altered the SR Ca²⁺ load–leak relationship. Epac activation was also not able to alter the spark frequency. Others, however, found a clear link between Epac activation and an increased SR Ca²⁺ leak via CaMKII‐dependent phosphorylation of RyR‐S2814 (Pereira et al., 2007, 2013). Pereira et al. (2015) recently showed that in different subcellular regions specific Epac isoforms are present, with Epac2 located near TTs and involved in the regulation of SR Ca²⁺ leak. The difference of the aforementioned study with the present data may be a species‐dependent difference. In addition, the specificity of both Epac inhibitors could not be assessed in the present study.

Lastly, the signalling to CaMKII could be related to local β‐AR subtypes. In rat myocytes β2‐ARs have been shown to preferentially reside within TTs in contrast to the β1‐ARs, which are located across the entire cell surface and may be expected to mediate a more ‘global’ response compared to β2‐ARs (Nikolaev et al. 2010).

Integrated CaMKII and PKA signalling during β‐adrenergic stimulation

CaMKII and PKA pathways both contribute to the β‐adrenergic response. CaMKII signalling appears to have its major effects on diastolic events with more modest impact on systolic function, whereas PKA has major impact on systolic function through the modulation of SR Ca²⁺ content.

While this opens interesting perspectives as discussed in the next section, there are also several unanswered questions about the interaction and relative importance of the two pathways in modulating RyR activity.

A simultaneous requirement of CaMKII and PKA for the enhancement of RyR activity has recently been proposed by Polakova et al. (2015), using mouse myocytes, where presumably all RyRs are coupled. If we only consider coupled RyRs, the conclusions of Polakova are consistent with the present data (Fig. 7) showing that PKA inhibition reduces activity of coupled RyRs. A specific explanation here is, as stated above, that PKI inhibition reduces cellular Ca²⁺ levels (because it lowers the SR Ca²⁺ load) and thereby prevents CaMKII activation. Nevertheless, with the reduced Ca²⁺ load in the presence of PKA, sparks from coupled RyRs were still elevated and sensitive to CaMKII inhibition. An alternative explanation may therefore be that PKA has a direct co‐activating effect.

Not much is known about such a potential molecular mechanism. The role of PKA‐dependent phosphorylation at S2808/2809 in itself and in contrast to S2814 has been debated (Wehrens et al. 2004; Xiao et al., 2006, 2007; MacDonnell et al. 2008; Morimoto et al. 2009; Shan et al. 2010). Molecular modelling found even more phosphorylation sites on RyR2 (Van Petegem, 2012) but their relative functional roles remain largely unknown. Our initial hypothesis was for a dual phosphorylation of coupled RyRs at dyadic sites with cumulated stimulating effects and highest spark frequency for coupled RyRs (PKA and CaMKII) compared to non‐coupled RyRs (PKA only). However, the data presented in this study do not show additive effects at coupled sites despite the fact that there is a site‐specific CaMKII effect.

The most puzzling aspect to our data is why inhibition of CaMKII results in lower spark frequency in coupled RyRs vs. non‐coupled RyRs despite maintained PKA activation. If the role of PKA is predominantly through increased SR Ca²⁺ load, there should not be a difference between coupled and non‐coupled RyRs, assuming SR Ca²⁺ load is the same throughout the SR network. We have previously investigated whether there were standing SR gradients after stopping the conditioning pulses and at the time of sparks recording, but could not detect any (Dries et al. 2013). The remaining explanation, and hypothesis, is then that PKA modulation of RyRs at coupled sites is less predominant than at non‐coupled sites. Differences in AKAP could also play a role and explain the differential contribution of PKA but are presently beyond study.

Finally, the data from Fig. 9 C suggest that there is a load‐independent effect of PKA presumably through phosphorylation of RyR clusters.

Limitations

A number of potential mechanisms were not further investigated, some due to methodological limitations. We failed to find suitable NOS1 antibodies to work in the pig. So far we also have not been able to measure spatially confined signals through ROS, because of low signals. Immunostaining is not optimal for quantification of P‐CaMKII, although the signals appear to be quite specific (supported by KN‐93 inhibition of immunostaining – pilot data, not shown) but measuring localized CaMKII signals in living cells requires fluorescence resonance energy transfer probes that can only be introduced during cell culture (Erickson et al. 2011).

There are additional hypotheses regarding microdomain signalling for future studies. These include the distribution of β‐ARs (Nikolaev et al. 2010) as well as the macrocomplex properties of the RyR clusters, coupled vs. non‐coupled. So far biochemical approaches are not feasible due to the difficulties in extracting the two RyR populations but we are currently developing methods for in situ characterization.

In the present study we could not resolve differences in the total level of CaMKII phosphorylation despite the fact that the functional evidence clearly indicates CaMKII‐dependent phosphorylation at coupled sites, which make up about 50% of all sites. Several factors may contribute to this. The data in Fig. 6 suggest there is a baseline degree of phosphorylation of CaMKII. Against this background small changes may be difficult to detect. For the data on total P‐CaMKII, we also compare populations of cells, introducing more variability and the inherent variability of the assay decreases the ability to detect small changes in fluorescence between cells. In our analysis, a ratio of staining in coupled vs. non‐coupled RyRs within each cell is determined on a cell by cell basis and reduces noise on these measurements. Furthermore, we applied only a low concentration of ISO to avoid massive stimulation and Ca²⁺ overload, which probably resulted in submaximal activation of CaMKII.

Perspectives and conclusions

The increase in diastolic events with β‐adrenergic stimulation, at the level of the whole cell, is both PKA‐ and CaMKII‐driven. Yet, the sparks that are specifically CaMKII‐dependent, i.e. at coupled sites, probably have the highest potential for triggering Ca²⁺ wave activity and deleterious membrane depolarizations. First, as shown in Fig. 4 these are the sites that have more repetitive sparks and CaMKII increases the number of coupled sites being recruited for repetitive activation, thereby facilitating summation in space and time to form propagating Ca²⁺ waves. Secondly, the location at the membrane promotes direct activation of the NCX current. Lastly, in vivo β‐adrenergic stimulation is a combination of increased frequency (chronotropy at the SA node) and direct β‐AR‐mediated effects as studied here at low frequency. As we have previously shown that frequency by itself specifically enhances diastolic events at coupled RyRs (Dries et al. 2013), these two mechanisms are likely to reinforce each other and predict wave formation from coupled RyRs. Data on Ca²⁺ waves using myocytes conditioned at a high frequency in the presence of ISO support this concept: more Ca²⁺ waves originated at coupled RyRs than at non‐coupled RyRs (Fig. 12 A). A preliminary data set further shows that this phenomenon is under control of CaMKII, as inhibition with AIP decreased Ca²⁺ wave frequency at coupled sites only (Fig. 12 B).

Figure 12. More Ca²⁺ waves originate from coupled RyRs after ISO .

A, incidence of Ca²⁺ waves occurring in coupled vs. non‐coupled RyRs when stimulating cells at high frequency (2 Hz) and in the presence of ISO (n _cells = 33; N _pigs = 8) and ISO with AIP (n _cells = 20; N _pigs = 3). B, preliminary data set showing the Ca²⁺ waves frequency in coupled vs. non‐coupled RyRs in the presence of AIP (n _cells = 20; N _pigs = 3). ^* P < 0.05.

The present results also indicate that CaMKII activation has limited impact on the global Ca²⁺ handling that underlies the positive inotropic and lusitropic actions of β‐adrenergic stimulation. Therefore, the present data further support the concept that CaMKII may be a preferential target for reducing arrhythmic events as inhibition would have minor effects on contraction and relaxation.

Finally, although the current study has concentrated on ventricular myocytes, there is also the possibility that similar RyR subpopulation modulation is present in other cell types. Previous work has shown the presence of a large fraction of non‐coupled RyRs in atrial cells of large mammals including humans (Dibb et al. 2009; Lenaerts et al. 2009). Whether these RyR subpopulations have different modulation is the subject of ongoing work.

In conclusion, the present data extend the concept of microdomain signalling in the dyadic cleft and open perspectives for selective modulation of subpopulations of RyRs.

Additional information

Competing interests

None.

Author contributions

K.R.S., E.D. and H.L.R. were involved in designing the experiments and interpreting the data; E.D., D.M.J., S.M.K., G.G. and P.H. collected the data and E.D., D.J.S., G.G. and P.H. analysed the data; K.R.S., E.D., D.J.S., D.M.J., H.L.R., G.G., S.M.K. and P.H. were involved in writing the paper. All authors approved the final version of the manuscript.

Funding

This work was funded by the Interuniversity Attraction Poles P7/10 to K.R.S. and the Research Foundation Flanders (FWO) (project grant to K.R.S.; postdoctoral fellowship to D.M.J. and PhD fellowship to E.D.).