Inhibition of Cardiac Kir Current (IK1) by Protein Kinase C Critically Depends on PKCβ and Kir2.2

Background Cardiac inwardly rectifying Kir current (IK1) mediates terminal repolarisation and is critical for the stabilization of the diastolic membrane potential. Its predominant molecular basis in mammalian ventricle is heterotetrameric assembly of Kir2.1 and Kir2.2 channel subunits. It has been shown that PKC inhibition of IK1 promotes focal ventricular ectopy. However, the underlying molecular mechanism has not been fully elucidated to date. Methods and Results In the Xenopus oocyte expression system, we observed a pronounced PKC-induced inhibition of Kir2.2 but not Kir2.1 currents. The PKC regulation of Kir2.2 could be reproduced by an activator of conventional PKC isoforms and antagonized by pharmacological inhibition of PKCβ. In isolated ventricular cardiomyocytes (rat, mouse), pharmacological activation of conventional PKC isoforms induced a pronounced inhibition of IK1. The PKC effect in rat ventricular cardiomyocytes was markedly attenuated following co-application of a small molecule inhibitor of PKCβ. Underlining the critical role of PKCβ, the PKC-induced inhibition of IK1 was absent in homozygous PKCβ knockout-mice. After heterologous expression of Kir2.1-Kir2.2 concatemers in Xenopus oocytes, heteromeric Kir2.1/Kir2.2 currents were also inhibited following activation of PKC. Conclusion We conclude that inhibition of cardiac IK1 by PKC critically depends on the PKCβ isoform and Kir2.2 subunits. This regulation represents a potential novel target for the antiarrhythmic therapy of focal ventricular arrhythmias.


Methods and Results
In the Xenopus oocyte expression system, we observed a pronounced PKC-induced inhibition of Kir2.2 but not Kir2.1 currents. The PKC regulation of Kir2.2 could be reproduced by an activator of conventional PKC isoforms and antagonized by pharmacological inhibition of PKCβ. In isolated ventricular cardiomyocytes (rat, mouse), pharmacological activation of conventional PKC isoforms induced a pronounced inhibition of I K1 . The PKC effect in rat ventricular cardiomyocytes was markedly attenuated following co-application of a small molecule inhibitor of PKCβ. Underlining the critical role of PKCβ, the PKC-induced inhibition of I K1 was absent in homozygous PKCβ knockout-mice. After heterologous expression of Kir2.1-Kir2.2 concatemers in Xenopus oocytes, heteromeric Kir2.1/Kir2.2 currents were also inhibited following activation of PKC.

Conclusion
We conclude that inhibition of cardiac I K1 by PKC critically depends on the PKCβ isoform and Kir2.2 subunits. This regulation represents a potential novel target for the antiarrhythmic therapy of focal ventricular arrhythmias.

Introduction
Cardiac inwardly rectifying potassium current I K1 mediates the terminal repolarization phase of the cardiac action potential and determines diastolic membrane conductance [1]. The molecular basis of cardiac I K1 is formed by heterotetrameric assembly of three related channels: Kir2.1, Kir2.2 and Kir2.3. Ventricular I K1 has been shown to be predominantly based on heteromeric assembly of Kir2.1 and Kir2.2 channels, whereas Kir2.3 channels are more relevant on the atrial level [2][3][4][5][6].
In Andersen's syndrome (Long QT syndrome 7), loss-of-function mutations in Kir2.1 are associated with reduced I K1 density leading to QT interval prolongation, prominent U-waves and triggered activity in the form of complex ventricular ectopy and polymorphic ventricular tachycardia [7]. Interestingly, catecholaminergic stress can trigger ventricular arrhythmia in a subset of those patients and there is considerable phenotypical overlap with catecholaminergic polymorphic ventricular tachycardia (CPVT) [7,8]. In chronic heart failure, downregulation of cardiac I K1 is a hallmark of the electrical remodeling process and has been shown to facilitate triggered ventricular arrhythmias [9][10][11][12].
It has been demonstrated that pathophysiologically relevant neurohumoral pathways (e.g. via α 1 -adrenoreceptors and ET A -endothelin receptors) inhibit cardiac I K1 through protein kinase C (PKC) dependent signalling, thereby favouring focal arrhythmogenesis by destabilisation of the resting membrane potential [13][14][15][16][17]. To date, twelve isoforms of PKC have been characterised and are classified based on their structure and cofactor regulation in "conventional", "novel" and "atypical" isoforms. Targeting specific PKC isoforms for therapeutic purposes is under extensive evaluation. Kong et al. (2008) demonstrated that PKCβ is a crucial mediator of myocardial ischemia-reperfusion injury and hence suggested inhibition of PKCβ as a therapeutic strategy [18]. In line with this argumentation, Bowman et al. (1997) found that over-expression of PKCβ in the myocardium induced sudden death in neonatal mice and pathological hypertrophy in adult mice [19].
It has been recognized in various species and in human cardiomyocytes that activation of protein kinase C induces a marked inhibition of native cardiac I K1 [13,16,20]. Notably, it has also been shown that stimulation of α-adrenergic receptors strongly promotes focal ventricular tachycardia in a dog model of inherited arrhythmia by a decrease of I K1 in Purkinje fibers [15]. However, the precise molecular mechanisms underlying this regulation and the responsible cardiac PKC isoform have not been elucidated to date.
In this study, we provide evidence for the crucial role of the PKCβ isoform for the inhibition of I K1 by protein kinase C. We further demonstrate that Kir2.2 most likely determines PKCsensitivity of Kir2.1/Kir2.2 heteromers, the main molecular substrate of ventricular I K1 . This regulation may contribute to arrhythmogenesis in focal ventricular tachycardia.
For patch-clamp measurements of rat and mice ventricular I K1 , pharmacological agents were added to the bath solution to block L-type calcium currents (10 μmol/l nisoldipine), the slow component of delayed rectifier K + current (I Ks , 10 μmol/l chromanol), the rapid component of delayed rectifier K + current (I Kr , 10 μmol/l E4031) and ATP-dependent K + channels (K ATP , 1 μmol/l glibenclamide).

Isolation of rat and mouse cardiomyocytes
Adult rat ventricular cardiomyocytes were isolated by enzymatic dispersion by use of collagenase. In brief, rats were anesthetized with sodium pentobarbital (60 mg/kg i.p.) and heparinized and the hearts were rapidly removed and perfused in a Langendorff perfusion system. Meanwhile the heart was subsequently digested by collagenase to break down the extracellular matrix. Afterwards, the ventricles were minced and the resulting cellular digest was washed, filtered and resuspended in Medium 199 (Sigma Aldrich Chemie GmbH, Germany). Freshly isolated cardiomyocytes were attached to laminin coated cover slips by one hour incubation at 37°C and 5% CO 2 . After one hour, the medium was changed to separate the living, attached cells from the dead and further experiments were started.
Adult mice were sacrificed by decapitation. Single cells from mouse ventricular tissue were obtained using the protocol described in O´Connel et al. (2007) [23]. Again, the isolated cardiomyocytes were attached to laminin coated cover slips by one hour incubation and after medium change patch-clamp experiments were started.
The investigation conforms to the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH publication No. 85-23, revised 1996). Approval for the animal experiments reported in this study was given by the Regierungspräsidium Karlsruhe, Germany, reference numbers G-221/12, T-78/15 and T-18/15.

Electrophysiology and statistics
The whole-cell patch-clamp configuration was used to record currents from isolated rat and mouse ventricular cardiomyocytes. The two microlelectrode voltage-clamp configuration was used to record currents from Xenopus laevis oocytes as published previously [22]. All measurements were carried out at room temperature (20-22°C). Only recordings where the initial leak currents did not significantly change within the observation period were considered for data analysis. Statistical data are presented as mean ± standard error of the mean (SEM). In order to compare the statistical significance of the results, unpaired Student´s t-tests or ANOVA followed by t-test were performed: p<0.05 was considered statistically significant and p<0.01 was considered highly significant.

Activation of PKC inhibits Kir2.2 but not Kir2.1 currents
In order to investigate regulatory effects of PKC on homomeric Kir2.1 and Kir2.2 channels, heterologous expression of these channels was performed in Xenopus oocytes. A standardised voltage protocol was applied which evoked characteristic inwardly rectifying currents. Test pulses to potentials ranging from -120 mV to +40 mV were applied in 10 mV-increments for a period of 400 ms each. The holding potential was -80 mV. Steady state inward currents at -120 mV were measured to quantify effects. The observation time was 30 minutes in all experiments using Xenopus oocytes to ensure steady-state conditions. Fig 1A and 1D display typical recordings of Kir2.1 and Kir2.2 currents under control conditions. Under control conditions, i.e. during superfusion with the bath solution, currents exhibited a small time-dependent run-up as reported previously [16,20]. During 30 minutes of superfusion with the bath solution, Kir2.1 currents increased to 111.3 ± 4.9% (n = 5) of the initial values. Kir2.2 channels showed a current run-up to 117.3 ± 6.1% (n = 6). Biophysical current characteristics remained unaltered (data not shown). Next, experiments were repeated with pharmacological activation of PKC by application of the phorbol ester PMA at a concentration of 1 μmol/l. This high concentration of PMA was chosen to achieve maximal PKC activation. During the observation period of 30 minutes, Kir2.1 currents showed a small but significant decrease by 10.3 ± 5.6% (n = 5; p<0.05; Fig 1B and 1G). In contrast, PMA caused a pronounced inhibition of Kir2.2 currents with a decrease by 80.7 ± 4.7% (n = 5; Fig 1E and 1H; p<0.01).
Then, we examined the effects of more specific pharmacological activators of protein kinase C with different pharmacological profiles. Thymeleatoxin (TMTX) predominantly activates conventional PKC isoforms, i.e. PKCα, PKCβ and PKCγ [24], whereas ingenol (ING) predominantly activates novel PKC isoforms, i.e. PKCδ and PKCε [25]. Notably, the presence of all known conventional as well as novel PKC isoforms in Xenopus laevis oocytes was shown recently [26,27]. The experiments were repeated under conditions identical to those described above, i.e. with the same voltage protocol and the same observation time. Inward currents measured before and after application of the PKC activators were compared to obtain relative effects.
In summary, these data point to a pronounced inhibitory regulation of Kir2.2 currents by protein kinase C that predominantly depends on conventional PKC isoforms. In contrast, the mild inhibition of Kir2.1 channels by PMA rather depends on phospholipase C activation without direct involvement of PKC.

Role of PKCβ in the regulation of Kir2.2 channels
As mentioned above, conventional PKC isoforms seem to play a major role in the regulation of Kir2.2 currents. In order to identify the PKC isoform involved in the inhibitory regulation of Kir2.2 channels by PKC, we applied a small molecule inhibitor of PKCβ (3-(1-(3-Imidazol-1-ylpropyl)-1H-indol-3-yl)-4-anilino-1H-pyrrole-2,5-dione; for simplification termed "PKCβ inhibitor" subsequently) together with TMTX (100 nmol/l). As shown in the previous paragraph, application of TMTX alone at this concentration induced a pronounced Kir2.2 current inhibition (-78.1 ± 3.5%, n = 6, Fig 1H). This regulation was strongly attenuated by co-application of the PKCβ inhibitor in a concentration-dependent manner with a resulting current inhibition by 43.4 ± 7.5% (n = 6) at a concentration of 3 μmol/l and by merely 14.7 ± 8.6% (n = 8) at a concentration of 10 μmol/l (p<0.05 and p<0.01 in comparison to the effect of TMTX alone, respectively, Fig 1H). Thus, these pharmacological data support a central role of PKCβ in the regulation of Kir2.2 channels by protein kinase C.

Pharmacological inhibition of PKCβ abolishes I K1 suppression by PKC in rat ventricular cardiomyocytes
In order to examine the significance of native I K1 regulation by PKCβ, we performed whole-cell patch-clamp experiments with isolated rat ventricular cardiomyocytes (Fig 2). The holding potential was set at -80 mV and test pulses ranging from -130 mV to -80 mV in 10 mV-increments were applied for 200 ms each. Compared to voltage-clamp experiments using Xenopus oocytes, the patch-clamp measurements were restricted to inward currents because of not entirely blocked sodium and calcium currents at more positive potentials. After recording a control measurement, we repeated this protocol each 60 seconds for five minutes during superfusion with the bath solution to ensure current stability. Then, TMTX (100 nmol/l) was perfused into the bath and currents were recorded in intervals of five minutes. Recordings from a typical experiment are displayed in Fig 2A-2C. I K1 inward currents were markedly inhibited after application of cPKC activator TMTX (Fig 2A and 2B). Summary data of the time course of effect are shown in Fig 2D. After an observation period of 15 minutes, currents were reduced to 49.9 ± 8.6% of the respective initial values (n = 5; p<0.05). At the end of each experiment, 50 μmol/l barium chloride (BaCl 2 ) was applied to block residual Kir currents. Interestingly, BaCl 2 induced only a small additional current inhibition to 40.3 ± 10% (n = 5), indicating that approximately 84% of BaCl 2 -sensitive current had already been inhibited previously by TMTX ( Fig 2D). Next, we repeated these experiments using TMTX (100 nmol/l) in combination with the pharmacological PKCβ inhibitor (3 μmol/l) that was already used in the expression system as reported in the preceding paragraph. Notably, co-application of the PKCβ inhibitor almost completely abolished the effect of TMTX. After the observation period of 15 minutes, current amplitudes were almost unchanged compared to the initial values at 97.4 ± 11.5% (n = 5; p>0.05). However, the following application of BaCl 2 to the bath induced a fast block of inward currents to 53.5 ± 18.3% (n = 5), indicating that only 6% of the BaCl 2 -sensitive current had been inhibited previously with the co-application of TMTX and PKCβ inhibitor (Fig 2E). All experiments are summarised in Fig 2F.

Inhibitory regulation of ventricular I K1 is absent in PKCβ-knockout mice
In order to further validate the molecular specificity of our findings, we used homozygous PKCβ-knockout mice (PKCβ null (-/-)) [18] and wild type mice (wt) for comparative experiments. From these mice, ventricular cardiomyocytes were isolated and I K1 was measured as described in the previous paragraph. Initially, I K1 was elicited each 60 seconds for five minutes under control conditions to ensure current stability. Afterwards, cPKC activator TMTX (100 nmol/l) was applied (Fig 3A-3C). During the observation period of 15 minutes, I K1 from wild type mice decreased to 48.7 ± 8.6% (n = 5) of the initial values (p<0.05). The subsequent application of BaCl 2 (50 μmol/l) only induced an additional reduction of inward currents to final In homozygous PKCβ-knockout mice the ventricular I K1 is not inhibited after PKC activation with thymeleatoxin. Isolated ventricular cardiomyocytes from wild type (A-C) and PKCβ-knockout mice (D-F) were examined using the whole-cell patch-clamp configuration. Thymeleatoxin (TMTX; 100 nmol/l) was used for PKC activation. The same voltage protocol as in the experiments with rat ventricular cardiomyocytes was used. (G-H) Summary data from the experiments using cardiomyocytes from wild type and PKCβknockout mice. BaCl 2 induced a rapid inhibition of inward currents in PKCβ-knockout mice but not in wild type mice, pointing to a nearly absent PKC regulation of I K1 in PKCβ-knockout mice. * indicates statistical significance. doi:10.1371/journal.pone.0156181.g003 Inhibition of Cardiac Kir Current by PKCβ amplitudes of 27 ± 4.2% (n = 5) of the initial values, demonstrating that approximately 70% of BaCl 2 -sensitive current had been inhibited by the preceding application of TMTX (Fig 3G and 3H).
Ventricular cardiomyocytes from PKCβ-knockout mice were subjected to the same protocol (Fig 3D-3F). Interestingly, in those cells the effect of TMTX was almost completely abolished. After 15 minutes of wash-in of TMTX (100 nmol/l), Kir current amplitudes remained at 96.3 ± 11% (n = 11) of the respective initial values (p<0.01 in comparison to the effect of TMTX in wild type cardiomyocytes). The subsequent application of BaCl 2 into the bath to block residual inward currents induced a rapid reduction of inward currents to 29.4 ± 6.7% of the respective initial values (Fig 3G and 3H) indicating that in cardiomyocytes from PKCβknockout mice, TMTX inhibited only 5% of the barium-sensitive current.
Heteromeric Kir2.1/2.2 channels are inhibited after activation of PKC It has been reported that Kir2.1 and Kir2.2 are the predominant Kir2.x subunits in human cardiac ventricle, and that heteromeric assembly of Kir2.1 and Kir2.2 is likely to be the main molecular correlate of human ventricular I K1 [4]. Furthermore, it has been shown that coexpression of Kir2.1 and Kir2.2 in Xenopus oocytes gives rise to currents with distinct biophysical properties that resemble those of human native I K1 better than those of homomeric Kir2.1 or Kir2.2 currents [3,4]. Based on the presented data showing that Kir2.2 subunits are inhibited by protein kinase C whereas Kir2.1 subunits are not, we were interested in investigating the PKC regulation of Kir2.1/Kir2.2 heteromers. Preisig-Müller and co-workers (2002) expressed heteromeric Kir2.1/Kir2.2 channels generated by concatenating the channels´RNA and expressing them as a single RNA chain resulting in "concatemers" [3]. We also used this approach to express heteromeric Kir2.1/Kir2.2 channels in Xenopus oocytes. Kir2.1/Kir2.2 heteromeric currents were measured and effects were quantified as described for homomeric channels above. Typical currents after heterologous expression of Kir2.1-Kir2.2 concatemers in Xenopus oocytes under control conditions and after application of TMTX (100 nmol/l) and the respective current-voltage relationship are shown in Fig 4A-4C. Under control conditions, heteromeric Kir2.1/Kir2.2 currents increased to 114.4 ± 6.9% (n = 8) which is comparable to the previously described run-up of the homomeric Kir2.1 and Kir2.2 channels. After exposure to TMTX, currents decreased to 61 ± 8.3% (n = 10, p<0.01, Fig 4D). In view of the fact that Kir2.1 is insensitive to regulation by PKC as shown above, we therefore propose that co-heteromerization of Kir2.2 subunits confers PKC-sensitivity to the Kir2.1/Kir2.2 heterotetramers.

Discussion
Although the inhibitory regulation of cardiac I K1 by the protein kinase C system has been demonstrated to be of major pathophysiological relevance for focal arrhythmogenesis, the underlying molecular mechanisms have only partially been elucidated to date [13][14][15][16][17]. To the best of our knowledge, this is the first study providing comparative data on the regulation of homomeric and heteromeric Kir2.1 and Kir2.2 channels and native I K1 by protein kinase C isoforms. Our data demonstrate that protein kinase C signalling induces a pronounced inhibition of homomeric Kir2.2 channels but not of Kir2.1 channels. This regulation is predominantly mediated by conventional PKC isoforms with a crucial role of the PKCβ isoform as demonstrated in cardiomyocytes from PKCβ-knockout mice. Kir2.2 channel subunits most likely mediate PKCsensitivity of Kir2.1/Kir2.2 heteromers.

Differential regulation of homomeric Kir2.1 and Kir2.2 channels by PKC
We observed a pronounced inhibition of Kir2.2 channels by PKC while Kir2.1 channels were insensitive to this regulation. In Xenopus oocytes, Kir2.2 currents showed a marked inhibition after application of the phorbol ester PMA that is commonly used as a potent activator of PKC. However, also the more specific PKC activator thymeleatoxin induced an inhibition of Kir2.2 currents. In contrast, Kir2.1 currents showed only a mild inhibition by PMA but neither by thymeleatoxin nor ingenol. However, the PMA-induced Kir2.1 inhibition was abolished after co-application of D609, a specific phospholipase C inhibitor. Therefore, we propose that a PKC-sensitivity of Kir2.1 channels is unlikely and that the effect of PMA is probably caused by an activation of phospholipase C with a subsequent reduction of the membrane PIP 2 content independent of PKC [28,29].
Different studies of the modulation of Kir2.1 channel activity by PKC have been performed with contradicting results. Initially, Fakler and co-workers (1994) observed a Kir2.1 channel inhibition by PKC which was later questioned due to the non-physiological experimental setting in excised inside-out patches [32]. Accordingly, in whole cell-experiments two other groups did not observe regulatory effects of PKC on homomeric Kir2.1 channels [33,34].
In an earlier work from our group, using the computer program HUSAR Prosite (DKFZ, Heidelberg, Germany) to search for key sequences for PKC phosphorylation (amino acids S/ T-X-R/K) in Kir2.2 channels, 4 putative sites (T38, S64, T353, S357) were identified with S64 and T353 have been emerged to be of functional relevance [20]. We performed an alignment between human Kir2.1 and Kir2.2 channel sequences (Clustal Omega software, European Molecular Biology Laboratory-European Bioinformatics Institute, Cambridge, UK), demonstrating that the PKC consensus sites T38, S64 and T353 are lacking in Kir2.1 channels (Fig 5A  and 5B). Unless the lack of these consensus sites provides a plausible mechanism why Kir2.1 channels are insensitive towards PKC regulation we did not show direct phosphorylation of PKC at Kir2.2 residues and therefore other mechanisms, leading to Kir2.2 channel inhibition downstream PKC signalling are not fully excluded.
Inhibition of ventricular I K1 depends on conventional PKC isoforms with an essential role of PKCβ examined in further detail in the respective studies. Based on the pharmacological data obtained in the expression system we hypothesized that this regulation is mediated by conventional PKC isoforms with a central role of PKCβ. In line with this notion, we found in rat ventricular cardiomyocytes that cPKC activator thymeleatoxin induced a pronounced inhibition of I K1 and that this effect was suppressed by co-application of a small-molecule PKCβ inhibitor. Finally, we obtained ventricular cardiomyocytes from PKCβ-knockout mice to test our hypothesis at a level of molecular specificity superior to pharmacological approaches. Supporting our hypothesis, we observed a marked inhibitory regulation of ventricular I K1 by protein kinase C in wild type cardiomyocytes that was completely absent in cardiomyocytes from PKCβ-knockout mice indicating that PKCβ is essential for this regulation.
Kir2.2 channels mediate PKC-sensitivity of heteromeric Kir2.1/Kir2.2 channels The main molecular correlate of ventricular I K1 is heterotetrameric assembly of Kir2.x channels with Kir2.1 and Kir2.2 being the predominant subunits [3,6]. As Kir2.2 showed a pronounced regulation by PKC pathways but Kir2.1 channels did not, we were interested to study the PKC regulation of heteromeric channels. We used Kir2.1-2.2 concatemers provided by Preisig-Müller and co-workers (2002) to generate heteromeric channels [3]. Expression of these concatemers results in heterotetrameric channels consisting of two Kir2.1 and Kir2.2 subunits, respectively. Kir2.1/Kir2.2 heteromers exhibited a pronounced PKC-sensitivity in spite of the fact that homomeric Kir2.2 channels were lacking in the membrane. Hence, we propose that the presence of Kir2.2 channel subunits within a heteromer determines its sensitivity to PKC regulation.

Potential implications on cellular electrophysiology
In cardiomyocytes, proarrhythmic signalling through adrenergic α 1A receptors, angiotensin receptors and endothelin receptors is associated with PKC translocation, leading to a subsequent inhibition of I K1 [13,15,16]. The subsequent destabilization of the resting membrane potential strongly facilitates the generation of delayed afterdepolarisations, triggered activity and ventricular tachyarrhythmias [7,12]. Hence, the PKCβ dependent inhibition of I K1 may be an antiarrhythmic target in the therapy of focal ventricular tachycardia.

Conclusion
In this study, we provide evidence for the regulation of cardiac I K1 by conventional protein kinase C isoforms with a central role of the PKCβ isoform. As Kir2.1 channels are insensitive to PKC regulation, the regulation of heteromeric channels by PKC depends on Kir2.2 subunits. Due to its proarrhythmic potential, this regulation represents a potential target for the therapy of focal ventricular tachycardia.