Activation of Protein Kinase C (cid:1) Inhibits the Two-pore Domain K (cid:2) Channel, TASK-1, Inducing Repolarization Abnormalities in Cardiac Ventricular Myocytes*

Activation of the platelet-activating factor (PAF) receptor leads to a decrease in outward current in murine ventricular myocytes by inhibiting the TASK-1 channel. TASK-1 carries a background or “leak” current and is a member of the two-pore domain potassium channel family. Its inhibition is sufficient to delay repolarization, causing prolongation of the action potential duration, and in some cases, early after depolarizations. We set out to determine the cellular mechanisms that control regulation of TASK-1 by PAF. Inhibition of TASK-1 via activation of the PAF receptor is protein kinase C (PKC)-dependent. Using isoform-specific PKC inhibitor or activator peptides in patch clamp experiments, we now demonstrate that activation of PKC (cid:1) is both necessary and sufficient to regulate murine TASK-1 current in a heterologous expression system and to induce repolarization abnormalities in isolated myocytes. Further-more, site-directed mutagenesis studies have identified threonine 381, in the C-terminal tail of murine TASK-1, as a critical residue in this regulation. Regulation of cardiac function depends on the appropriate cumulative activity of numerous ion channels in individual cardiomyocytes that are responsible for the sequential depolar-ization-repolarization cycle known as the action potential (AP). 1 the

Regulation of cardiac function depends on the appropriate cumulative activity of numerous ion channels in individual cardiomyocytes that are responsible for the sequential depolarization-repolarization cycle known as the action potential (AP). 1 Although the major currents contributing to the AP have been described (1), additional small currents at specific points of low conductance in the AP cycle are sufficient to induce repolarization abnormalities in isolated cells (2,3) and there-fore arrhythmias in situ. These arrhythmias may contribute to the electrical abnormalities that lead to sudden death after myocardial infarction, which persists as the number one cause of death in the United States. We have focused on one channel that has been proposed to contribute to cardiac arrhythmias, TASK-1, a member of the recently described family of two-pore domain potassium channels (4).
The two-pore domain K channel family is composed of at least 15 different members. These channels are widely distributed in excitable tissue, primarily in the brain and heart, and in general are responsive to environmental cues such as temperature, pH, and stretch (5,6). Several are also regulated by lipids such as arachidonic acid or platelet-activating factor (PAF) (7)(8)(9). PAF is an inflammatory phospholipid that has been linked to arrhythmogensis in isolated canine ventricular myocytes (10). We have recently shown that PAF regulates the TASK-1 channel and determined that the arrhythmogenic effect of the stable PAF analog, carbamyl-platelet-activating factor (C-PAF) in mouse cardiomyocytes, is due to the inhibition of TASK-1 current in a protein kinase C (PKC)-dependent manner (2).
In this study, we elucidate the molecular mechanism of the C-PAF effect on TASK-1 current by identifying the ⑀ isoform of PKC (PKC⑀) as a critical component in PAFR signaling. In addition, using site-directed mutagenesis, we have tentatively identified the critical residue that is the target for PKC in the murine channel.

EXPERIMENTAL PROCEDURES
Myocyte Preparation-Mouse ventricular myocytes were isolated using a retrograde coronary perfusion method published previously (11). All the experiments were carried out according to the guidelines issued by the IACUC of Columbia University. Adult mice, 2-3 months old, were anesthetized with a xylazine and ketamine mix and heparinized, the heart was quickly removed, and the ascending aorta was connected to the outlet of a Langendorff column and perfused with 20 -25 ml of a buffer solution (37°C) containing (in mM): NaCl, 112; KCl, 5.4; NaHCO 3 , 4.2; MgCl 2 , 1.6; HEPES, 20; glucose, 5.4; NaH 2 PO 4 , 1.7; taurine, 10; L-glutamine, 4.1; minimum essential medium amino acids solution, 2%; minimum essential medium vitamin solution, 1%; adjusted to pH 7.4, and equilibrated with 100% O 2 . Next, the heart was perfused with an enzyme solution containing collagenase (0.2 mg/ml; Worthington Type II) and trypsin (0.04 mg/ml) at 35°C for 10 -12 min. After this perfusion, the atria were removed, and the ventricles were minced and transferred to a 50-ml flask with an enzyme solution containing collagenase (0.45 mg/ml), trypsin (0.08 mg/ml), Ca 2ϩ (0.75 mM), and bovine serum albumin (4.8 mg/ml). The flask was shaken vigorously for 5-10 min at 32°C before the supernatant was removed and the cells were collected by centrifugation; this operation was repeated two or three times, and additional disaggregated cells were collected. After centrifugation, the myocytes were resuspended in the buffer solution containing Ca 2ϩ (0.75 mM) and bovine serum albumin and stored at room temperature until use. Rod-shaped, Ca 2ϩ -tolerant myocytes, obtained with this procedure, were used within 6 h of dissociation. Measurements were performed only on quiescent myocytes with clear striations.
Cell Culture and Transfection-Chinese hamster ovary cells (CHO) were grown in F-12 medium supplemented with 10% fetal bovine serum. Twenty-four hours prior to transfection, cells were seeded into 6-well plates at 80 -90% of confluence. Transfections were carried out with the GeneJammer transfection reagent (Stratagene) according to the manufacturer's instructions. Briefly, cells were washed with phosphate-buffered saline, and their medium was replaced with supplemented F-12 medium (900 l/well). For each well, GeneJammer (6 l) was incubated with Opti-MEM (90 l) followed by the addition of DNA (1 g). This mixture was then added to the wells, and 3 h later, an additional 2 ml of supplemented F-12 medium was added. After incubating overnight, the cells were washed, and their medium was replaced.
Cells were either co-transfected with pCMV-TASK1 together with pEGFP-C1 (1 g total, 3:1) or transfected with pTIE or T381A-pTIE (1 g). 48 h after the transfection, the cells were checked under the microscope for green fluorescence. Approximately 20% of the cells were positive for EGFP, and these were then used for patch clamp experiments. Due to the culture-to-culture variability in the expression of TASK-1 current, most comparisons were made on matched controls from the same transfection. Summary results were then obtained by pooling data from several different culture preparations.
Solutions and Recording Apparatus-The myocyte suspension or the coverslip with CHO cells was placed into a perfusion chamber, mounted on the stage of an inverted microscope. Unless otherwise indicated, CHO cells were superfused at room temperature with standard external Tyrode's buffer, containing (in mM): NaCl, 140; KCl, 5.4; CaCl 2 , 1; MgCl 2 , 1; HEPES, 5; glucose, 10; adjusted to pH 7.4. Recordings were begun after the current reached a stable baseline (usually 3-4 min after initial cell rupture). In myocytes, TASK-1 current is small and exists in the presence of numerous larger K ϩ currents. To increase the inward component of TASK-1 current and to block other potassium currents in myocytes, we used a modified high K ϩ external solution (modified Tyrode's) to reduce outward rectification of TASK-1 current. The composition of this solution was (in mM): NaCl, 100; KCl, 50; CaCl 2, 1; MgCl 2 , 1; HEPES, 5; glucose, 10; tetraethylammonium, 1; CsCl, 5; adjusted to pH 7.4. Membrane potential and current were measured in the whole cell configuration using borosilicate glass pipettes with a tip resistance between 3 and 5 megaohms and filled with a pipette solution containing (in mM): aspartic acid, 130; KOH, 146; NaCl, 10; CaCl 2 , 2; EGTA, 5; HEPES, 10; MgATP, 2; pH 7.2. The stock solutions of C-PAF and of the PKC inhibitor, bisindolylmaleimide (BIM-I; Calbiochem), were prepared in water and diluted to the final concentrations in Tyrode's or modified Tyrode's, as appropriate. The PKC activator, phorbol 12-myristate 13-acetate (PMA), was prepared in Me 2 SO and then diluted in Tyrode's. The final Me 2 SO concentration did not exceed 0.1%, and the same concentration was present in the control solution. The peptides, ⑀V1-2 (EAVSLKPT (13)) and ⑀V1-7 (HDAPIGYD (14, 15)), PKC⑀-specific inhibitor and activator, respectively, and an inactive scrambled peptide (LSETKPAV (13)) were synthesized by the Columbia University Protein Core. Peptides were prepared in water and then diluted in the pipette solution to a final concentration of 100 nM. Myocytes treated with the peptides were monitored continuously beginning immediately after rupture to detect the occurrence of any arrhythmias during dialysis of the peptide. The application of C-PAF to cells treated with the inhibitor peptide was started after the peptide had been permitted to dialyze into the cell (8 -10 min after rupture for CHO or 10 -12 min after rupture for myocytes).
The current and the voltage protocols were generated using Clampex 8.0 software applied by means of an Axopatch 200-B and a Digidata 1200 interface (Axon Instruments). In current clamp mode, for recording action potentials, the signals were filtered at 1 KHz (low pass Bessel filter) and acquired at a sampling rate of 5 KHz. In voltage clamp mode, the current signals were filtered at 1 KHz and acquired at 500 Hz.
Data Analysis and Statistics-Data were analyzed using pCLAMP 8.0 (Axon) and Origin 6.0 (Microcal) and are presented as mean Ϯ S.E. Records have been corrected for the junction potential, which was measured to be Ϫ9.8 mV. Steady state currents were determined by computer calculation of average current over at least 1 min. Unless otherwise stated, current density comparisons were determined at a voltage of ϩ30 mV. Current density changes are expressed as percentage of inhibition in CHO cell experiments in which TASK-1 is essentially the only current and a pretreatment baseline current can be readily recorded. In myocytes, TASK-1 is measured as the drug-sensitive current, and thus, it is not possible to measure a baseline current to normalize the result when studying the effect of C-PAF or PMA on TASK-1. Therefore, changes in this current in myocytes are expressed in absolute values (pA/pF). Fisher's exact test was used to test the significance of frequency data, and Student's t test was used to compare paired or independent data; a value of p Յ 0.05 was considered statistically significant.

C-PAF Inhibition of TASK-1 Current in CHO Cells Requires
Activation of PKC-Untransfected CHO cells have no significant endogenous K ϩ currents (data not shown); thus, all of the current measured in transfected cells was carried by TASK-1. Therefore, we expressed TASK-1 in CHO cells to test the effect of C-PAF (185 nM) on the current in whole cell patch clamp experiments. During a slow ramp protocol (Ϫ110 mV to ϩ30 mV in 6 s), C-PAF rapidly induced a reversible decrease in TASK-1 current that reached steady state within 2 min. When quantified at the maximal current (at ϩ30 mV), this set of cells expressed 68.6 Ϯ 16.4 pA/pF in control solution versus 60.2 Ϯ 14.3 pA/pF in the presence of C-PAF, a 12% decrease in the mean current density ( Fig. 1A; n ϭ 9, p ϭ 0.01). We next tested whether the effect of C-PAF on TASK-1 current was due to PKC activation by perfusing the cells with BIM-I (100 nM), a nonisoform-specific PKC inhibitor for 2 min before applying C-PAF. In the presence of BIM-I, there was no measurable C-PAF-sensitive current (Fig. 1B, n ϭ 12).
To determine whether activation of PKC alone was sufficient to reduce TASK-1 current, we treated CHO cells expressing TASK-1 with a nonspecific activator of PKC, PMA (100 nM). PMA significantly inhibited TASK-1 current in a manner that was similar to the effect of C-PAF ( Fig. 1C; n ϭ 11, p Ͻ 0.01). The specificity of the PMA effect was verified by exposing cells to an inactive PMA analogue, 4␣-phorbol 12-myristate 13-acetate (␣PMA; 100 nM). ␣PMA had no detectable effect on TASK-1 current expressed in CHO cells (Fig. 1D). In all TASK-1-expressing cells tested, the mean control current was 71.8 Ϯ 12.3 pA/pF, whereas in the presence of PMA, the current fell to 59.2 Ϯ 10.1 pA/pF. The PMA inhibition (19.8 Ϯ 2.7%, n ϭ 17) was significantly greater than that of C-PAF (12.1 Ϯ 1.0%, n ϭ 20; p Ͻ 0.01) when measured at the maximum test voltage of ϩ30 mV, and was irreversible.
The Activation of PKC⑀ Decreases TASK-1 Current in CHO Cells-Having shown that the activation of PKC by either C-PAF or PMA was sufficient to cause a decrease of the TASK-1 current, we next asked whether one specific isoform of PKC was responsible for this effect. We initially discounted the role of the classical PKC isoforms since preliminary studies had suggested that the C-PAF effect on TASK-1 was not calciumdependent (data not shown). Given the prominent role of PKC⑀ in cardiac physiology, we tested the ability of a PKC⑀-specific inhibitor peptide to block the drug-induced reduction in TASK-1 current. A scrambled peptide was used as a control (13). The peptides were introduced to the cells by dialysis through the patch pipette at a final concentration of 100 nM, and recordings were initiated 8 -10 min after the rupture of the membrane to allow the peptide to equilibrate in the cell. C-PAF failed to inhibit TASK-1 current in the presence of the PKC⑀ inhibitor peptide (25.6 Ϯ 12.2 pA/pF before C-PAF versus 25.4 Ϯ 12.4 pA/pF after C-PAF, n ϭ 8, not significant; Fig. 2A). On the contrary, in the presence of the scrambled peptide, C-PAF-induced inhibition of TASK-1 (8.4 Ϯ 1.5%, n ϭ 10) did not differ from control trials in the absence of any peptide (data not shown). Similarly, the addition of the PKC⑀ inhibitor peptide to the pipette completely blocked the PMA-sensitive current in CHO cells expressing TASK-1 ( Fig. 2B; 42.4 Ϯ 12.7 pA/pF before PMA versus 41.2 Ϯ 12.3 pA/pF after PMA, n ϭ 10, not significant), whereas the PMA effect was still present with the scrambled peptide (45.1 Ϯ 7.0 pA/pF before PMA versus 36.6 Ϯ 6.2 pA/pF after PMA, n ϭ 11, p Ͻ 0.01). Summary data for C-PAF and PMA are shown in Fig. 2C.
Does C-PAF Inhibition of TASK-1 Current in Ventricular Myocytes Depend upon Activation of PKC⑀?-We next asked whether the C-PAF-sensitive current in murine ventricular myocytes, defined previously as a TASK-1 current (2), was also mediated by activation of PKC⑀. Recordings were done either with the PKC⑀ inhibitor peptide or with the scrambled peptide in the patch pipette while cells were held at Ϫ10 mV. Ten to twelve min after the rupture of the membrane and when the holding current was stable for at least 1 min, C-PAF (185 nM) was superfused over the myocytes. In the presence of the scrambled peptide, C-PAF caused a decrease in outward current that was indistinguishable from the effect of C-PAF in the absence of peptide (a typical trace is shown in Fig. 3A). The effect of C-PAF was absent, however, when the PKC⑀ inhibitor peptide was included in the patch pipette (a typical trace is shown Fig. 3B). Results from numerous trials showed that the inhibitor peptide significantly inhibited the ability of C-PAF to reduce TASK-1 current in isolated mouse ventricular myocytes, whereas the scrambled peptide had no effect (Fig. 3C).
To further verify that the C-PAF-sensitive current identified in voltage clamp studies was carried by the TASK-1 channel, the I-V relation in myocytes was studied with a slow ramp protocol (Ϫ50 mV to ϩ30 mV over 6 s) in the presence of modified Tyrode's. These conditions minimize the contamination of the TASK-1 current by other K ϩ currents and should allow the calculation of the C-PAF-sensitive current over a wide voltage range by minimizing the outward rectification. To confirm this, we first examined the expressed TASK-1 current in CHO cells in modified Tyrode's. As expected, the I-V relation was markedly less rectifying (data not shown), and the reversal potential was less negative (Ϫ24.4 Ϯ 1.5 mV, as compared with a calculated value of Ϫ27.5 mV in modified Tyrode's for a K ϩ -selective current). The C-PAF inhibition in the presence of elevated K ϩ (10.2 Ϯ 1.8% inhibition, n ϭ 16) was indistinguishable from the previously reported effect of the lipid on TASK-1 in CHO cells recorded in normal Tyrode's (p ϭ 0.33).
In modified Tyrode's solution, myocytes exposed to the scrambled peptide in the patch pipette had a significant decrease in net current in response to C-PAF (a typical cell is shown in Fig. 4A1; n ϭ 8; p Ͻ 0.01) that was essentially identical to the effect measured in the absence of peptide in the pipette (data not shown). Typical of TASK-1 in high K ϩ , the  7). All recordings were made in whole cell configuration using a ramp protocol (Ϫ110 to ϩ30 mV over 6 s) in normal Tyrode's solution at pH 8 and corrected for the junction potential. Drugs were applied when the current was stable for at least 1 min and perfused for 2 min for C-PAF or 6 min for PMA. The drug-sensitive current was measured as the difference between the mean current at steady state (averaged from four successive ramps) in control and in the presence of the drug. The drug-sensitive currents were normalized by cell capacitance and expressed as current density (pA/pF).  n ϭ 11, open symbols). The percentage of inhibition in each case was measured at ϩ30 mV by comparison of each cell before and after drug (C). Both C-PAF and PMA significantly inhibit TASK-1 current in the presence of the scrambled peptide (*, p Ͻ 0.05, t test, comparing control with drug-treated in the presence of scrambled peptide). Neither C-PAF nor PMA had a significant effect on the current in the presence of the inhibitor peptide (not significant versus control), and the effect of both drugs on TASK-1 current was significantly reduced by the inhibitor peptide (**, p Ͻ 0.05, t test, comparing drug in the presence of scrambled peptide with drug in the presence of inhibitor peptide). All the recordings started 8 -10 min after the rupture of the membrane, and the drugs were applied after the current was stable for at least 1 min. Drug treatment and calculation of the drug-sensitive currents were done as described in the legend for Fig. 1. C-PAF-sensitive current is nearly linear and has a reversal potential of Ϫ26.1 Ϯ 1.9 mV (Fig. 4A2). In the presence of the inhibitor peptide, however, the C-PAF had virtually no effect on net current (Fig. 4B1), and the C-PAF-sensitive current was abolished (Fig. 4B2), indicating that PKC⑀ also plays a crucial role in the regulation of TASK-1 current by PAFR in myocytes. Summary data are shown in Fig. 4C.

Does PKC⑀ Play a Role in C-PAF-induced Repolarization Abnormalities in Isolated Myocytes
?-We previously showed that C-PAF induced abnormal automaticity in paced ventricular mouse myocytes and elicited spontaneous activity in quiescent myocytes (2). Now, we asked whether this abnormal automaticity could be due to PKC⑀ activation. To test this, action potential recordings were done on mouse ventricular myocytes paced at 1 Hz with either the PKC⑀-specific inhibitor peptide or an inactive scrambled peptide in the pipette (100 nM). Action potentials were continuously monitored, from the rupture of the membrane until the end of the protocol. C-PAF was applied 10 -12 min after the rupture. When the scrambled peptide was in the pipette, C-PAF induced abnormalities during repolarization in 14 of 19 cells (Fig. 5A; not different from the response of cells treated with C-PAF in the absence of any peptide). In contrast, C-PAF failed to induce repolarization abnormalities in any of the eight cells that were exposed to the PKC⑀-specific inhibitor peptide (Fig. 5B). The difference in observed responses was significant (p Ͻ 0.001, Fisher's exact test).
Further confirming that activation of PKC⑀ is sufficient to alter the electrical activity of the myocyte, we observed that a specific activator peptide of this kinase included in the patch pipette induced prolongation of repolarization, early after depolarizations, and additional spontaneous beats in eight of nine cells tested in the absence of any added C-PAF. In these trials, recordings were begun immediately after the rupture of the membrane, and abnormal rhythm occurred 5-6 min later. Under similar conditions but with the scrambled peptide in the pipette, abnormal automaticity was observed in only 2 of 10 cells tested (Fig. 6; p Ͻ 0.006; Fisher's exact test).
An analysis of the murine TASK-1 sequence revealed a single PKC consensus site that included threonine (residue 381) as the kinase target. Therefore, we constructed a site-directed mutant at this site, converting Thr-381 to alanine. The mutant construct, named T381A-pTIE, was expressed in CHO cells, and when tested by our typical ramp protocol, it demonstrated activity that was comparable with the wild-type channel. However, the mutant channel was no longer sensitive to C-PAF inhibition (maximal current recorded at ϩ30 mV in the absence of C-PAF was 45.5 Ϯ 7 pA/pF versus the current in the presence of C-PAF, 44.2 Ϯ 7 pA/pF; n ϭ 10; not significant, Fig. 7). Similar results were obtained when mutant TASK-1 current was tested in the presence of PMA (Fig. 7C, right). DISCUSSION Previous studies have shown that the abnormalities of repolarization induced by PAF in ventricular myocytes are due to alterations of the background potassium current carried by TASK-1 (2). Shortly after the channel was cloned, heterologous expression studies showed that TASK-1 was inhibited by PMA and that the inhibition could be blocked by BIM-I (16), suggesting a role for PKC in the regulation of channel function. We now show that both overexpressed and native TASK-1 are inhibited by activation of the PAFR and that this inhibition is dependent upon the activation of the ⑀ isoform of PKC. The activation of PKC⑀ is not only necessary but also sufficient to alter repolarization in isolated myocytes. This sufficiency is evident both by the ability of PMA to inhibit TASK-1 current in CHO cells and by the ability of a PKC⑀ activator peptide to induce abnormal automaticity in myocytes in the absence of added PAF. The results obtained when the TASK-1 channel is overexpressed in a heterologous system support the myocyte data by confirming that PAF inhibits TASK-1 in a PKC⑀-dependent manner. Furthermore, in the heterologous system, PKC⑀ appears to be the only PKC isoform involved in the regulation of murine TASK-1 since blocking PKC⑀ is sufficient to fully block the PMA effect on the channel.
Murine TASK-1 has a single consensus PKC site, which is threonine-381, a residue in the C-terminal cytoplasmic tail. Using site-directed mutagenesis, we mutated this site, replacing threonine with the nonphosphorylatable residue, alanine. The T381A mutant expresses normally in CHO cells but is not inhibited by the addition of C-PAF, nor is it sensitive to PMA treatment. Although we have yet to show direct biochemical evidence of channel phosphorylation at this site, the mutagen-

FIG. 3. The C-PAF-dependent inhibition of TASK-1 current in mouse ventricular myocytes requires activation of PKC⑀, steady-state current measurement.
A, in voltage clamp, myocytes were held at Ϫ10 mV, dialyzed with scrambled peptide, and superfused with C-PAF (185 nM) for 2 min. This treatment causes an inhibition of an outward K ϩselective current identified previously as TASK-1 (2). B, in the presence of the PKC⑀ inhibitor peptide (100 nM in the pipette solution), C-PAF was unable to affect the current. C, the C-PAF-sensitive current was not different from zero (*, p Ͻ 0.05, comparing the C-PAF-sensitive current in the presence of inhibitor peptide, n ϭ 4, with no peptide, n ϭ 25, or scrambled peptide, n ϭ 4). In the typical traces shown in A and B, the baseline outward holding current was adjusted to zero to illustrate the C-PAF-sensitive current. The holding current in A and B was 125 and 76 pA, respectively. The recordings started 10 -12 min after the rupture of the membrane. C-PAF was applied after the current was stable for at least 1 min. esis studies certainly allow the recognition of Thr-381 as a critical residue in the PKC-dependent regulation of murine TASK-1 and are supportive of the hypothesis that this site is phosphorylated by PKC⑀, resulting in regulation of the channel. Although human TASK-1 is 83% identical to the murine channel, the PKC site is not in a region that is highly conserved. In fact, the cytoplasmic tail of human TASK-1 contains two putative PKC consensus sequences. It will be interesting to compare the regulation of the human TASK-1 with the murine homologue and determine its sensitivity to regulation by PAF and PKC.
In addition to TASK-1, several other two-pore domain channels are regulated by kinase activity although the molecular mechanisms that underlie the regulation are not entirely clear. For example, TREK-1 (9) and its putative invertebrate homologue, the Aplysia S-K ϩ channel (17), are inhibited by a cyclic AMP-dependent protein kinase phosphorylation in the C-terminal cytoplasmic tail (18,19). In both channels, the effect is due to a change in the open probability of the channel. Human TWIK-1 and TWIK-2 are activated by the application of PMA when expressed in oocytes (20,21). There does not appear to be any change in the single channel conductance. Rather, PMA appears to recruit previously silent channels within the cellattached patch. In this case, however, there is no direct evidence of TWIK channel phosphorylation, and thus, the possibility that the altered channel function may be mediated by kinase action on a second protein cannot be discounted.
Single channel studies of the Drosophila two-pore domain channel, Kcnk0, have described three gating states: one open and two closed. The two closed states are typified by either short or long intraburst closures. When the channel is phosphorylated, the open probability of the channel increases due to a decrease in the frequency and duration of the long-lasting closed state, resulting in an increase in the total current (22).
Thus, kinase-dependent modulation of two-pore domain channels is generally associated with altered open probability rather than a change in single channel conductance. In the case of TASK-1, four gating states have been proposed: two open (one principal and one substate with different conductance) and two closed (16,23). By analogy to other two-pore domain channels, phosphorylation of murine TASK-1 at Thr-381 might decrease the total current by favoring gating of the substate relative to the principal conductance state, decreasing mean open time, or increasing mean closed time. Single channel studies will be needed to reach a clear conclusion on this mechanism. Nevertheless, it does seem clear that channel regulation through activation of PKC⑀ differs fundamentally from inhibition induced by methanandamide since neither PMA nor PAF reduces the current more than 20%, whereas methanandamide inhibition typically reaches ϳ60% (2).
The role of PKC⑀ in cardiac function is complicated by observations that this isoform can mediate the cardioprotective events of ischemic preconditioning (Ref. 24, and reviewed in FIG. 4. The C-PAF-dependent inhibition of TASK-1 current in mouse ventricular myocytes requires activation of PKC⑀, current-voltage relation. C-PAF-sensitive current was recorded in whole cell configuration using a ramp protocol (Ϫ50 to ϩ30 mV over 6 s) in modified Tyrode's solution. The recordings started 10 -12 min after the rupture of the membrane, and C-PAF (185 nM) was applied for 2 min after the current was stable for at least 1 min. C-PAF-sensitive current was obtained as the difference between the mean current (average of four successive ramps) at steady state in control and in the presence of C-PAF; the current was normalized by the capacitance of the cell and expressed as current density (pA/pF). A1 depicts the net current from a typical cell before and after C-PAF treatment in the presence of scrambled peptide. A2 depicts the mean C-PAF-sensitive current recorded from myocytes in the presence of scrambled peptide (100 nM in the pipette; n ϭ 8). B1 depicts the net current from a typical cell before and after C-PAF treatment in the presence of inhibitor peptide. B2 illustrates that in the presence of the inhibitor peptide, the mean C-PAFsensitive current was abolished (100 nM in the pipette, n ϭ 7; *, p Ͻ 0.05). The mean C-PAF-sensitive current quantified at ϩ30 mV is summarized in C.
FIG. 5. The inhibition of PKC⑀ prevents repolarization abnormalities in paced mouse ventricular myocytes exposed to C-PAF. Action potentials were recorded in current clamp mode from myocytes paced at 1 Hz in regular Tyrode's solution. With no peptide in the pipette, perfusion with C-PAF for 2 min induced repolarization abnormalities in five of seven cells (data not shown), which was similar to the result with the scrambled peptide in the pipette in which 14 of 19 cells exhibited repolarization abnormalities during C-PAF perfusion (A shows the record from a typical cell). In the presence of the inhibitor peptide, the effect of C-PAF was completely absent (B shows a cell typical of eight studied). Specific areas of interest are expanded to the right of the record as indicated from control pacing (hexagon) or during C-PAF application (star). The recordings were started 10 -12 min after rupture of the membrane. The heavy horizontal line indicates 0 mV in each case.
Ref. 25), and under other conditions, plays a lead role in the development of hypertrophy and failure (26). Some of the explanations for these dichotomous results may lie in the variability of the level of expression of the kinase and in the subsequent control of its subcellular localization and formation of signaling complexes. For example, it has been shown that PKC⑀ localizes in complexes at mitochondrial membranes after brief repeated episodes of ischemia. Could this sequester enough of the kinase to prevent its association with TASK-1 in the plasma membrane and thereby prevent the arrhythmogenic reduction in this background K ϩ current? Pharmacological antagonism of the PAFR or ischemic preconditioning are both able to significantly reduce the occurrence of ventricular ectopic beats after coronary occlusion (27) but likely work by different mechanisms. The effect of the PAFR antagonist is consistent with the known sequence of events that includes cardiac generation of PAF during ischemia, leading to inhibition of TASK-1 via a PKC⑀-dependent pathway and subsequent generation of abnormal repolarization in ventricular myocytes. This pathway may not occur after preconditioning if the repeated ischemic events lead to movement of PKC⑀ away from the site where it may interact with TASK-1.
We have previously noted the transient nature of the C-PAF induced current in isolated myocytes (2). This is also evident in Fig. 3 and is presumably due to desensitization of the signaling cascade. It is not known whether the response is equally transient in the in situ heart. However, even a transient repolarization abnormality, if induced on the appropriate myocardial substrate as might be found in a diseased heart, could initiate a sustained arrhythmic event. In this regard, the outward rectifying nature of the TASK-1 I-V relation makes it particularly relevant to the plateau phase of the action potential. The plateau represents a period of high membrane resistance in which even small currents can exert a significant effect. It is well recognized that reduction in net outward current during the action potential plateau can lead to action potential prolongation and subsequent arrhythmias through the activation of other currents (3). Further, in the setting of cardiac disease, down-regulation of outward K ϩ currents can result in the reduction of "repolarization reserve" (28) such that even a small further decrease in net outward current can lead to marked action potential prolongation and arrhythmogenesis. In our experiments, it is likely that there is a progressive inhibition of TASK-1 current either by C-PAF or by the activator peptide activating PKC⑀. However, due to the repolarization reserve, a marked failure of repolarization and subsequent arrhythmias does not occur until the current is reduced beyond a critical threshold level. This accounts for the delay in the onset of arrhythmias during C-PAF superfusion and suggests that PAF-induced inhibition of TASK-1 current is likely to be particularly arrhythmogenic in the context of cardiac disease, in which other K ϩ currents are already compromised.
FIG. 6. The activation of PKC⑀ mimics the effect of C-PAF to induce repolarization abnormalities during the action potential in mouse ventricular myocytes. APs were recorded in current clamp mode from myocytes paced at 1 Hz in regular Tyrode's solution. When a scrambled peptide was included in the pipette, only 2 of 10 cells showed repolarization abnormalities (a typical recording is shown in A). In contrast, the presence of the PKC⑀-specific activator peptide alone, without the perfusion of C-PAF, was able to induce early after depolarizations and abnormalities during the repolarization of the AP in eight of nine cells tested (a typical recording is shown in B). Specific areas of interest are expanded to the right of the record as indicated from control pacing (hexagon) or during the effect of the peptide (star). The recordings were started immediately after rupture of the membrane. The heavy horizontal line indicates 0 mV in each case.

FIG. 7.
Mutation of threonine 381 removes the sensitivity of murine TASK-1 to C-PAF and PMA when the channel is expressed in CHO cells. A TASK-1 mutant in which Thr-381 was converted to alanine (T381A) was generated and expressed in CHO cells and compared with the wild-type (WT) channel. The C-PAF-sensitive current was obtained in Tyrode's at pH 8 using a ramp protocol in whole cell configuration. The mutant channel displayed normal current (in amplitude, sensitivity to pH, reversal potential, and shape), but C-PAF (185 nM) did not inhibit the current (A; n ϭ 10). In each experiment, cells transfected with the wild-type channel were used as control for current and C-PAF effect (B; n ϭ 11). The drug-sensitive currents are calculated as the difference between mean current (average of four successive ramps) at steady state in control and in the presence of C-PAF or PMA as noted. C-PAF was applied for 2 min after the current was stable for at least 1 min. PMA was applied for 6 min after the current was stable for at least 1 min. The current was normalized by cell capacitance and expressed as current density (pA/pF). The percentage of control TASK-1 current was calculated, and the data were summarized (C; *, p Ͻ 0.05).