Phosphatidylinositol 4,5-Bisphosphate Is Acting as a Signal Molecule in α1-Adrenergic Pathway via the Modulation of Acetylcholine-activated K+ Channels in Mouse Atrial Myocytes*

We have investigated the effect of α1-adrenergic agonist phenylephrine (PE) on acetylcholine-activated K+ currents (I KACh). I KACh was recorded in mouse atrial myocytes using the patch clamp technique.I KACh was activated by 10 μm ACh and the current decreased by 44.27 ± 2.38% (n = 12) during 4 min due to ACh-induced desensitization. When PE was applied with ACh, the extent of desensitization was markedly increased to 69.34 ± 2.22% (n = 9), indicating the presence of PE-induced desensitization. I KAChwas fully recovered from desensitization after a 6-min washout. PE-induced desensitization of I KACh was not affected by protein kinase C inhibitor, calphostin C, but abolished by phospholipase C (PLC) inhibitor, neomycin. When phophatidylinositol 4,5-bisphosphate (PIP2) replenishment was blocked by wortmannin (an inhibitor of phophatidylinositol 3-kinase and phophatidylinositol 4-kinase), desensitization ofI KACh in the presence of PE was further increased (97.25 ± 7.63%, n = 6). Furthermore, the recovery from PE-induced desensitization was inhibited, and the amplitude of I KACh at the second exposure after washout was reduced to 19.65 ± 2.61% (n = 6) of the preceding level. These data suggest that the KAChchannel is modulated by PE through PLC stimulation and depletion of PIP2.

Phosphatidylinositol 4,5-bisphosphate (PIP 2 ) 1 is well known as a central molecule in the phosphoinositide cycle, by serving as the precursor of important signaling molecules such as inositol trisphosphate (IP 3 ), diacylglycerol, or phosphatidylinositol 3,4,5-trisphosphate. Recently, it was shown that PIP 2 is not just a precursor, but also exerts a direct role in the regulation of various ion transporters, such as Na ϩ /Ca 2ϩ exchanger (1), IP 3 receptor Ca 2ϩ channel (2), Na ϩ -activated nonselective cation channel (3), and several inwardly rectifying K ϩ channels including ROMK1 (4,5), IRK1 (4, 6), K ATP channels (1,4,(7)(8)(9), and G protein-gated inward rectifying K ϩ (GIRK) channels (4,10). The underlying mechanism of PIP 2 action was investigated in detail for GIRK, and it was shown that the activation of GIRK channels by G␤␥ depends on the presence of PIP 2 (4, 6, 10 -12). This result may imply that PIP 2 is a final regulator molecule for GIRK channel activity and that G␤␥ exerts its effect by strengthening the interaction of the channel with PIP 2 .
The K ACh channel in cardiac myocytes is believed to be heterotetrameric complex formed by GIRK1 and GIRK4. Slowing of heart rate by vagal stimulation is known to be mediated by the activation of K ACh channels, which leads to hyperpolarization of membrane potentials in pacemaking cells and in atrial myocytes (13)(14)(15)(16)(17). It is also well known that the effect of vagal stimulation fades gradually (vagal escape), due to desensitization of K ACh currents (18 -22). Molecular mechanisms of K ACh activation and desensitization have been subjects of intense researches for more than a decade. However, it is not yet clear whether the recent view on the mechanism of GIRK channel regulation is valid for native K ACh channels. It has been generally believed that direct binding of G␤␥ to the channel results in activation (23,24), but the role of PIP 2 in this process is still controversial. In rat atrial myocytes, exogenously applied PIP 2 and other phospholipid were reported to block agonist-mediated K ACh channel activation (25).
The aim of the present study is to investigate the role of PIP 2 in normal signaling pathway for the regulation of native K ACh channels. Considering that PIP 2 content in the membrane can be controlled by the PLC-linked receptor (26,27), we used mouse atrial myocytes that possess both PLC-linked receptors, such as ␣ 1 -adrenergic receptor and K ACh channels. The results of the present study show that the ␣ 1 -adrenergic agonist accelerates desensitization of the K ACh channel, possibly through the depletion of the PIP 2 pool in plasma membrane, supporting the hypothesis that PIP 2 is acting as a signal molecule in the regulation of K ACh channels through ␣ 1 -adrenergic pathway.

EXPERIMNTAL PROCEDURES
Cell Isolation-The isolation of single atrial myocytes from mice was performed as described by Harrison et al. (28) with minor modifications. Mice were killed by cervical dislocation, and the heart was quickly removed. The heart was cannulated by a 24-gauge needle and then retrogradely perfused via the aorta on a Langendorff apparatus. During coronary perfusion all perfusates were maintained at 37°C and equilibrated with 100% O 2 . Initially the heart was perfused with normal Tyrode solution for 2-3 min to clear the blood. The heart was then perfused with Ca 2ϩ -free solution for 2 min. Finally the heart was perfused with enzyme solution for 14 -16 min. Enzyme solution contains 0.14 mg ml Ϫ1 collagenase (Sigma Type 5) in Ca 2ϩ -free solution. After perfusion with enzyme solution, the atria were separated from the ventricles and chopped into small pieces. Single cells were dissociated in high K ϩ , low Cl Ϫ storage medium from these small pieces using blunt-tip glass pipette. Cells were stored at 4°C until use. , phenylephrine (Sigma), and neomycin (Biomol) were dissolved in deionized water to make a stock solution and stored at Ϫ20°C. On the day of experiments one aliquot was thawed and used. Calphostin C (Biomol) and wortmannin (WMN; Biomol) were first dissolved in dimethyl sulfoxide (Me 2 SO) as a stock solution and then used at the final concentration in the solution. All experiments were conducted at 35 Ϯ 1°C. In the presence of ACh, 10 M glibenclamide was applied to inhibit the ATP-sensitive K ϩ channel. Cells were superfused with solution by gravity at ϳ5 ml/min. Approximately 30 s were required to change completely the bath contents.

Materials and Solutions-Normal
Voltage Clamp Recording and Analysis-Membrane currents were recorded in nystatin-perforated patch configuration using an Axopatch-1C amplifier (Axon Instruments). Nystatin forms voltage-insensitive ion pores in the membrane patch that are somewhat selective for cations over anions but are impermeant to Ca 2ϩ and other multivalent ions or molecules Ͼ0.8 nm in diameter (29). This method, therefore, minimizes dialysis of intracellular constituents with the internal pipette solution. Nystatin was dissolved in Me 2 SO at a concentration of 50 mg/ml and then added to the internal pipette solution to yield a final nystatin concentration of 200 g/ml. The patch pipettes were pulled from borosilicate capillaries (Clark Electromedical Instruments, Pangbourne, United Kingdom) using a Narishige puller (PP-83; Narishige, Tokyo, Japan). We used patch pipettes with a resistance of 2-3 megaohms when filled with the above pipette solutions. The electrical signals were displayed during the experiments on an oscilloscope (Tektronix, TDS 210) and a chart recorder (Gould). Data were digitized with pClamp software 5.7.1 (Axon Instruments) at a sampling rate of 1-2 kHz and filtered at 5 kHz. Voltage clamp and data acquisitions were performed by a digital interface (Digidata 1200, Axon Instruments) coupled to an IBM-compatible computer using pClamp software 5.7.1 (Axon Instruments) at a sampling rate of 1-2 kHz and filtered at 5 kHz.
Statistics and Presentation of Data-The results in the text and in the figures are presented as means Ϯ S.E., n ϭ number of cells tested. Statistical analyses were performed using the Student's t test. The difference between two groups was considered to be significant when p Ͻ 0.01 and not significant when p Ͼ 0.05.

RESULTS
Activation and Desensitization of I KACh -Acetylcholine-activated K ϩ current (I KACh ) was activated by adding 10 M acetylcholine (ACh) to the bath solution, while the cell was volt-age-clamped at the holding potential of Ϫ40 mV (Fig. 1A). Upon the application of ACh, a rapid increase in outward current was observed. Despite the continuous presence of ACh, the activation of I KACh was not sustained at its peak, but the amplitude of I KACh decreased slowly. We regarded this decrease as a result of ACh-induced desensitization of I KACh . The ACh-induced desensitization was recovered after washout of ACh, so that the amplitude of I KACh at the second exposure to ACh in a 6-min interval was similar to that at the first exposure. In subsequent experiments, such a paired application of ACh was used for investigating the effect of various signal molecules on regulation of K ACh channel, regarding the I KACh at the first response as the control.
Characteristics of I KACh activation and desensitization were further investigated from the current-voltage (I-V) curves. I-V curves were obtained from the current response induced by voltage ramps between ϩ60 and Ϫ120 mV (at a speed of Ϯ 0.6 V s Ϫ1 ) from the holding potential of Ϫ40 mV. The ramps were applied before ACh application (a), at peak (b), 4 min in ACh (c), and washout of ACh (d), as indicated in Fig. 1A. Corresponding I-V curves were plotted in Fig. 1B: the reversal potentials were shifted slightly to negative potentials toward K ϩ equilibrium potential, and the shape of I-V curves was changed by ACh. The degree of inward rectification was less strong in the presence of ACh (b and c) compared with that in control (a). A strong inward rectification in "a" is considered to be typical for inward rectifying K ϩ currents, IRK (30). The I-V curves for net I KACh were obtained by subtracting the control curve (a) from the I-V curves in the presence of ACh, as shown in Fig. 1C: "b-a" represents I KACh at peak (I KACh , peak), and "c-a" represents I KACh at 4 min in ACh (I KACh , 4 min). The shape of inward rectification and the reversal potential of two curves were not different, indicating that the decrease in current amplitude during exposure to ACh occurs uniformly over the voltage range tested.
Desensitization of I KACh Is Accelerated by ␣ 1 -Adrenergic Agonist-In Fig. 2A, 100 M phenylephrine (PE) was applied together at the second exposure to ACh. From the continuous recording of current trace at Ϫ40 mV, it was noticed that the process of desensitization was markedly accelerated by PE, resulting in a greater reduction of I KACh after the same period ( Fig. 2A). But, the amplitude of I KACh at peak was not signifi- cantly affected by PE. We regarded this increased desensitization by PE as PE-induced desensitization. The effect of PE on I KACh desensitization was well illustrated in I-V curves: the difference between I KACh , peak (bЈ-aЈ) and I KACh , 4 min (cЈ-aЈ) in the presence of PE was significantly greater compared with the control (Fig. 2, B and C). But PE did not affect the degree of inward rectification and the reversal potential, indicating that PE modulates I KACh itself, rather than modulating other current systems. PE-induced desensitization was also reversible after a 6-min recovery period, and the third exposure to ACh elicited an outward current with a similar peak amplitude and desensitization (Fig. 2A).
The data were summarized in Fig. 2D. The amplitude of I KACh was measured at Ϫ40 mV to minimize the possible contamination of IRK and voltage-activated K ϩ currents. The amplitude of I KACh at peak was not significantly different between control (775.90 Ϯ 90.29 pA, n ϭ 12) and PE (644.96 Ϯ 94.21 pA, n ϭ 9). However, the amplitude of I KACh at 4 min in ACh was significantly smaller in PE, indicating that desensitization was increased by PE. When the extent of desensitization was determined as a proportion of the current decrease during 4 min, it was 44.27 Ϯ 2.38% (n ϭ 12) in control and increased significantly to 69.34 Ϯ 2.22% (n ϭ 9) by PE.
PLC, but Not PKC, Is Involved in the Increased Desensitization by PE-To elucidate the mechanisms for PE-induced desensitization, we blocked each step of the signal transduction pathway related with PE. When PKC inhibitor, calphostin C (2.5 M), was pretreated before the application of PE at the second exposure to ACh, acceleration of I KACh desensitization by PE was still observed (Fig. 3A). To focus on the change in desensitization, I-V curves for desensitized current were plotted in Fig. 3B. It was noticed that desensitized current in the presence of PE and calphostin C (bЈ-cЈ) was almost completely overlapped by that in the presence of PE only (b-c). The extent of desensitization in the presence of PE and calphostin C was 79.20 Ϯ 4.46% (n ϭ 4), indicating no significant difference from that in the presence of PE only (69.34 Ϯ 2.22%). Furthermore phorbol 12-myristate 13-acetate (100 nM), a specific PKC activator, did not mimic the effect of PE on the desensitization of I KACh (n ϭ 5, data not shown). The amplitude of I KACh at peak was affected neither by calphostin C nor by phorbol 12-myristate 13-acetate.
We then tested the effect of PLC inhibitor, neomycin. When neomycin (500 M) was pretreated before the application of PE at the second exposure to ACh, the increase in desensitization by PE was no longer observed (Fig. 3C). This finding suggests that PE-induced desensitization of I KACh is antagonized by neomycin. The I-V curve for desensitized current in the presence of neomycin and PE (bЈ-cЈ) was significantly smaller than that in the presence of PE (b-a). The extent of desensitization in the presence of PE and neomycin was 48.79 Ϯ 3.95% (n ϭ 9), showing a significant difference from the extent of desensitization in the presence of PE (69.34 Ϯ 2.22%), but not different from the control (44.27 Ϯ 2.38%). Neomycin itself did not significantly affect the activation of I KACh and ACh-induced desensitization (I KACh, peak : 759.20 Ϯ 93.43 pA; desensitization: 48.65 Ϯ 3.76%, n ϭ 7).
The extent of desensitization obtained in various conditions was summarized in Fig. 3E. Based on these results, it is suggested that PE-induced desensitization of I KACh is the result of the activation of PLC, but not through the activation of PKC.
Effect of the Depletion of PIP 2 Pool by Wortmannin on I KACh -We postulated that PLC involvement is related with PIP 2 , and this possibility was tested by using WMN (an inhibitor of PI 3-kinase and PI 4-kinase). It was reported that WMN inhibits the replenishment of PIP 2 after the depletion of PIP 2 by the receptor-mediated activation of PLC (27). Therefore, we examined whether the inhibition of PIP 2 replenishment by WMN affects PE-induced desensitization of K ACh current and its recovery. In Fig. 4A, we first applied ACh and PE simultaneously to induce the increased desensitization of I KACh and confirmed the full recovery of I KACh from desensitization after a 6-min washout with normal Tyrode solution. Then the same series of experiment was performed in the presence of 100 M WMN (Fig. 4B). PE-induced desensitization of I KACh was greatly accelerated by WMN, and I KACh, 4 min became very smaller. The extent of desensitization in the presence of WMN was 97.25 Ϯ 7.63% (n ϭ 6), and this value was significantly greater than 64.22 Ϯ 5.70% (n ϭ 7) in the absence of WMN. In contrast, WMN alone without PE did not significantly affect the activation and desensitization of I KACh (I KACh, peak : 603.14 Ϯ 80.63 pA; desensitization: 45.84 Ϯ 4.07%, n ϭ 6, data not shown). These results suggest that blockade of PIP 2 synthesis by WMN facilitated PE-induced desensitization, but not ACh-induced desensitization.
I KACh, peak was not different, but I KACh, peak was significantly smaller in the presence of WMN, indicating that PE-induced desensitization of I KACh was further accelerated by WMN. The extent of desensitization in the presence of WMN was 97.25 Ϯ 7.63% (n ϭ 6), and it was significantly greater than 64.22 Ϯ 5.70% (n ϭ 7) in the absence of WMN. In contrast, WMN alone without PE did not significantly affect the activation and desensitization of I KACh (I KACh, peak : 603.14 Ϯ 80.63 pA; desensitization: 45.84 Ϯ 4.07%, n ϭ 6, data not shown). These results suggest that blockade of PIP 2 synthesis by WMN facilitated PE-induced desensitization, but not ACh-induced desensitization.
The recovery from PE-induced desensitization was also affected by WMN (Fig. 4B). In the continuous presence of WMN, the amplitude of I KACh, peak measured at the second exposure to ACh was only 19.65 Ϯ 2.61% (n ϭ 6) of the preceding level (Fig.  4D), indicating that the recovery from desensitization was significantly inhibited by WMN. WMN also inhibited the recovery from ACh-induced desensitization, but to a lesser extent: I KACh, peak measured at the second exposure to ACh in the absence of PE was 70.92 Ϯ 9.18% of preceding level (n ϭ 6, data not shown). The degree of inhibition is comparable with the magnitude reduction of basal PIP 2 levels in unstimulated cells by WMN in this period time (27), suggesting that the incomplete recovery of ACh-induced desensitization was the result of basal reduction of PIP 2 level by WMN.
Facilitation of PE effect by WMN was further confirmed in Fig. 4C, where PIP 2 pool in the membrane was depleted before the first exposure to ACh by pretreating WMN and PE for 3 min. The amplitude of I KACh, peak was only 178.33 Ϯ 85.36 pA (n ϭ 3), and this value was significantly smaller than I KACh, peak in control (775.90 Ϯ 90.29 pA, n ϭ 12). Above results imply that depletion of PIP 2 pool by PE causes a decrease in I KACh and that the increase in desensitization by PE can be explained by the same mechanism.
Effect of PE on IRK-As well as K ACh channel, IRK is known to be regulated by PIP 2 (4, 6). We examined whether IRK is affected by the various substances used in the present study. IRK was determined from the I-V curve, as described previously in Fig. 1B. The amplitude was measured at Ϫ120 mV, where IRK is large. IRK was not affected either by PE (100 M) or by WMN (100 M). However, the addition of PE in the presence of WMN decreased IRK by 31.41 Ϯ 1.93%. These results suggest that the change of PIP 2 level induced by normal signal molecule such as PE may not contribute to IRK regulation, although IRK is inhibited when PIP 2 level is reduced further down by inhibiting its replenishment. DISCUSSION The main question addressed in the present study is whether PIP 2 is acting as a signal molecule for the regulation of native K ACh channels. The results obtained can be summarized as follows: 1) PE, ␣ 1 -adrenergic agonist, accelerates desensitization of the I KACh ; 2) PE-induced desensitization was inhibited by PLC inhibitor, neomycin, but not by PKC inhibitor, calphostin C; 3) when wortmannin, an inhibitor PI 3-kinase and PI 4-kinase was applied with PE, desensitization of I KACh was further accelerated, and the recovery from desensitization was inhibited. From these results it was suggested that K ACh channels are regulated by ␣ 1 -adrenergic agonist through the depletion of the PIP 2 pool in plasma membrane.
Classical signal molecules produced by the activation of PLClinked receptors are IP 3 and diacylglycerol. The role of IP 3 on ion channels has not been reported in cardiac myocytes, but diacylglycerol may possibly contribute to the regulation of ion channels via PKC. Desensitization of K ACh channel was known to be caused by phosphorylation of m2 muscarinic receptor (31)(32)(33)(34), so it is possible that PKC is involved in the increased desensitization of K ACh channel by PE. We tested this possibility, but calphostin C, a PKC-specific inhibitor did not inhibit PE action (desensitization of 79.20 Ϯ 4.46%, Fig. 3, A and E). Furthermore, the effect of PE was not mimicked by direct pharmacological activation of PKC with phorbol 12-myristate 13-acetate. These results support the idea that the mechanism for PE-induced desensitization was the depletion of PIP 2 rather than the production of PIP 2 metabolites that may inhibit the K ACh channel. Although we did not carry out the biochemical measurements of PIP 2 concentration in the present experiment, the decrease in PIP 2 concentration in the plasma membrane by PLC-linked receptor has been reported previously in Chinese hamster ovary cells and human neuroblastoma cells (26,27). The involvement of PIP 2 in the PE effect was further supported by the finding that the PE-induced desensitization became irreversible when the replenishment of PIP 2 was blocked by WMN (Fig. 4B). Furthermore when the PIP 2 pool was depleted by preincubation of PE and WMN, activation of the K ACh current was reduced (Fig. 4C).
IRK channels, on the other hand, showed a different response. In the case of IRK, PE did not affect the channel activity, although PE with WMN affected the channel activity. These data suggest that interaction of IRK with PIP 2 is stronger than that of K ACh channel, as suggested previously (4,6,35), and that the effect of PIP 2 depletion on IRK occurs at much lower concentration. Another possibility that should be tested in future studies is a co-localization of a specific PLC-linked receptor and a specific ion channel. In this view, the PIP 2 pool, which is regulated by PE, may not be uniformly distributed over the whole membrane, but localized closely with K ACh channels.
It has previously been reported by other studies that several PLC-linked receptor can inhibit K ACh channel. Braun et al. (36) reported that the selective ␣ 1 -adrenergic agonist methoxamine reduced both the IK 1 and K ACh current in rabbit atrial myocytes. Yamaguchi et al. (37) reported that endothelin-1 and endothelin-3 inhibited K ACh current in guinea pig atrial myocytes. Their observation is similar to the effect of PE presented in this paper, but they failed to identify the mechanism of the ␣ 1 -adrenergic agonist or endothelin induced inhibition. They only demonstrated that these effects were not mediated by PKC or IP 3 . But it now seems to be very likely that PIP 2 is involved in those effects. Recently, channel expression studies have demonstrated that PLC-linked receptors inhibit GIRK1/GIRK4 channels (38) or K ATP channels (39), and these effects were mediated by depletion of the PIP 2 pool in membrane.
The functional consequence of accelerated desensitization of I KACh by ␣ 1 -adrenergic receptor may be an early cessation of parasympathetic effect in the continuous presence of ACh. This discovery may be of particular importance, since it provides a novel pathway for sympathetic-parasympathetic interaction. Interaction between sympathetic and parasympathetic system was recognized early in various experimental conditions, and this interaction is also considered to be of clinical importance. However, the precise signal transduction pathways involved in this interaction are not fully understood, except the inhibition of adenylate cyclase by acetylcholine as a mechanism of parasympathetic antagonizism to sympathetic stimulation. To our knowledge, the pathway presented in the present study seems to be the first report of a reciprocal pathway through which sympathetic stimulation can antagonize parasympathetic activity.
In conclusion, ␣ 1 -adrenergic agonist accelerates the desensitization of K ACh channel through the regulation of PIP 2 pool, suggesting that the receptor mediated regulation of the PIP 2 pool may play an important role in the control of cellular function through the modulation of ion channels.