Angiotensin II inhibits bTREK-1 K+ channels in adrenocortical cells by separate Ca2+- and ATP hydrolysis-dependent mechanisms.

Bovine adrenocortical cells express bTREK-1 K+ channels that set the resting membrane potential (V(m)) and couple angiotensin II (AngII) and adrenocorticotropic hormone (ACTH) receptors to membrane depolarization and corticosteroid secretion. In this study, it was discovered that AngII inhibits bTREK-1 by separate Ca2+- and ATP hydrolysis-dependent signaling pathways. When whole cell patch clamp recordings were made with pipette solutions that support activation of both Ca2+- and ATP-dependent pathways, AngII was significantly more potent and effective at inhibiting bTREK-1 and depolarizing adrenal zona fasciculata cells, than when either pathway is activated separately. External ATP also inhibited bTREK-1 through these two pathways, but ACTH displayed no Ca2+-dependent inhibition. AngII-mediated inhibition of bTREK-1 through the novel Ca2+-dependent pathway was blocked by the AT1 receptor antagonist losartan, or by including guanosine-5'-O-(2-thiodiphosphate) in the pipette solution. The Ca2+-dependent inhibition of bTREK-1 by AngII was blunted in the absence of external Ca2+ or by including the phospholipase C antagonist U73122, the inositol 1,4,5-trisphosphate receptor antagonist 2-amino-ethoxydiphenyl borate, or a calmodulin inhibitory peptide in the pipette solution. The activity of unitary bTREK-1 channels in inside-out patches from adrenal zona fasciculata cells was inhibited by application of Ca2+ (5 or 10 microM) to the cytoplasmic membrane surface. The Ca2+ ionophore ionomycin also inhibited bTREK-1 currents through channels expressed in CHO-K1 cells. These results demonstrate that AngII and selected paracrine factors that act through phospholipase C inhibit bTREK-1 in adrenocortical cells through simultaneous activation of separate Ca2+- and ATP hydrolysis-dependent signaling pathways, providing for efficient membrane depolarization. The novel Ca2+-dependent pathway is distinctive in its lack of ATP dependence, and is clearly different from the calmodulin kinase-dependent mechanism by which AngII modulates T-type Ca2+ channels in these cells.

Adrenocortical cells also express losartan-insensitive AT 2 receptors, which comprise about 20% of the AngII receptors in these cells (15,17). The signaling pathways and function of these receptors in the physiology of corticosteroid-secreting cells is unknown.
In whole cell patch clamp recordings from bovine AZF and AZG cells, AngII maximally inhibited bTREK-1 by 72-77% with an IC 50 of ϳ145 pM, provided that ATP was present in the recording pipette at millimolar concentrations (1,11,12). An-gII-mediated inhibition of bTREK-1 was eliminated when ATP in the pipette solution was replaced by the non-hydrolyzable ATP analog AMP-PNP, or UTP (12,21). The kinase or ATPase that mediates AngII inhibition through this ATP hydrolysisdependent pathway has not been identified.
In these whole cell recording studies, [Ca 2ϩ ] i was buffered to 22 nM with 11 mM BAPTA. In a separate study, we found that bTREK-1 in AZF cells was inhibited by raising the [Ca 2ϩ ] i in the pipette solution from 22 nM to 2 M (22). bTREK-1 was also inhibited by superfusing cells with the Ca 2ϩ ionophore ionomycin, provided that [Ca 2ϩ ] i was buffered by 2 mM, rather than 11 mM BAPTA. At the single channel level, in excised inside-out patch recordings, the application of saline containing 35 M Ca 2ϩ to the cytoplasmic face of the membrane markedly inhibited bTREK-1 channel activity. In these experiments, 5 mM ATP was present at the cytoplasmic face of the membrane.
Taken together, the results of these studies could suggest that AngII inhibits bTREK-1 through a single Ca 2ϩ -and ATP hydrolysis-dependent signaling pathway that requires the activation of a Ca 2ϩ -dependent kinase. In bovine AZG cells, AngII enhances the activity of T-type Ca 2ϩ channels through the activation of Ca 2ϩ /calmodulin-dependent kinase II (23)(24)(25). Alternatively, the effective inhibition of bTREK-1 by AngII in whole cell recordings with [Ca 2ϩ ] i strongly buffered to 22 nM using 11 mM BAPTA suggests that AngII may inhibit bTREK-1 by multiple signaling pathways, only one of which is Ca 2ϩ -dependent.
In whole cell and single channel patch clamp recordings from bovine adrenocortical cells, we discovered that AngII does inhibit bTREK-1 through separate Ca 2ϩ and ATP hydrolysis-dependent signaling pathways. Simultaneous activation of both pathways by AngII leads to near complete inhibition of bTREK-1 current and pronounced depolarization of bovine adrenocortical cells. Molecular features of the new Ca 2ϩ -dependent mechanism were identified.
Isolation and Culture of AZF Cells-Bovine adrenal glands were obtained from steers (age 2-3 years) at a local slaughterhouse. Isolated AZF and AZG cells were obtained and prepared as previously described (26). After isolation, cells were either resuspended in Dulbecco's modified Eagle's medium/F-12 (1:1) with 10% fetal bovine serum, 100 units/ml penicillin, 0.1 mg/ml streptomycin, and the antioxidants 1 M tocopherol, 20 nM selenite, and 100 M ascorbic acid (Dulbecco's modified Eagle's medium/F-12ϩ) and plated for immediate use, or resuspended in fetal bovine serum, 5% Me 2 SO, divided into 1-ml aliquots, and stored in liquid nitrogen for future use. For patch clamp experiments, cells were plated in Dulbecco's modified Eagle's medium/F-12ϩ in 35-mm dishes containing 9-mm 2 glass coverslips. Coverslips were treated with fibronectin (10 g/ml) at 37°C for 30 min then rinsed with warm, sterile phosphate-buffered saline immediately before adding cells. Cells were maintained at 37°C in a humidified atmosphere of 95% air, 5% CO 2 .
Cells were used for patch clamping 24 -48 h after transfection with bTREK-1. Transfected cells were plated on 9-mm glass coverslips as described above. Fifteen min before initiating a patch clamp experiment, anti-CD8 antibody-coated beads were added to the culture dish. Upon transferring coverslips to the recording chamber, transfected cells were identified based on decoration with the beads. Whole cell and single channel bTREK-1 currents were recorded from transfected cells as described below for AZF cells.
Patch Clamp Experiments-Patch clamp recordings of K ϩ channel currents were made in the whole cell and inside-out patch configuration from bovine AZF cells. Although the results reported in this study were obtained by recording currents from AZF cells, preliminary recordings from AZG cells showed no difference in bTREK-1 currents with respect to modulation by AngII. For whole cell recordings, the standard external solution consisted of 140 mM NaCl, 5 mM KCl, 2 mM CaCl 2 , 2 mM MgCl 2 , 10 mM HEPES, and 5 mM glucose, with pH adjusted to 7.3 using NaOH. The standard pipette solution consisted of (in mM): 120 KCl, 2 MgCl 2 , 10 HEPES, and 0.2 GTP, with pH titrated to 6.8 using KOH. The buffering capacity of pipette solutions was varied by adding combinations of Ca 2ϩ and BAPTA or EGTA using the Bound and Determined software program (28). Low and high capacity Ca 2ϩ buffering solutions contained 0.5 mM EGTA and 11 mM BAPTA, respectively. The low capacity Ca 2ϩ buffering solution was normally Ca 2ϩ -free. [Ca 2ϩ ] i was buffered to 22 nM in the high capacity buffering solution. Nucleotides, including MgATP, NaUTP, and AMP-PNP were added to pipette or bath solutions as noted in the text. For inside-out patch recordings, the standard external and pipette solutions used for whole cell recordings were switched.
Recording Conditions and Electronics-AZF cells were used for patch clamp experiments 2-12 h after plating. Typically, cells with diameters Ͻ15 m and capacitances of 10 -15 pF were selected. Coverslips were transferred from 35-mm culture dishes to the recording chamber (volume: 1.5 ml) that was continuously perfused by gravity at a rate of 3-5 ml/min. For whole cell recordings, patch electrodes with resistances of 1.0 -2.0 M⍀ were fabricated from Corning 0010 glass (World Precision Instruments, Sarasota, FL). These electrodes routinely yielded access resistances of 1.5 Ϫ4.0 M⍀ and voltage-clamp time constants of Ͻ100 s. For single channel recordings, patch electrodes with higher resistances (3-5 M⍀) were used. K ϩ currents were recorded at room temperature (22-25°C) according to the procedure of Hamill et al. (29) using a List EPC-7 patch clamp amplifier.
Pulse generation and data acquisition were done using a personal computer and PCLAMP software with TL-1 interface (Axon Instruments, Inc., Burlingame, CA). Currents were digitized at 2-10 KHz after filtering with an 8-pole Bessel filter (Frequency Devices, Haverhill, MA). Linear leak and capacity currents were subtracted from current records using summed scaled hyperpolarizing steps of 1 ⁄2 to 1 ⁄4 pulse amplitude. Data were analyzed using PCLAMP (CLAMPFIT 9.2, FETCHAN 6.04, and PSAT 6.04) and SigmaPlot (version 8.0) software. Drugs were applied by bath perfusion, controlled manually by a six-way rotary valve. p values were calculated using Student's t test.
The absence of time-and voltage-dependent bTREK-1 inactivation allows the corresponding membrane current to be easily isolated in whole cell recordings, using either of two voltage clamp protocols. When voltage steps of 300 ms duration are applied from a holding potential of Ϫ80 mV, bTREK-1 can be measured near the end of a voltage step when the transient Kv1.4 current has inactivated (Fig. 1, A-C, left traces). Alternatively, bTREK-1 can be selectively activated by an identical voltage step, after a 10-s prepulse to Ϫ20 mV has fully inactivated Kv1.4 channels (Fig. 1, A-C, right traces). Measurement of bTREK-1 by either method yielded nearly identical results.
In previous whole cell recording studies on AZF cells using patch electrodes containing 2-5 mM MgATP and [Ca 2ϩ ] i strongly buffered to 22 nM with 11 mM BAPTA, AngII inhibited bTREK-1 by a maximum of 77-82% with an IC 50 of 145 pM. However, when MgATP in the pipette was replaced with UTP, which is not a substrate for kinases or ATPases, AngII was ineffective (1,12,21).
These results suggest that AngII-mediated inhibition of bTREK-1 may occur through separate Ca 2ϩ and ATP hydrolysisdependent signaling pathways. If AngII modulates bTREK-1 by parallel mechanisms, each of which produces only partial inhibition, it is likely that simultaneous activation of both pathways would more effectively suppress this current. Accordingly, when bTREK-1 currents were recorded with a pipette solution whose composition supported activation of both ATPand Ca 2ϩ -dependent pathways (0.5 mM EGTA, 5 mM MgATP), AngII (2 nM) was significantly more effective, inhibiting bTREK-1 by 92.9 Ϯ 0.7% (n ϭ 8) (Fig. 1, C and D).
Not only was AngII more effective at inhibiting bTREK-1 when both the Ca 2ϩ and ATP pathways were available, it was also more potent. To compare the potency of AngII as an inhibitor of bTREK-1 through Ca 2ϩ and combined pathways, AZF cells were superfused with AngII at various concentrations and bTREK-1 currents were recorded with pipette solutions containing 0.5 mM EGTA and either 2 nM NaUTP or 5 mM MgATP. Inhibition curves constructed from data obtained in these experiments clearly showed that AngII was more potent, as well as more effective, at inhibiting bTREK-1 when Ca 2ϩ -and ATPdependent pathways were available ( Fig. 2A, left).
With both pathways available for activation, AngII inhibited bTREK-1 almost completely, with an IC 50 of 63 pM. By comparison, with only the Ca 2ϩ pathway available for activation, AngII inhibited bTREK-1 by a maximum of ϳ70%, with an IC 50 of 149 pM ( Fig. 2A). The marked enhancement of bTREK-1 inhibition by AngII when both Ca 2ϩ -and ATP-dependent pathways were activated was most pronounced at lower AngII concentrations. At low AngII concentrations, bTREK-1 inhibition was partially reversible with washing ( Fig. 2A).
Activation of Ca 2ϩ -and ATP-dependent Pathways Correlate with Membrane Depolarization-bTREK-1 channels are largely responsible for setting the resting potential of bovine adrenocortical cells (1,12). In the absence of a significant inward current, the Goldman-Hodgkin-Katz voltage equation predicts that a large fraction of the bTREK-1 channels would have to be inhibited to effectively depolarize these cells. The simultaneous activation of both the Ca 2ϩ -and ATP-dependent pathways should therefore provide a more efficient means of depolarizing adrenocortical cells. Accordingly, in combined current and voltage clamp experiments, AngII inhibited bTREK-1 and strongly depolarized AZF cells only when both the Ca 2ϩ -and ATP-dependent pathways were available for activation.
ATP, but Not ACTH or cAMP, Inhibits bTREK-1 through Both Ca 2ϩ -and ATP-dependent Pathways-Previously, we demonstrated that externally applied ATP and UTP inhibited bTREK-1 channels through activation of a G protein-coupled purinergic receptor with a P2Y 3 agonist profile (32). In that study where [Ca 2ϩ ] i was strongly buffered with 11 mM BAPTA, ATP (100 M) inhibited bTREK-1 by a maximum of 71.3 Ϯ 3.2% and this inhibition was eliminated by substituting AMP-PNP for MgATP in the pipette solution. Thus, similar to AngII, externally applied ATP inhibits bTREK-1 through a ATP hydrolysis-dependent mechanism. P2Y 3 receptors are coupled to PLC activation and the release of [Ca 2ϩ ] i (33,34). To determine whether externally applied ATP also inhibited bTREK-1 through a Ca 2ϩ -dependent mechanism, inhibition was studied with a pipette solution designed to permit activation of the Ca 2ϩ , but not the ATP-dependent pathway. With [Ca 2ϩ ] i weakly buffered by 0.5 mM EGTA, ATP at 10 and 100 M inhibited bTREK-1 by 26.2 Ϯ 9.3 (n ϭ 7) and 42.8 Ϯ 4.4% (n ϭ 6), respectively (Fig. 3, A and D).
When considered in conjunction with the previously mentioned study (32), these results indicate that similar to AngII, external ATP inhibits bTREK-1 through separate Ca 2ϩ -and ATP hydrolysis-dependent mechanisms. Accordingly, when AZF cells were superfused with ATP under conditions where Ca 2ϩ -and ATP-dependent pathways could be activated, bTREK-1 was inhibited by 85.6 Ϯ 5.8% (n ϭ 4) (Fig. 3D).
ACTH also potently inhibits bTREK-1 in bovine AZF cells with an IC 50 of ϳ5 pM (1). Although the signaling pathways are only partially understood, inhibition likely involves both protein kinase A-dependent and -independent actions of cAMP (5,35). Inhibition by ACTH or cAMP is eliminated when AMP-PNP or UTP is substituted for ATP in the pipette solution with [Ca 2ϩ ] i strongly buffered by 11 mM BAPTA (26,35). Experiments were done to determine whether, in addition to the ATP-dependent mechanism, ACTH or cAMP could also inhibit bTREK-1 by a Ca 2ϩ -dependent process.
Because cAMP is the principle intracellular messenger for ACTH, the failure of ACTH to inhibit bTREK-1 in the above experiments could be because of a requirement for cAMP synthesis, previous to inhibition through a Ca 2ϩ -dependent pathway. However, when AZF cells were superfused with the membrane-permeable cAMP analog 8-pcpt-cAMP (250 M) while recording bTREK-1 with pipettes containing 0.5 mM EGTA and 2 mM UTP, no significant inhibition of bTREK-1 was observed (Fig. 3, C and D). Pre-exposure to 8-pcpt-cAMP also did not affect inhibition of bTREK-1 by AngII (Fig. 3D).
Properties of the Ca 2ϩ -dependent Inhibitory Pathway-Results to this point have identified a new Ca 2ϩ -dependent pathway by which hormones and paracrine factors inhibit a specific background K ϩ current and depolarize bovine adrenocortical cells. Because AngII physiologically regulates the secretion of aldosterone and cortisol, it is important to identify the molecular components of the signaling pathway that links AngII receptors to TREK-1 channels and membrane depolarization.
The Ca 2ϩ -dependent Inhibition of bTREK-1 Is Voltage-independent and -specific-Although bTREK-1 K ϩ channels are active at negative membrane potentials, they display weak voltage dependence (35). The Ca 2ϩ -dependent inhibition of bTREK-1 by AngII was independent of voltage over a wide range of test potentials. The experiment illustrated in Fig. 4 shows current-voltage relationships recorded before and after superfusing the cell with AngII (2 nM) under conditions where only the Ca 2ϩ -dependent pathway could be activated. In this experiment, bTREK-1 was inhibited by greater than 90% over the range of test potentials where it could be accurately measured (Ϫ20 to ϩ40 mV). Inhibition through the Ca 2ϩ -dependent pathway was entirely selective for bTREK-1. The voltage-gated Kv1.4 current was not altered in this and three similar experiments.
Ca 2ϩ -dependent Inhibition of bTREK-1 Is Mediated through an AT 1 Receptor and Is G Protein-dependent-Bovine adrenocortical cells express both AT 1 and AT 2 AngII receptors (15,17). In a previous study, we demonstrated that AngII-mediated inhibition of bTREK-1 through the ATP-dependent pathway occurred through activation of an AT 1 receptor (12). In the present study, we found that AngII-dependent inhibition of bTREK-1 through the Ca 2ϩ pathway is also mediated through where B is the AngII concentration, K d is the equilibrium dissociation constant, and X is the Hill coefficient. Values are mean Ϯ S.E. of from 4 to 23 determinations. B and C, AngII-mediated bTREK-1 inhibition and membrane depolarization. K ϩ currents were recorded from AZF cells using the voltage protocols described in the legend of Fig. 1 with pipette solutions designed to allow activation of the Ca 2ϩ (B), or the combined Ca 2ϩ -/ATP-dependent pathways (C). When bTREK-1 reached a stable maximum value, membrane potential was then recorded after switching to current clamp, and AngII (2 nM) was superfused as indicated. When membrane potential reached a stable value, K ϩ currents were recorded to determine the extent of bTREK-1 inhibition. Membrane potential was plotted against time at the right. Numbers on traces correspond to currents recorded before (1) and after (2) superfusing AngII. a losartan-sensitive AT 1 receptor and is G protein-dependent.
When whole cell recordings were made with pipette solution that permitted selective activation of the Ca 2ϩ -dependent pathway, the AT 1 receptor antagonist losartan suppressed inhibition of bTREK-1 by AngII (Fig. 5, A and D). In the experiment illustrated in Fig. 5A, bTREK-1 grew to a stable maximum before sequentially superfusing the cell with saline containing 500 nM losartan, followed by one containing losartan and 2 nM AngII. Whereas losartan alone produced a small inhibition of bTREK-1, it nearly completely prevented inhibition of the current by AngII. After washing with control saline, a second application of AngII in the absence of the antagonist inhibited bTREK-1 by 80%.
Although AT 1 receptors are typically coupled to PLC activation through a pertussis toxin-insensitive G protein, direct interactions of AT 1 receptors with PLC isoforms have been observed (36). To determine whether the Ca 2ϩ -dependent pathway for AngII-mediated inhibition of bTREK-1 involves a G protein intermediate, whole cell recordings were made with pipette solutions containing the non-hydrolyzable GTP analog GDP␤S (500 M) in place of GTP. With GDP␤S in the pipette, AngII (2 nM) inhibited bTREK-1 by only 23.4 Ϯ 5.7% (n ϭ 10), compared with the control value of 66.0 Ϯ 2.9% (n ϭ 22) (Fig.  5, B and D). The diminished effectiveness of AngII in inhibiting bTREK-1 through this Ca 2ϩ -dependent path was not because of a nonspecific action of GDP␤S, because the Ca 2ϩ ionophore ionomycin remained very effective at inhibiting bTREK-1 under these conditions (Fig. 5D).
PLC and Ca 2ϩ Dependence of AngII-mediated Inhibition of bTREK-1-Activation of AT 1 receptors on adrenocortical cells is coupled to PLC␤ activation and the synthesis of IP 3 and diacylglycerol (37)(38)(39). PLC␤ itself displays a dependence on Ca 2ϩ for full activity (40,41). The suppression of AngII-mediated inhibition of bTREK-1 by strongly buffering [Ca 2ϩ ] i to 22 nM with 11 mM BAPTA could have resulted from a failure to activate PLC rather than inhibition of an IP 3 -stimulated Ca 2ϩ transient. If so, then inhibition through this pathway could be mediated through a PLC-dependent second messenger, such as diacylglycerol rather than by a direct effect of Ca 2ϩ on the channel.
However, when whole cell recordings were made with [Ca 2ϩ ] i buffered to a resting physiological concentration of 100 nM with 11 mM BAPTA, AngII (2 nM) remained ineffective, inhibiting bTREK-1 by 10.6 Ϯ 5.9% (n ϭ 4). Thus, at physiological [Ca 2ϩ ] i , which supports the activation of PLC, AngII does not inhibit bTREK-1 in the presence of strong Ca 2ϩ buffering.
In addition, it was observed that raising [Ca 2ϩ ] i in the pipette to 100 nM did not alter the expression of bTREK-1. At [Ca 2ϩ ] i of 22 and 100 nM, maximum bTREK-1 current densities were 50.8 Ϯ 7.9 and 57.1 Ϯ 8.9 pA/pF (n ϭ 5), respectively. Apparently, Ca 2ϩ does not affect bTREK-1 activity at concentrations up to 100 nM.
To determine whether AngII-mediated inhibition of bTREK-1 through the Ca 2ϩ pathway is dependent on [Ca 2ϩ ] ext , AZF cells were superfused with saline containing AngII and no added Ca 2ϩ . In the absence of added Ca 2ϩ , AngII-mediated inhibition of bTREK-1 was blunted, but not eliminated (Fig. 6, B and C).
In the experiment illustrated in Fig. 6B, currents were initially recorded in control saline, after which the cell was sequentially superfused with saline without added Ca 2ϩ , followed by this "zero Ca 2ϩ " solution containing 2 nM AngII, and finally by one containing AngII as well as 2 mM Ca 2ϩ . The amplitude of the bTREK-1 current rapidly increased in the absence of external Ca 2ϩ , and then decreased by 30% upon superfusion of AngII. Superfusion of saline containing 2 nM AngII and 2 mM Ca 2ϩ further reduced bTREK-1 current amplitude to 29% of its maximum value within 10 min.
Overall, in 6 cells, AngII (2 nM) inhibited bTREK-1 by an average of 45.1 Ϯ 8.7% (n ϭ 6) in the absence of added external Ca 2ϩ , compared with 66.0 Ϯ 2.9% (n ϭ 22) in control saline (Fig. 6C). In three cells where inhibition of bTREK-1 by AngII was first measured in the absence, and then in the presence of external Ca 2ϩ , inhibition increased from 28.4 to 66.2%.
Mechanism of Ca 2ϩ -dependent Inhibition of bTREK-1-The results presented thus far indicate that Ca 2ϩ entering through the plasma membrane and released from intracellular stores may both contribute to AngII-mediated inhibition of bTREK-1. The molecular mechanism by which Ca 2ϩ inhibits the activity of bTREK-1 channels is not known. Because the Ca 2ϩ pathway functions in the absence of hydrolyzable ATP, it is likely that Ca 2ϩ interacts directly with the bTREK-1 channel or an associated subunit or protein. If Ca 2ϩ interacts directly with the bTREK-1 channel protein, then agents that increase intracellular Ca 2ϩ should effectively inhibit the activity of cloned bTREK-1 channels expressed in a cell line. The Ca 2ϩ ionophore ionomycin (1 and 5 M) inhibited the activity of bTREK-1 channels expressed in CHO-K1 cells by 54.5 Ϯ 18.6 and 75.6 Ϯ 5.6%, respectively (Fig. 7, A and C).

FIG. 4. Ca 2؉ -dependent inhibition is voltage-independent and
specific. K ϩ currents were activated at 30-s intervals by voltage steps of varying size from a holding potential of Ϫ80 mV, before and after superfusing the AZF cell with AngII (2 nM). Top, current records at test potentials between Ϫ60 and ϩ40 mV (in 10-mV increments) before and after AngII as indicated. Bottom, bTREK-1 current amplitudes from experiments shown are plotted against test voltage.
The inhibition of bTREK-1 channels expressed in CHO-K1 cells by ionomycin suggests a direct interaction with the channel or a Ca 2ϩ -dependent protein expressed in a wide range of cells. The Ca 2ϩ -dependent modulation of a number of ion channels requires a calmodulin intermediate. In some instances Ca 2ϩ -free apocalmodulin is constitutively bound to the channel, whereas in others, only the Ca 2ϩ -calmodulin complex binds to the channel with high affinity (44,45). To determine whether the Ca 2ϩ -dependent inhibition of bTREK-1 is calmodulin dependent, the inhibition of bTREK-1 by AngII was studied in whole cell recordings with pipettes containing a CIP. This 17-residue peptide mimics the calmodulin-binding domain of myosin light chain kinase and binds to the calmodulin with a K d of 6 pM (46).
Effect of Ca 2ϩ on Unitary bTREK-1 Currents-Whole cell patch clamp experiments indicated that Ca 2ϩ -mediated inhibi- FIG. 5. Ca 2؉ -dependent inhibition of bTREK-1 by AngII through G protein-dependent AT 1 receptor. A, receptor subtype: whole cell K ϩ currents were recorded from AZF cells at 30-s intervals in response to voltage steps to ϩ20 mV applied from a holding potential of Ϫ80 mV with or without depolarizing prepulses. After bTREK-1 current reached a stable maximum amplitude, cells were superfused with losartan (500 nM) or PD123319 (500 nM), followed by saline containing the antagonist and AngII Effect of ionomycin and CIP on cloned and native bTREK-1 channels. A, inhibition of cloned bTREK-1 channels by ionomycin. K ϩ currents were recorded from bTREK-1-transfected CHO-K1 cells in response to voltage steps to ϩ20 mV applied at 30-s intervals from a holding potential of Ϫ80 mV. After recording currents in control saline, cell was superfused with ionomycin (5 M). Numbers on current traces correspond to those on the plot of bTREK-1 amplitudes at the right. B, effect of CIP on bTREK-1 inhibition by AngII. Whole cell K ϩ currents were recorded from an AZF cell at 30-s intervals in response to voltage steps to ϩ20 mV with or without depolarizing prepulses using a pipette solution that contained 500 nM CIP. After bTREK-1 reached a stable maximum, cell was superfused with 2 nM AngII. Numbers on current traces correspond to those on the plot of bTREK-1 amplitudes at the right. C, summary of experiments as in A and B. Bars indicate fraction of bTREK-1 remaining after steady state block by ionomycin (1 or 5 M) or AngII (2 nM). Values are mean Ϯ S.E. for the indicated number of separate determinations. *, p value for AngII ϩ CIP ϭ 0.0046. bTREK-1 amplitudes, excluding the possibility that inhibition occurs through a pore-blocking mechanism. Ca 2ϩ (10 M) also inhibited the activity of unitary bTREK-1 channels expressed in COS-7 cells, although clear, reversible inhibition was observed in only 2 of 4 cells (Fig. 8C).
AngII Does Not Modulate the T-type Ca 2ϩ Current through the Ca 2ϩ -dependent Pathway-AngII has been reported to enhance the T-type Ca 2ϩ current of bovine AZG cells by a Ca 2ϩdependent mechanism that includes a leftward shift in the voltage dependence of channel activation, and requires the activation of a Ca 2ϩ /calmodulin-dependent protein kinase (23,25). Experiments were done to determine whether AngII could modulate T-type Ca 2ϩ currents in bovine adrenal cortical cells by the Ca 2ϩ -dependent, but kinase-independent pathway.
Nearly all freshly plated bovine AZF cells express only low voltage-activated, slowly deactivating T-type Ca 2ϩ currents (47). In the experiment illustrated in Fig. 9A, T-type deactivating "tail" currents were recorded with pipette solutions containing 2 mM NaUTP in place of MgATP, with Ca 2ϩ weakly buffered by 0.5 mM EGTA. Currents were recorded at Ϫ80 mV after activation by 10-ms depolarizing pulses to 0 mV. AngII (2 nM) failed to increase the amplitude of the T-type Ca 2ϩ current in this or any of six similar experiments. AngII also failed to significantly shift the voltage dependence of T channel activation under similar conditions (Fig. 9B).

DISCUSSION
In this study, we identified a new Ca 2ϩ -dependent pathway by which AngII inhibits bTREK-1 K ϩ channels and depolarizes bovine adrenocortical cells. The simultaneous activation of the novel Ca 2ϩ -dependent pathway and a previously described ATP hydrolysis-dependent pathway by AngII provides a failsafe mechanism for potent and effective bTREK-1 inhibition and depolarization of AZF cells (Fig. 10). Other paracrine factors, including ATP, which activate PLC-coupled receptors and stimulate corticosteroid secretion from bovine adrenocortical cells also inhibit bTREK-1 K ϩ channels through these parallel pathways. In contrast, ACTH which functions through G s and cAMP inhibits bTREK-1 through ATP hydrolysis, but not through Ca 2ϩ -dependent mechanisms.
In addition to AngII, ATP also inhibited bTREK-1 through the Ca 2ϩ -dependent pathway at concentrations that stimulate large increases in cortisol production from bovine AZF cells (13). In a previous study, we showed that ATP also inhibited bTREK-1 by an ATP hydrolysis-dependent mechanism through a nucleotide receptor with a P2Y 3 agonist profile (14). These two results indicate that activation of the PLC-coupled P2Y 3 receptor in AZF cells inhibits bTREK-1 by the same two pathways utilized by AngII. Accordingly, externally applied ATP was significantly more effective at inhibiting bTREK-1 when recordings were made under conditions that supported activation of both the Ca 2ϩ -and ATP-dependent paths.
In contrast to AngII and external ATP, ACTH and cAMP failed to inhibit bTREK-1 when only the Ca 2ϩ -dependent mechanism was available for activation. Thus, although ACTH and cAMP both inhibit bTREK-1 by protein kinase A-dependent and -independent mechanisms, neither of these is mediated by Ca 2ϩ (35).
The Ca 2ϩ -dependent Signaling Pathway-The effective inhibition of bTREK-1 by AngII in whole cell recordings with pipettes containing UTP or AMP-PNP in place of MgATP precluded the involvement of a kinase or ATPase in the Ca 2ϩdependent response, because UTP and AMP-PNP are not FIG. 9. AngII does not modulate T-type Ca 2؉ current through the Ca 2؉ -dependent pathway. The effect of AngII on T-type Ca 2ϩ current amplitude and voltage-dependent activation were measured in AZF cells. A, Ca 2ϩ currents were activated by 10-ms voltage steps to Ϫ5 mV applied at 30-s intervals from a holding potential of Ϫ80 mV. After recording Ca 2ϩ currents in standard saline, cell was superfused with saline containing 2 nM AngII. "Tail current" amplitudes recorded at Ϫ80 mV are plotted at the right. B, activation curves. The effect of AngII on the voltage dependence of T-type Ca 2ϩ channel activation was measured by recording tail currents at a constant potential (Ϫ80 mV) after applying activating voltage steps of 10-ms duration in 10-mV increments to potentials between Ϫ60 and ϩ30 mV before (•) and after (E) superfusing cells with 2 nM AngII. Tail  substrates for these enzymes. However, the molecular components of the Ca 2ϩ -dependent signaling pathway remained to be identified.
Experiments in this study in combination with those from a previous report (12) demonstrate that inhibition of bTREK-1 by AngII through both the Ca 2ϩ -and ATP hydrolysis-dependent pathways are mediated by a losartan-sensitive AT 1 receptor through a G-protein-dependent process. Although most physiological responses mediated through AT 1 receptors require a G-protein intermediate, several studies have shown that activated AT 1 receptors interact directly with different effector proteins (17,36,48). AT 1 receptors in many cells, including adrenocortical cells, are coupled to PLC through the GTP-binding protein G q (17). U73122 (3 M) significantly reduced, but did not eliminate, AngII-mediated inhibition of bTREK-1. This result is consistent with a model where the Ca 2ϩ -dependent inhibition of bTREK-1 depends on PLC activation. U73122 inhibits agonistinduced PLC␤ activation in intact cells with IC 50 values in the micromolar range (42). At concentrations above 3 M, U73122 significantly inhibited the expression of bTREK-1 in whole cell recordings as has been previously reported for other K ϩ channels, precluding the use of higher concentrations in our experiments (49). In this regard, it is significant that the inactive U73122 analog U73343 (3 M) had no effect on AngII-mediated inhibition of bTREK-1.
PLC activation in adrenocortical cells leads to IP 3 -induced release of intracellular Ca 2ϩ , as well as Ca 2ϩ influx through plasma membrane channels (20,37,39). Superfusing AZF cells with Ca 2ϩ -free saline initially increased bTREK-1 current and significantly reduced the inhibition of this current by AngII, indicating that Ca 2ϩ influx through plasma membrane channels contributes to the regulation of channel activity. Bovine adrenocortical cells express TRPC4 non-selective cation channels that may be activated by AngII (37,38). Ca 2ϩ entering through these TRPC4 channels could be linked to the inhibition of nearby bTREK-1 channels.
The rapid increase in bTREK-1 current observed upon superfusion of Ca 2ϩ -free saline could result from a reduction in the basal Ca 2ϩ influx through the plasma membrane or by a change in the effective transmembrane voltage. Removal of extracellular Ca 2ϩ shifts the voltage dependence of channel activation in the hyperpolarizing direction by effectively reducing the electric field strength across the membrane (50). Consequently, because bTREK-1 channels display weak voltage dependence, the absence of [Ca 2ϩ ] o could enhance current by increasing channel open probability (26).
Bovine adrenocortical cells also express T-type voltage-gated Ca 2ϩ channels that may represent the other major pathway for Ca 2ϩ entry in these cells (10,47,51). However, because AngII failed to increase Ca 2ϩ current through T-type Ca 2ϩ channels under the conditions of our experiment, it is unlikely that Ca 2ϩ entering through these channels contributed to inhibition of bTREK-1.
The reduction in AngII-stimulated inhibition of bTREK-1 current produced by 2-APB (80 M) is consistent with a model wherein inhibition is mediated in part by IP 3 -stimulated release of Ca 2ϩ from intracellular stores. In this regard, 2-APB inhibits IP 3 -stimulated Ca 2ϩ release with an IC 50 of 42 M, but increases Ca 2ϩ release from microsomes at concentrations above 90 M (43). Therefore, 2-APB was not used at concentrations above 80 M in our studies. Overall, the results indicate that the Ca 2ϩ -dependent inhibition of bTREK-1 by AngII involves both PLC-dependent Ca 2ϩ influx and release from intracellular stores.
Molecular Mechanism for Ca 2ϩ Inhibition of bTREK-1-The results of experiments using ionomycin, CIP, and Ca 2ϩ applied directly to excised membrane patches provide insight into the molecular mechanism by which Ca 2ϩ inhibits bTREK-1. The absence of MgATP in all of these experiments ruled out the participation of protein kinases, simplifying the problem. However, Ca 2ϩ modulates the gating of a number of K ϩ channels by multiple kinase-independent mechanisms. Several Ca 2ϩ -actvated K ϩ channels are gated by Ca 2ϩ binding directly to the channel (52,53). For other K ϩ channels, modulation by Ca 2ϩ proceeds through a calmodulin intermediate. Calmodulin can be constitutively associated with the channel, or it may bind to the channel only after first binding Ca 2ϩ (44,45,54,55).
Ionomycin inhibited bTREK-1 K ϩ current in transfected cells by a maximum of 75.6 Ϯ 5.6%, indicating that as many as three-fourths of these channels retain their Ca 2ϩ sensitivity even when expressed in a foreign cell type. Similarly, Ca 2ϩ (10 M) inhibited the activity of bTREK-1 channels expressed in COS-7 cells in half of the patches tested. If Ca 2ϩ -mediated inhibition of bTREK-1 required an AZF cell-specific subunit or other protein not expressed in other eukaryotic cells, ionomycin and Ca 2ϩ would have been ineffective as inhibitors of the transfected channels. However, if the Ca 2ϩ -dependent inhibition of bTREK-1 were mediated through a ubiquitously expressed protein such as calmodulin, ionomycin and Ca 2ϩ would likely retain their effectiveness in transfected cells. The incomplete inhibition of transfected bTREK-1 channels by ionomycin and Ca 2ϩ in whole cell and single channel recordings may suggest that not all bTREK-1 channels become associated with the Ca 2ϩ -dependent protein in CHO-K1 or COS-7 cells.
The reduced inhibition of bTREK-1 observed when the patch pipette contained CIP supports a role for calmodulin in the inhibition of this channel by AngII. However, because this peptide binds to calmodulin with extremely high affinity (K d ϭ 6 pM), it is perhaps surprising that CIP did not more effectively suppress the AngII response. This may suggest that apocalmodulin is constitutively and tightly bound to the native bTREK-1 channel in a Ca 2ϩ -independent fashion. Near its carboxyl-terminal end (amino acids 340 -354), bTREK-1 protein contains a sequence that conforms partially to the consensus "IQ" motif calmodulin-binding site found in several other K ϩ channels where calmodulin is constitutively associated (54,56). The inhibition of bTREK-1 by Ca 2ϩ in excised inside-out AZF cell patches indicates that if Ca 2ϩ -dependent inhibition of bTREK-1 is mediated through calmodulin, then calmodulin must be constitutively associated with the bTREK-1 channel.
Comparison of Signaling Pathways for AngII Inhibition of bTREK-1 and Activation of T-type Ca 2ϩ Channels-In bovine AZG cells, AngII enhances current through Ca v 3.2 T-type Ca 2ϩ channels by activation of Ca 2ϩ /calmodulin-dependent kinase II, which phosphorylates these channels at a specific site (25,57). Interestingly, we have shown that AngII inhibits bTREK-1 channels in bovine adrenocortical cells by two entirely different pathways. Furthermore, AngII does not modulate Ca v 3.2 Ttype Ca 2ϩ channels in bovine adrenocortical cells by either of the two mechanisms that inhibit bTREK-1 channels. In a previous study, we showed that AngII has no effect on the T-type Ca 2ϩ current in bovine AZG cells through the ATP-dependent pathway (11). Finally, the results of the current study do not rule out the possibility that AngII may also inhibit bTREK-1 by a third pathway involving a Ca 2ϩ -dependent kinase such as Ca 2ϩ /calmodulin kinase or kinase C.
Modulation of Cloned Neuronal TREK-1 Channels-The inhibition of rat brain TREK-1 channels by G q -coupled metabotropic glutamate receptors has also been reported, and the signaling pathways mediating rat brain TREK-1 regulation have been explored in transfected cell lines (58). Glutamatemediated inhibition of TREK-1 was blocked by U73122, indicating the involvement of PLC. However, in contrast with our studies, these whole cell experiments were done with [Ca 2ϩ ] i buffered by 5 mM EGTA and without ATP in the pipette solution. In this system, evidence is presented indicating that inhibition occurred through a direct effect of the PLC-generated second messenger diacylglycerol, rather than through Ca 2ϩ -or ATP-dependent pathways. It appears that TREK-1 K ϩ channels in various cell types are modulated by a range of signaling pathways, involving kinases, diacylglycerol, cAMP, and Ca 2ϩ .
The differences that we have observed in the signaling pathways that modulate native bTREK-1 channels compared with cloned TREK-1 channels from other species could arise through differences in channel-associated proteins, or signaling pathways present in the native or host cells. For example, the availability of calmodulin for interaction with TREK-1 may be quite different in a transfected cell, when compared with a native cell.
Physiological Significance-bTREK-1 channels function as a central control point in bovine adrenocortical cells where hormonal and paracrine signals originating at the cell membrane are coupled to depolarization-dependent Ca 2ϩ entry. We have shown that bovine AZF cells have developed a fail-safe mechanism for potent and effective inhibition of bTREK-1 current and membrane depolarization by AngII through the activation of parallel Ca 2ϩ -and ATP hydrolysis-dependent signaling pathways (Fig. 10).
In neurons and other excitable cells, bTREK-1 and other four transmembrane/two-pore K ϩ channels act as a brake on electrical activity (5,59). It remains to be seen whether, under physiological conditions, bTREK-1 inhibition by AngII triggers Ca 2ϩ -dependent action potentials in bovine adrenocortical cells. Action potentials have been recorded in adrenocortical cells from several species (60 -62). In the absence of bTREK-1, opposing T-type Ca 2ϩ currents and A-type K ϩ currents could give rise to action potentials.
In addition to AngII, external ATP and perhaps other para-crine factors inhibit bTREK-1 through these same two pathways. Inhibition of specific ion channels by simultaneous activation of multiple signaling pathways may be a common theme when near total inactivation is required. Accordingly, ACTHand cAMP-mediated inhibition of bTREK-1 by protein kinase A-dependent and -independent mechanisms may be yet another example (5,35). Overall, ACTH and AngII inhibit bTREK-1 through at least four different signaling pathways.
Although it is clear that AngII acts through an AT 1 receptor by a G protein-dependent mechanism to inhibit bTREK-1 via separate Ca 2ϩ -and ATP-dependent pathways, the underlying molecular events are incompletely understood. Specifically, our results do not demonstrate that the inhibition of bTREK-1 by Ca 2ϩ is entirely dependent on PLC␤. Furthermore, although Ca 2ϩ entry through the plasma membrane contributes to bTREK-1 inhibition, the molecular identity of the plasma membrane channel has not been determined. Finally, the specific role of calmodulin in TREK-1 inhibition by Ca 2ϩ remains to be defined.