Bovine adrenocortical zona fasciculata (AZF) cells express a novel ATP-dependent K+-permeable channel (I AC). Whole cell and single-channel recordings were used to characterize I AC channels with respect to ionic selectivity, conductance, and modulation by nucleotides, inorganic phosphates, and angiotensin II (ANG II). In outside-out patch recordings, the activity of unitaryI AC channels is enhanced by ATP in the patch pipette. These channels were K+ selective with no measurable Na+ or Ca2+ conductance. In symmetrical K+ solutions with physiological concentrations of divalent cations (M2+), I ACchannels were outwardly rectifying with outward and inward chord conductances of 94.5 and 27.0 pS, respectively. In the absence of M2+, conductance was nearly ohmic. Hydrolysis-resistant nucleotides including AMP-PNP and NaUTP were more potent than MgATP as activators of whole cell I AC currents. Inorganic polytriphosphate (PPPi) dramatically enhancedI AC activity. In current-clamp recordings, nucleotides and PPPi produced resting potentials in AZF cells that correlated with their effectiveness in activatingI AC. ANG II (10 nM) inhibited whole cellI AC currents when patch pipettes contained 5 mM MgATP but was ineffective in the presence of 5 mM NaUTP and 1 mM MgATP. Inhibition by ANG II was not reduced by selective kinase antagonists. These results demonstrate that I AC is a distinctive K+-selective channel whose activity is increased by nucleotide triphosphates and PPPi. Furthermore, they suggest a model for I AC gating that is controlled through a cycle of ATP binding and hydrolysis.
- potassium channel
- adenosine 5′-triphosphate
- angiotensin II
bovine adrenal zona fasciculata (AZF) cells express a novel K+-permeable channel (I AC) that functions pivotally in the regulation of cortisol secretion. I AC channels may set the resting potential of AZF cells, while inhibition of these channels by ACTH and angiotensin II (ANG II) is coupled to membrane depolarization, Ca2+ entry, and cortisol secretion (15, 39, 40).
Whole cell patch-clamp studies have shown thatI AC combines properties unique among ionic currents described thus far. Specifically, I ACis activated when the patch electrode contains ATP at millimolar concentrations (14), while it is inhibited by cAMP through an A-kinase-independent mechanism (16). I AC is also inhibited by antagonists of both cyclic nucleotide-gated (CNG) cation channels and K+ channel blockers (21).
Inhibition of I AC through multiple G protein-coupled receptors, including those activated by ACTH, ANG II, and multiple P1 and P2 nucleotide receptors, requires ATP hydrolysis (16, 39, 40, 47, 48). For at least one of these receptors and its associated second messenger (ACTH and cAMP), inhibition of I AC is independent of any known protein kinases (16).
Together, these results identify I AC as a novel channel combining properties of K+-selective and CNG cation channels. The overall sequence similarity between voltage-gated K+ channels and CNG cation channels suggests a common origin (26). Several such intermediate forms have been identified, including the ether-à-go-go(eag) family of K+ channels, which are modulated by cAMP. Some eag channels may also display Ca2+permeability (11, 24).
A large family of K+ channels directly gated by ATP exists. However, these inwardly rectifying K+ channels are uniformly inhibited by ATP (5). I ACchannels are the first K+-permeable channel whose activity depends on the presence of ATP in either hydrolyzable or nonhydrolyzable forms. In this regard, our results are consistent with a model in which I AC channel gating is coupled to a cycle of ATP binding and hydrolysis with similarities to that proposed for the cystic fibrosis transmembrane conductance regulator (CFTR) Cl− channel (7). In the proposed scheme, I AC channel open probability is enhanced upon ATP binding, while activation of G protein-coupled receptors promotes channel closing subsequent to ATP hydrolysis.
In the present study, single-channel recording from outside-out patches was used to characterize these novel K+-permeable channels with respect to ionic selectivity, conductance, and rectification. In whole cell recordings, the modulation of I AC by nucleotides, inorganic phosphates, and ANG II was studied to determine whether I AC gating might be controlled through an ATP hydrolysis cycle.
MATERIALS AND METHODS
Tissue culture media, antibiotics, fibronectin, and fetal bovine serum were obtained from GIBCO (Grand Island, NY). Coverslips were from Bellco Glass (Vineland, NJ). Enzymes, ANG II, MgATP, NaATP, NaUTP, NaCTP, KATP, 5-adenylylimidodiphosphate (AMP-PNP, lithium salt), guanosine 5′-O-(2-thiodiphosphate) (GDPβS), 1,2-bis(2-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid (BAPTA), the inorganic phosphates polytriphosphate (PPPi), pyrophosphate (PPi), Pi, and phosphatidylinositol 4,5-bisphosphate (PIP2), and staurosporine were obtained from Sigma Chemical (St. Louis, MO). Penfluridol was purchased from Jansen Pharmaceuticals (Beerse, Belgium). Calphostin C, AG-490, genistein, herbimycin, and PD-98059 were purchased from Calbiochem (San Diego, CA).
Isolation and culture of AZF cells.
Bovine adrenal glands were obtained from steers (age range: 1–3 yr) within 60 min of slaughter at a local slaughterhouse. Fatty tissue was removed immediately, and the glands were transported to the laboratory in ice-cold PBS containing 0.2% dextrose. Isolated AZF cells were prepared as previously described (16). Cells were plated in 35-mm dishes containing 9-mm2 glass coverslips that had been treated with fibronectin (10 μg/ml) at 37°C for 30 min and then rinsed with warm, sterile PBS immediately before cells were added. DMEM/F-12 medium supplemented with 10% fetal bovine serum, 100 U/ml penicillin, 0.1 mg/ml streptomycin, and the antioxidants 1 μM tocopherol, 20 nM selenite, and 100 μM ascorbic acid was used. Dishes were maintained at 37°C in a humidified atmosphere of 95% air-5% CO2.
Patch-clamp recordings of K+ channel currents were made in the whole cell and outside-out patch configuration. For whole cell recordings, the standard pipette solution consisted of 120 mM KCl, 2 mM MgCl2, 1 mM CaCl2, 20 mM HEPES, 11 mM BAPTA, 200 μM GTP, and 5 mM MgATP, with pH buffered to 7.2 using KOH. Titration of pH to 7.2 with KOH raised the total K+concentration to 160 mM. Pipette solution of this composition yielded a free Ca2+ concentration of 2.2 × 10−8 M, as determined by the Bound and Determined software program (10). In many experiments, MgATP was replaced with other nucleotides or an inorganic phosphate, as described in the text. The external solution consisted of 140 mM NaCl, 5 mM KCl, 2 mM CaCl2, 2 mM MgCl2, 10 mM HEPES, and 5 mM glucose, with pH adjusted to 7.4 using NaOH.
The standard external and pipette solutions used for single-channel recording from outside-out patches were identical to those used for whole cell recordings. A number of other external or pipette solutions were also used and are described in the text. All solutions were filtered through 0.22-μm cellulose acetate filters.
AZF cells were used for patch-clamp experiments 2–12 h after they were plated. Typically, cells with diameters of <15 μm and capacitances of 8–12 pF were selected. Coverslips were transferred from 35-mm culture dishes to the recording chamber (volume: 1.5 ml), which 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 MΩ were fabricated from Corning 0010 glass (World Precision Instruments, Sarasota, FL). These electrodes routinely yielded access resistances of 1.5–4 MΩ and voltage-clamp time constants of <100 μs. For single-channel recordings, patch electrodes with higher resistances of 3–5 MΩ were used. K+ currents were recorded at room temperature (22–25°C) according to the procedure of Hamill et al. (25) with the use of an Axopatch 1-D patch-clamp amplifier.
Pulse generation and data acquisition were performed with the use of a personal computer and pCLAMP software with a TL-1 interface (Axon Instruments, Burlingame, CA). Currents were digitized at 5–20 kHz after filtering with an eight-pole Bessel filter (Frequency Devices, Haverhill, MA). Linear leak and capacity currents were subtracted from current records by using scaled hyperpolarizing steps of 1:3–1:4 amplitude. Data were analyzed and plotted with pCLAMP 5.5 and 6.02 (Clampan, Clampfit, Fetchan, and pSTAT) and SigmaPlot 4.0. Drugs were applied by bath perfusion, controlled manually by a six-way rotary valve.
Calculation of IAC channel activity.
Because of uncertainty about the number of channels in any given patch (N) and the nonstationary character ofI AC activity in outside-out patches, channel activity was expressed in terms of NP o rather than P o (open probability).NP o was calculated from the expressionI = NP o ı̂, where I is the measured mean current, N is the number of active channels in the patch, ı̂ is the single-channel current, and P o is the open probability for samples of 90–100 consecutive traces, each 400 ms in duration.
ATP-dependent activity of unitary IAC current in outside-out patches.
Previously, in whole cell recordings from AZF cells, we found thatI AC K+ current is present initially but grows 10-fold or more to a stable amplitude over a period of minutes, provided that ATP is present in the recording pipette at millimolar concentrations (14, 39). In excised outside-out patch recordings, single I AC channels showed a similar time- and ATP-dependent increase in channel open probability.
In Fig. 1, unitary currents were recorded with a pipette containing 2 mM MgATP in response to voltage steps applied at 4-s intervals from a holding potential of −40 mV to a test potential of +30 mV. Under these conditions, a single type of unitary current was present with a measured mean amplitude of 3.95 ± 0.34 pA. In this experiment, I AC channel activity (NP o) increased 5.5-fold from an initial value of 0.17 at time 0 to a value of 0.95 after 7 min of recording. At this time, it is clear that the membrane patch contains at least two functioning channels. In contrast to whole cell recordings, where I AC reaches a stable maximum amplitude, unitary I AC channel activity typically increased continuously for the life of the patch.
When outside-out patch recordings were made with pipettes containing MgATP at concentrations below 1 mM, I AC channel activity was markedly reduced and NP o failed to increase even during prolonged recordings. In the experiment shown in Fig. 2 A,I AC activity in an outside-out patch was recorded with a pipette containing 0.1 mM MgATP. Analysis of an amplitude histogram constructed from channel openings indicated the presence of a single type of channel with a mean amplitude of 3.82 ± 0.28 pA. NP o did not increase from its initial value of 0.01 during 20 min of recording. Similar results were obtained from eight separate patches under these conditions.
By comparison, when recordings were made with pipettes containing 5 mM MgATP, NP o increased dramatically during the course of an experiment. In the experiment shown in Fig. 2 B, amplitude histograms constructed from idealized channel openings showed the presence of three nearly equally spaced peaks indicative of at least three active channels in the patch. Gaussian fits of these histograms yielded means of 3.96 ± 0.34, 7.87 ± 0.32, and 11.95 ± 0.42 pA. NP o increased continuously during this experiment and was estimated to be 0.83 after 6 min of recording from this outside-out patch.
In other experiments, we attempted to study the direct activation ofI AC channels in inside-out patches upon superfusion of the cytoplasmic membrane surface with internal solution containing MgATP. Surprisingly, in this configuration, 5 mM MgATP failed to activate I AC channels in any of nine cells tested.
K+ selectivity and conductance of IAC channels.
In whole cell recordings, I AC appears as a noninactivating, weakly voltage-dependent, ATP-activated K+current (14, 16, 39). Properties of unitaryI AC current, including ionic selectivity, conductance, and rectification, in symmetrical K+ solutions have not been described.
To obtain a measure of the K+ selectivity and conductance of I AC channels, unitaryI AC currents were recorded from outside-out patches at voltages ranging from −90 to +60 mV by using external solutions containing either Na+(160 mM) or K+(160 mM) and standard pipette solution supplemented with 5 mM MgATP. External and pipette solutions also contained divalent cations (M2+) as in standard solutions. With Na+-containing external solution,I AC channels were outwardly rectifying. No inward Na+ current through I ACchannels was detectable at potentials as negative as −90 mV (Fig.3, A, left, andB).
When the external Na+-containing solution was replaced with K+ solution, I AC channels were still outwardly rectifying but reversed near 0 mV and became clearly inward at negative potentials (Fig. 3, A, right, andB). Although no systematic study of voltage-dependent open probability was done, I AC channels remained active at potentials as negative as −80 mV. In these symmetrical K+ solutions, at test potentials between 0 and +60 mV,I AC channels had a measured chord conductance of 78.2 pS. By comparison, at potentials between 0 and −60 mV, chord conductance was 30.1 pS (Fig. 3 B).
These results show that, in the presence of physiological concentrations of M2+, I AC channels are highly selective for K+ with negligible conductance to Na+. Furthermore, in symmetrical K+ solutions containing approximately physiological concentrations of M2+, I AC channels are outwardly rectifying with a 2.5-fold ratio of outward relative to inward K+ conductance.
Effect of M2+ on unitary conductance and rectification of IAC channels.
The rectifying properties of some K+-selective channels are caused by the unidirectional block of K+ flux by M2+. In particular, the rectification of many inward rectifier K+ channels occurs because intracellular Mg2+ blocks the outward flow of K+(37).
In outside-out patch recordings made in symmetrical K+solutions, it was discovered that the presence of M2+ on either side of the membrane dramatically altered unidirectional K+ flow through these channels. Specifically, with standard external and pipette solutions containing Ca2+ and Mg2+ at approximately physiological concentrations (external: 2 mM Ca2+, 2 mM Mg2+; pipette: 22 nM Ca2+, 2 mM Mg2+),I AC channels were outwardly rectifying in symmetrical K+ with chord conductance of 94.5 pS measured between 0 and +80 mV, compared with 27 pS between 0 and −80 mV (Fig.4, A and E).
Deletion of M2+ from the external solution markedly increased the amplitude of inward unitary currents amplitudes measured at negative potentials but only slightly increased the amplitude of outward currents measured at positive voltages (Fig. 4 B). In the absence of external M2+, the chord conductance between 0 and +80 mV increased only to 98.5 pS, while the chord conductance measured between 0 and −80 mV was 154.8 pS, a 5.7-fold increase over that measured in the presence of external Ca2+ and Mg2+.
Selective removal of internal Mg2+ in the presence of external M2+ enhanced the amplitude of unitary outward currents, while inward currents measured at negative potentials were similar to those recorded with M2+ present on both sides of the membrane (Fig. 4 C). Specifically, the chord conductance measured between 0 and +80 mV was 143.3 pS, a value 52% larger than that observed with Mg2+ in the pipette solution. In contrast, chord conductance between 0 and −80 mV was 26.9 pS, a value nearly identical to that measured with M2+present on both sides of the membrane.
The results of experiments in which M2+ were selectively deleted from either the external or pipette solutions indicate that the rectifying properties of the I AC channel are caused by bidirectional inhibition of K+ flux by M2+. When present on the external side of the membrane, M2+ reduce the influx of K+, andI AC appears as an outward rectifier. When M2+ are deleted from the external solution but present intracellularly, I AC channels become inwardly rectifying.
When M2+ were omitted from the external solution and pipette solutions contained no added Mg2+ and nominal Ca2+ (22 nM), the I AC channels became nearly ohmic. Under these conditions, the chord conductances measured between 0 and +80 mV and between 0 and −80 were 161 and 175 pS, respectively (Fig. 4, D and E). Although inward K+ conductance increased dramatically in the absence of M2+, inward Na+ current remained undetectable (data not shown).
The inhibition of K+ flux throughI AC channels by M2+ suggests an interaction in the permeation pathway. CNG cation channels and at least one K+ channel are permeable to Ca2+ (11,28). To determine whether I AC channels display measurable Ca2+ conductance, we measured unitaryI AC currents in the presence of external solutions containing 107 mM Ca2+ in the absence of Na+ or K+ at potentials as negative as −100 mV. No inward Ca2+ current throughI AC channels was detected in any of three tested cells (data not shown).
Activation of IAC by poorly hydrolyzable nucleotides.
Two types of K+ channels expressed by bovine AZF cells are easily distinguished in whole cell recordings. In addition to the rapidly inactivating A-type K+ current (I A), the noninactivatingI AC current grows continuously over a period of minutes in whole cell recordings (14, 16, 39, 41). The absence of time-dependent inactivation of I ACallows it to be isolated for measurement with the use of either of two voltage-clamp protocols. When voltage steps of 300-ms duration are applied from a holding potential of −80 mV to a test potential of +20 mV, I AC can be selectively measured near the end of a step, at a point where I A has completely inactivated (Fig. 5 A,left). With the second protocol, I ACcan be selectively activated by an identical voltage step after a 10-s prepulse to −20 mV has fully inactivated I A(Fig. 5 A, middle).
Previous studies have led to the hypothesis thatI AC activity is linked to a cycle of ATP-binding and hydrolysis, wherein I AC open probability increases with ATP binding while its hydrolysis or dissociation leads to channel closing (14). If this were the case, then other poorly hydrolyzable nucleotides might prove to be more potent and/or effective activators of I AC than ATP itself.
In whole cell patch-clamp recordings, we compared NaUTP with MgATP with respect to potency and effectiveness as activators ofI AC K+ current. At concentrations between 0.1 and 5 mM, NaUTP was significantly more potent than MgATP at enhancing I AC expression (Fig. 5). In the experiment shown in Fig. 5 A, the time-dependent increase inI AC amplitude was monitored by using patch pipettes containing either 1 mM MgATP or 1 mM NaUTP. With 1 mM MgATP in the recording pipette, I AC increased <2-fold from its initial value (trace 1) to a stable maximum (trace 2) during 14 min of recording (Fig. 5 A). By comparison, with 1 mM NaUTP in the recording pipette,I AC increased >12-fold, reaching a stable maximum amplitude within 10 min. Overall, with 1 mM NaUTP applied intracellularly through the pipette, I AC reached a maximum current density of 50.6 ± 5.5 pA/pF (n= 10), a value nearly fourfold greater than that observed with 1 mM MgATP in the pipette (13.0 ± 1.0 pA/pF, n = 24) (Fig. 5 B).
At concentrations of 0.1 and 0.4 mM, NaUTP was also significantly more effective than 1 mM MgATP at activating I AC(Fig. 5 B). Activation of I AC by NaUTP reached a maximum at concentrations between 2 and 5 mM. At these higher concentrations, NaUTP was only slightly more effective than 5 mM MgATP at enhancing I AC activity (Fig. 5 B).
In addition to NaUTP, other nucleotides, including the nonhydrolyzable ATP analog AMP-PNP and the pyrimidines NaTTP and NaCTP, were more potent activators of I AC than MgATP. With AMP-PNP, NaTTP, or NaCTP present in the pipette solution at a concentration of 1 mM, I AC reached maximum current densities that were 2.5- to 3.5-fold greater than that observed with 1 mM MgATP (Fig. 6, A andB).
Activation of IAC by inorganic phosphates.
Inorganic phosphates inhibit ATPases by occupying sites on the enzyme from which phosphate is released after ATP is cleaved, interrupting further cycles of ATP hydrolysis (7, 22). Inorganic phosphates such as PPi and PPPi may promote the opening of CFTR Cl− channels via this mechanism (7,22, 23).
The effect of inorganic phosphates, including Pi, PPi, and PPPi, on I ACexpression was studied in whole cell recordings. Of the three phosphates tested, PPPi was clearly most effective at enhancing I AC expression. In the experiment shown in Fig. 7 A, K+ currents were recorded with a pipette containing standard pipette solution supplemented with 5 mM PPPi. In the presence of PPPi, a noninactivating current resemblingI AC grew continuously over a period of ∼10 min, increasing more than 10-fold from an initial value of <100 pA (Fig. 7 A, trace 1) to a stable maximum of ∼1,150 pA (trace 2).
To further establish the identity of the PPPi-activated K+ current, the cell was superfused with the diphenylbutylpiperidine penfluridol, a selective and potent antagonist of I AC. Penfluridol inhibitsI AC with an IC50 of 187 nM, while the transient I A K+ current is inhibited only at 200-fold higher concentrations (20). Penfluridol (1 μM) inhibited the noninactivating current (Fig.7 A, trace 3) by ∼80%, while the inactivating K+ current I A was not reduced.
When I AC is activated by MgATP, ANG II inhibits this current with an IC50 of ∼150 pM (39,40). However, when the nonhydrolyzable nucleotide AMP-PNP is used to activate I AC, ANG II is completely ineffective (40). ANG II was also ineffective as an inhibitor of I AC when PPPi activated this current. In the experiment shown in Fig. 7 B,I AC was activated with a pipette solution containing 2 mM PPPi. Over a 10-min period,I AC grew from an initial amplitude of <25 pA (Fig. 7 B, trace 1) to a maximum of ∼600 pA (trace 2). The superfusion of 10 nM ANG II failed to significantly reduce I AC (trace 3).
Overall, at a concentration of 5 mM, PPPi was slightly less effective than MgATP as an activator of I AC. With pipette solution containing PPPi,I AC reached a maximum current density of 44.4 ± 14.7 pA/pF (n = 5) compared with 52.4 ± 7.1 pA/pF (n = 17) obtained in the presence of 5 mM MgATP (Fig. 8 B).
At concentrations lower than 2 mM, MgATP is relatively ineffective as an activator of I AC (14). However, at a concentration of 1 mM, MgATP significantly potentiated activation of I AC by PPPi. With pipette solution containing 1 mM MgATP and 5 mM PPPi in combination, I AC reached a maximum current density of 80.2 ± 9.6 pA/pF (n = 17) (Fig. 8,A and B).
PPi and Pi were far less effective than PPPi as activators of I AC. With pipette solutions containing 5 mM PPi or Pi in addition to 1 mM MgATP, I AC reached maximum current densities of 22.3 ± 7.3 pA/pF (n = 8) and 15.7 ± 3.5 pA/pF (n = 6), respectively, compared with 13.0 ± 1.0 pA/pF (n = 24) observed in the presence of 1 mM MgATP alone (Fig. 8, A and B).
Activation of unitary IAC currents by NaUTP and PPPi.
NaUTP and PPPi both activated I ACchannels in outside-out patches at the same concentrations that were effective in whole cell recordings. Figure9 shows representative recordings made with pipettes containing 1 mM NaUTP (Fig. 9 A) or 2 mM PPPi (Fig. 9 B). As in outside-out patch recordings made with MgATP, NP o typically increased with time in the presence of UTP and PPPi. However, with UTP and PPPi, channel activity was unstable and sometimes disappeared abruptly.
Amplitude analysis of unitary currents recorded in the presence of NaUTP and PPPi showed that these two agents activated identical channels. In Fig. 9 A, Gaussian fits of amplitude histograms generated from idealized openings recorded in the presence of NaUTP yielded two peaks with mean amplitudes of 3.60 ± 0.39 and 7.13 ± 0.43 pA, indicative of two active channels. Similarly, Gaussian fits of amplitude histograms generated from recordings with PPPi in the pipette yielded peaks with means of 3.61 ± 0.36 and 7.05 ± 0.45 pA (Fig. 9 B). Unitary currents activated by UTP and PPPi were blocked nearly completely by the I AC-selective antagonist penfluridol (500 nM), further establishing the identity of these channels (data not shown).
Effect of nucleotides and PPPi on AZF cell membrane potential.
I AC channels may set the resting membrane potential (V m) of AZF cells. If so, thenV m should depend on the presence of nucleotides and polyphosphates that have been shown to activate these channels. We measured V m of AZF cells in whole cell current-clamp recordings with patch electrodes containing standard saline supplemented with MgATP, PPPi, or NaUTP.
The membrane potential of AZF cells was strongly dependent on the presence of nucleotides or polyphosphates in the recording pipette and was well correlated with their potency as activators ofI AC. With 1 mM MgATP in the recording pipette, average V m reached −25.6 ± 5.3 mV (n = 8) after 5–20 min of recording. Raising the MgATP concentration to 5 mM increased V m to −72.7 ± 1.9 mV (n = 3) (Table1).
At a concentration of 1 mM, UTP was much more effective than ATP at producing a negative membrane potential in AZF cells. With NaUTP (1 mM) in the patch electrode, the average V m was measured to be −72.0 ± 3.0 mV (n = 7), a value very similar to that recorded in the presence of 5 mM MgATP (Table 1).
The combination of 5 mM PPPi and 1 mM MgATP was very effective in activating I AC current in AZF cells (see Fig. 8). Accordingly, the presence of these two agents in the patch electrode in current-clamp experiments yielded averageV m of −67.9 ± 4.2 (n = 8) in AZF cells (Table 1).
Anionic phospholipid PIP2 does not activate IAC channels.
The plasmalemmal phospholipid PIP2 activates a number of ATP-sensitive channels by altering their sensitivity to ATP or by direct interaction with the channels (8, 17, 29, 44). PIP2 failed to enhance the expression ofI AC current in whole cell recordings from AZF cells. In these experiments, PIP2 was added to pipette solutions at several different concentrations along with 1 mM MgATP and 200 μM GTP. At PIP2 concentrations of 5, 10, and 50 μM,I AC reached maximum current densities of 9.2 ± 1.4 (n = 4), 13.6 ± 2.6 (n = 8), and 9.4 ± 2.0 pA/pF (n = 2), respectively, values not significantly different from the control value of 13.0 ± 1.0 pA/pF (n = 24) obtained in the absence of PIP2. The addition of 10 μM PIP2 to the patch electrode in addition to 5 mM MgATP also failed to alter the inhibition of I AC current by ANG II (10 nM) (data not shown).
ANG II-mediated inhibition of IAC and ATP hydrolysis.
The activity of I AC K+ channels is promoted by the binding of hydrolyzable and poorly hydrolyzable nucleotide triphosphates. However, inhibition ofI AC through activation of a number of G protein-coupled receptors, including ANG II receptors, requires the presence of hydrolyzable ATP (16, 39, 40, 48). These results are consistent with a model in which I ACopening and closing are controlled through an ATP hydrolysis cycle involving ATP binding and metabolism by an ATPase. However, the requirement for hydrolyzable ATP might implicate a protein kinase rather than an ATPase in I AC inhibition.
Experiments were done to determine whether ATP-dependent inhibition ofI AC by ANG II was mediated through a mechanism requiring an ATPase or, alternatively, a protein kinase. The inhibition of I AC by ANG II was studied by using pipette solution containing 5 mM NaUTP and various concentrations of MgATP. Although UTP is more potent than ATP as an activator ofI AC channels, it is a poor substrate for phosphate transfer enzymes, including protein kinases and ATPases (6, 33). ATP, on the other hand, is a substrate for both protein kinases and ATPases, although typically at different concentrations. Protein kinases are fully activated by 50 μM ATP, whereas cellular ATPases frequently display significantly higher Michaelis-Menten constant (K m) values for ATP (18, 34).
As previously reported, ANG II effectively inhibitsI AC when the pipette solution contains MgATP (5 mM) and GTP (200 μM) as the only nucleotides. Under these conditions, ANG II reduced I AC by 82 ± 5% (n = 6) (Fig. 10,A and B). When NaUTP replaced ATP in the pipette, ANG II was much less effective, inhibiting I ACby only 10 ± 5% (n = 8). The addition of 50 μM or even 1 mM MgATP to pipette solution containing 5 mM NaUTP failed to restore I AC inhibition by ANG II. Under these conditions, ANG II inhibited I AC by only 14 ± 4% (n = 3) and 14 ± 7% (n = 3), respectively (Fig. 10, A and B). Raising intracellular MgATP to 2 mM in the presence of 5 mM NaUTP partially restored I AC inhibition by ANG II to 57.6% (n = 5).
The failure of ATP, at concentrations up to 1 mM, to restore inhibition of I AC by ANG II suggests that a protein kinase does not mediate this inhibition. ANG II combines with losartan-sensitive AT1 receptors to activate a number of different protein kinases in various cells, including adrenocortical cells (36, 42, 46, 49). To further explore the possibility that ANG II-mediated inhibition of I AC involves a protein kinase, we studied the effect of six different protein kinase inhibitors on ANG II-mediated inhibition of I AC.
The activation of AT1 receptors on adrenocortical cells by ANG II leads to the phospholipase C-dependent synthesis of diacylglycerol, which activates protein kinase C (3, 13). Calphostin C is a potent and specific protein kinase C antagonist (IC50 = 50 nM) (45). When applied directly to the cytoplasm through patch electrodes at a concentration of 500 nM, calphostin C failed to prevent inhibition ofI AC by ANG II (10 nM). In the presence of this antagonist, ANG II reduced I AC by 92 ± 2% (n = 6) (Fig. 11,A and B).
ANG II activates tyrosine kinases in a number of cells through interaction with AT1 receptors (42). However, the tyrosine kinase inhibitors genistein (20 μM) and herbimycin (10 μM) both failed to reduce I AC inhibition by ANG II, even when used at concentrations at least 10-fold higher than their reported IC50 values (Fig. 11 B) (2,35).
In at least one type of cell, AT1 receptors are physically associated with Janus kinase 2 (JAK2) intracellular protein kinases, which are activated by ANG II (36). However, the specific JAK2 inhibitor AG-490 (5 μM) (38) failed to produce any inhibition of I AC at a concentration 50 times its reported IC50 (Fig. 11 B).
ANG II also activates the mitogen-activated protein (MAP) kinase pathway in a variety of cells, including bovine adrenal cortical cells (46, 49). However, the MAP kinase-selective antagonist PD-98059 (50 μM) was also completely ineffective at suppressingI AC inhibition when applied directly through the patch pipette at a concentration at least 10 times greater than the reported IC50 (Fig. 11 B).
Staurosporine is a potent nonselective protein kinase antagonist. This microbial alkaloid inhibits most serine/threonine protein kinases with IC50 values of <20 nM (45). At a concentration of 1 μM, staurosporine partially reduced ANG II-mediated inhibition of I AC from its control value of 82 ± 5% (n = 6) to 62 ± 11% (n = 8) (Fig. 11 B).
We have described the properties and regulation of a unique ATP-activated ion channel in bovine AZF cells. UnitaryI AC channels were shown to be K+selective with no measurable Na+ or Ca2+conductance. I AC channels were outwardly rectifying in the presence of physiological concentrations of M2+, but conductance became almost ohmic in the absence of M2+ on either side of the membrane.
The potent activation of I AC channels by poorly hydrolyzable nucleotides and PPPi and the failure of ANG II to inhibit these channels when these agents replace ATP in the pipette solution are consistent with a model for I ACgating coupled to an ATP hydrolysis cycle. Overall, these results identify I AC as a distinctive new type of K+ channel with properties that would allow it to couple hormonal and metabolic signals to membrane potential and cortisol production.
Nucleotide-dependent activation of IAC channels in membrane patches.
Nucleotide-dependent activation of unitary I ACchannels in outside-out patch recordings resembled, in time course and concentration dependence, previous results obtained in whole cell recordings (14). Specifically, I ACchannels were active in outside-out patches only when ATP or UTP were present in the patch pipette at millimolar concentrations. These results are consistent with the idea that nucleotide-dependent activation of these channels is independent of cytoplasmic proteins that are absent in outside-out patches and of protein kinases for which UTP is a poor substrate.
The failure of MgATP to activate I AC channels in excised inside-out patches suggests that key regulatory factors or associated membrane proteins are lost during this form of patch excision. This result is not surprising in view of the rich combination of metabolic factors that control the activity of this K+channel (14, 16, 19, 39). In particular, because the molecular identity of I AC channels has not been determined, we do not know whether nucleotide binding sites are located on the pore-forming proteins or an associated subunit. The inability to directly activate I AC K+ channels in inside-out patches limits studies exploring the modulation of these channels. For example, we have been unable to describe the temporal pattern and reversibility of ATP-mediated I ACactivation.
K+ selectivity, conductance, and rectification of IAC channels.
Although the molecular structure of I AC and its relationship to other channels is unknown, the results of the present study identify it as a true K+-selective channel with no measurable conductance to Na+ or Ca2+. In symmetrical K+ solutions, unitaryI AC currents reversed at potentials slightly positive to 0 mV. This probably occurred through a reduction in pipette K+ activity through binding to ATP or BAPTA.
The identification of I AC as an authentic K+-selective channel is significant in view of the aforementioned similarities to CNG cation channels with respect to pharmacology and gating by cAMP (16, 20, 43). The lack of measurable Na+ and Ca2+ conductance clearly distinguishes I AC from CNG cation channels andDrosophila eag K+ channels (11).I AC K+ channels may represent a new intermediate form in a continuum linking true K+-selective and CNG cation channels.
M2+ and rectification.
Inwardly rectifying ATP-sensitive K+ (KATP) channels comprise a major class of K+-selective channels that include two rather than six membrane-spanning regions. Although KATP channels resemble I AC channels with respect to gating by ATP, KATP channels are uniformly inhibited, rather than activated, by the nonhydrolytic binding of ATP (5). In addition, KATP channels are sensitive to sulfonylurea agonists and antagonists, whereasI AC channels are not (1, 5, 20).
In the present study, a third fundamental difference betweenI AC K+ channels and KATPchannels was identified. Specifically, in symmetrical solutions, KATP channels are inwardly rectifying due to the unidirectional block of K+ efflux by Mg2+ at physiological concentrations (5, 28). By comparison, in the presence of physiological concentrations of internal and external M2+ and symmetrical K+,I AC channels are outwardly rectifying. Removal of intracellular Mg2+ showed that outflow of K+is only weakly blocked by this divalent cation, an effect that is more than matched by block of K+ influx by Ca2+ and Mg2+. Accordingly, when these two M2+ are deleted from external and pipette solutions, I ACchannels conduct K+ almost equally well in either direction.
The relative contributions of Ca2+ and Mg2+ to block of K+ influx were not determined in our experiments. K+ efflux through I AC channels is dramatically inhibited by intracellular Ca2+ at a concentration of only 2 μM (19). However, this effect appears to occur through an action on I AC gating rather than permeation.
Overall, although they are both ATP-gated and K+-selective channels, I AC and KATP channels differ in several fundamental aspects. It is unlikely that they belong to the same family of K+ channels.
Activation of IAC channels by nucleotides and inorganic phosphates.
At concentrations ≤1 mM, nucleotides including AMP-PNP, NaUTP, NaCTP, and NaTTP were much more effective than MgATP as activators ofI AC. In a model for I ACgating in which nucleotide binding leads to channel opening and nucleotide hydrolysis or dissociation allows the channel to close, nonhydrolyzable nucleotides could be more potent through a reduction in the effective “off rate,” lowering the dissociation constant. Accordingly, NaUTP was at least 10-fold more potent than MgATP as an activator of I AC.
The activation of I AC channels by PPPi is also consistent with a model forI AC gating controlled by an ATP hydrolysis cycle. ATP hydrolysis by ATPases involves cleavage of the phosphate bond, followed by liberation of Pi. Inorganic phosphates bind to sites from which Pi is released after cleavage of ATP, interrupting further ATP hydrolysis cycles (7, 12,22). If I AC closing is coupled to the hydrolysis of ATP and subsequent dissociation of organic phosphate from its binding site, then these same agents may stabilizeI AC channels in the open or bursting configuration.
I AC K+ channels appear to be the first K+ channel yet described whose activity is enhanced by ATP and other nucleotides, as well as inorganic phosphates, through a mechanism not involving protein kinases. In this regard, the gating of I AC channels resembles that of the CFTR Cl− channel in that hydrolysis-resistant ATP analogs and polyphosphates stabilize the open state of the channel, locking it into a prolonged open or bursting state (7, 22). Thus the gating of this ATP-activated Cl− channel may also be tightly coupled to an ATP hydrolysis cycle. However, the biochemical mechanisms controlling the activity of CFTR Cl− channels are quite complex and not well understood. While an ATP hydrolysis cycle seems to be involved, previous phosphorylation by A-kinase is a requirement for activation (7, 30). The overall similarity to I AC gating is not yet clear.
Although the activation of I AC K+channels by poorly hydrolyzable nucleotides and polytriphosphates is consistent with a model for I AC gating involving an ATP hydrolysis cycle, the results raise questions regarding the nature of the binding sites involved. First, if the nucleotide binding site is specifically designed to accommodate ATP, it is surprising that pyrimidine nucleotide triphosphates including UTP, CTP, and TTP also bind to the site with similar affinity.
Second, if inorganic phosphates bind to a channel-associate site on the ATPase that is normally occupied by Pi, why is PPPi much more effective than Pi or PPi as an activator of I AC channels? The structure-activity results involving nucleotides and inorganic phosphates indicate that it is the triphosphate group itself that is the critical moiety in the activation of I ACchannels. Currently, there is no satisfactory explanation for this order of effectiveness. PPPi and PPi have also been shown to be quite effective in activating CFTR Cl−channels (7, 22).
The failure of the anionic phospholipid PIP2 to activateI AC channels indicates that the presence of multiple phosphate groups in a molecule does not ensure that it will activate this current. It represents a further distinction betweenI AC and other ATP-sensitive K+channels that are uniformly activated by PIP2 (8, 17,29, 44).
Nucleotides and membrane potential.
Good correlation exists between the activation ofI AC K+ channels by nucleotides and polyphosphates and the magnitude of the V minstalled by these same agents in AZF cells. This result provides evidence that I AC channels are primarily responsible for setting V m. Furthermore, single-channel recordings showed that I ACchannels remain active at very negative membrane potentials, as expected for a channel that sets V m near the K+ equilibrium potential.
A direct relationship exists among nucleotide triphosphate concentration, I AC activity, and membrane potential. This suggests a specific mechanism whereby membrane potential, Ca2+ entry, and cortisol secretion could be linked to the metabolic state of the cell and, therefore, to other variables such as blood glucose concentration. Cortisol is a glucose counterregulatory hormone that acts in opposition to insulin in maintaining blood glucose levels (9).
Mechanism for ANG II-mediated inhibition of IAC.
The failure of ANG II to inhibit I AC when NaUTP or PPPi replaced ATP in the pipette is consistent with our previous observation that ANG II was ineffective in the presence of the nonhydrolyzable ATP analog AMP-PNP (40). However, the requirement for ATP hydrolysis could signal the involvement of either an ATPase or a protein kinase. The failure of the addition of 0.05 or 1 mM MgATP to the pipette (in addition to 5 mM UTP) to restore inhibition by ANG II argues for the involvement of an ATPase rather than a kinase in this response. Nearly all kinases are fully activated by the substrate MgATP at concentrations of 50 μM, whereas cellular ATPases have higher K m values for ATP (18, 27,34). Thus, if ANG II-mediated inhibition ofI AC occurred through activation of a protein kinase, low concentrations of ATP should have been sufficient to restore this effect.
Results of experiments with multiple protein kinase antagonists support the conclusion that no protein kinase known to be activated by ANG II mediates inhibition of I AC. In various cells, including those of the adrenal cortex, ANG II acts through AT1 receptors to activate protein kinase C, tyrosine kinases, JAK/STAT (signal transducers and activators of transcription) kinases, and MAP kinases (4, 36, 38, 46, 49). The failure of specific antagonists of each of these kinases to attenuateI AC inhibition by ANG II indicates that none of these is involved in this response.
The nonselective protein kinase antagonist staurosporine (1 μM) reduced the inhibition of I AC by ANG II from 82% to only 62%. At this concentration, staurosporine completely inhibits a wide range of protein kinases, including serine/threonine kinases, and tyrosine kinases (45). It is possible that an unidentified staurosporine-sensitive protein kinase contributes toI AC inhibition by ANG II.
Identity and function of IACK+ channels.
Although the results of this study identify I ACas a true K+-selective ion channel, its molecular structure and relationship to other K+ channel gene families is unknown. Its weak voltage dependence, insensitivity to sulfonylureas and PIP2, and lack of inward rectification suggest that it does not belong to the six-membrane-spanning, voltage-gated channels or to the two-membrane-spanning inward rectifiers. In this regard, a new family of K+-selective channels with two pore domains in tandem has been identified in organisms ranging from yeast to humans (31, 32). A number of these noninactivating, outwardly rectifying channels display properties similar to those ofI AC.
Regardless, the convergent inhibition of I AC by multiple G protein-coupled receptors through second messengers, including Ca2+ and cAMP, and the activation of these K+ channels by ATP at physiological concentrations indicate that I AC is a central control point where hormonal and metabolic signals are transduced to electrical events involved in cortisol secretion. In this scheme, the control ofI AC K+ channel activity through a cycle of ATP binding and hydrolysis may be a fundamental mechanism linking biochemical signals to AZF cell membrane potential.
In this regard, ATP-activated I AC K+channels provide an interesting contrast to KATP channels of insulin-secreting cells, which are inhibited, rather than activated, by ATP. The opposing actions of ATP on the activity of these two metabolic sensors are consistent with their function in regulating insulin and cortisol secretion. In pancreatic β-cells, high blood glucose levels lead to elevated ATP, KATP inhibition, membrane depolarization, Ca2+ entry, and insulin secretion (1, 5). In bovine AZF cells, elevated glucose would be associated with I AC activation, suppressing Ca2+ entry and cortisol secretion. Accordingly, cortisol is secreted under conditions of metabolic stress, where insulin secretion is suppressed (9).
This work was supported by National Institute of Diabetes and Digestive and Kidney Diseases Grant DK-47875 (to J. J. Enyeart).
Address for reprint requests and other correspondence: J. J. Enyeart, Dept. of Neuroscience, The Ohio State Univ., College of Medicine, 5190 Graves Hall, 333 W. 10th Ave., Columbus, OH 43210-1239 (E-mail:).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2001 the American Physiological Society