G-protein regulation of outwardly rectified epithelial chloride channels incorporated into planar bilayer membranes.

Experiments were designed to test if immunopurified outwardly rectified chloride channels (ORCCs) and the cystic fibrosis transmembrane conductance regulator (CFTR) incorporated into planar lipid bilayers are regulated by G-proteins. pertussis toxin (PTX) (100 ng/ml) + NAD (1 mM) + ATP (1 mM) treatment of ORCC and CFTR in bilayers resulted in a 2-fold increase in single channel open probability (Po) of ORCC but not of CFTR. Neither PTX, NAD, nor ATP alone affected the biophysical properties of either channel. Further, PTX conferred a linearity to the ORCC current-voltage curve, with a slope conductance of 80 +/- 3 picosiemens (pS) in the +/- 100 mV range of holding potentials. PKA-mediated phosphorylation of these PTX + NAD-treated channels further increased the Po of the linear 80-pS channels from 0.66 +/- 0.05 to >0.9, and revealed the presence of a small (16 +/- 2 pS) linear channel in the membrane. PTX treatment of a CFTR-immunodepleted protein preparation incorporated into bilayer membranes resulted in a similar increase in the Po of the larger conductance channel and restored PKA-sensitivity that was lost after CFTR immunodepletion. The addition of guanosine 5'-3-O-(thio)triphosphate (100 mum) to the cytoplasmic bathing solutions decreased the activity of the ORCC and increased its rectification at both negative and positive voltages. ADP-ribosylation of immunopurified material revealed the presence of a 41-kDa protein. These results demonstrate copurification of a channel-associated G-protein that is involved in the regulation of ORCC function.

Cystic fibrosis (CF) 1 is an autosomal recessive disease that is common in North America. A characteristic of the disease is impaired Cl Ϫ transport across several tissues including those of the airway epithelia. The gene responsible for impaired Cl Ϫ transport encodes the cystic fibrosis transmembrane conductance regulator (CFTR), a protein that acts as a small linear Cl Ϫ channel at the plasma membrane. A second Cl Ϫ channel in the apical membrane of affected tissue (the so-called outwardly rectified chloride channel or ORCC) is also affected in CF such that it cannot be activated by PKA and ATP in cells with the CF phenotype. The presence of a functional CFTR in the membrane is required for the PKA/ATP activation of the ORCC (1)(2)(3). The biophysical properties of ORCC and CFTR Cl Ϫ channels are distinct and well established. ORCCs have a nonlinear current-voltage (I/V) relationship with a 20 -40-pS single-channel conductance at hyperpolarizing voltages and a 60 -80-pS conductance at depolarizing voltages (3)(4)(5)(6)(7)(8). ORCCs are blocked by a wide variety of molecules including DIDS and the calixarenes (9) and have a halide permeability sequence of I Ϫ Ͼ Cl Ϫ Ͼ Br Ϫ . ORCCs can be activated by PKA and protein kinase C (10 -13). Conversely, the CFTR Cl Ϫ channel has a linear I/V relationship with an 8 -16-pS single-channel conductance. Channel activity can be blocked by diphenylamine-2-carboxylic acid (DPC) and glibenclamide (13), but not by DIDS. The halide permeability sequence for CFTR is Br Ϫ Ͼ Cl Ϫ Ͼ I Ϫ (14,15).
Heterotrimeric GTP-binding proteins (G-proteins) regulate ion channels in a variety of tissues including respiratory epithelia (16). Regulation can be direct or indirect through a cytoplasmic pathway involving second messengers and protein kinases (16,17). Heterotrimeric G-proteins can also regulate intracellular vesicle trafficking (18). Both regulatory mechanisms (direct regulation and regulation of endo-and exocytosis) have been implicated in cAMP-dependent Cl Ϫ secretion in normal and CF epithelia (17,18). It was shown in whole-cell patch clamp studies (18) that pertussis toxin (PTX), which uncouples heterotrimeric G i proteins from their receptors, increases Cl Ϫ transport and restores cAMP-activated Cl Ϫ currents in airway epithelial cells isolated from CF patients. Additional studies suggested that the heterotrimeric G-protein G␣i-2 regulates CFTR Cl Ϫ conductance in human airway epithelial cells by modulating vesicle trafficking and the delivery of CFTR Cl Ϫ channels from an intracellular vesicular pool to the plasma membrane (18). The same authors also demonstrated that the only pertussis toxin-sensitive G-protein expressed in human airway cells was G␣i-2 (18). As the apical membrane of airway epithelial cells contains two cAMP-activated Cl Ϫ channels (ORCC and CFTR), the possibility exists that both of them are regulated by G-proteins. G␣i-2 inhibits CFTR function solely by preventing trafficking of the protein to the apical membrane. However, G␣i-2 may regulate ORCCs by a direct mechanism independent of second messenger involvement (17,18).
We recently reported the simultaneous isolation and functional reconstitution of an ORCC and CFTR from apical membrane vesicles of bovine tracheal epithelial cells (3). Isolated channels were functionally preserved and exhibited similar regulatory features as native channels recorded in patch clamp studies from airway epithelia. In light of these observations, the goals of this study were to determine whether a PTXsensitive G-protein copurified with these Cl Ϫ channels and to * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18  examine the regulatory relationship between the isolated channels and the copurified G-protein.

Materials
[ 32 P]NAD was obtained from NEN-DuPont. A hydrazidederivatized Acti-Disk was obtained from FMC Corporation (Pine Brook, NJ). Peroxide-free Triton X-100, and SM2-Biobeads were from Bio-Rad. Phospholipids were obtained from Avanti Polar Lipids (Birmingham, AL). All other reagents were of analytical grade and were purchased from Sigma, Bio-Rad, or Fisher (Pittsburgh, PA).

Methods
Tracheal Apical Membrane Preparation-Apical membrane vesicles were prepared by differential centrifugation using a procedure first described by Langridge-Smith et al. (20), modified as described previously (21). Aliquots of vesicles (average protein concentration, 5 mg/ml) were stored in liquid nitrogen until use. Extraction of peripheral proteins from native membrane vesicles was achieved by incubation of vesicles for 30 min in KCl buffer (100 mM KCl, 5 mM Tris/Hepes, and 0.5 mM MgCl 2 ) titrated to pH 10.8 with 0.1 M NaOH. Alkaline-stripped vesicles were recovered by centrifugation at 35,000 ϫ g for 35 min. These peripheral protein-extracted apical membrane vesicles were next solubilized with Triton X-100 (0.8%) in the presence of KCl buffer (at pH 7.4). Quantitation of solubilized protein was performed using the BCA method (Pierce).
Immunopurification of Cl Ϫ Channel Proteins with p38 Antibodies-A polyclonal rabbit antibody generated against a previously purified, reduced 38-kDa anion channel protein (p38; Ref. 21) was covalently linked to a hydrazide-activated disk. Purified immune IgG was oxidized by sodium metaperiodate (0.02 M) in the dark at room temperature for 1 h and coupled to the disk by recirculating overnight at room temperature to reach maximum binding capacity (15-20 mg of IgG/disk). Prior to use, the disk was washed extensively with 1 M NaCl, 10 mM Na 2 HPO 4 , pH 7.4. Solubilized prewashed apical membrane vesicles or fractions from the cation exchange CM column were diluted in 40 ml of the above buffer and recirculated over the disk for 1 h to allow binding of the protein to the antibody. The disk was then extensively washed with NaCl/phosphate buffer before elution with 100 mM glycine, pH 3. Successive 2-ml fractions were collected and immediately neutralized to pH 7.4 with 100 l of 1.5 M Tris. The first 15 fractions contained the most protein and were pooled and concentrated to 100 l and used for reconstitution in liposomes or for further biochemical characterization.
Polyacrylamide Gels-Protein separation on polyacrylamide gels was performed using the method of Laemmli (22). Each gel was calibrated by simultaneously running M r standards in a parallel lane. The apparent M r of the unknown proteins was calculated from appropriately constructed log M r versus relative mobility (R f ) curves.
Immunoprecipitation of CFTR-Immunoprecipitation of CFTR from p38-immunopurified bovine tracheal proteins was performed as described previously (3,19) using a polyclonal anti-CFTR NBD1 antibody. Subsequent labeling and analysis of precipitated proteins with [␥-32 P]ATP was done as described previously (3,19,23). Inhibitory anti-CFTR 505-511 antibodies were obtained as a kind gift from Drs. J. Dedman and M. Kaetzel (University of Cincinnati).
Pertussis Toxin-induced ADP-ribosylation-Solubilized tracheal apical membrane (20 g of protein) or immunopurified material (p38) with or without CFTR was subjected to ADP-ribosylation as described previously (24). Ribosylation was carried out in a buffer containing 10 mM Tris base (pH 7.5), 1 mM EDTA, 20 mM thymidine, 1 mM ATP, and 0.01 mM GTP. The assay was started by the addition of [ 32 P]NAD to a final concentration of 10 M and was carried out at 32°C for 45 min. 20 l of 2 ϫ Laemmli's sample buffer was added to 20 l of the incubation mixture and electrophoresed. Gels were dried and exposed to Kodak X-Omat AR film at Ϫ70°C for appropriate lengths of time (12-48 h).
Planar Lipid Bilayer Experiments-Reconstitution of immunopurified Cl Ϫ channel proteins into proteoliposomes was performed as described previously (25,26). Concentrated protein samples were mixed with a phospholipid mixture (phosphatidylethanolamine, phosphatidylserine, phosphatidylcholine at a ratio of 50:30:20 (w/w/w)) in the presence of 400 mM KCl, 0.5 mM MgCl 2 , and 5 mM Tris/HCl, pH 7.4. The final volume was 600 l, and the protein:phospholipid ratio was 1:10 (w/w). Triton X-100 was removed by incubation of the samples with 150 mg of Bio-Beads SM-2 at room temperature for 45 min, followed by incubation at 4°C overnight. Proteoliposomes were separated from the beads using a 1-ml syringe fitted with a 27-gauge needle.
Planar lipid bilayers, composed of a mixture of 25 mg/ml diphytanoyl-phosphatidylethanolamine, diphytanoyl-phosphatidylserine, and oxidized cholesterol in a 2:1:2 (w/w/w) ratio, were painted with a fire-polished glass capillary over a 200-m hole drilled in a polystyrene chamber, as described previously (26). Bilayer formation was ascertained by an increase in membrane capacitance to a final value of 300 -400 picofarads. Artificial liposomes containing anion channel proteins were incorporated into bilayers bathed with symmetrical solutions of 100 mM KCl and 10 mM MOPS adjusted to pH 7.4. The observation of a stepwise increase in current was taken as an indication of channel incorporation into the lipid bilayer. Current measurements were performed with a high gain amplifier circuit based on a design previously described (25). Steady-state single channel current-voltage (I/V) curves were measured after channel incorporation by applying a known voltage and measuring individual channel current. Single channel records were analyzed using pCLAMP software (Axon Instruments, CA), as described previously (25,26). Single channel data were stored digitally and for analysis were filtered at 500 Hz with an 8-pole Bessel filter and acquired at 1 ms/point. The dashed line in the figures represents the zero current level.

Identification of a Pertussis-sensitive G Protein in Tracheal
Membranes-To determine if an inhibitory G-protein was present in our preparation, we examined the ability of PTX to ADP-ribosylate both solubilized tracheal membrane vesicles and the partially purified tracheal material that was used for incorporation into planar bilayers. ADP-ribosylation with PTX revealed the presence of a single band migrating at 40 -41 kDa in solubilized tracheal membrane vesicles and in the protein material that contained both the ORCC and CFTR ( Fig. 1; n ϭ  4). A band at 40 -41 kDa was also detected in the remaining material after CFTR precipitation (Fig. 1).
Effects of GTP␥S, PTX, and PKA on ORCC and CFTR in Planar Lipid Bilayers-In order to determine if G-proteins were involved in ORCC and CFTR channel regulation, we examined whether the addition of agents affecting the biochemical and functional state of G-proteins influenced the behavior of tracheal anion channels incorporated into planar lipid bilayers. In the absence of PKA ϩ ATP (control), only a single, outwardly rectified channel could be identified (Figs. 2-4 to DIDS but was sensitive to DPC, a known blocker of CFTR (Fig. 2). When inhibitory anti-CFTR 505-511 antibodies were added subsequent to cis addition of PKA ϩ ATP, the P o of the incorporated ORCC was 0.26 Ϯ 0.05 (n ϭ 9), not significantly different (p Ͼ 0.01) from that before PKA ϩ ATP addition. The channel activity of CFTR (under conditions where the ORCC was inhibited by DIDS) was not appreciably altered in separate experiments by the addition of GTP␥S (P o ϭ 0.58 Ϯ 0.07, n ϭ 9, as opposed to P o ϭ 0.63 Ϯ 0.08, n ϭ 9). We previously demonstrated that inhibitory anti-CFTR 505-511 antibodies inhibited CFTR channel activity and did not affect channel activity of the ORCC (19). Thus, these antibodies were used to facilitate the calculation of the P o of the ORCC by eliminating CFTR channel activity.
The addition of PTX (100 ng/ml), an agent known to inactivate G i proteins, together with NAD (1 mM) and ATP (1 mM), also significantly increased the P o of the ORCC from 0.38 Ϯ 0.03 to 0.66 Ϯ 0.05 (n ϭ 11, Fig. 3). Furthermore, PTX conferred a linearity to the ORCC current-voltage curve, which had a slope conductance of 80 Ϯ 3 pS at a holding potential of Ϯ100 mV (see Fig. 5). As with GTP␥S, PKA-mediated phosphorylation of these PTX ϩ NAD-treated channels revealed the presence of a small (16 Ϯ 2-pS), linear channel in the bilayer membrane (Fig. 3). Inhibition of the ORCC by DIDS afforded us the opportunity to compare the activity of phosphorylated CFTR following PTX treatment. There was no significant dif-ference in the channel P o of CFTR for either the GTP␥S-treated channel (P o ϭ 0.63 Ϯ 0.08 (n ϭ 9); Fig. 2) or the PTX-treated channel (P o ϭ 0.58 Ϯ 0.05 (n ϭ 9); Fig. 4).
To further explore the regulatory relationship between Gproteins and the ORCC, we immunodepleted CFTR from the immunopurified tracheal preparation prior to reconstitution of this material into the lipid bilayer. As previously shown, immunodepletion of CFTR did not affect the biophysical properties of the ORCC, although under these conditions the ORCC could not be activated by PKA ϩ ATP (3). The addition of PTX to the presumed cytoplasmic face of the incorporated channel in the absence of CFTR resulted in a similar increase in P o of the ORCC as had been observed in the presence of CFTR (Fig. 4). Moreover, PTX restored sensitivity to PKA ϩ ATP to the ORCC, as P o was increased from 0.63 Ϯ 0.06 to 0.91 Ϯ 0.08 (n ϭ 12) under these conditions, even in the absence of CFTR (Fig.  4). Furthermore, PTX treatment also conferred a linearity to ORCC, as was the case in the presence of CFTR. I/V curves, derived from these experiments are shown in Fig. 5. It is clear from these plots that PKA-mediated phosphorylation of ORCC partially restored rectification properties after the addition of PTX. This effect of PTX was independent of the presence or absence of CFTR. GTP␥S-treated ORCC, in contrast, was significantly more rectified at both negative and positive holding potentials (100 pS at ϩ40 to ϩ100 mV and 16 pS at Ϫ40 to Ϫ100 mV). Interestingly, the negative branch of the I/V curve of ORCC in the absence of CFTR was almost identical to the I/V curve of ORCC treated with GTP␥S in the presence of CFTR.

DISCUSSION
The results presented in this study show that a pertussis toxin-sensitive G-protein copurifies with bovine tracheal ORCC and CFTR channels and is involved in the direct regulation of the ORCC. We have previously demonstrated that solubilized apical membrane vesicles from bovine trachea, immunopurified with an antibody raised against an anion channel protein, contain both an ORCC and CFTR (3). The addition of PKA ϩ ATP to the cis (or cytoplasmic) bathing solution of a bilayer containing this material activated a large (80 pS at ϩ80 mV) conductance, outwardly rectified anion channel that was sensitive to DIDS. Inhibition of this 80-pS channel with 100 M DIDS in the presence of PKA and ATP revealed a second, low conductance anion channel (16 pS) that exhibited a linear I/V relationship and that was sensitive to DPC. These data suggested that the immunopurified tracheal material contained both a functional ORCC and a CFTR. Moreover, these channels maintained a regulatory relationship even after this harsh purification procedure, e.g. the presence of CFTR was required for PKA activation of ORCC (3). In contrast, the recent model proposed to explain regulation of Cl Ϫ transport via ORCC and CFTR in epithelial cells (31,32) includes an amplification cascade initiated by an initial interaction of extracellular ATP with a G-protein-coupled P 2U receptor. Our studies with immunopurified proteins reconstituted into planar lipid bilayers demonstrated that the ORCC could be activated by PKA in the presence of G551D mutant CFTR, but only when ATP was added to both sides of the channel-containing bilayer, consistent with the external ATP stimulation part of this model (19).
Several other studies report copurification of channels or receptors with associated G-proteins (24,(27)(28)(29). In our preparation, a PTX-sensitive, ADP-ribosylated G-protein remained with the ORCC complex after CFTR precipitation. A direct regulatory relationship between the ORCC and a G-protein was confirmed by the addition of GTP␥S or PTX to the ORCC either by itself or incorporated together with CFTR. The addition of pertussis toxin, which prevents the dissociation of the heterotrimeric G-protein complex, activated the ORCC. The addition of GTP␥S decreased the activity of the ORCC in the planar lipid bilayer. Uncoupling of the ORCC from the Gprotein also altered the rectification properties of the channel, suggesting that channel rectification is influenced at least in part by its interaction with a G-protein. Interestingly, the ORCC regained PKA sensitivity after the addition of PTX, even in the absence of CFTR. These results are consistent with the findings of Schwiebert et al. (18) that G-proteins are involved in the regulation of Cl Ϫ transport in airway epithelia. It would appear that uncoupling of the effector (in our case ORCC) from its associated G-protein restores PKA sensitivity to the channel. It is known that phosphorylation of the ␤-receptor plays a role in desensitization of this receptor (30). Phosphorylation in this case prevents interaction between a G-protein and the phosphorylated ␤-adrenergic receptor and stops further mediation of signals from the stimulated ␤-receptor (30). A similar mechanism between the ORCC and a G-protein may also exist. Phosphorylation of the channel may uncouple a G inhibitory (G i ) protein from the channel and thus activate it. Dephosphorylation of the channel would be predicted to permit the G i protein to interact with the channel again and to inhibit it. However, after activation with PTX (uncoupling of the channel from the G-protein; Ref. 31), additional stimulation with PKA and ATP was possible, suggesting that complete uncoupling of the channel from a G-protein may uncover additional phosphorylation sites on the channel protein. Moreover, these findings indicate that there are at least two major independent regulatory inputs that converge on the ORCC.
Recently, Schwiebert et al. suggested that ATP transported through CFTR acts as an autocrine stimulator of the ORCC (32). The proposed mechanism of regulation is via a P 2U receptor that, either through a direct coupling to the ORCC or through a G-protein-coupled signaling pathway, stimulates the ORCC. We have previously shown that the active form of CFTR is required for PKA-mediated activation of ORCC (19), while in the present study we have demonstrated the presence of a G-protein in our preparation. However, in contrast to the findings of Guggino and co-workers (32), namely that extracellular ATP at nanomolar concentrations stimulates the ORCC in normal or CF cells, we found that under our experimental conditions ATP had no effect on channel activity if CFTR was not present in the membrane. Moreover, ATP had no effect on the ORCC if PKA ϩ ATP was not present on the cytoplasmic side of incorporated channels. There are several possible explanations for our results. One possibility is that ATP (on the cis side of the bilayer) may bind to the ORCC, which subsequently interacts with CFTR and is thereby activated. Interaction of the ORCC with CFTR may allosterically diminish interaction of the ORCC with its G-protein, as both the G-protein (inhibition) and CFTR (activation) regulate the ORCC and have opposite effects on the channel. Rectification of the ORCC may be in part due to G-protein coupling, because treatment with PTX confers a linearity to the otherwise outwardly rectified behavior of the channel. CFTR also affects rectification, because when CFTR was present, the ORCC exhibited less rectification in the negative voltage quadrant. This finding supports the hypothesis that both molecules (CFTR and a G-protein) are coupled to the ORCC and have opposite effects on rectification (i.e. G-protein coupling induces rectification, whereas CFTR decreases rectification).
A second possibility to consider is that ATP may bind to CFTR directly. Alternatively, a receptor for ATP (purinergic receptor) could be strongly associated with CFTR. If this hypothesis is correct, we may have copurified a purinergic receptor together with CFTR. However, it would seem unlikely that the appropriate signaling pathway would be maintained under our conditions. If, on the other hand, CFTR is both coupled to ATP and bound to a G-protein, ATP binding could uncouple the G-protein from its effector (in this case ORCC). This would allow the involvement of CFTR in the regulation of different processes through G-proteins. In either case, we have demonstrated that both a G-protein and CFTR are involved in the regulation of ORCC.
In summary, we have shown that a pertussis toxin-sensitive G-protein copurifies with ORCC and CFTR and that, after precipitation of CFTR, the G-protein remains in the material that contains the ORCC. The effects of PTX and GTP␥S on the ORCC were independent of the presence or absence of CFTR. PTX and GTP␥S did not have any effect on the CFTR Cl Ϫ channel activity under our experimental conditions (in presence of PKA and ATP from cytoplasmic side). These observations are consistent with the hypothesis that G-proteins directly regulate the ORCC, but not CFTR. Additionally, whereas we have previously demonstrated that precipitation of CFTR prevents the PKA ϩ ATP-dependent activation of the ORCC (3), the data presented in this study have shown that even in the absence of CFTR, the ORCC can be activated by PKA ϩ ATP as long as PTX is present on the cis side. This observation suggests a possible regulatory role of CFTR on the ORCC through a G-protein interaction.