Regulation of the Epithelial Na+ Channel by the RH Domain of G Protein-coupled Receptor Kinase, GRK2, and Gαq/11*

The G protein-coupled receptor kinase (GRK2) belongs to a family of protein kinases that phosphorylates agonist-activated G protein-coupled receptors, leading to G protein-receptor uncoupling and termination of G protein signaling. GRK2 also contains a regulator of G protein signaling homology (RH) domain, which selectively interacts with α-subunits of the Gq/11 family that are released during G protein-coupled receptor activation. We have previously reported that kinase activity of GRK2 up-regulates activity of the epithelial sodium channel (ENaC) in a Na+ absorptive epithelium by blocking Nedd4-2-dependent inhibition of ENaC. In the present study, we report that GRK2 also regulates ENaC by a mechanism that does not depend on its kinase activity. We show that a wild-type GRK2 (wtGRK2) and a kinase-dead GRK2 mutant (K220RGRK2), but not a GRK2 mutant that lacks the C-terminal RH domain (ΔRH-GRK2) or a GRK2 mutant that cannot interact with Gαq/11/14 (D110AGRK2), increase activity of ENaC. GRK2 up-regulates the basal activity of the channel as a consequence of its RH domain binding the α-subunits of Gq/11. We further found that expression of constitutively active Gαq/11 mutants significantly inhibits activity of ENaC. Conversely, co-expression of siRNA against Gαq/11 increases ENaC activity. The effect of Gαq on ENaC activity is not due to change in ENaC membrane expression and is independent of Nedd4-2. These findings reveal a novel mechanism by which GRK2 and Gq/11 α-subunits regulate the activity ENaC.

involve multiple phosphorylated effector targets, which are not GPCRs.
The present study was inspired by our observation that a GRK2 mutant that lacks kinase activity was effective in up-regulating the basal activity of ENaC. We found that this stimulatory effect of GRK2 on ENaC can be attributed to the ability of the RH domain of GRK2 to interact with and inhibit the ␣-subunit of Gq/11, which acts as a negative regulator of ENaC.

EXPERIMENTAL PROCEDURES
DNA Constructs-Wild-type mouse ␣-, ␤and ␥-ENaC and ␣-ENaC clones containing HA at the N-terminal and V5 at the C-terminal, all of which are in pcDNA3.1, were a gift from Thomas R. Kleyman (University of Pittsburgh). ␣-, ␤-, and ␥-ENaC subclones containing FLAG at the C-terminal in pcDNA3.1 were provided by Angeles Sanchez-Perez (University of Sydney, Australia). WtGRK2, K220R GRK2, D110A GRK2, RH domain of GRK2, wtG␣q, and Q209L G␣q in pcDNA3 were gifts from Philip B. Wedegaertner (Thomas Jefferson University). ⌬RH-GRK2 in pcDNA3.1 was a gift from Jerrold M. Olefsky (University of California). Q209L G␣11, Q205L G␣14, and Q212L G␣15/16 in pcDNA3.1 were purchased from cDNA Resource Center (University of Missouri). The C-terminaltruncated ENaC mutants were generated by PCR-based methods performed using primers, which have stop codons at specific sites to generate Pro 646 stop ␣-ENaC, Cys 594 stop ␤-ENaC or Phe 610 stop ␥-ENaC.
Cell Culture and Transfection-Fischer rat thyroid (FRT) cells were a gift from Lucio Nitsch (University of Naples), M1 mouse collecting duct cells, originally generated by Stoos et al. (23), were a gift from Christoph Korbmacher (Universität Erlangen-Nürnberg, Germany) and HEK293T cells were purchased from the American Type Culture Collection. These cell types were cultured in Dulbecco's modified Eagle's/Ham's F-12 media with 100 units/ml penicillin and 100 g/ml streptomycin at 37°C. The media for FRT cells contained 5% fetal bovine serum, the media for HEK293T cells contained 10% fetal bovine serum whereas the media for M1 cells contained 10% fetal bovine serum and 100 nM dexamethasone. FRT and M1 cells were seeded onto Millicell-PCF permeable filter supports (Millipore, catalogue number PIHP01250). One day after seeding, FRT cells were co-transfected with cDNA of ␣-, ␤-, and ␥-mENaC-FLAG (0.7 g/ml each). When appropriate, FRT and M1 cells were transfected with cDNA of wtGRK2, mutants of GRK2, mutants of G␣q, G␣11 or G␣14 (3 g/ml), siRNA against GRK2, siRNA against Nedd4-2 or siRNA against G␣q, G␣11, or G␣14 (0.5 g/ml). siRNAs against GRK2 (5Ј-GGA CAC AAA AGG AAT CAA GTT-3Ј) and G␣q (5Ј-GGC TCA TGC ACA ATT GGT-3Ј) were obtained from Ambion and siRNAs against G␣11 (5Ј-CAC AAC TGG CAT CAT CGA GTA-3Ј), G␣14 (5Ј-TTG GTT GCT CTG AGT GAA TAT-3Ј) and Nedd4-2 (5Ј-AAC CAC AAC ACA AAG TCA CAG-3Ј) were obtained from Qiagen. In short, cDNA or siRNA were mixed with Lipofectamine-2000 (Invitrogen) in Opti-MEM reduced serum media (Invitrogen) and incubated for 20 min at room temperature before being transferred to the apical side of the monolayer and further incubated for 4 h at 37°C. The transfection medium was then replaced with Dulbecco's modified Eagle's/Ham's F-12 media containing fetal bovine serum and antibiotics. In addition, the medium also contained dexamethasone (100 nM) for M1 cells or amiloride (10 M) for FRT cells.
RT-PCR and Quantitative RT-PCR-Total RNA was isolated from FRT cells using TRI Reagent (Sigma). 2 g of total RNA was used for reverse transcription (RT) reactions. The following primers were used: G␣11 (forward, 5Ј-GAC CAG AAC AAG GCC AAT GC-3Ј; reverse, 5Ј-GAA GAT GAT GTT CTC CAG GTC-3Ј), G␣14 (forward, 5Ј-CAG AAT AAG GAA AAT GCC CAG-3Ј; reverse, 5Ј-CCT GGT CAT ATT CAC TCA GAG-3Ј), and GAPDH (forward, 5Ј-TGG CCT TCC GTG TTC CTA CC-3Ј; reverse, 5Ј-TGT AGG CCA TGA GGT CCA CCA C-3Ј). PCR amplification was performed for 30 cycles. A 10 l aliquot of the PCR product was then separated by electrophoresis on a 1.5% agarose gel and stained with ethidium bromide. Quantitative RT-PCR (QRT-PCR) reactions of newly synthesized RT products were carried out using KAPA SYBR FAST qPCR Kit (Kapabiosystems). Fluorescence detection was performed using the RG-3000 real-time PCR system (Corbett Research). Amplification conditions consisted of 40 cycles of 95°C for 15 s and 60°C for 1 min after incubation at 95°C for 5 min for enzyme activation. Reaction efficiencies and relative gene expression were analyzed using REST 2009 software program from Qiagen.
Immunoblot-Two days after transfection, cells were washed twice with phosphate-buffered saline before treatment with a lysis buffer containing (in mM) Tris-HCl (50), NaCl (150), EDTA (10), with 10% glycerol and 1% Triton X-100 plus Complete Protease Inhibitor mixture (Roche Applied Science). After the protein concentration of each lysate had been determined, equal amounts of protein lysate were loaded onto a 12% SDS-polyacrylamide gel. Following electrophoresis, the protein was transferred to a nitrocellulose membrane and incubated overnight with anti-GRK2 monoclonal antibody, anti-G␣q polyclonal antibody, or anti-␤-actin monoclonal antibody (Santa Cruz Biotechnology), anti-FLAG M2 monoclonal antibody (Sigma), or anti-HA monoclonal antibody (Cell Signaling). The blots were washed to remove unbound antibodies before incubation with a horseradish peroxidase-conjugated secondary antibody for 1 h. The blots were washed with a TBS buffer containing 0.1% Tween-20. The proteins of interest were then visualized using an ECL TM Western blotting kit (GE Healthcare) and quantitated by densitometric analysis using ImageJ (NIH). The data are representative of at least three experiments.
Quantitation of Amiloride-sensitive Na ϩ Current-After the monolayers became confluent, normally within 2-3 days after transfection, the Millicell-PCF insert was transferred to a modified Ussing chamber. Apical and basolateral surfaces of the monolayer were simultaneously perfused with a solution containing (in mM) NaCl (130), CaCl 2 (1), KCl (1), MgCl 2 (1), glucose (5), HEPES (10), pH 7.4, maintained at 37°C. The chamber was connected to a VCC MC6 multichannel voltage/current clamp amplifier, controlled and monitored using the Acquire & Analyze software (V2.3.177, Physiologic Instruments, San Diego, CA). Experiments were carried out under current clamp (open circuit) conditions as previously described (24). Transepithelial potential was measured with reference to the baso-lateral side of the epithelium and current pulses were injected to assess the transepithelial resistance. The equivalent shortcircuit current was calculated using Ohm's law and plotted by the software. Amiloride-sensitive equivalent short-circuit current (I ami ) was determined as the change in current following the addition of amiloride (10 M) to the apical bathing solution. Because of large variation of I ami between each batch of cells, all data were normalized by dividing the amiloride-sensitive short circuit current of each experiment by an average of the amiloride-sensitive short circuit current of at least 3 control experiments obtained from the same batch of cells in the same day. This ratio is reported as normalized amiloride-sensitive equivalent short-circuit current (I ami(normalized) ). The mean value of I ami of each control group is reported in the figure legends. Data for each experiment were obtained from at least three different batches of cells and are reported as mean Ϯ S.E. with the number of experiments in parentheses. Statistical significance was assessed using Student's t test.
Quantitation of Na ϩ /K ϩ ATPase Activity-M1 cell monolayers were mounted in a modified Ussing chamber bathed symmetrically with the physiological solution and the equivalent short-circuit current monitored. Activity of the Na ϩ /K ϩ -ATPase was determined as previously described (24). In short, amiloride (10 M) was added to the apical bathing solution to inhibit activity of ENaC. Nystatin (360 g/ml) was then added to the apical solution to permeabilize the apical membrane and to eradicate any rate limitation associated with the apical Na ϩ entry. Short circuit current (I sc ) was allowed to stabilize for 30 min before the addition of 1 mM ouabain, an inhibitor of the Na ϩ /K ϩ -ATPase, to the basolateral solution. The change in I sc following the addition of ouabain was used to estimate the activity of the Na ϩ /K ϩ -ATPase.
Surface Expression of ENaC-HEK293 cells were transfected with FLAG-tagged ␣-, ␤-, and ␥-mENaC. Two days after transfection, the cells were washed three times with ice-cold phosphate-buffered saline and then incubated for 30 min in 5 ml of cell-impermeant sulfo-NHS-SS-biotin solution (0.5 mg/ml; Pierce) at 4°C. The reaction was stopped by quenching with Tris-buffered saline. The cells were solubilized in lysis buffer, and the lysate centrifuged at 14,000 rpm for 5 min at 4°C. The supernatant was collected and mixed with 250 l of NeutrAvidin TM gel slurry (Pierce) before incubating, with gentle rocking, for 60 min at room temperature. After incubation, the sample was centrifuged at 1,000 rpm for 2 min. The precipitant containing the biotinylated proteins was washed five times with lysis buffer. Finally, the biotinylated proteins were eluted by the addition of SDS sample buffer and then analyzed by Western blot using anti-FLAG M2 monoclonal antibody or anti-HA monoclonal antibody.

GRK2 Activates ENaC by a Phosphorylation-independent
Mechanism-Our previous study in isolated mouse mandibular duct cells suggested that kinase activity of GRK2 up-regulates activity of ENaC and prevents down-regulation of the channel by Na ϩ feedback inhibition (17). To further investigate the role of GRK2 in regulation of ENaC activity, we transiently expressed GRK2 in mouse renal collecting duct (M1) cells that endogenously express ENaC by transfecting the cells with a plasmid containing wild-type GRK2 (wtGRK2) (Fig. 1A, lower panel). Activity of ENaC was determined from transfected cell monolayers in Ussing chamber experiments (Fig. 1, A and B). We found that over-expression of GRK2 significantly increased the normalized amiloride-sensitive Na ϩ current (1.53 Ϯ 0.17, n ϭ 6 versus 1.00 Ϯ 0.07, n ϭ 6 in untransfected cells; p Ͻ 0.05), consistent with GRK2 positively regulating the activity of ENaC as reported in our previous study (17). We, therefore, assessed whether the effect of GRK2 on the basal activity of ENaC in M1 cells is mediated by its kinase activity. We did so by overexpressing a kinase-dead GRK2 mutant ( K220R GRK2) (Fig. 1A), in which the invariant lysine residue 220 in the protein kinase catalytic domain was converted to an arginine to terminate kinase activity (25). We found that activity of ENaC in cells expressing the K220R GRK2 mutant (1.40 Ϯ 0.14, n ϭ 6) was significantly higher than that of the control group (p Ͻ 0.05) and that the effect of the K220R GRK2 mutant on ENaC was not significantly different from that of the wtGRK2 (Fig. 1B). Thus, the mechanism by which GRK2 stimulates the basal activity of ENaC is not associated with its kinase activity. Furthermore, expression of a GRK2 construct that lacks the N-terminal RH domain (⌬RH-GRK2) (26) was without any effect on the activity of ENaC (Fig. 1B). Together, these data suggest that the effect of GRK2 on the activity of ENaC in M1 collecting duct cells is independent of its action as a protein kinase but dependent on its RH domain.
To determine the mechanism by which the RH domain of GRK2 regulates ENaC, we first established a mammalian expression system that could be used for further investigation. We reconstituted ENaC in Fischer Rat Thyroid (FRT) cells by transiently expressing the ␣-, ␤-, and ␥-ENaC subunits. A similar approach has previously been used successfully to investigate regulation of ENaC (24,27). The cells were co-transfected with the wtGRK2 construct or with siRNA directed against GRK2 (Fig. 1C). In agreement with our studies in M1 cells ( Fig.  1, A and B), overexpression of GRK2 enhanced normalized amiloride-sensitive current in FRT cells (1.63 Ϯ 0.10, n ϭ 11 versus 1.00 Ϯ 0.05, n ϭ 6 in control ENaC-only transfected cells; p Ͻ 0.001), while siRNA-mediated knockdown of GRK2 expression significantly attenuated the amiloride-sensitive current compared with that of cells transfected with a scrambled siRNA (0.53 Ϯ 0.04, n ϭ 7 versus 1.02 Ϯ 0.05, n ϭ 6; p Ͻ 0.001, Fig. 1C). In addition, the siRNA directed against GRK2 inhibited expression of endogenous GRK2 by 50% (p Ͻ 0.05), whereas the scrambled siRNA was without any affect. As observed in M1 cells, the effect of GRK2 on the activity of ENaC was independent of its kinase activity. We found that normalized amiloride-sensitive current in cells expressing the K220R GRK2 mutant (1.58 Ϯ 0.09, n ϭ 6) was significantly higher than that of the control group (p Ͻ 0.001). Expression of the K220R GRK2 mutant increased normalized amiloride-sensitive current in ENaC-transfected FRT cells to a value that is not significantly different from that of cells expressing wtGRK2 (Fig. 1C). Thus, kinase activity is not essential for GRK2 to exert its stimulatory effect on the activity of ENaC in FRT cells. To confirm the role of the RH domain, cells were co-transfected with ENaC and the ⌬RH-GRK2 mutant (Fig. 1C). Under these conditions, the normalized amiloride-sensitive current (0.89 Ϯ 0.09, n ϭ 4) observed in the co-tranfected cells was not significantly different from that of the control cells expressing ENaC alone (1.00 Ϯ 0.05, n ϭ 6; p Ͼ 0.05). Subsequently, ENaC and a GRK2 mutant ( D110A GRK2), in which the aspartic acid at the position 110 in its RH domain was mutated to alanine, were co-expressed in FRT cells. This GRK2 mutant has been shown to be unable to interact with G␣q/11/14 (28,29). We found that, the D110A GRK2 mutant failed to alter activity of ENaC (Fig.  1C), supporting the conclusion that the observed effect of GRK2 on ENaC may be due to GRK2 preventing G␣q from inhibiting the activity of ENaC.
Gq/11 Family Proteins Inhibit Activity of ENaC-We then investigated the effect of G␣q/11 on ENaC activity by suppressing expression of G␣q using an siRNA in FRT cells co-transfected with ENaC ( Fig. 2A). We found that normalized amiloride-sensitive current in G␣q siRNA-transfected cells (1.25 Ϯ 0.08, n ϭ 6) was significantly higher than that of the control cells (1.00 Ϯ 0.05, n ϭ 8; p Ͻ 0.01). Next, FRT cells were co-transfected with ENaC and a constitutively active G␣q mutant, Q209L G␣q (30,31), which decreased the normalized amiloride-sensitive current in FRT cells (0.15 Ϯ 0.05, n ϭ 8; p Ͻ 0.001; Fig. 2A), suggesting that G␣q is a negative regulator of the channel. We then confirmed these findings in M1 collecting duct cells, in which ENaC is endogenously expressed, by showing that over-expression of the constitutively active Q209L G␣q mutant abolished activity of ENaC and the siRNA directed against G␣q significantly increased activity of the channel (Fig. 2B).
To determine whether ENaC is also regulated by the ␣-subunits of G␣11 and G␣14, which are known to interact with GRK2, we expressed constitutively active Q209L G␣11 and Q205L G␣14 mutants in FRT cells. Expression of Q209L G␣11 and Q205L G␣14 was confirmed by RT-PCR (Fig. 2, C and D, upper panel). qPCR analysis data (Fig. 2, C and D, lower panel) suggest that relative expression of Q209L G␣11 and Q205L G␣14 was increased by 4 -5 fold, whereas siRNAs decreased expression of their respective targets by over 80%. Moreover, we found that over-expression of Q209L G␣11 and Q205L G␣14 mutants inhibited the amiloride-sensitive Na ϩ current (Fig. 2E). Conversely, siRNAs directed against G␣11 and G␣14 attenuated expression of G␣11 and G␣14 (Fig. 2, C and E) and significantly increased activity of ENaC (Fig. 2E). In addition, expression of the constitutively active Q212L G␣15/16 mutant also inhibited the amiloride-sensitive Na ϩ current (data not shown). Hence, the G␣ of all of the Gq/11 family act as negative-regulators of ENaC activity.
To eliminate a possibility that the inhibitory effect of Gq on ENaC activity may be due to Gq inhibiting Na ϩ /K ϩ -ATPase function, the activity of which is required for generating a driving force for Na ϩ absorption via ENaC (32), we determined the effect of Gq on the ouabain-sensitive current which represents activity of the Na ϩ /K ϩ -ATPase. We found that expression of Q209L G␣q has no effect on the ouabain-sensitive current in M1 cells (Fig. 3, A and B).
It has been suggested that proteolysis of ENaC subunits by serine proteases is important for maturation and activation of ENaC (33)(34)(35). It is, therefore, conceivable that G␣q inhibits ENaC by attenuating proteolysis of the channel. To address this issue, we co-expressed, in HEK293T cells, an ␣-ENaC clone that has HA at the N-terminal and V5 at the C-terminal together with ␤and ␥-ENaC, which have FLAG at the C-terminal, and Q209L G␣q. It was previously shown that the ␣-ENaC is cleaved by serine proteases, yielding a 30 kDa N-terminal fragment (33). Our immunoblot analysis of the biotinylated membrane proteins, using anti-HA antibody, revealed the presence of N-terminal fragments of ␣-ENaC (Fig. 3C with control biotinylation in Fig. 3E). No biotinylated proteins were detected in the cytoplasmic fraction (Fig. 3F). A ratio between the density of the 30 kDa ␣-ENaC N-terminal fragment and that of 95 kDa uncleaved ENaC (Fig. 3D) suggests that overexpression of Q209L G␣q does not significantly change proteolysis of ␣-ENaC, hence, it is unlikely that inhibition of ENaC by G␣q observed in this study is due to the effect of G␣q on proteolysis of the channel.
The Effect of G␣q on ENaC Is Nedd4-2-independent-Several regulators of ENaC, including the hormones aldosterone, insulin, and vasopressin, and non-hormone regulators, such as intracellular Na ϩ , modulate activity of the channel by regulating the abundance of ENaC expressed at the cell membrane  (20). The mechanisms by which these regulators control membrane expression of ENaC converge at Nedd4-2 (20), an ubiquitin protein-ligase that triggers ubiquitin-dependent internalization of ENaC. Of particular interest is that cytosolic Na ϩ increases Nedd4-2-dependent internalization of ENaC (14) by a mechanism that involves activation of the ␣-subunit of the G protein Go (13). This suggests that Nedd4-2 may be involved in the Gq/11 protein-mediated signaling pathways that down-regulate ENaC activity. We investigated this possibility by generating mutant ENaC subunits, ␣ T , ␤ T -, and ␥ T -ENaC, which were truncated at the amino acids Arg-646, Cys-594, and Phe-610, respectively, to remove their C termini, including the proline-rich PY-motifs, which are the binding sites for Nedd4-2. Each of these mutated ENaC subunits was co-tranfected into FRT cells with appropriate wt-ENaC subunits, i.e. ␣ T ␤␥, ␣␤ T ␥, or ␣␤␥ T , or co-transfected i.e. ␣ T ␤ T ␥ T, with or without the FIGURE 3. G␣q has no effect on the activity of Na ؉ /K ؉ ATPase or proteolytic cleavage of ENaC. A, short circuit current recording of M1 cell monolayers transfected with Q209L G␣q or without (control). Amiloride (10 M) was applied to the apical bath solution to eliminate activity of ENaC before the apical membrane was permeabilized by nystatin (360 g/ml). constitutively active Q209L G␣q mutant. Activity of ENaC in these cell monolayers was then determined in Ussing chamber experiments. Consistent with the role of Nedd4-2 in regulating ENaC activity, the normalized amiloride-sensitive current in cells expressing the mutated ENaC subunits, which cannot interact with Nedd4-2, was significantly higher than that of the cells expressing wild-type ENaC subunits (Fig. 4A). We further found that co-expression of the Q209L G␣q mutant was equally as effective in inhibiting the amiloride-sensitive current in cells expressing the C-terminal-truncated mutant of ENaC as in those expressing wtENaC (Fig. 4A), which indicates that the inhibition of ENaC activity by G␣q is independent of the presence of the PY-motifs. This implies that it is not mediated by Nedd4-2. To confirm this finding, we inhibited expression of Nedd4-2 in FRT cells by transfecting the cells with siRNA directed against Nedd4-2. As was shown in our earlier study (27), this maneuver significantly increased the amiloride-sensi-tive current in the transfected cells (Fig. 4B). Expression of Q209L G␣q together with siRNA against Nedd4-2 decreased the normalized amiloride-sensitive current by 73% of that observed in a control group which was not transfected with the siRNA (Fig. 4B), hence, we conclude that the inhibitory effect of G␣q on ENaC is independent of Nedd4-2.
To determine whether G␣q inhibits ENaC by reducing membrane expression of the channel, we investigated the effect of overexpression of G␣q on the abundance of ENaC at the cell membrane by co-transfecting HEK293T cells with ENaC and wild-type G␣q or constitutively active Q209L G␣q mutant. Analysis of the abundance of ENaC present in membrane fractions isolated from these cells revealed that expression of the wtG␣q, Q209L G␣q mutant or wtGRK2 had no effect on membrane expression of ENaC (Fig. 4C).
To further investigate the mechanism by which G␣q/11 regulates the activity of ENaC, we used different pharmacological FIGURE 4. The effect of G␣q on ENaC is not dependent on Nedd4-2 or known common pathways. A, I ami (normalized) in FRT cells transfected with wild-type ␣-, ␤-, and ␥-ENaC (␣␤␥), wild-type ␤␥-ENaC with mutant ␣-ENaC that is truncated at amino acid Pro 646 (␣ T ␤␥), wild-type ␣␥-ENaC with mutant ␤-ENaC that is truncated at amino acid Cys 594 (␣␤ T ␥), wild-type ␣␤-ENaC with mutant ␥-ENaC that is truncated at amino acid Phe 610 (␣␤␥ T ) or co-transfected with ␣ T ␤ T ␥ T ENaC. The solid bar indicates that the cells were co-transfected with Q209L G␣q. The average I ami for the control group is 6.18 Ϯ 1.42 A/cm 2 (n ϭ 6). B, I ami (normalized) in FRT cells transfected with three subunits of ENaC (control) or co-transfected with ENaC and Q209L G␣q. The cells were co-transfected with siRNA against Nedd4 -2 (siRNA Nd4 -2) as indicated. The average I ami for the control group is 3.87 Ϯ 0.49 A/cm 2 (n ϭ 6). C, representative immunoblot analysis for FLAG-tagged membrane protein in HEK293T cells co-transfected with ENaC-FLAG and an empty pcDNA3.1 vector to keep the plasmid concentration the same for all experiments or co-transfected with ENaC-FLAG together with wtG␣q, Q209L G␣q, or wtGRK2. Proteins at the membrane surface of HEK293T cells were biotinylated and pulled down with NeutrAvidin and detected by immunoblotting using anti-FLAG M2 monoclonal antibody. The relative density of the FLAG-tagged membrane protein to that of the FLAG-tagged protein in total cell lysates was determined by densitometric analysis and shown in the lower panel. D, I ami (normalized) in FRT cells transfected with ENaC or co-transfected with ENaC and Q209L Gq as indicated. Solid bars indicate that cells were treated with U73122 (10 M), BIM (1 M), PD98059 (20 M), or staurosporine (100 nM). The average I ami for the control group is 4.25 Ϯ 1.01 A/cm 2 (n ϭ 5). ** and *** indicate p Ͻ 0.01 and 0.001, respectively. The number of experiments is shown in parentheses.
blockers of signaling proteins known to be downstream to G␣q/ 11, including phospholipase C␤ (PLC␤) inhibitor, U73122, protein kinase C (PKC) inhibitor, bisindolylmaleimide-1 (BIM), and ERK1/2 inhibitor, PD98059, to assess the effect of G␣q on ENaC (Fig. 4D). We found that these blockers do not change the inhibitory effect of the constitutively active Q209L G␣q mutant on the activity of exogenous ENaC expressed in FRT cells (Fig.  4D). Neither was the nonspecific kinase inhibitor, staurosporine, able to prevent the effect of Q209L G␣q on ENaC (Fig. 4D).
GRK2 Increases Activity of ENaC by Inhibiting G␣q-Our data suggest that the binding site for G␣q/11 on GRK2 is critical to the non-kinase mediated effect of GRK2 on ENaC. To confirm that GRK2 increases the activity of ENaC by binding G␣q, we suppressed expression of G␣q/11 in FRT cells by co-transfecting the cells with a combination of siRNAs directed against G␣q, G␣11, and G␣14. These cells were then tranfected with ENaC with or without wtGRK2. As shown in Fig. 5A, suppress-ing G␣q, G␣11, and G␣14 significantly increased activity of ENaC. Co-expression of wtGRK2 did not increase the activity of ENaC in the cells further. The lack of effect of GRK2 in cells in which expression of G␣q, G␣11, and G␣14 was knocked down suggests that the stimulatory effect of GRK2 on ENaC may be associated with activity of G␣q/11.
To determine whether GRK2 can overcome the inhibitory effect of G␣q on ENaC, we co-tranfected FRT cells with ENaC and the constitutively active Q209L G␣q mutant together with a construct containing the RH domain of GRK2 (GRK2-RH) (36). Expression of GRK2-RH significantly increased normalized amiloride-sensitive current in this cell type when compared with that of a control group without GRK2-RH (1.78 Ϯ 0.13, n ϭ 9 versus 1.00 Ϯ 0.04, n ϭ 8; p Ͻ 0.001, Fig. 5B). Expression of Q209L G␣q (0.05 g/ml) decreased the normalized amiloridesensitive current to 0.58 Ϯ 0.08 (n ϭ 6; p Ͻ 0.001, Fig. 5B). Co-expression of GRK2-RH restored the activity of ENaC to a FIGURE 5. RH domain of GRK2 up-regulates ENaC activity by inhibiting G␣q. A, I ami (normalized) in FRT cells transfected with ENaC or co-transfected with ENaC and wtGRK2. The solid bar indicates that the cells were co-transfected with siRNA directed against G␣q, G␣11, and G␣14 together. The average I ami for the control group is 14.31 Ϯ 2.27 A/cm 2 (n ϭ 5). B, I ami (normalized) in FRT cells co-transfected with ENaC and with (ϩ) or without (Ϫ) a plasmid containing the RH domain of GRK2 (GRK2-RH, 0.05 g/ml). The solid bar indicates that the cells were co-transfected with Q209L G␣q. The average I ami for the control group is 8.81 Ϯ 2.92 A/cm 2 (n ϭ 8). C, I ami (normalized) in FRT cells co-transfected with ENaC and various concentrations of Q209L G␣q (0.05, 0.1, 0.25, 0.5, 1, and 3 g/ml, closed circles) or Q209L G␣q and GRK2-RH (3 g/ml, open circles) (n ϭ 5-6). D, I ami (normalized) in M1 cells transfected with (ϩ) and without (Ϫ) wtGRK2 and siRNAs directed against G␣q, G␣11, and G␣14. The solid bar indicates that cells were treated with UTP (100 M). The average I ami for the control group is 4.47 Ϯ 0.47 A/cm 2 (n ϭ 23). ** and *** indicate p Ͻ 0.01 and 0.001, respectively. The number of experiments is shown in parentheses. NS indicates no significant difference between groups.
level similar to that observed in another group of cells that were transfected with GRK2-RH in the absence of Q209L G␣q mutant (Fig. 5B). These findings suggest that GRK2-RH can completely overcome the inhibitory effect of G␣q on ENaC.
As mentioned in the Introduction, the RH domain of GRK2 contains a binding site that selectively binds to and inhibits G␣q, G␣11, and G␣14. We envisage that, if binding to GRK2 prevents G␣q from inhibiting ENaC, the ability of GRK2-RH to prevent ENaC inhibition by G␣q should be attenuated at high concentration of G␣q. To investigate this, we co-transfected FRT cells with ENaC with different concentrations of Q209L G␣q. We found that G␣q inhibits activity of ENaC in a concentration-dependent manner (Fig. 5C). Co-expression of GRK2-RH restored activity of ENaC in Q209L G␣q-transfected cells and shifted the concentration-response curve of ENaC to Q209L G␣q expression to the right (Fig. 5C). When transfected with 0.05 g/ml Q209L G␣q plasmid, the normalized amiloridesensitive current in FRT cells was 0.58 Ϯ 0.08 (n ϭ 6), representing a 42% inhibition compared with that of the control ENaC-only-transfected cells (1.00 Ϯ 0.04, n ϭ 8; p Ͻ 0.001). Co-expression of GRK2-RH with Q209L G␣q increased the normalized amiloride-sensitive current to 1.74 Ϯ 0.25 (n ϭ 6), a value that is not significantly different to that observed in cells co-tranfected with GRK2-RH but without Q209L G␣q (1.78 Ϯ 0.13, n ϭ 9; p Ͼ 0.05). On the other hand, expression of Q209L G␣q at a concentration of 0.3 g/ml reduced the normalized amiloride-sensitive current to 0.05 Ϯ 0.01 (n ϭ 6), an inhibition of 95% when compared with that of the control untransfected cells. Under these conditions, GRK2-RH failed to fully restore activity of ENaC, only increasing the normalized amiloride-sensitive current to 0.15 Ϯ 0.02 (n ϭ 6). The reciprocal relationship between Q209L G␣q concentration and the ability of GRK2 to restore activity of ENaC is consistent with the effect of GRK2 on ENaC being dependent upon the ability of GRK2 to directly bind to, and prevent, G␣q from inhibiting the channel, which becomes less effective in the presence of higher concentrations of activated G␣q.
It has been suggested that P2Y 2 purinoceptors are coupled with Gq/11 (37) and that activation of the receptors inhibits activity of ENaC (38,39). We found that siRNAs directed against G␣q, G␣11, and G␣14 attenuated the inhibitory effect of the P2Y 2 receptor agonist, uridine-tri-phosphate (UTP), on the activity of ENaC (Fig. 5D), suggesting that the mechanism by which P2Y 2 receptor activation inhibits the activity of ENaC is mediated, in part, via Gq/11. Since expression of wtGRK2 blunts the effect of UTP on ENaC in a similar manner to the effect of the siRNAs against Gq/11/14 and the effect of GRK2 and the siRNAs are not additive (Fig. 5D), it is most likely that GRK2 exerts its effect on UTP signaling by inhibiting Gq/11, hence disrupting part of the signaling pathway activated during P2Y 2 receptor activation.

DISCUSSION
The present study has provided evidence that GRK2 functions as an endogenous activator of ENaC in epithelia. Inhibition of GRK2 by siRNA significantly decreased activity of ENaC in cells in which the channel is endogenously expressed. Our data also elucidate the motifs in GRK2, which are responsible for this function. The ability of a kinase-inactivated GRK2, K220R GRK2, to increase the activity of ENaC similarly to that of the wild-type GRK2, indicates that this effect of GRK2 is independent of its intrinsic kinase activity. The effect of GRK2 is, however, mediated by its RH domain, because the ⌬RH-GRK2 mutant which lacks this domain was ineffective in regulating activity of ENaC. An investigation using a construct containing only the RH domain of GRK2 further revealed that the presence of the RH domain alone increased activity of the channel.
The phosphorylation-independent effect mediated by the RH domain of GRK2 revealed in the current study identifies an additional stratum of regulation provided by this kinase to control epithelial Na ϩ absorption. Unlike its phosphorylation-dependent mechanism, which impinges upon Nedd4-2 and acts downstream in the G␣o-mediated signaling system that regulates ENaC (13,22,40), the RH domain of GRK2 mediates its effect by inhibiting the effect of G␣q on the channel. The lack of kinase effect of GRK2 on activity of ENaC reported in this study may reflect the absence of the Nedd4-2-dependent GRK2 signaling system that regulates ENaC in cultured epithelia.
The RH domain of GRK2 has been implicated in the regulation of several G protein signaling events (3) including insulinstimulated glucose transporter 4 translocation in adipocytes (26) and phosphoinositide hydrolysis during activation of several GPCRs, including the endothelin receptor (41), metabotropic glutamate receptor (42), parathyroid hormone receptor (43), angiotensin receptor (44), and thromboxane A 2 receptor (7). Together, these findings suggest that the RH domain plays a central role in mediating the physiological effects of GRK2. In the present study, the failure of the D110A GRK2 mutant, in which an amino acid in its RH domain was mutated to prevent binding of GRK2 to G␣q/11 (28), to regulate the basal activity of ENaC suggests that direct binding of GRK2 to G␣q/11 is essential for the mechanism by which GRK2 activates ENaC in this experimental system. The reciprocal relationship between the concentration of G␣q and the ability of the RH domain of GRK2 to prevent the Q209L G␣q mutant from inhibiting ENaC further suggests that the RH domain of GRK2 may act as an effector antagonist. This phosphorylation-independent effect of GRK2 on G␣ is likely to act in concert with its intrinsic catalytic activity to maximize the negative effect of GRK2 on G protein signaling that controls activity of ENaC.
One important contribution of this study is that it is the first to demonstrate the negative regulatory effect of G␣q/11 on the channel. Data from our siRNA studies suggest that, under basal conditions, there is a tonic effect of endogenous G␣q/11 that suppresses activity of ENaC in epithelia. Similar activation of ENaC was observed with siRNA directed against G␣q, G␣11 or G␣14. These G␣ subunits share over 80% homology in their amino acid sequences (45), hence their ability to inhibit the channel might reflect a promiscuous capacity to activate the same downstream signaling pathways, one of which inhibits ENaC. Gq/11-coupled GPCRs are known to regulate several cellular events in epithelial cells, including intracellular Ca 2ϩ mobilization (46), mucin secretion (47,48), and anion secretion (49). Our data suggest that, Gq/11 plays a small, but significant, part in the P2Y 2 receptor-mediated signaling mechanism that inhibits activity of ENaC. The fact that GRK2 rescues part of ENaC activity during P2Y 2 receptor activation, most likely by inhibiting Gq/11, further emphasizes the physiological significance of GRK2 in the regulation of Na ϩ transport in epithelia.
The nature of these signaling pathways remains unclear, as our data exclude the pathways activated by G␣q/11 which are known to inhibit ENaC. For example, activation of phospholipase C␤ by G␣q/11 can initiate hydrolysis of inositol lipids, especially phosphatidylinositol 4,5-bisphosphate (PIP 2 ), (45) leading to inhibition of ENaC (38,50). Our finding that the inhibitor of PLC, U73122, does not prevent inhibition of ENaC by the Q209L G␣q mutant makes it unlikely that U73122-sensitive PLC␤ isoforms are involved in the mechanism by which G␣q inhibits ENaC. In this regard, a role for PLC isoforms that are less sensitive to U73122 (51) cannot be excluded. Similarly, it is unlikely that G␣q exerts its effect on ENaC via PKC, ERK1/2 or some other staurosporine-sensitive protein kinase, as expression of G␣q is effective in inhibiting activity of ENaC in cells pretreated with pharmacological inhibitors of these kinases. It has also been suggested that expression of the constitutively active Q209L G␣q inhibits phosphatidylinositol 3-kinase (PI3K), a known activator of ENaC (52), by a mechanism that is independent of PLC activation (53). The effect of Gq on PI3K, however, cannot explain the negative effect of G␣q on ENaC in our study since the inhibitory effect of G␣q could still be observed in cells over-expressing a dominant-negative mutant of PI3K, ⌬p85 PI3K (54) (data not shown). Finally, Chen et al. (55) suggested that the inhibitory effect of G␣q on the activity of the two-pore-domain K ϩ channel may be due to a direct binding of the G␣ to the channel. This effect of G␣q is independent of PLC or PIP 2 . Our co-immunoprecipitation assay, however, has failed to detect any interaction between G␣q and ENaC subunits (data not shown), hence, a direct inhibitory effect of G␣q on the activity of ENaC cannot be confirmed in this experimental system.
We have shown previously that the activity of ENaC is regulated by ␣of Go and Gi2 (13,56). G␣o is a component of the Na ϩ feedback signaling pathway that inhibits activity of ENaC in the presence of high cytosolic Na ϩ concentration (13,15). This effect of Go is mediated by ubiquitin protein ligases of the Nedd4 family (14,18,57), which increase internalization of membrane ENaC, hence decreasing membrane expression and activity of the channel (58). G␣i2 is a component of the anion feedback pathway, which is activated by high cytosolic Cl Ϫ concentrations (13), however, unlike the effect of G␣o, its inhibitory effect on ENaC does not involve Nedd4 family proteins (14). Our data indicate that the effect of G␣q on ENaC does not depend upon the presence of the PY-motifs at the C-terminal of ENaC subunits, which are the binding sites of Nedd4 proteins, nor is it attenuated in cells in which siRNA had been used to knock down Nedd4-2 activity. Moreover, overexpression of Q209L G␣q has no effect on membrane expression of ENaC. Hence, like G␣i2, the mechanism by which G␣q inhibits activity of ENaC does not involve Nedd4 proteins. At present, the signaling pathway mediated by G␣q to regulate activity of ENaC remains elusive.