Cell surface expression of the ROMK (Kir 1.1) channel is regulated by the aldosterone-induced kinase, SGK-1, and protein kinase A.

The Kir1.1 (ROMK) subtypes of inward rectifier K+ channels mediate potassium secretion and regulate sodium chloride reabsorption in the kidney. The density of ROMK channels on the cortical collecting duct apical membrane is exquisitely regulated in concert with physiological demands. Although protein kinase A-dependent phosphorylation of one of the three phospho-acceptors in Kir1.1, Ser-44, also a canonical serum-glucocorticoid-regulated kinase (SGK-1) phosphorylation site, controls the number of active channels, it is unknown whether this involves activating dormant channels already residing on the plasma membrane or recruiting new channels to the cell surface. Here we explore the mechanism and test whether SGK-1 phosphorylation of ROMK regulates cell surface expression. Removal of the phosphorylation site by point mutation (Kir1.1, S44A) dramatically attenuated the macroscopic current density in Xenopus oocytes. As measured by antibody binding of external epitope-tagged forms of Kir1.1, surface expression of Kir1.1 S44A was inhibited, paralleling the reduction in macroscopic current. In contrast, surface expression and macroscopic current density was augmented by a phosphorylation mimic mutation, Kir1.1 S44D. In vitro phosphorylation assays revealed that Ser-44 is a substrate of SGK-1 phosphorylation, and expression of SGK-1 with the wild type channel increased channel density to the same level as the phosphorylation mimic mutation. Moreover, the stimulatory effect of SGK-1 was completely abrogated by mutation of the phosphorylation site. In conclusion, SGK-1 phosphorylation of Kir1.1 drives expression on the plasmalemma. Because SGK-1 is an early aldosterone-induced gene, our results suggest a possible molecular mechanism for aldosterone-dependent regulation of the secretory potassium channel in the kidney.

From the Department of Physiology, University of Maryland School of Medicine, Baltimore, Maryland 21201 The Kir1.1 (ROMK) subtypes of inward rectifier K ؉ channels mediate potassium secretion and regulate sodium chloride reabsorption in the kidney. The density of ROMK channels on the cortical collecting duct apical membrane is exquisitely regulated in concert with physiological demands. Although protein kinase A-dependent phosphorylation of one of the three phospho-acceptors in Kir1.1, Ser-44, also a canonical serum-glucocorticoid-regulated kinase (SGK-1) phosphorylation site, controls the number of active channels, it is unknown whether this involves activating dormant channels already residing on the plasma membrane or recruiting new channels to the cell surface. Here we explore the mechanism and test whether SGK-1 phosphorylation of ROMK regulates cell surface expression. Removal of the phosphorylation site by point mutation (Kir1.1, S44A) dramatically attenuated the macroscopic current density in Xenopus oocytes. As measured by antibody binding of external epitope-tagged forms of Kir1.1, surface expression of Kir1.1 S44A was inhibited, paralleling the reduction in macroscopic current. In contrast, surface expression and macroscopic current density was augmented by a phosphorylation mimic mutation, Kir1.1 S44D. In vitro phosphorylation assays revealed that Ser-44 is a substrate of SGK-1 phosphorylation, and expression of SGK-1 with the wild type channel increased channel density to the same level as the phosphorylation mimic mutation. Moreover, the stimulatory effect of SGK-1 was completely abrogated by mutation of the phosphorylation site. In conclusion, SGK-1 phosphorylation of Kir1.1 drives expression on the plasmalemma. Because SGK-1 is an early aldosterone-induced gene, our results suggest a possible molecular mechanism for aldosterone-dependent regulation of the secretory potassium channel in the kidney.
Extracellular potassium homeostasis, maintained by the regulation of renal potassium excretion, is dependent on the activity of weakly inward rectifying ''small conductance'' potassium channels (SK) 1 that are expressed on the apical membrane of epithelial cells in the distal nephron (1,2). Encoded by the ROMK (Kir 1.1 or KCNJ1) gene (3,4), these Kir channels are thought to be the major, but not exclusive (5,6), route for potassium transport into the tubule lumen and constitute a final regulated component of the potassium secretory machinery of the kidney (7,8). Indeed, aldosterone, vasopressin, and other factors precisely regulate SK activity, controlling potassium excretion in accord with the demands of potassium balance. Because ROMK channels normally exhibit a very high open probability, near unity, physiologic augmentation of channel activity, as controlled by hormones and dietary potassium (9), is achieved largely by regulated changes in the number of active channels on the plasmalemma.
Although the precise molecular mechanisms responsible for physiological augmentation of ROMK channel surface density have remained unclear, a growing body of evidence has pointed to an important role of protein kinase A (PKA) phosphorylation. Like the native secretory channel (10), ROMK activity is dependent on direct PKA phosphorylation (11,12), possibly facilitated by A-kinase-associated proteins (13). In fact, activation of ROMK by PKA is thought to underlie the regulation of renal potassium transport by vasopressin (14). All three PKA phospho-acceptor sites in ROMK, embedded within the cytoplasmic NH 2 -(Ser-44) and COOH termini (Ser-219 and Ser-313), must be phosphorylated for full channel function (15). Interestingly, single channel experiments revealed that the different PKA phospho-acceptors regulate the channel through different mechanisms (15). Phosphorylation of the two COOHterminal sites are required to maintain the channel in a high open probability state (15), controlling both pH-dependent gating (16) and phosphatidylinositol 4,5-biphosphate-dependent activation of the channel (17). Phosphorylation of the NH 2terminal site, on the other hand, has no effect on channel open probability. Instead, it appears to control the number of active channels on the cell surface (15), reminiscent of the way vasopressin affects the activity of the native secretory channel (14). Just how this is achieved remains a fundamental question in the field. In particular, it is unknown whether phosphorylation of the serine 44 residue simply switches a pre-existing pool of inactive channels on the membrane into an active gating mode or drives cell surface expression by a vesicular trafficking process.
Close inspection of the NH 2 -terminal PKA site in ROMK1 reveals that it also falls within a canonical serum-and glucocorticoid-regulated kinase (SGK1) phosphorylation sequence (recognized by a RXRXX(S/T) (18,19)), suggesting that the channel, and serine 44 in particular, might also be a target of SGK1. This observation has potentially important physiologi-cal ramifications. SGK1, a member of the PKB/Akt family of serine/threonine kinases (20), is an immediate early aldosterone-induced gene product (21,22) in the renal collecting duct (22)(23)(24), which has been shown recently to regulate epithelial Na ϩ channel (ENaC) cell surface expression (25) and, consequently, act as a key mediator of the early aldosterone effect on sodium transport in the distal nephron. Conceivably, SGK1 might also control the number of functional ROMK channels at the plasmalemma and explain how aldosterone regulates renal potassium secretion. It should be pointed out that an early test of this idea, using a co-expression assay with the Xenopus orthologue of SGK-1 and the mammalian ROMK channel, failed to detect a permissive action of the kinase (26). Nevertheless, it has remained uncertain whether the negative result was a consequence of the xenogeneic nature of the experimental design, or the requirement of A-kinase-associated proteinlike proteins, or the high level of basal phosphorylation state of ROMK in oocytes. Interestingly, a recent report that SGK activation of the sodium-hydrogen exchanger-3 is facilitated by the sodium-hydrogen exchange regulatory factor-2 (NHERF-2) (27), a PDZ protein found in the collecting duct (28) that directly interacts with ROMK, 2 provided further incentive to evaluate if mammalian SGK-1 regulates the ROMK channel.
The goals of the present study were to determine the mechanism by which phosphorylation of serine 44 in ROMK controls channel activity and to determine whether this residue is a substrate for the aldosterone-induced kinase, SGK-1. Here we show that SGK-1 phosphorylates serine 44 to drive cell surface expression, providing a potential mechanistic explanation for aldosterone-dependent regulation of the secretory potassium channel.

MATERIALS AND METHODS
Molecular Biology-A hemagglutinin (HA) epitope tag was introduced into the extracellular loop of ROMK by overlap extension PCR at position 113 of ROMK1. Both ends of the epitope were flanked by two glycine residues to enhance accessibility and flexibility of the extracellular HA tag, creating a sequence which reads 111 EFGGYPYDVPDY-AGGYP. ROMK1 serine 44 was changed to alanine or aspartic acid using Stratagene's QuikChange Site-directed mutagenesis kit. The mouse SGK-1 cDNA, encoding the entire reading frame, was amplified from reverse-transcribed mouse kidney RNA with the PCR by using SGK-1-specific primers (5Ј-GGGACGATGACCGTCAAAGC-3Ј, and 5Ј-AGCACTCAGAGGAAGGAATCCAC-3Ј). All constructs used for studies in Xenopus oocytes were subcloned between the 5Ј-and 3Ј-untranslated region of the Xenopus ␤-globin gene in the modified pSD64 vector to increase expression efficiency (29). This vector also contains a polyadenylate sequence in the 3Ј-untranslated region (dA23dC30). With the exception of EGFP-ROMK, all constructs used for mammalian expression were subcloned into pcDNA 3.1ϩ (Invitrogen). EGFP was engineered onto the NH 2 terminus of HA-ROMK by subcloning the epitopetagged Kir1.1 in-frame with EGFP in the pEGFP-Cl (Clontech). The sequence of all amplified or modified cDNAs was confirmed by dye termination DNA sequencing (University of Maryland School of Medicine Biopolymer Core).
cRNA Synthesis-Complementary RNA was transcribed in vitro in the presence of capping analogue G(5Ј)ppp(5Ј) from linearized plasmids containing the cDNA of interest using SP6 RNA polymerase (mMessage Machine, Ambion Inc.). cRNA was purified by phenol/chloroform extraction and precipitated with ammonium acetate/isopropyl alcohol. Yield was quantified spectrophotometrically and confirmed by agarose gel electrophoresis.
Oocyte Isolation and Injection-Oocytes from female Xenopus laevis (Xenopus Express, Homosassa, FL) were isolated and maintained using the standard procedures as described previously (30). Briefly, frogs were anesthetized with 0.15% 3-aminobenzoate, and a partial oophorectomy was performed through an abdominal incision. Oocyte aggregates were manually dissected from the ovarian lobes and then incubated in OR-2 medium (82.5 mM NaCl, 2 mM KCl, 1 mM MgCl 2 , and 5 mM HEPES, pH 7.5) containing collagenase (type 3, Worthington) for 2 h at room temperature to remove the follicular layer. After extensive washing with collagenase-free OR-2, oocytes were stored at 19°C in OR-3 medium (50% Leibovitz's medium, 10 mM HEPES, pH 7.4). 12-24 h later, healthy looking Dumont stage V-VI oocytes were pneumatically injected with 50 nl of diethyl pyrocarbonate-treated water containing 0 -1 ng of cRNA and then stored in OR-3 medium at 19°C for 2-6 days.
Electrophysiology-Whole cell currents in Xenopus oocytes were monitored using a two-microelectrode voltage clamp as described previously (30,31). Briefly, oocytes were bathed in a 45 mM K solution (45 mM KCl, 45 mM N-methyl-D-glucamine-Cl, 1 mM MgCl 2 , 1 mM CaCl 2 , 5 mM HEPES, pH 7.4). Voltage sensing and current injecting microelectrodes had resistances of 0.5-1.5 megohms when backfilled with 3 M KCl. Once a stable membrane potential was attained, oocytes were clamped to a holding potential of Ϫ20 mV, and currents were recorded during 500-ms voltage steps, ranging from Ϫ100 to ϩ40 mV in 20-mV increments. Data were collected using an ITC16 analogue to digital, digital to analogue converter (Instrutech Corp.), and filtered at 1 kHz and digitized on line at 2 kHz using Pulse software (HEKA Electronik) for later analysis. ROMK currents are taken as the barium-sensitive inward current (2 mM barium acetate) as we have done before (31,32). Values reported in the text are the barium-sensitive inward currents at Ϫ100 mV. Cation permeability ratios (P K /P Na ) were estimated from the change in reversal potential induced by replacing extracellular potassium with an equivalent concentration of sodium as we have done before (30).
Surface Expression-Plasmalemma expression of the external HAtagged ROMK1 channel was measured in single oocytes following procedures outlined by Zerangue et al. (33) with slight modifications. In these studies, oocytes were washed two times in cold OR-2 medium, fixed with 4% formaldehyde in OR-2 for 15 min at 4°C, and washed 3 times in OR-2. To block spurious antibody binding, oocytes were then incubated for 1 h at 4°C in OR-2 containing 1% bovine serum albumin (BSA). Exposed HA epitopes on the surface of intact oocytes were labeled with a rat monoclonal anti-HA antibody (Roche Applied Science, 1 g/ml, 3F10, 1% BSA, 4°C overnight), and then oocytes were washed at 4°C with OR-2 and incubated with HRP-coupled goat anti-rat (The Jackson Laboratories, 1 g/ml, 1% BSA, 4°C, 3 h). Cells were washed for 1 h at 4°C with OR-2 containing 1% BSA then again for 10 min in OR-2 medium without BSA. Individual oocytes were placed in 50 l of enhanced chemiluminescence substrate (Amersham Biosciences) and incubated for 1 min at room temperature. Luminescence from single oocytes was measured for 10 s in a Sirius luminometer and reported as relative light units per s.
Western Blot Analysis-Oocytes were processed following the protocol described by Kamsteeg and Deen (34) to isolate proteins from total membrane. In brief, oocytes were washed twice in homogenization buffer (80 mM sucrose, 5 mM MgCl 2 , 5 mM NaH 2 PO 4 , 1 mM EDTA, 20 mM Tris, pH 7.4) containing a protease inhibitor mixture (5 g/ml leupeptin, 1 mM phenylmethylsulfonyl fluoride, and 5 g/ml pepstatin A) and then broken by trituration with a 25-gauge syringe. To pellet yolk proteins and nuclei, homogenates were spun twice at low speed (100 ϫ g) for 10 min. Supernatants were then spun at high speed (14,000 ϫ g) for 20 min at 4°C to collect the total membrane fraction. Pellets were washed once in the homogenization buffer and spun at top speed again for 10 min and then placed in solubilization buffer (4% sodium deoxycholate, 20 mM Tris, pH 8.0, 5 mM EDTA, 10% glycerol, containing the protease inhibitors) and rocked for 2 h at 37°C. Particulate material was pelleted (14,000 ϫ g for 20 min at 4°C), and the solubilized proteins in the supernatant were resolved by SDS-PAGE and transferred to nitrocellulose membranes.
Blots were blocked in Tris-buffered saline with Tween 20 (0.1%) (TBS-T) containing 5% non-fat dry milk for 1 h at room temperature. Rat anti-HA monoclonal antibody (3F10, Roche Applied Science) was diluted in 5% non-fat dry milk to 0.1 g/ml and incubated at room temperature for 1 h, washed for 15 min in TBS-T, incubated with goat anti-rat antibody coupled to horseradish peroxidase (HRP) (1:5000, 5% non-fat dry milk), and washed extensively for 20 min in TBS-T. Bound antibodies were then revealed using enhanced chemiluminescence reagent (Amersham Biosciences) and fluorography (X-Omat, Eastman Kodak Co.). Fluorographs were assessed by densitometry using NIH Image. Integrated optical density of each band is reported in arbitrary units where the control is considered to be 1 unit.
Immunofluorescence of COS-7-Cells, grown on glass coverslips, were transfected with DNA (2 g total) using Gene Juice (Novagen). After 2 days, cells were washed in ice-cold modified Ringer's solution (144 mM NaCl, 5 mM KCl, 1 mM CaCl 2 , 5.5 mM CaCl 2 , 5.5 mM glucose, 1.2 mM NaH 2 PO 4 , 5 mM HEPES, pH 7.4) and fixed with 4% formaldehyde in Ringer's solution for 15 min at 4°C. One group of cells was 2 D. Yoo and P. A. Welling, unpublished observations. permeabilized with 0.1% Triton X-100 for 30 min at room temperature and washed three times in the modified Ringer's solution before blocking. Non-permeabilized cells were not treated with Triton. Both groups where blocked with 1% BSA in the modified Ringer's solution for 30 min at room temperature and labeled with primary antibody in 0.1% BSA (rat anti-HA (Roche Applied Science) at 10 ng/ml; mouse anti-PDI 1:100) for 3 h at room temperature and then washed in modified Ringer's solution three times. Next, cells were incubated with secondary antibody conjugated to either Alexa 488 or 568 (goat anti-rat at 1:100; goat anti-mouse 1:100, Molecular Probes) in 0.1% BSA for 1 h at room temperature in the dark. Slides were then washed and mounted onto slides in VectaShield and sealed with nail polish. Cells were visualized using a 410 Zeiss laser-scanning confocal microscope under a 63ϫ oil immersion lens. Images were acquired at a zoom of 2 and a pinhole size of 18 using 16-frame averaging and were processed with Adobe Photoshop.
In Vitro Phosphorylation of ROMK by SGK-1-To determine whether ROMK is a substrate of SGK-1 phosphorylation, in vitro phosphorylation assays were conducted using a constitutively active, recombinant form of SGK-1 (⌬1-60, S422D, SGK-1, Upstate Biotechnology Inc.). Either immunopurified ROMK or ROMK synthetic peptides were used as substrates. For the ROMK immunoprecipitation study, COS-7 cells were transfected with the HA epitope-tagged Kir1.1 using Gene Juice (Novagen). 48 h post-transfection, cells were sequentially washed with ice-cold modified Ringer's solution and lysis buffer (150 mM NaCl, 20 mM Tris-Cl, pH 7.4), harvested in cold lysis buffer with protease inhibitor mixture, and resuspended at 5ϫ cell volume with cold lysis buffer supplemented with protease inhibitor mixture and 1% Triton X-100. Cells were passed through a 27 1 ⁄2-gauge needle, incubated on ice for 30 min, and then centrifuged at 15,000 ϫ g for 15 min at 4°C. Lysates containing the soluble proteins were precleared with 100 l of 50% Sepharose CL-4B slurry for 2 h rotating at 4°C. Precleared COS lysates were then rotated overnight with 100 l of 20% protein-G Sepharose slurry with 2 g of rabbit anti-HA antibody (Upstate Biotechnology, Inc.). Immunoprecipitates were washed 3 times with lysis buffer containing 0.1% Triton. SGK-1 phosphorylation of immunopurified ROMK bound on Sepharose beads was carried out with an active recombinant SGK-1 purchased from Upstate Biotechnology, Inc. (Lake Placid, NY), according to their recommendations as described below. Once the phosphorylation reaction was complete (30°C, 60 min), immunoprecipitates were washed 3 times in ADBI buffer (20 mM MOPS, pH 7.2, 25 mM ␤-glycerol phosphate, 5 mM EGTA, 1 mM sodium orthovanadate, 1 mM dithiothreitol). Bound proteins were eluted by incubating at room temperature for 30 min in 20 l of SDS sample buffer and run on 12% acrylamide gels. Gels were dried using gel drying sheets (Promega) and exposed to autoradiography film (Kodak, X-Omat) for 1-3 h. In parallel immunoprecipitation reactions, eluates were run on 12% acrylamide gels followed by Western blot analysis as described earlier with rat anti-HA antibody (Roche Applied Science).
A synthetic peptide corresponding to the wild type ROMK NH 2 terminus (residues, 34 -48) and a mutant negative control peptide containing a serine to alanine substitution at position 44 were synthesized by University of Maryland Baltimore Biopolymer core facility. In vitro phosphorylation of the peptides by SGK-1 was carried out according to the manufacturer's recommendation. Briefly, reactions (50 l), containing the peptide (80 M), SGK-1 (1-25 ng), [␥-32 P]ATP (1 Ci/l), protein kinase C inhibitor (20 M), and PKA inhibitor (2 M) in ADBI buffer, were incubated at 30°C for 60 min. Aliquots of the reaction were transferred onto the center of a P81 filter paper square and then washed 3 times with 0.75% phosphoric acid followed by another 3 washes in acetone. The P81 squares were placed in scintillation fluid, and bound ratio activity was quantified in a liquid scintillation counter. Aliquots of the reaction were also resolved on Tris-Tricine acrylamide gels (16.5%) followed by autoradiography.
To determine the apparent K m values for SGK and PKA phosphorylation of the Ser-44 site in ROMK, in vitro phosphorylation reactions were carried out as described above with varying amounts of the wild type ROMK peptide (0 -200 M) over the initial, linear rate of phosphorylation (0 -10 min). Constitutively active recombinant SGK-1, as above, and purified bovine heart PKA catalytic subunit (Upstate Biotechnology, Inc.) were used. Parallel phosphorylation reactions, containing identical specific activities of either SGK-1 or PKA, were performed in triplicate and compared. Specific activity was independently determined by measuring rates of phosphate transfer to the optimum phosphorylation motif of each kinase (SGK (RPRAATF); PKA (GRTGRRNSI)).
Quantitation of SGK mRNA Expression by Real Time PCR-Mice, maintained on a standard rodent chow, were placed on a high potas-sium diet containing 10% KCl for 0.5, 1, 2, and 4 days. Kidneys were harvested, and total RNA was isolated using Trizol reagent (Invitrogen) according to the manufacturer's recommendations. Two micrograms of total kidney RNA was reverse-transcribed by using Omniscript (Qiagen). Serum-and glucocorticoid-regulated kinase-1 or actin cDNAs were separately amplified in parallel real time PCRs, containing an aliquot of the reverse-transcribed reaction, 0.2ϫ SYBR Green I (Molecular Probes), and 50 pmol of either serum-and glucocorticoid-regulated kinase-1 (SGK)-specific primers (forward 5Ј-TGTCTTGGGGCTGTCCT-GTATG-3Ј; reverse 5Ј-GCTTCTGCTGCTTCCTTCACAC-3Ј) or actinspecific primers (forward 5Ј-GGCATTGTTACCAACTGGGACG-3Ј; reverse 5Ј-GCTTCTCTTTGATGTCACGCACG-3Ј) in separate wells using the Bio-Rad iCycler real time PCR detection system. Each sample was amplified in triplicate. Specificity of the amplified product was determined using melting curve analysis and agarose gel electrophoresis. For quantification, real time PCR of serially diluted SGK cDNA was carried out as a positive control and to establish an optimal threshold cycle detection level; threshold cycle values, C t , of the test samples were then measured from cycle-dependent product amplification curves (Bio-Rad software package iCycler version 3.0a). The change in SGK transcript, reported as the fold induction over the actin control, was calculated using the 2(Ϫ⌬⌬C t ) method (35).
Statistical Analysis-Statistical analysis was performed using GB-Stat TM for Macintosh statistics package (Dynamic Microsystems, Inc). Analysis of variance followed by Fisher's LSD Post Hoc test was performed to verify significance. Values of at least p Ͻ 0.05 were considered significant. All data are given as mean Ϯ S.E.

RESULTS
Surface Expression Assay-Phosphorylation of the serine 44 residue regulates ROMK either by activating dormant channels already residing on the plasma membrane or by recruiting new channels to the cell surface, two mechanisms that can be distinguished readily by differences in cell surface expression. Accordingly, we incorporated a hemagglutinin epitope tag into the extracellular loop of the ROMK channel so that plasmalemma channel levels could be quantified accurately by antibody binding. When expressed in Xenopus oocytes, HA-tagged ROMK carried weakly inward-rectifying, K ϩ -selective (P K /P Na Ͼ15) and barium-sensitive currents identical to those observed for the wild type channel (Fig. 1). To measure relative amounts of surface HA-tagged ROMK, we employed a method that combines enzyme amplification with the sensitivity and linearity of analytical luminometry as originally described by Zerangue et al. (33). In this assay, exposed HA epitopes on the surface of intact oocytes were labeled with a monoclonal antibody to HA and then with an HRP-conjugated secondary antibody. Antibody bound to the cell surface was quantified by luminometer measurements of single oocyte chemiluminescence. As shown in Fig. 2, the luminescence response is linear over a wide range of expression levels, permitting an accurate appraisal of relative ROMK channel density at the cell surface.
Phosphorylation of Serine 44 Controls ROMK Expression at the Plasmalemma-To determine the mechanism by which phosphorylation of the serine 44 residue regulates ROMK, we employed the assay, described above, to measure to cell surface channel density and the two-microelectrode voltage clamp to measure the macroscopic potassium channel current activity in Xenopus oocytes separately injected with equal amounts of either wild type, HA-tagged ROMK, or HA-tagged ROMK channels bearing phosphorylation null (S44A) or phosphorylation mimic (S44D) mutations (Fig. 3). Xenopus oocytes are uniquely suited for these studies because endogenous levels of PKA generally support high constitutive activity of the ROMK channel (11). Replacement of the critical serine with alanine (S44A) caused a dramatic fall in macroscopic current and a parallel reduction in cell surface expression, without change in total cellular abundance of ROMK protein, revealing that the phospho-acceptor, serine 44, is required for ROMK expression at the plasmalemma. This finding is confirmed and extended by the observation that channels bearing the phosphorylation mimic mutation (S44D) exhibited a greater macroscopic current and cell surface channel density than the wild type channel. Although the S44D mutation tended to produce a slightly larger surface channel density than functional activity, it was not statistically significant, indicating that surface expression and channel activity increase in parallel. Western blot analysis confirmed that the response was not related to dramatically different protein expression levels; no differences in band density could be detected between the different groups (band density of WT, 1 unit; S44A, 0.86 Ϯ 0.13 units; S44D, 0.9 Ϯ 0.15 units). Collectively, these data indicate that a high degree of basal phosphorylation drives cell surface expression of ROMK in Xenopus oocytes. As inferred by the enhanced activity and surface expression of S44D channels, a small pool of wild type ROMK must exist in an intracellular compartment in a nonphosphorylated form, available for recruitment to the plasma membrane.
SGK-1 Is a Potential Physiological Regulator of ROMK in Potassium Adaptation-This hypothesis predicts that phosphorylation of serine 44 by an exogenous kinase should increase wild type ROMK activity to the same extent as the phosphorylation mimic mutation, S44D. Analysis of the ROMK primary structure revealed that serine 44 not only lies within one of three PKA phosphorylation sites but is also uniquely embedded within an SGK phosphorylation consensus sequence, recognized by the RXRXX(S/T) motif (18,19). Furthermore, we found that the abundance of SGK-1 transcript is rapidly increased in the mouse kidney following augmentation of dietary potassium intake (Fig. 4). Because the response is coincident with the early increase in ROMK channel density in the renal collecting duct (36), SGK-1 might be a physiological mediator of ROMK in the potassium adaptation response, further prompting us to test whether that mammalian form of SGK-1 controls ROMK activity by direct phosphorylation of serine 44.
SGK-1 Regulates ROMK Activity and Surface Expression-In light of a recent report (27) that SGK activation of the sodium-hydrogen exchanger-3 is facilitated by NHERF-2, we

FIG. 2. Extracellular HA epitope ROMK displays a linear relationship between surface expression and functional expression.
Oocytes were injected with varying amounts of extracellular HA-tagged ROMK cRNA (250, 500, 750 pg, and 1 ng). Macroscopic ROMK activity, reported as the whole cell, barium-sensitive K ϩ current at Ϫ100 mV in 45 mM K ϩ , was measured and correlated with levels of tagged channel on the surface of non-permeabilized oocytes as assessed by labeling with an anti-HA epitope antibody and an HRP-conjugated secondary antibody. Single oocyte chemiluminescence was detected by luminometry and is reported in relative light units/s (RLU/Sec). Simple linear regression analysis reveals an r value of 0.98. Data shown are from a representative batch of oocytes (Ͼ5 oocytes at each data point). Dose-response curves were performed in all subsequent studies, and oocytes from frogs with r values greater than or equal to 0.98 were used for further analysis. Error bars are expressed as Ϯ S.E. tested whether SGK-1 regulates ROMK activity in an NHERF-2-dependent manner. In these studies, summarized in Fig. 5, SGK-1 and NHERF-2 cRNA were injected separately or together into oocytes with the ROMK channel cRNA, and macroscopic ROMK currents were measured by two-microelectrode voltage clamp 72-84 h post-injection. We observed that SGK-1 co-expression consistently stimulated ROMK in oocytes isolated from 4 of 5 frogs; increasing the macroscopic current to a level similar to the one exhibited by the phosphorylation mimic (S44D) (see below). No stimulatory effect was observed with a catalytically inactive mutant of SGK-1, K127M (19) (ROMK only, Ϫ9 Ϯ 0.36 A versus ROMK Ϯ SGK-1 K127M, Ϫ8.9 Ϯ 0.50 A). In all cases with wild type SGK-1, no further stimulation was seen with NHERF-2 (Fig. 5). Although it remains to be tested whether by A-kinase-associated proteins (37), or scaffolding molecules like NHERF, are required for efficient regulation of ROMK in the native kidney, our studies indicate that exogenous NHERF is not required for full SGK-1-dependent regulation of ROMK1 under these experimental conditions. Western blot analysis confirmed that the response was not  Macroscopic potassium currents, I/I WT , and cell surface expression, RLU/RLU WT , were measured as above in oocytes injected with HA-ROMK alone, co-injected with HA-ROMK cRNA and SGK-1 cRNA, or injected with cRNA encoding HA-ROMK bearing the phospho-mimic mutation, S44D (750 pg each). Data are normalized to the mean wild type current, I WT , or surface luminescence (RLU WT ) and represent a mean Ϯ S.E. of cohorts of at least seven oocytes. *, p Ͻ 0.05. related to dramatically different protein expression levels; no differences in band density could be detected between the different groups (band density of the ROMK only, 1 unit; ROMK ϩ NHERF-2, 1.03 Ϯ 0.16 units; ROMK ϩ NHERF-2 and SGK-1, 0.99 Ϯ 0.12 units; ROMK ϩ SGK-1, 1.03 Ϯ 0.04 units). The particular batch of oocytes, in which SGK-1 had no effect on ROMK, was notable in that macroscopic current densities of the wild type and S44D ROMK channels were indistinguishable, suggesting maximal phosphorylation was already achieved. In any case, the usual SGK-1 effect on ROMK activity was paralleled by an increase in the number of channels at the plasmalemma as measured by the surface antibody binding chemiluminescence assay (Fig. 6).
SGK-1 Directly Phosphorylates ROMK-To determine whether SGK-1 regulates ROMK1 directly, we first tested whether ROMK1 is a substrate for SGK-1 phosphorylation. In these studies, HA-ROMK1 was transfected into COS-7 cells, immunopurified on anti HA-bound protein G-Sepharose beads, and incubated with [␥-32 P]ATP in the presence and absence of constitutively active recombinant SGK-1. Fig. 7A shows a typical study. Autoradiography revealed a phosphorylated ϳ45-kDa species in reactions containing SGK protein but not in those in which SGK or ROMK was omitted. Western blotting of the immunoprecipitate with anti-HA antibodies confirmed the identity of the phosphorylated ϳ45-kDa protein as HA-ROMK. Thus, ROMK is indeed a substrate for SGK phosphorylation.
To determine whether SGK phosphorylates serine 44, we carried out in vitro phosphorylation assays using a synthetic peptide corresponding to the wild type ROMK NH 2 -terminal residues, 34 -48, and a mutant NH 2 -terminal peptide (residues 34 -48) in which serine 44 was replaced with alanine. As shown in Fig. 7B, we found that SGK strongly phosphorylates the wild type ROMK peptide and that replacement of serine 44 with alanine caused a complete inhibition of the SGK-mediated phosphorylation. Thus, SGK1 directly phosphorylates ROMK at serine 44.
As a further test of SGK-1 specificity, phosphorylation reaction velocity was measured as a function of ROMK synthetic peptide concentration (wild type ROMK NH 2 -terminal residues, 34 -48) so that the apparent K m could be determined (Fig.  8). SGK phosphorylation was compared in parallel reactions with PKA, well known to phosphorylate Ser-44 in the physiological regulation of the ROMK channel. As shown in Fig. 8, concentration-dependent phosphorylations of the ROMK peptide by SGK-1 and PKA are both well described by Michaelis-Menten kinetics as predicted for a specific phosphorylation process. More importantly, these studies reveal that the K m for SGK-1 (42 Ϯ 8 M) is identical to PKA (42 Ϯ 8 M). These Peptides were incubated with [␥-32 P]ATP in the presence and absence of constitutively active recombinant SGK-1. Reactions were either run on 16.5% Tris-Tricine acrylamide gels, or 32 P incorporation was measured by scintillation counting. Only WT peptides in the presence of SGK were phosphorylated, as confirmed by both autoradiography and scintillation counts (counts per min, cpm). values are comparable with the micromolar affinities reported for most well known PKA substrates (see Ref. 38 for address to web-based catalog) such as pyruvate kinase (39) (ϳ20 M), providing strong support for SGK-1 specificity.
SGK-1 Activation of ROMK Requires Serine 44 -If phosphorylation of serine 44 is the mechanism by which SGK stimulates ROMK, mutation of the phosphorylation site should abrogate the stimulatory effect of SGK. To test this hypothesis, macroscopic channel activity was measured under voltage clamp as above in Xenopus oocytes expressing either wild type, phosphorylation null (S44A), or phospho-mimic (S44D) ROMK channels in the presence and absence of SGK-1. As shown in Fig. 9, removal of the phosphoacceptor, S44A, completely abrogated the stimulatory effect on the channel. Furthermore, the phosphorylation mimic mutation increased channel activity to a level that was identical to the SGK-1-stimulated wild type channel. Moreover, SGK-1 co-expression did not cause a further significant augmentation of S44D ROMK channel activity.

SGK-1 Phosphorylation-dependent ROMK Plasmalemma
Expression in Mammalian Cells-Our studies above in Xenopus oocytes indicate that high levels of basal ROMK phosphorylation drive cell surface expression in this system. These observations contrast mammalian expression systems where active PKA is normally low and ROMK is inefficiently expressed on the plasmalemma (40). In fact, as shown in COS-7 cells, ROMK largely accumulates in an intracellular compartment resembling the endoplasmic reticulum (Fig. 10). Because the closely related channel, Kir 2.1 (41), is largely expressed on the plasmalemma in the same cells using identical transfection and detection conditions, it is unlikely that the intracellular expression of ROMK is an artifact of overexpression. Moreover, co-localization of the ROMK channel with protein-disulfide isomerase reveals that ROMK is largely expressed in the endoplasmic reticulum.
To determine whether SGK-1 phosphorylation increases cell surface expression in a system where the channel is expressed in the endoplasmic reticulum under basal conditions, COS-7 cells were transfected with a double-tagged ROMK construct, containing EGFP on the NH 2 terminus and the extracellular HA epitope, in the absence and presence of SGK-1. The EGFP tag permits detection of ROMK-transfected cells under nonpermeabilized conditions, and the external epitope allows detection of plasmalemma channel expression by HA antibody binding. As shown in Fig. 11, SGK-1 co-transfection dramatically increased the number of cells that express ROMK on the cell surface. This effect was absolutely dependent on serine 44 (Fig. 11C). Removal of the phosphoacceptor, S44A, completely abrogated the stimulatory effect of SGK-1 on plasmalemma channel expression. Furthermore, the phosphorylation mimic mutation increased cell surface expression to a level that was identical to the SGK-1-stimulated wild type channel.

DISCUSSION
The density of active ROMK channels on the renal cortical collecting duct apical membrane is exquisitely controlled by hormones and other factors, helping to ensure that renal potassium excretion matches dietary intake and potassium homeostasis is maintained (7). Down-regulation of ROMK channel density in states of potassium deprivation is thought to involve Src-dependent channel endocytosis (42,43), whereas phosphorylation of one of the three PKA phosphorylation sites in ROMK (12), serine 44, has been implicated in physiological augmentation of active ROMK channel density (15). In the present study, we elucidate the molecular mechanism and identify a pathway for the hormonal regulation of the channel by vasopressin and aldosterone. By measuring cell surface density and functional activity of wild type and phosphorylation mutant ROMK channels, we discovered that phosphorylation of serine 44 dramatically increases functional channel density by recruiting or retaining new channels on the plasmalemma rather than by activating silent channels already present on the plasma membrane. The observation implies that vasopressin, and other procedures that cause PKA-mediated phosphorylation of serine 44, enhance the number of ROMK channels at the cell surface by phosphorylation-dependent intracellular trafficking processes, reminiscent of the way vasopressin controls the AQP-2 water channel in the collecting duct (44). Moreover, our discovery that serine 44 is also a substrate for SGK-1 illuminates a potential molecular mechanism for the regulation of ROMK density by dietary potassium.
A series of adaptive changes in the collecting duct principal cell take place in response to an increase in dietary potassium, allowing a more effective and enhanced excretion of potassium after an acute potassium load (7). Although it is generally believed that the response, called potassium adaptation, depends on elevated aldosterone levels and other synergistic factors, triggering an increase in apical macroscopic potassium conductance (45) and parallel increase in the number of active SK channels on the apical plasmalemma (2,8,46), the molecular mechanisms and pathways connecting aldosterone to the channel have remained vague. At most, observations that the increase in SK density can be observed on the cortical collecting duct apical membrane soon after dietary potassium load is increased (ϳ6 h) (8), and without changes in ROMK transcript abundance (47), have suggested that the response, at least in the acute phase, is mediated by a post-translational modification of a pre-existing pool of channels or closely associated proteins. Now, following recent discoveries that aldosterone rapidly induces the SGK-1 kinase (21,22) to control the epithelial sodium channel (ENaC) density (25), the present study provides evidence that SGK-1 may also serve as an arbitrator of the aldosterone-dependent limb of potassium adaptation process, directly regulating ROMK surface expression. We do not want to suggest our findings rule out the role of other yet-to-be identified signaling pathways in potassium adaptation, rather our data point to a potential role of SGK-1. First, SGK-1 transcript in the kidney is rapidly induced by dietary potassium loading, precisely correlating with the early increase in ROMK channel density as measured in the renal collecting duct by Palmer and Frindt (36). Second, the ROMK channel contains a single consensus site for phosphorylation by SGK-1 (serine 44), recognized by the sequence RXRXX(S/T) (18,19 by this observation, in vitro phosphorylation studies revealed that ROMK serine 44 is indeed a substrate for phosphorylation by SGK-1. In fact, we found that SGK-1 phosphorylates the Ser-44 site in ROMK with an affinity that is identical to that exhibited by PKA. Third, co-expression of SGK-1 with ROMK1 stimulated channel activity and surface expression to about the same level observed with channels bearing a phosphorylation mimic mutation, S44D. Finally, the stimulatory effect of SGK-1 was completely abrogated by removal of the critical phosphorylation acceptor, serine 44. It should be pointed out that although aldosterone might be necessary for potassium adaptation, it is not sufficient to increase SK density on the CCD apical membrane (48). Indeed, elevation of aldosterone levels by osmotic minipumps or dietary sodium restriction has no effect on potassium secretory channel activity in the rat cortical collecting duct, even though the increase in SK channel density that is normally observed in response to dietary potassium loading (1, 46) requires intact adrenal glands and presumably elevated aldosterone levels. To explain these observations, Palmer and Frindt (48) reasonably proposed that the potassium adaptation response requires unidentified kaliuretic factor(s), which work synergistically with aldosterone to augment SK channel density. The identification of SGK-1 as a regulator of ROMK surface expression provides a clue about how this might work. Although aldosterone induces the expression of SGK-1 (22,24), the kinase must be subsequently activated by phosphorylation (20), involving 3-phosphoinositide-dependent kinases 1 and 2 (18) or the mitogen-activated protein kinase family member extracellular signal-regulated protein kinase 5 (49,50), and perhaps other unidentified kinases. Being sequentially activated at transcriptional and post-transcriptional levels, SGK-1 is poised to integrate the effects of aldosterone and a number of extracellular signals, such as those that might emanate from the kaliuretic factor, and thereby play a key role in regulating ROMK channel density in potassium adaptation.
Our results with mouse SGK-1 and ROMK-1 clearly differ from those initially reported by Chen et al. (21), who failed to detect a stimulatory effect of the Xenopus SGK-1 on ROMK2, a splice variant that lacks the first 19 amino acids of ROMK1. Furthermore, while the present paper was under revision, Yun et al. (51) reported that SGK-1 does not directly act by itself but instead synergizes with the Na ϩ /Hϩ exchange regulating factor-2 to stimulate ROMK. Although there are a number of methodological differences between these studies and ours that could potentially account for the disparity, we believe that the most likely explanation lies in a variable degree of basal ROMK phosphorylation in Xenopus oocytes. This concept has precedence with the AQP2 water channel in which high, but variable, levels of channel phosphorylation in Xenopus oocytes can yield contrasting effects of exogenous cAMP on the PKA-dependent surface expression response (52,53). It is well known that Xenopus oocytes are developmentally arrested at the first meiotic prophase by high intracellular cAMP and active PKA levels and that meiotic maturation depends on a progesteroneinduced reduction in cAMP (54). The high levels of active PKA, in turn, generally drive a high basal activity of ROMK in this expression system (12,15). Nevertheless, the fraction of ROMK channels that exist in the fully phosphorylated, or active, form is predicted to be variable and highly sensitive to the levels of channel expression relative to the amounts of endogenous cAMP and active PKA. Although oocytes were not exposed to progesterone to lower cAMP in the present study, we did fortuitously overexpress ROMK at a level that was nearly an order of magnitude greater than those reported earlier, pre-sumably increasing the probability that some ROMK channels escape phosphorylation by endogenous PKA and thus be available for SGK-1. Indeed, as inferred from the increased activity and surface expression of ROMK channels bearing the phosphorylation mimic mutation, S44D, compared with the wild type channel, a small population of wild type channels must exist in a non-phosphorylated form, open for phosphorylation by exogenous SGK-1, at the expression levels in the present study. Although augmentation averaged 40 -50% of the basal activity, similar to what was observed with SGK-1, it is important to note that some variation between batches of oocytes was observed, presumably a consequence of variable basal phosphorylation levels. In cases where maximum phosphorylation of ROMK has all ready been achieved by endogenous PKA, no effect of SGK-1 will be observed, even though SGK-1 is capable of phosphorylating and regulating the channel.
Just how phosphorylation of serine 44 controls cell surface expression, regulating ROMK traffic to and/or from the plasmalemma, remains to be determined. Phosphorylation-dependent intracellular trafficking processes are thought to be coordinated by phosphorylation-dependent interaction or dissociation of intracellular trafficking machinery. For instance, recent studies indicate that SGK-1 phosphorylation of the NEDD4 -2 ubiquitin ligase inhibits the PY-WW domain (26)-mediated NEDD-4 interaction with the epithelial sodium channel, preventing channel ubiquitination, endocytosis, and degradation and thereby increasing ENaC surface expression (25). Unlike ENaC, ROMK does not have the capacity to bind to NEDD4 -2 through a PY-WW interaction, making it unlikely that SGK-1-dependent regulation of ROMK is similar to ENaC. Instead, our observations in COS cells, where ROMK channels are usually inefficiently processed to the plasmalemma and reside predominantly in the endoplasmic reticulum until SGK-1 co-expression, provide reason to speculate that phosphorylation of the serine 44 residue in ROMK may regulate ROMK traffic from the endoplasmic reticulum. Intriguingly, serine 44 juxtaposes a putative endoplasmic reticulum retention signal, RXR, similar to those initially identified in other Kir-type channels (33).
In summary, phosphorylation of ROMK at serine 44 by PKA or SGK-1 is necessary for cell surface expression of the channel. The observation may provide a mechanistic explanation for vasopressin and aldosterone-dependent regulation of the secretory potassium channel in the kidney.