sgk Is an Aldosterone-induced Kinase in the Renal Collecting Duct

The early phase of the stimulatory effect of aldosterone on sodium reabsorption in renal epithelia is thought to involve activation of apical sodium channels. However, the genes initiating this effect are unknown. We used a combination of polymerase chain reaction-based subtractive hybridization and differential display techniques to identify aldosterone-regulated immediate early genes in renal mineralocorticoid target cells. We report here that aldosterone rapidly increases mRNA levels of a putative Ser/Thr kinase,sgk (or serum- andglucocorticoid-regulated kinase), in its native target cells, i.e. in cortical collecting duct cells. The effect occurs within 30 min of the addition of aldosterone, is mediated through mineralocorticoid receptors, and does not require de novo protein synthesis. The full-length sequences of rabbit and mouse sgk cDNAs were determined. Both cDNAs show significant homology to rat and human sgk (88–94% at the nucleotide level, and 96–99% at the amino acid level). Coexpression of the mouse sgk in Xenopus oocytes with the three subunits of the epithelial Na+ channel results in a significantly enhanced Na+ current. These results suggest that sgk is an immediate early aldosterone-induced gene, and this protein kinase plays an important role in the early phase of aldosterone-stimulated Na+ transport.

Aldosterone is a key regulator of Na ϩ homeostasis. Its effects are mediated through the mineralocorticoid receptor, a member of the steroid-thyroid receptor superfamily. Although the central role of aldosterone in the control of renal Na ϩ transport has been known for decades, the exact molecular steps of this regulation are still unknown.
It is generally accepted that aldosterone increases Na ϩ reabsorption by affecting both the apical amiloride-sensitive epithelial Na ϩ channel (ENaC) 1 and the basolateral Na,K-ATPase (reviewed in Ref. 1). Aldosterone's effects on Na ϩ transport can be divided into early and late phases. The early response, a stimulated apical Na ϩ influx, takes place about 0.5-3 h after hormone addition and is thought to be mediated by activation of pre-existing apical channels (2), whereas the late phase of aldosterone action (several hours to days) probably involves synthesis of new Na ϩ channels (3)(4)(5)(6)(7) and Na,K-ATPase molecules (8).
There is a strong indication that activation of Na ϩ channels pre-existing in the membrane occurs through post-synthetic modifications such as phosphorylation or methylation. Regulation of Na ϩ channel activity by protein kinases has been described by several laboratories (9 -12). Importantly, aldosterone was found to increase phosphorylation of specific residues on the ␤and ␥-ENaC subunits in vivo in Madin-Darby canine kidney cells (13). The kinase that mediates this effect has not yet been identified. Aldosterone-enhanced Na ϩ channel activity seems to involve methylation of apical membrane proteins (9,14). A recent report demonstrated that aldosterone increases carboxymethylation of the ␤-ENaC in A6 cells, and carboxymethylation of ENaC reconstituted in planar lipid bilayers leads to an increase in open probability (15).
Both the early and the late effects of aldosterone are transcription-and translation-dependent. However, the specific genes involved in the early response to aldosterone are still unknown. Such immediate early genes might encode enzymes that directly modify ENaC or increase the activity of regulatory proteins that, in turn, activate ENaC. Although previous observations by several laboratories including ours strongly indicate that aldosterone increases the mRNA levels of ENaC subunits (3)(4)(5)(6)(7), this effect does not occur until several hours or days and is thus unlikely to be responsible for the early effects of aldosterone.
Previous attempts to identify aldosterone-induced proteins using two-dimensional gel electrophoresis were largely unsuccessful (16,17). Although these studies yielded several putative aldosterone-induced proteins, these proteins were expressed during the late phase and were probably not specific for aldosterone (16,17). More recently, Attali et al. (18) identified a dexamethasone-induced gene (CHIP) in rat colon; however, it is not induced by aldosterone in the kidney (19). Spindler et al. (20) recently described early induced genes in A6 cells. The high concentrations of aldosterone (1 M) used in that study, combined with the fact that A6 cells do not have functional mineralocorticoid receptors (MRs) (21), suggest that those genes are most likely regulated through glucocorticoid receptors (GR). In addition, the functional significance of those genes in mammalian mineralocorticoid target cells remains to be determined.
The goal of this study was to identify and characterize early induced (or repressed) aldosterone-regulated genes in the native target cells, i.e. in renal cortical collecting duct (CCD) cells. Primary cultures of CCD cells express functional MRs and respond to aldosterone with enhanced Na ϩ transport and in-creased ENaC mRNA expression (5,22). Thus, this model system is ideally suited to identify aldosterone-induced genes.
Here we report that in rabbit CCD cells, aldosterone rapidly induces the levels of sgk mRNA, and this induction does not require de novo protein synthesis, indicating that sgk is a primary aldosterone-induced gene. In addition, when co-expressed in Xenopus oocytes, sgk significantly increases the amiloride-sensitive Na ϩ current, suggesting that it plays an important role in the physiological effect of aldosterone.

EXPERIMENTAL PROCEDURES
CCD Cell Isolation and Culture-CCD cells were isolated from rabbit renal cortex by solid-phase immunoadsorption, using a monoclonal antibody against an ectoantigen on these cells, as described previously (23,24). CCD cells were grown in primary cultures on porous-bottomed dishes as described in Refs. 23 and 24. After reaching confluence, the cultures were incubated in Dulbecco's modified Eagle's medium/F-12 medium with 10% fetal bovine serum, which was charcoal-stripped four times, to avoid the influence of possible corticosterone and cortisol contamination from fetal bovine serum. (According to our measurements, this procedure eliminates Ͼ99.0% of the glucocorticoids present in serum.) After 24 h, the medium was changed to serum-free medium containing 10 nM aldosterone, the specific GR agonist RU28362, the GR-antagonist RU486 (25), or vehicle for 15 min to 24 h. The effect of corticosteroids on transepithelial voltage and amiloride-sensitive current was determined as described previously (22). To inhibit protein synthesis, cultures were pre-incubated for 30 min with 5 g/ml cycloheximide before the addition of the steroid; cycloheximide was continuously present during incubation with steroids.
RNA Isolation and PCR-based Suppression Subtractive Hybridization-From the cultured cells, poly(A) RNA was isolated using Ambion's Poly(A)Pure kit, and total RNA was isolated using TRI Reagent TM (Molecular Research Center, Inc.). PCR-based subtractive hybridization and suppression PCR were performed using methods described in Refs. 26 and 27 and the PCR-Select TM kit (CLONTECH). Two cDNA pools were synthesized using 2 g of poly(A) RNA from each control and aldosterone-treated cells. cDNAs were digested with RsaI for 2 h, and subtractive hybridization and suppression PCR were performed as described previously (26,27) and in the protocol included with the PCR-Select TM kit, with the following modifications. We used modified nested primers by adding a RsaI half-site (AC) plus two different nucleotides at each end. The use of combinations of 16 upper and 16 lower nested primers modified in this way resulted in the subdivision of the resulting cDNAs into 256 subpools. In addition, we used a touchdown PCR for the second round of PCR with three cycles each of annealing at 75°C, 74°C, and 73°C for 2 min, followed by eight cycles at 72°C for 2 min. The final extension was at 72°C for 20 min. The resulting DNA fragments were separated on a sequencing size nondenaturing acrylamide gel and visualized on a FluorImager TM 575 after staining with Sybr-Green I.
Sequencing and Cloning of Differentially Expressed cDNAs-cDNA fragments that were differentially expressed in two control or aldosterone-treated cDNA pools originating from separate RNA preparations and from different subtractive hybridizations were excised from the gel and re-amplified using the same nested primers as in the first amplification. After verifying that the right cDNA fragment was amplified, DNA sequencing was performed by the Dye Deoxy Terminator chemistry on an ABI 373A automated sequencer. The cDNAs were subcloned into pCR-Blunt vector (Invitrogen) and transformed into TOP10 One Shot Competent cells (Invitrogen). Plasmid DNAs were re-sequenced. Gene-specific PCR primers were designed and used for quantitative RT-PCR.
Quantitative RT-PCR-To determine the relative abundance of sgk mRNA in control and steroid-treated CCD cells, we used quantitative RT-PCR methods as described previously (5, 28 -31). cDNA was synthesized using 2 g of total RNA from control or steroid-treated CCD cells (5). Sense (5Ј-GAACCACGGGCTCGTTTCTAT-3Ј) and antisense (5Ј-GCAGGCCATACAGCATCTCAT-3Ј) PCR primers were selected based on the sequence of rabbit sgk. These primers amplify a 298-base pair PCR product. Reactions were performed under standard conditions with four different amounts (10, 2.5, 0.625, and 0.156 ng) of cDNA originating from control or steroid-treated cells. After a 2-min denaturation at 96°C, PCR was carried out for 25 cycles (95°C for 45 s, 57°C for 45 s, and 72°C for 1 min), and then a final extension was done at 72°C for 8 min. The relative abundance of ␤-actin mRNA in each CCD cell sample was determined using primers and conditions as described previously (28). cDNA samples derived from control and steroid-treated cells were always amplified simultaneously. The PCR products were separated on a 5% polyacrylamide gel and quantitated by densitometry using a FluorImager TM 575 (Molecular Dynamics). The slope of the amount of PCR products versus the amount of template cDNA was derived by linear regression. These values were normalized for the amount of ␤-actin mRNA.
Isolation of the Full-length Rabbit sgk by 5Ј and 3Ј Rapid Amplification of cDNA Ends-mRNA isolated from CCD cells was reverse transcribed with a lock-docking oligo(dT) primer, using a Marathon kit (CLONTECH). After second strand synthesis and adaptor ligation, two rounds of PCR were carried out. The first PCR was primed with the appropriate gene-specific antisense primers (primer 1, GCAGGCCATA-CAGCATCTCAT; primer 2, GAACCACGGGCTCGTTTCTAT) and the adaptor-specific sense primer (AP-1; CLONTECH). In the second PCR, a nested adaptor primer (AP-2; CLONTECH) and nested gene-specific primers (primer 1N, CACACCCGAGTATCTTGCACCTGAG; primer 2N, TGGGTTACCTGCACTCTCTGAACATC) were used. Products from the second PCR were fractionated on agarose, purified, reamplified, and sequenced.
Northern Blot Analysis-Northern blot analysis was carried out using standard protocols. In brief, 2 g of total RNA from control or aldosterone-treated (30 min) CCD cells was fractionated on a 1.2% agarose gel containing 1.1% formaldehyde. RNA was transferred to a BrightStar-Plus positively charged nylon membrane (Ambion) and probed with 32 P-labeled antisense rabbit sgk RNA probe generated using the Lig'Scrib and Strip-EZ kits (Ambion) with T7 RNA polymerase. Prehybridization was performed at 65°C for 2 h in Northern-MaX TM Prehybridization/Hybridization Buffer (Ambion). Hybridization was done for 16 h using the same conditions as described for prehybridization. Three washes were carried out at 65°C for 30 min each with 0.1ϫ SSC and 1% SDS, and then two washes were carried out with 0.1ϫ SSC and 0.1% SDS for 20 min each. After the final wash, the blot was exposed to x-ray film.
Generation of 3Ј-UTRϪ Mouse sgk/pSP64poly(A) Construct-The entire 3Ј-UTR of the mouse sgk was removed by HindIII digestion, and the plasmid was ligated into pSP64poly(A) vector (Promega). The ligated DNA was transformed into Escherichia coli DH5␣ cells, and plasmid from the resulting clones was verified by sequencing and linearized with EcoRI.
Expression of ENaC and sgk in Xenopus Oocytes and Measurements of Amiloride-sensitive Currents with the Two-electrode Voltage Clamp-cRNA from ␣-, ␤-, and ␥-subunits of rat ENaC and mouse sgk were synthesized with T7 or SP6 mMESSAGE mMACHINE kit (Ambion). Stage V-VI oocytes were isolated from Xenopus laevis, defolliculated by collagenase treatment, and injected with 1 ng of cRNA of each subunit of ENaC with or without 1 ng of sgk cRNA. Oocytes were kept at 19°C in amphibian solution supplemented with 1 M amiloride for 36 h before experiments. Whole cell currents were measured with constant perfusion (100 mM sodium gluconate, 4 mM KCl, 2 mM CaCl 2 , and 10 mM Hepes, pH 7.4) with a two-electrode voltage clamp (Oocyte Clamp C-725B; Warner Instrument Corp.). The magnitude of the amiloridesensitive current was calculated as the difference of whole cell currents in the absence and presence of 10 M amiloride in the perfusate. Membrane voltage was held at Ϫ100 mV, and IV curves were obtained by changing the voltage in 20 mV steps from Ϫ180 to 80 mV using an ITC-16 A/D converter (Instrutech), the Pulse software (HEKA Elektronik, Lambrecht, Germany), and a Power Macintosh computer. Data were filtered at 250 Hz and sampled at 1 kHz.

Use of the PCR Select and Differential Display Method to
Identify Aldosterone-regulated Genes-Rabbit CCD cells cultured on permeable filters (22,24) were subjected to a shortterm (1-h) aldosterone treatment. mRNAs differentially expressed in the control and aldosterone-treated cells were identified using a combination of suppression PCR-based subtraction hybridization (26,27) and differential display (32) methods. Only those cDNA fragments that showed differential expression (control versus aldosterone) in two separate cDNA pools originating from different animals were further studied.
This method resulted in a highly reproducible pattern of amplified DNA fragments from the subtracted cDNA pools. Despite enrichment for genes preferentially expressed in the control or aldosterone-treated cells, many cDNAs were shared in the two pools. This is to be expected, because the two cell populations differed only in the presence or absence of aldosterone for 1 h, and subtraction of common cDNAs is not complete. However, about 30 -40% of cDNA fragments seemed unique.
Aldosterone Induces the Expression of sgk, a Ser/Thr Kinase in Rabbit CCD Cells-One of the cDNA fragments (ϳ350 base pairs) that showed increased expression in the aldosteronetreated mRNAs was excised, reamplified, and sequenced. BLAST search indicated that this cDNA shows very high homology to human and rat sgk (which is a serum-and glucocorticoid-regulated putative serine/threonine kinase (33,34). Gene-specific PCR primers were then used to amplify a 298base pair amplicon. The identity of this PCR product as the rabbit sgk was verified by sequencing. The nucleotide sequence of this cDNA was 91% and 88% identical to the human and rat sgk (33,34), respectively. The sequence of the rabbit sgk was determined using DNAs obtained by 3Ј and 5Ј rapid amplification of cDNA ends. The predicted amino acid sequence is 98% and 97% identical to the human and rat sgk (Fig. 1), indicating that the differentially expressed mRNA encodes the rabbit sgk. In addition, a significant homology was found between the 5Ј-UTRs of the rabbit, human, and rat sgk.
A representative RT-PCR is shown in Fig. 2A, demonstrating that the level of sgk mRNA is markedly increased in aldosterone-treated cells as compared with control. At the same time, aldosterone treatment did not affect the level of ␤-actin mRNA ( Fig. 2A). Thus, in additional experiments, the relative abundance of sgk mRNA was always normalized to the level of ␤-action mRNA in the same cDNA sample to correct for varia-tion in RNA integrity and efficiency of reverse transcription.
The time course of aldosterone induction of sgk mRNA expression was determined by quantitative RT-PCR, using RNA originating from CCD cultures incubated with vehicle or 10 nM aldosterone for different periods (15 min to 24 h) in serum-free medium. We and others have previously shown that quantitative RT-PCR is a sensitive and accurate method to determine changes in the relative abundance of mRNAs (5, 28 -31, 35, 36). Fig. 2B shows a representative time curve of sgk mRNA levels in control and aldosterone-treated CCD cells. Aldosterone rapidly increased the levels of sgk mRNA; the difference versus control was already significant at 30 min (242 Ϯ 30% of control, p Ͻ 0.005; n ϭ 7). sgk mRNA levels increased further until 4 h and then declined (Fig. 2B), although mRNA levels after 24 h of aldosterone treatment were still significantly higher than control values. Mean values of sgk mRNA levels were 280 Ϯ 30% (n ϭ 11), 385 Ϯ 60% (n ϭ 7), 302 Ϯ 50% (n ϭ 10), and 235 Ϯ 43% (n ϭ 4) of the paired time-control after 60 min, 120 min, 240 min, and 24 h after aldosterone treatment, respectively. The difference versus control samples at the same times was statistically significant at each time point (p Ͻ 0.01; Student's two-tailed t test.) Northern analysis confirmed the early induction because a ϳ 2.4-kb mRNA transcript hybridizing with the rabbit sgk RNA probe was significantly increased in RNA originating from CCD cells after a 30-min treatment with 10 nM aldosterone (Fig. 2C).
sgk is an immediate early gene regulated at the transcriptional level by glucocorticoids and serum, and these stimuli did FIG. 1. Alignment of the deduced amino acid sequence of rabbit sgk with the human, rat, and mouse sequences. The rabbit and mouse sequences were determined in this study, the rat sequence was determined in Ref. 33, and the human sequence was determined in Ref. 34. Amino acid differences between species are highlighted. not require de novo protein synthesis (37). The rapid time course of induction in CCD cells strongly suggests that the induction is also a direct effect. This conclusion was confirmed by testing the effect of aldosterone on sgk mRNA levels during inhibition of protein synthesis. The induction of sgk by aldosterone was unaffected by cycloheximide (Fig. 3), indicating that it does not require de novo protein synthesis.
Aldosterone, particularly at high concentrations, can bind GRs, although its affinity for the GR is significantly lower than for the MR (38). To verify that the effect of aldosterone is mediated through MRs, we studied the effect of aldosterone in the presence of saturating concentrations of the GR antagonist RU486. The induction of sgk mRNA by aldosterone was unchanged when GRs were blocked by RU486 (Fig. 4), indicating that the effect of aldosterone was indeed mediated through the MR. Fig. 4 also shows that in CCD cells, activation of GRs with a pure GR agonist, RU28362, also elevated sgk mRNA expression.
Sequence Determination of the Full-length Mouse sgk-Searching of the GenBank TM expressed sequence tag database with the rabbit sgk sequence for matching sequences revealed several sequences with high homology to fragments of the rabbit sgk. Multiple alignments indicated that the IMAGE Consortium construct clone ID 570181 is likely to contain the complete coding sequence of the mouse sgk. Sequence analysis of this clone revealed a 2429-base pair sequence that showed a 91% nucleotide identity to the rat sgk, and an 88% identity to human sgk. To verify the identity of this clone as the mouse sgk, a restriction analysis was carried out, which revealed the predicted patterns (data not shown). The longest open reading frame of the mouse sgk sequence predicts a 431-amino acid protein with 97%, 96%, and 86% identity to the rat (33), human (34), and shark (39) sgk, respectively (Fig. 1).
sgk Increases the Activity of ENaC when Expressed in Xenopus Oocytes-To examine the functional effect of sgk on ENaC, we co-expressed the mouse sgk with ENaC subunits in oocytes. Previous studies indicated that the half-life of sgk transcripts is extremely short (Ͻ30 min in mammary epithelia; Ref. 37), and its 3Ј-UTR contains AU-rich regions (Ref. 37 and our data) characteristic for short-lived transcripts. Therefore, we generated a construct by eliminating the entire 3Ј-UTR of the mouse sgk cDNA and inserting it into the SP64poly(A) vector. Oocytes were injected with either the three subunits of ENaC alone or in combination with the stabilized mouse sgk. After 36 h of incubation, the amiloride-sensitive components of whole cell currents were measured with the two-electrode voltage clamp technique.
In four independent experiments using different batches of oocytes, we observed significantly larger amiloride-sensitive The levels of sgk mRNA were determined using quantitative RT-PCR and normalized for ␤-actin mRNA as described under "Experimental Procedures." C, Northern blot analysis of the expression of sgk mRNA in control-and aldosterone-treated CCD cells. Rabbit CCD cells were incubated with 10 nM aldosterone for 30 min at 37°C. Two g of total RNA were analyzed by hybridization using a 32 P-labeled antisense sgk cRNA probe.
FIG. 3. Cycloheximide does not prevent aldosterone-induction of sgk. CCD cells were pre-incubated with vehicle or 5 g/ml cycloheximide (CHX) for 30 min at 37°C, and then 10 nM aldosterone or vehicle was added to the cells, and incubation continued for 60 min. sgk mRNA levels were determined by quantitative RT-PCR and normalized for ␤-actin mRNA levels. Ⅺ, vehicle; f, aldosterone; stripped bar: aldosterone ϩ cycloheximide; u, cycloheximide. n ϭ 4 for each group. *, p Ͻ 0.05 using Student's paired t test (two-tailed) when compared with control. Aldosterone versus aldosterone ϩ cycloheximide was not significant. currents in oocytes co-injected with ENaC and sgk than in oocytes injected with ENaC alone. In three of the four experiments, the difference was highly significant (p Ͻ 0.0001). The results obtained with all oocytes (34 in the ENaC group and 29 in the ENaC ϩ sgk group) are summarized in Fig. 5. The mean amiloride-sensitive current measured at Ϫ100 mV in oocytes injected with ENaC alone was 4.42 Ϯ 0.66 A. Oocytes coinjected with ENaC and sgk expressed significantly larger currents with a mean of 9.78 Ϯ 0.90 A (p Ͻ 0.0001) (Fig. 5A). The difference between the two groups was observed at all membrane voltages as shown in the current-voltage relations of Fig.  5B. These results indicate that sgk, directly or indirectly, stimulates the activity of ENaC when both proteins are expressed in the same cell.

DISCUSSION
Several lines of evidence suggest that the early increase in Na ϩ transport by aldosterone involves activation of pre-existing Na ϩ channels in the apical membrane of Na ϩ -transporting epithelia (2,9). However, the gene products responsible for this effect have yet to be identified. The question of whether ENaC or regulatory proteins are direct targets of post-synthetic modifications initiated by aldosterone still remains. Previous studies suggested that both carboxymethylation and phosphorylation of the ENaC or associated regulatory proteins might be involved in this regulation (9 -11, 15, 40, 41).
The main finding of this study is that aldosterone rapidly increases the mRNA levels of a putative Ser/Thr kinase in the native mineralocorticoid target cells, and this kinase is able to activate ENaC when co-expressed in Xenopus oocytes.
sgk is a member of the Ser/Thr protein kinase family. It belongs to a subfamily of Ser/Thr kinases, which, unlike other protein kinases, are predominantly regulated at the transcriptional level (33,34,37). sgk was first cloned from rat mammary epithelial cells (33) and was found to be regulated by glucocorticoids and serum (37). Additional studies revealed that folliclestimulating hormone and alterations in cell volume also regulate sgk expression (34,42). While this paper was under review, Chen et al. (43) reported that dexamethasone, a glucocorticoid, rapidly induces sgk in a Xenopus kidney cell line (A6), and in situ hybridization indicated that sgk mRNA levels are also increased in the kidney of rats treated with aldosterone. These authors also found that the A6 sgk stimulated the current of oocytes injected with Xenopus ENaC. Importantly, sgk is an immediate early gene (33,37,44,45), and there is a functional glucocorticoid response element in its promoter.
The finding that aldosterone rapidly induces the expression of a Ser/Thr kinase in its target cells and that this induction seems to be a direct effect is interesting for several reasons. First, previous studies suggested that protein kinases are involved in the hormonal regulation of ENaC (9 -12). Most importantly, aldosterone increases the phosphorylation of Ser and Thr residues of the carboxyl-terminal region of ␤and ␥-ENaC in vivo (13). Furthermore, the finding that sgk significantly increases the amiloride-sensitive current through ENaC in oocytes strongly indicates that this kinase is important in aldosterone-stimulated Na ϩ reabsorption.
The mechanism by which sgk activates Na ϩ current has not been examined in this study but will be determined in additional experiments. In light of the present results, a plausible mechanism is direct phosphorylation of Ser and Thr residues on ENaC subunits that results in the activation of ENaC. In this respect, it is interesting to note that sgk is also induced by serum (37), and insulin increases phosphorylation of the same Ser and Thr residues on ␤and ␥-ENaC that are phosphorylated after aldosterone (13). Because fetal bovine serum is a rich source of growth factors, it is conceivable that the insulinmediated increase in ENaC activity is also mediated through phosphorylation via sgk. Alternatively, sgk could phosphorylate and thereby activate regulatory proteins, such as enzymes mediating post-synthetic modifications of ENaC. In this respect, it is tempting to speculate that sgk might activate a carboxymethyl transferase, which can be stimulated by aldosterone (15), and lead to Na ϩ channel activation this way. Phosphorylation of ENaC or a regulatory protein might also change the trafficking of ENaC. It is quite conceivable that these possibilities are not alternatives, but rather, aldosterone regulates multiple events simultaneously. For instance, phosphorylation of ENaC or modifying enzymes might proceed simultaneously with methylation, resulting in activation of ENaC due to changes in kinetics or open probability. At the same time, aldosterone, through sgk or other early induced/repressed gene products, may also regulate the trafficking of ENaC, thereby increasing the number of channels in the membrane.
An interesting observation is that in human hepatoma cells, sgk is induced by changes in cell volume: shrinkage increases sgk mRNA levels, whereas swelling reduces sgk mRNA levels (39). This would seem logical if sgk played an important role not only in the regulation of Na ϩ entry but also in the control of cell volume. On one hand, aldosterone induces sgk, which activates Na ϩ channels, thereby increasing intracellular Na ϩ , resulting in an increase in cell volume. On the other hand, cell swelling rapidly decreases the expression of sgk, creating a negative feedback in cell volume regulation. Such feedback could also explain the decline in sgk mRNA levels at 24 h despite the continuous presence of aldosterone (Fig. 2).
In this study, we also found that sgk mRNA is induced by glucocorticoids in CCD cells, similarly as in other cell types (37,43). Because aldosterone has some affinity for the GR, it was important to verify that aldosterone acted through the MRs. Data obtained with receptor-specific antagonists indicate that the induction of sgk by aldosterone in CCD cells is indeed mediated through MRs. Thus, activation of either the MR or the GR results in the activation of the same gene in CCD cells. This is not surprising, because in vitro MR and GR bind to the same consensus sequences (glucocorticoid response element), and no unique mineralocorticoid receptor elements have been identified thus far. Despite the ubiquitous expression of the GR, mineralocorticoid target cells in vivo respond only to aldosterone, due to the cell-specific expression of the enzyme 11␤-hydroxysteroid dehydrogenase-2 that eliminates endogenous glucocorticoids (46). However, if mineralocorticoid target cells are treated with synthetic glucocorticoids, which are not subject to degradation by 11␤-hydroxysteroid dehydrogenase-2, activation of the GR by glucocorticoids results in an increase in transepithelial Na ϩ transport similar to that seen following activation of the MR by aldosterone (22).
In summary, this study demonstrates that aldosterone rapidly stimulates the expression of sgk, a Ser/Thr protein kinase in the native target cells in the kidney. This effect is direct because it does not require de novo protein synthesis and is mediated through MRs. sgk, when coexpressed in Xenopus oocytes with all three ENaC subunits, leads to a significant increase in Na ϩ current, suggesting that this protein kinase plays an important role in the early phase of aldosteronestimulated Na ϩ transport.