Dexamethasone stimulates transcription of the Na+-K+-ATPase β1 gene in adult rat lung epithelial cells

Hong Hao, Christine H. Wendt, Gurpreet Sandhu, David H. Ingbar


Na+-K+-ATPase plays an essential role in active alveolar epithelial fluid resorption. In fetal and adult alveolar epithelial cells, glucocorticoids (GC) increase Na+-K+-ATPase activity and mRNA levels. We sought to define the mechanism of Na+-K+-ATPase gene upregulation by GC. In a rat alveolar epithelial cell line (RLE), dexamethasone (Dex) increased β1-subunit Na+-K+-ATPase mRNA expression two- to threefold within 3 h after exposure to the GC. The increased gene expression was due to increased transcription as demonstrated by nuclear run-on assays, whereas mRNA stability remained unchanged. Transient transfection of 5′ deletion mutants of a β1 promoter-reporter construct demonstrated a 1.5- to 2.2-fold increase in promoter activity by Dex. All of the 5′ deletion constructs contained partial or palindromic GC regulatory elements (GRE) and responded to GC. The increased expression of promoter reporter was inhibited by RU-486, a GC receptor (GR) antagonist, suggesting the involvement of GR. The palindromic GRE at -631 demonstrated Dex induction in a heterologous promoter construct. Gel mobility shift assays using RLE nuclear extracts demonstrated specific binding to this site and the presence of GR. We conclude that GC directly stimulate transcription of Na+-K+-ATPase β1 gene expression in adult rat lung epithelial cells through a GR-dependent mechanism that can act at multiple sites.

  • sodium pump
  • glucocorticoid receptor
  • glucocorticoid response element
  • promoter
  • alveolar fluid
  • ion transport
  • type II cells

na+-k+-atpase is a transmembrane protein essential for the transport of sodium and potassium across the plasma membrane in an ATP-dependent manner (17, 47). The enzyme is composed of an α-subunit (113 kDa) and a smaller glycosylated β-subunit (35 kDa). The two subunits consist of at least four isoforms α14 and β14 that are expressed in a tissue- and cell-specific manner (11, 24, 25, 31, 52). The α-subunit mediates the catalytic activity of the enzyme, and the β-subunit plays a role in trafficking and anchoring the enzyme to the plasma membrane (29, 34). In the lung, the β-subunit appears to be rate limiting for the assembly of the αβ-complex and regulates the level of enzyme transported to the plasma membrane (22, 32, 35).

Glucocorticoids (GC) regulate the expression of α-and β-subunits of the Na+-K+-ATPase gene in a variety of tissues including lung, kidney, liver, heart, cardiac muscle, and smooth muscle (4, 26, 36, 38, 46, 48). For example, GC increased renal Na+-K+-ATPase α1-and β1-subunit mRNA three- to sixfold and its corresponding activity two- to threefold in neonatal rats but not in adult rats (6, 7). In cultured rat aorta smooth muscle cells, Na+-K+-ATPase α1 and β1 mRNAs were augmented 2.5- and 10-fold by dexamethasone (Dex) and aldosterone (Aldo), respectively. Although Dex-mediated induction of the β1 mRNA occurred only through the glucocorticoid receptor (GR), Aldo-mediated induction of the β1 mRNA utilized both gluco- and mineralocorticoid receptors (33).

GC can regulate Na+-K+-ATPase gene expression through multiple complex mechanisms, including transcriptional, posttranscriptional, translational, and protein activity regulation. In some cells, one or more of these processes may coexist in the regulation of the sodium pump genes. Although transcriptional effects of GC are a standard paradigm, the effects of GC on mRNA stability and translation have been less appreciated until recently (10). In nontransformed rat liver cells, Dex differentially induced Na+-K+-ATPase α1 and β1 mRNA abundance 2- and 40-fold, respectively, but the increased mRNA content was not due to increased transcription of mRNA. The increased mRNA abundance was associated with only a small increase in Na+-K+-ATPase activity (by 9%), suggesting post-transcriptional and/or translational regulation also were involved (4). Devarajan and Benz (14) recently demonstrated that in vitro GC can directly enhance α1-and α3-subunit translation via a GC-modulatory element in the 5′ untranslated region of their mRNA.

GC induction of transcription commonly involves the binding of the GR to a glucocorticoid response element (GRE) site within the promoter. The classic GRE consists of inverted hexanucleotide repeats separated by three nucleotides (GGTACA nnn TGTTCT) (3, 21). In the human Na+-K+-ATPase β1 gene, a GRE site at -650 conferred Dex responsiveness, whereas two other GREs at -1,048 and -276 were necessary only for optimal activation of β1 promoter by the hormone (13).

In previous studies from our laboratory, maternal GC treatment influenced Na+-K+-ATPase mRNA and protein expression in fetal rat lung in vivo (26). Furthermore, we demonstrated that GC increased Na+-K+-ATPase transcription in a fetal rat lung epithelial cell line (FD18) (8). For the present report, we studied the regulation of Na+-K+-ATPase β1 gene expression by Dex in an adult rat lung epithelial cell line. The results indicate that Dex increased β1 mRNA via an increase in transcription, without a change in mRNA stability, similar to the FD18 cell line. Serial deletion studies of the β1 promoter reveal at least one GRE at position -631 that functions as a binding site for the GR. Transcriptional activation by Dex was inhibited by the GR inhibitor RU-486, supporting the role of the GR in Dex upregulation.


Materials. Dex (water soluble) was purchased from Sigma (St. Louis, MO). Dex stock solution (10-4 M) was prepared in PBS. RU-486 (mifepristone) was purchased from Biomol (Plymouth Meeting, PA). RU-486 stock solution (10-2 M) was prepared in DMSO. Actinomycin D (Act D) was purchased from Sigma. Streptavidin paramagnetic particles were purchased from Promega (Madison, WI).

Plasmids. Plasmids containing the rat Na+-K+-ATPase β1 gene promoter (a gift of G. Gick) (30) and various lengths of 5′ flanking region were fused to a promoterless luciferase gene in pGL3 reporter vector (Promega). We identified seven putative GRE sites in the 817-bp upstream transcription start site using the MatInspector V2.2 program on the website at (40). PCR primers around each GRE site were designed to delete one or two GRE binding sites. Amplified PCR fragments were ligated into Bgl II/HindIII sites in the polylinker region of pGL3 vector. These constructs were named β1-794, β1-714, β1-602, β1-390, β1-274, and β1-34. The same cloning strategy was used to clone further upstream 5′ flanking region (from -817 bp to -4.4 kb), and these resulting constructs were named β1-4.4, β1-3.2, β1-2.1, and β1-1.2 (Fig. 1). The plasmid containing 4.4 kb of 5′ flanking region of β1 Na+-K+-ATPase gene was previously generated in our laboratory (unpublished data, Z. Zhao, C. Wendt, and D. Ingbar, GenBank no. AF020685). An amplified PCR fragment from -714 to -602 was ligated into phRLTK vector (Promega) to make the phRLTK-β1(714-602) construct.

Fig. 1.

Structure of Na+-K+-ATPase β1 5′ deletion mutants in luciferase expression vector. The β1 5′ deletion mutants were linked with a promoterless luciferase expression vector pGL3 (Luc). The -817-bp construct contained six half-site glucocorticoid response element (GRE, gray-shaded ovals) and one full-site GRE (black-shaded ovals) at position -631 (A). The further upstream region from -817 bp to -4.4 kb contained 5 full-site GRE (B).

Cell culture. The RLE-6TN rat lung epithelial cell line derived from primary rat alveolar epithelial cells that were transformed with the SV40-T antigen gene was purchased from ATCC (16). The original cell line exhibited characteristics of alveolar type II cells such as lipid-containing inclusion bodies and expression of cytokeratin. RLE cells were maintained in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal calf serum (Invitrogen, Carlsbad, CA), 100 units/ml penicillin, and 100 μg/ml streptomycin. Dex-treated cells were grown in DMEM supplemented with endogenous GC-removed serum (stripped serum) plus indicated concentrations of Dex. Serum was mixed with Bio-Rad AG1-X8 resin (5 g resin/100 ml serum) three times and filtered to remove resin each time to remove endogenous GC (8, 43). This procedure previously was confirmed to remove >99% exogenously labeled trace GC.

RNA isolation and Northern analysis. RLE cells were grown in DMEM/10% FBS to 50% confluence and were then placed in serum-free medium for 24 h to ensure GC-free conditions. The RLE cells were then treated with or without Dex (10-5 and 10-7 M) for various time intervals. Total RNA was isolated by the Tri reagent method (Sigma) at time points of 0, 3, 6, 12, 24, and 48 h. Northern analysis was performed with 10 μg of total RNA in an agarose-formaldehyde gel and transferred to Hybond-XL membrane (Amersham Pharmacia, Piscataway, NJ) (1). The membrane was prehybridized in PerfectHyb Plus (Sigma) at 68°C for 2 h. The blots were hybridized with a Na+-K+-ATPase β1 cDNA probe labeled with Rediprime II Kit (Amersham Pharmacia) at 68°C for 18 h. The membranes were washed two times in each condition for 5 min at room temperature with 2× SSC, 30 min at 50°C with 2× SSC-1% SDS, and 1 h at 68°C with 0.1× SSC-0.1% SDS. The blots were exposed to Kodak X-film or phosphorimaging to quantitate the radioactive signal.

mRNA stability. Stability of Na+-K+-ATPase β1 mRNA was measured as done previously (8, 9). In brief, RLE cells were grown in the medium with or without Dex 10-5 M for 24 h before the addition of Act D (10 μg/ml) to inhibit mRNA synthesis. Cells were treated with Act D for 0, 3, 6, 9, 12, 15, 24, and 27 h. Total RNA was isolated at each time point, and Northern blots were analyzed as described in RNA isolation and Northern analysis. To calculate the mRNA half-life, we represented the intensity of densitometry at each time point as the percentage of the signal at time 0 (which was considered as 100%) and plotted on a log scale as described (9, 50). The data shown were from six or seven independent experiments.

Nuclear run-on assay. The nuclear run-on assay was performed as described by Schubeler et al. (44) and Tao et al. (45). In brief, nuclei were isolated from RLE cells (1 × 107) treated with or without 10-7 M Dex overnight and incubated with 100 μCi (10 μl) α-[32P]UTP (3,000 Ci/mmol) for 30 min in a 30°C water bath. The plasmid DNA containing either full-length β1 cDNA or β-actin (ATCC) was denatured and transferred to Hybond XL membrane at 10 μg/slot as described by Sambrook and Russell (42). The slot blots were prehybridized in PerfectHyb Plus overnight; 13 million counts of RNA were added and then hybridized at 60°C for 72 h. The blots were washed as previously described (45), and the RNA-integrated optical density was determined by Molecular Analysis Software (Bio-Rad).

Transfection and luciferase assays. RLE cells were grown to 40-50% confluence in DMEM supplemented with 10% stripped FBS. We transfected cells with Lipofectamine Plus following the standard method of Invitrogen. Each transfection contained 0.25 μg of plasmid DNA with various deletion mutants of Na+-K+-ATPase β1 gene, 0.1 μg pSV-β-Gal (Promega), 6 μl Plus reagent, and 4 μl Lipofectamine reagent (Invitrogen) in a total volume of 1 ml. After a 3-h incubation, the transfection solution was replaced with fresh DMEM with or without Dex (10-5-10-8 M). Forty-eight hours later, cells were lysed by luciferase assay lysis buffer, and the cell lysate was centrifuged and assayed for luciferase activity with the Luciferase/Renilla Luciferase Assay System (Promega). All luciferase activities were normalized to β-galactosidase activity that was measured using the Galacto-Star method (Tropix, Bedford, MA). For the RU-486 inhibition experiments, after transfection, cells were incubated with 5× 10-6 M RU-486 for 2 h before incubation with 10-8 M Dex. Incubations were continued for 48 h until luciferase assay was performed.

Preparation of cellular extracts. RLE cells were treated with 10-7 M Dex for 3 h before the preparation of nuclear extracts. Nuclear extracts were prepared according to a modified procedure described by Dignam et al. (15), with the omission of dialysis. Protein concentrations were determined by the bicinchoninic acid method (Pierce, Rockford, IL). Whole cell extracts were prepared from Dex-treated RLE cells according to the method of Derfoul et al. (13).

EMSA. A double-stranded synthetic oligonucleotide (ATTGTCCGGATTGAGGTGGTTCAAGC) corresponding to the rat Na+-K+-ATPase β1 gene from position -631 to -606 was used as a probe. Human tyrosine amino transferase gene GRE CTAGGCTGTACAGGATGTTCTGCCTAG was used as a GRE consensus DNA site for competition experiments (Affinity BioReagents, Golden, CO) (41). An oligonucleotide containing an AP2 consensus sequence (Promega) was used as a nonspecific competitor. Oligonucleotides were either end radiolabeled as probes or unlabeled for use as competitors.

Nuclear extracts (5 μg) and 32P-labeled β1-631 (0.3-0.4 ng) were incubated in 30 μl of buffer containing 10 mM Tris (pH 7.5), 4% glycerol, 1 mM MgCl2, 0.5 mM EDTA, 0.5 mM DTT, 50 mM NaCl, and 0.05 mg/ml poly(dI-dC)·poly(dI-dC) for 20 min at room temperature. In some reactions, a 50- to 115-fold excess of cold competitors was incubated for 10 min at room temperature before the addition of radiolabeled probe. The reaction mixture was applied to a 4% nondenaturing polyacrylamide gel and electrophoresed in 0.5× Tris-borate-EDTA buffer at 200 V for 2 h at room temperature. The gel was dried under vacuum and autoradiographed at -80°C.

Western blotting. The -631 oligonucleotide probe used in the EMSA was biotin-labeled for use in purification of DNA binding proteins. The biotinylated probe was mixed with prewashed streptavidin paramagnetic particles (SAPMP) in 1× gel shift binding buffer (see EMSA) at room temperature for 30-60 min. Probe-bound SAPMP were washed three times with 1× gel shift buffer to remove any free probe. The beads were then added to 1,000 μg of whole cell extract in 1× gel shift binding buffer and incubated at room temperature for 20-30 min. The SAPMP mixture was washed four times with 1× gel shift binding buffer, and the particles were magnetically separated from the solution. Proteins bound to SAPMP were loaded onto a 7.5% denaturing gel and separated by electrophoresis. Proteins on the gel were transferred electrophoretically to an Immobilon-P membrane (Millipore, Bedford, MA). The membrane was blocked with 5% dry milk in TBS (20 mM Tris, pH 7.5, 150 mM NaCl) with 0.1% Tween 20 and then incubated with anti-GR antibody 1:500 (Santa Cruz Biotechnology, Santa Cruz, CA) at room temperature for 1 h followed by incubation with an anti-rabbit IgG conjugated to horseradish peroxidase at 1:2,000 (Sigma). Protein detection was via an enhanced chemiluminescence reaction (Amersham Pharmacia).


Dex increased steady-state Na+-K+-ATPase β1 mRNA level. RLE cells were treated with Dex at concentrations of either 10-7 or 10-5 M for various time intervals. Total RNA was isolated for Northern study with β1 cDNA as a probe. Within 3 h of Dex exposure, β1 mRNA steady-state level increased dramatically, approaching peak levels. The maximal mRNA upregulation was achieved within 6 h, demonstrating approximately a threefold increase at both Dex concentrations (Fig. 2).

Fig. 2.

Dexamethasone (Dex) increased β1 mRNA level in a time-dependent manner. RLE cells were treated with Dex at either 10-5 or 10-7 M for various time courses. The relative mRNA levels from Northern blots were analyzed by Molecular Analysis Software (Bio-Rad). Values are means ± SE from 4 separate experiments.

Dex did not alter Na+-K+-ATPase β1 mRNA stability. In some systems, GC regulates gene expression posttranscriptionally. As an example, GC increased human growth hormone mRNA expression by increasing its RNA stability as well as RNA synthesis (37). In the fetal rat lung, GC increases fatty acid synthase mRNA stability (51). To investigate whether the increase in Na+-K+-ATPase steady-state β1 mRNA levels in RLE cells was due to an increase in mRNA stability, we measured half-lives of Na+-K+-ATPase β1 mRNA in the presence or absence of Dex (10-5 M). The calculated half-life for β1-subunit mRNA from multiple experiments was identical, comparing the control with Dex-treated cells (15.0 ± 0.9 h vs. 15.4 ± 0.7 h, Fig. 3). This suggests the increased steady-state mRNA level of β1-subunit with GC is secondary to an increase in transcription.

Fig. 3.

Effect of Dex on β1 mRNA half-life. RLE cells were exposed to 10-5 M Dex for 24 h before the addition of actinomycin D (Act D, 10 μg/ml). Total RNA was isolated at different time points and probed with Na+-K+-ATPase β1 cDNA. Values show the β1 mRNA level at each time point compared with that at time 0, which was designated to be 100% (A). β1 mRNA half-lives are calculated from 6-7 experiments, and values are shown in B.

Dex increased Na+-K+-ATPase β1 gene transcription rate. To directly measure whether Dex increased Na+-K+-ATPase transcription, we performed nuclear run-on assays in isolated nuclei comparing control cells to cells treated with Dex (10 -7 M) overnight. Full-length actin cDNA was used as an internal control for normalization. Dex increased the transcription rate of Na+-K+-ATPase β1 gene an average of 1.8-fold in two independent experiments with similar response in each (1.6- and 1.9-fold, Fig. 4). This correlated to an increase in mRNA steady-state levels of between 1.6- and 2.1-fold (Fig. 2).

Fig. 4.

Dex increased Na+-K+-ATPase β1 transcription. In nuclear run-on assays, each slot contained 10 μg of either Na+-K+-ATPase β1 cDNA or actin cDNA. Nuclei from cells exposed to either 10-7 M Dex overnight or stripped serum were isolated, and RNA was labeled as hybridization probe. The blots were hybridized with 13 million counts of RNA, and the results are shown in A. Data normalized to actin are the average of 2 experiments (B) I.O.D., integrated optical density.

Dex induced Na+-K+-ATPase β1 gene promoter activity. To analyze the regulation of the Na+-K+-ATPase β1 promoter activity by Dex, we transiently transfected RLE cells with DNA constructs containing the β1 promoter (from -794 to +123, Fig. 1A; -4.4 to +151, Fig. 1B) linked to the luciferase reporter gene in the pGL3 vector. For the -794-bp construct, luciferase activity increased 1.8-fold in Dex-treated cells compared with untreated cells (Fig. 5, A and C), and the -4.4-kb construct demonstrated a 1.5-fold increase in luciferase activity (Fig. 5B). The maximal increase in promoter activity by Dex induction was demonstrated in the β1-714 and β1-3.2 constructs, with a 2.2-fold increase of each. The results of 5′ deletion were confirmed in a second plasmid vector (phRL vector from Promega) to rule out vector-specific responses to Dex (data not shown). Serial deletion between -794 and -714 resulted in a fivefold increase in the basal promoter activity, suggesting the presence of strong negative cis-regulatory elements within this region.

Fig. 5.

Dex induction of the β1 promoter activity. RLE cells were transfected with 0.25 μg of luciferase reporter plasmid (pGL3) containing various length of 5′ flanking sequence of β1 gene (up to 794 bp in A and 4.4 kb in B) along with 0.1 μg of β-galactosidase expression plasmid (pSV-β-Gal). Cells were treated with or without Dex (10-5-10-8 M) for 48 h until luciferase assay. Values were normalized to β-Gal activity and shown as means ± SE from 4-6 independent experiments. C: fold increases of luciferase activity comparing Dex-treated cells with untreated control cells. #Significant changes in activity between β1-714 and β1-602, β1-794 and β1-714, and β1-714 and β1-390 (t-test, P < 0.05). D: Dex-induction of β1 (714-602) and its reversed sequence β1 (714-602)R in phRLTK vector.

Computer analysis with MatInspector to look for homologous GRE sites revealed multiple sites in the 5′ flanking region between -350 and -4.4 kb. In the proximal promoter we identified six half-site GREs at positions -788, -758, -700, -594, -589, and -350 (Fig. 1A). The GRE at position -631 was palindromic, consistent with a full GRE and was homologous to a functional GRE site at -650 to -630 in the human Na+-K+-ATPase β1 gene promoter (13). The search also revealed five full palindromic GRE sites in the upstream region from -817 bp to -4.4 kb (Fig. 1B). To determine whether these GREs were functional, we transfected plasmids containing various lengths of 5′ deletions of the β1 promoter into RLE cells and studied for Dex responses. Serial deletion constructs demonstrated variability in the basal transcription activity; however, the fold induction of Na+-K+-ATPase promoter activity by GC remained relatively unchanged (from 1.5- to 2.2-fold) (Fig. 5, A-C). The elimination of all putative GRE sites resulted in minimal promoter activity identical to that of the empty vector.

Statistical analysis revealed a small but statistically significant decrease in promoter induction by GC between β1-714 and β1-602 (P < 0.05, Fig. 5C). To confirm that the full palindromic GRE at -631 was functional, we transfected RLE cells with phRLTK-β1(714-602) construct that contained the -631 site. The results of transient transfection experiments demonstrate that the GRE at -631 activated transcription in response to Dex in a heterologous promoter vector (Fig. 5D). This induction was orientation specific, since the reverse sequence did not result in upregulation by Dex.

RU-486 inhibited Dex induced Na+-K+-ATPase β1 promoter activity. RU-486 exerts its anti-GC activity at several steps of receptor action such as prevention of complete GR transformation and inhibition of chromatin remodeling and transactivation (5). To examine whether Dex-mediated induction of the β1 gene occurs via the GR or alternative mechanism, we transfected RLE cells with reporter plasmids containing various lengths of the 5′ deletion constructs of the β1 promoter followed by pretreatment with RU-486 before the addition of Dex (10-8 M). RU-486 alone, in the absence of Dex, had no effect on β1 promoter expression (Fig. 6A); however, the Dex-induced promoter activity was eliminated by RU-486 for all constructs studied (Fig. 6). These findings demonstrating RU-486 inhibition of Dex-mediated induction of the β1 promoter suggest that the Dex induction occurs through a GR.

Fig. 6.

RU-486 inhibited Dex-induced β1 promoter activity. RLE cells were transfected with luciferase reporter plasmid (pGL3) containing various length of 5′ deletion mutant (up to 794 bp in A and 4.4 kb in B). After 3 h of transfection, cells were treated with 5× 10-6 M RU-486 for 2 h before incubation with 10-8 M Dex. Luciferase activity was assayed at 48 h and normalized to β-Gal activity. All values shown are means ± SE from 4-7 independent experiments.

GR bound to GRE at -631. We studied the full palindromic GRE at -631 in the rat β1 proximal promoter, since it had significant sequence homology to the human GRE at -650 to -630, matching 11 of 12 bases and demonstrated increased promoter activity in deletion analyses as well as the heterologous expression system (Fig. 5, C and D). We utilized gel mobility shift assays to identify protein binding of this site. The synthetic oligonucleotide from -631 to -606 was end labeled as a probe to perform gel mobility shift assays. When this probe was incubated with the nuclear extract from Dex-treated cells, one dominant band (shown by arrow) along with three weak bands were identified (Fig. 7A, lane 2). Competition experiments using a well-defined consensus GRE showed specific competition with the dominant band only (lanes 3 and 4), whereas a nonspecific oligonucleotide (AP2) was a weak competitor (lane 5). These results suggest that the GR was part of the DNA-protein complex that binds in the -631 region. For confirmation, the -631 GRE was incubated with whole cell extracts, and the bound proteins were subjected to Western analysis using a GR-specific antibody. A distinct single band on the Western blot demonstrated the presence of GR in the GRE-protein complexes (Fig. 7B). Supershift experiments using the same anti-GR antibody resulted in a smear of DNA-protein complex at higher molecular weight, rather than a distinct shift (data not shown). This result of the supershift assay suggests more than one protein may be present in this binding assay, resulting in the incomplete access of the GR antibody. The multiple protein complex with the GR has been previously reported (39). The other palindromic GRE located at -1,045, -2,063, -2,487, -2,620, and -4,267 also demonstrated specific binding in vitro with nuclear extracts at the same gel shift position as that of GRE at -631 but with relatively lower affinities than the GRE at -631 (data not shown).

Fig. 7.

Specific binding of -631 GRE to the rat glucocorticoid receptor (GR). A: gel mobility shift assays. 32P-labeled -631 oligonucleotide probe was incubated with nuclear extracts isolated from Dex-treated RLE cells (lane 2). We carried out competition analysis by incubating nuclear extracts with 50- or 115-fold excess cold consensus GRE (lanes 3 and 4) or AP2 (lane 5) for 10 min before addition of the -631 oligonucleotide probe. Lane 1, free probe alone. Arrow, specific binding. B: GR bound to -631 GRE. The -631 DNA-bound proteins purified from streptavidin paramagnetic particles were electrophoresed on SDS-PAGE and silver stained as shown in lane 1. Proteins on the gel were electrophoretically transferred to the membrane and probed with anti-GR antibody (lane 2). Arrow, the GR band.


Alveolar fluid clearance is critical for oxygenation at the time of birth and for reabsorption of edema fluid after lung injury or alveolar edema. Active solute resorption by alveolar epithelium occurs predominantly through apical sodium channels and basolateral sodium pumps. Recent physiological data in adult rats show that Dex increased alveolar epithelial liquid clearance (18). In addition, studies in a fetal alveolar epithelial cell line and adult rat type II cells show that Dex stimulated Na+-K+-ATPase transcription and activity, respectively (2, 8). However, apart from one study of mammalian cells (CV-1) lacking endogenous GR (13), no studies of the molecular mechanism of GC effects on Na+-K+-ATPase transcription in intact cells from any organ have been reported. With data supporting the physiological relevance of GC regulation of Na+-K+-ATPase, we sought the mechanism of this upregulation by Dex using an adult rat lung epithelial cell line. Our results demonstrate that Dex upregulated the transcription of the β1-subunit gene in a GR-dependent fashion and that the palindromic GRE at -631 of the promoter was functional and bound to GR. Understanding the mechanism of Na+-K+-ATPase gene regulation by Dex should provide insight into manipulation of sodium pump expression by GC during development and alveolar edema states.

The various Na+-K+-ATPase isoforms are differentially regulated by GC depending on cell type. In liver cells of adrenalectomized rats, Dex increased the β1-subunit mRNA markedly more than the α1-subunit mRNA (40- vs. 2-fold) (4). Another example of the differential regulation of the β1-subunit mRNA occurred in cultured adult rat alveolar epithelial type II cells (2); after 6, 12, and 24 h of incubation with Dex, β1 mRNA increased by 2-, 3-, and 1.5-fold, whereas the steady-state level of α1 mRNA remained unchanged. These studies suggest that the β1-subunit gene contains elements essential for the response to Dex, whereas the α1-subunit gene might lack these elements or may be indirectly activated. Additional supportive data derives from the mineralocorticoid receptor/GR induction of human α1 and β1 Na+-K+-ATPase gene expression. GR-dependent induction of the human Na+-K+-ATPase gene was greater for the β1-subunit gene than the α1-subunit gene (13, 27). In the lung, the β1-subunit is felt to be the rate-limiting step for functional sodium pumps. In rat fetal lung epithelial cells (FD18), Dex induced a more rapid initial increase in β1 compared with α1-subunit mRNA level (6 vs. 18 h) (8). Therefore, we focused our studies on the β1 gene regulation by Dex.

Our results demonstrate that Dex induced a rapid increase in Na+-K+-ATPase β1 mRNA abundance. A similar increase occurred in a rat liver cell line, with mRNA level peaking at 6 h after Dex induction (4). Our previous studies in a fetal rat lung epithelial cell line (FD18) demonstrated a delay in the maximum mRNA level for β1 gene occurring at 24 h. This discrepancy in the response to Dex of the β1 gene between fetal and adult lung epithelial cells may represent differences that reflect developmental stage or the presence of other transcription factors.

GC can regulate specific mRNA levels at the post-transcriptional level, including altering of RNA stability and stimulating of nuclear-cytoplasmic transport (10, 20). In our study, treatment with Dex did not change mRNA stability. This result from adult cells is consistent with our prior observations in fetal lung epithelial cells where Dex did not alter mRNA stability (8). We did not examine sodium pump protein levels or whether there is also an independent effect of Dex on translation.

Similar to other genes subject to GC regulation, the Na+-K+-ATPase promoter has multiple putative responsive elements for the GR. Transient transfection experiments using 5′ flanking sequences linked with the luciferase receptor gene showed a twofold increase in promoter activity by Dex. However, deletion of serial putative GRE sites had little effect on the magnitude of the Dex-dependent induction. The maximal induction by Dex occurred in construct β1-714, which included the full GRE site at -631. The physiological relevance of these multiple sites remains unknown. First, only some or possibly one site may be physiologically relevant. That is, only certain sites may be accessible for DNA protein binding due to chromatin conformation or the modification of core histones, such as acetylation. These conformational changes and/or histone modifications may not play a role in transient transfection experiments where relatively small pieces of promoter DNA are introduced into the cell. Because Na+-K+-ATPase is considered an essential housekeeping gene and the β1-subunit is rate limiting in lung epithelial cells, there may be sufficient redundancy that weaker GRE assume functional significance as 5′ deletion is augmented. This pattern was seen for the multiple SP1 sites involved in hyperoxic stimulation of Na+-K+-ATPase β1-subunit transcription (49). Similar observations also were reported in the GC induction of human insulin receptor gene (hIR) expression. Although each GRE in hIR was functional in vitro by GC, the contribution of multiple GREs expression was neither additive nor cooperative (28). There is mounting evidence that the GR is recycled within the nucleus between sites of sequestration and active DNA binding during transcriptional activation (12, 19). The presence of GC results in activation and a bias toward enhanced DNA binding over sequestration. This implies that GR interactions with DNA-specific elements are dynamic and the GR may occupy its target sites only transiently. In addition, these target sites may vary depending on DNA accessibility, which in turn may be dependent on chromatin structure and histone modification. This suggests that multiple sites, although not simultaneously, may participate in the GC response. It is noteworthy to point out that the level of transcriptional induction of Na+-K+-ATPase β1 promoter is consistent with the two- to threefold increase in steady-state β1 mRNA by GC observed in vivo. Although this degree of induction is not very large compared with stress-responsive mRNAs, such as heat shock proteins or heme oxygenese, it is quite significant for an essential housekeeping gene whose protein consumes large amounts of ATP.

GC often mediate their effect through the intracellular GR but can act through other signaling pathways. In the case of GRs, the hormone-receptor complexes are transported into the nucleus and bind to a GRE sequence in target genes. GR are known to be present in lung epithelial cells, thus enabling them to be targets for GC action (23). Although multiple GRE within the Na+-K+-ATPase β1 promoter conferred activity in our deletion studies, we chose to study the GRE at -631 since it is a full palindromic GRE, confers the highest activity, and is homologous to the functional mineralocorticoid response element/GRE at positions -650 to -630 on the human β1-subunit promoter (13). Our RU-486 inhibition data indicate that the GR was involved in the transcriptional stimulation by GC. Our in vitro DNA binding assay also showed specific binding of the -631 GRE with nuclear extracts. This specific DNA-protein complex was inhibited by cold consensus GRE sequences, and we demonstrated the presence of GR by a GR antibody in Western blotting. In addition, we confirmed that the GRE at -631 was induced by Dex in the absence of its native gene when this promoter region (-631 to -606) was placed in a heterologous thymidine kinase promoter. Similar full-site GRE located upstream (-1,045, -2,063, -2,487, -2,620, and -4,267) also showed specific binding with nuclear extracts in gel mobility shift assays. However, they had relatively lower affinities compared with the GRE at -631 (data not shown). These distal full-site GRE demonstrated similar Dex induction as the proximal GRE, showing a 1.5- to 2.2-fold increase in promoter activity (Fig. 5B). Again, serial deletion of the distal GRE had little effect on the magnitude of the Dex-dependent induction. The relative contribution of the other GRE sites to transcription in vivo remains unknown.

In summary, we report that GC stimulate the transcription of the Na+-K+-ATPase β1 gene as demonstrated by an increase in the steady-state β1 mRNA levels, no alteration in mRNA stability, and an increase in β1 promoter activity in promoter-reporter transfection experiments. Direct measurements of transcription via nuclear run-on assay confirm this finding. In addition, GR plays a role in this induction, since transcription is inhibited by the GR antagonist RU-486, and GR is present in binding assays. Understanding the molecular mechanism of GC regulation of sodium pump and other transport proteins may improve our therapeutic approaches to treating lung edema during development and lung injury.


This work is supported by the Will Rogers Institute and National Heart, Lung, and Blood Institute Acute Lung Injury Specialized Center of Research Grant P50-HL-50152.


We are grateful to the colleagues in our laboratories for suggestions and help and to Z. Zhou for the cloning of the 4.4-kb Na+-K+-ATPase β1 gene. We thank Drs. Howard Towle and Michele Sanders for critical reading of the manuscript.


  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked ”advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

  • * C. H. Wendt and D. H. Ingbar served as joint senior authors with equal contributions to this paper.


View Abstract