Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

The amiloride-sensitive epithelial sodium channel (ENaC) is a transmembrane complex made up of at least three distinct subunits (, , ) that have now been cloned from a number of species, including rat and human (h) (1). Injection of complementary DNA (cDNA) or messenger RNA (mRNA) encoding ENaC (but not  or ) into Xenopus oocytes is sufficient to induce an amiloride-sensitive sodium current (1), but coexpression of all three subunits greatly increases the current amplitude (2). The effects of point mutations in each of the subunits on amiloride sensitivity suggest that all three subunits are involved in pore formation (6), although the exact stoichiometry remains controversial (7, 8). Each subunit contains two transmembrane domains, short cytoplasmic amino and carboxyl termini, and a large extracellular loop (9, 10). ENaCs belong to a superfamily of ion channels (reviewed by Barbry and Hoffman [11]) that now includes mammalian homologs ENaC, BNaC1, and BNaC2; Caenorhabditis elegans degenerins such as UNC-8; and ligand-gated channels such as FaNaC in Helix aspersa and the mammalian acid-sensing ion channel. The cytoplasmic domains of ENaC subunits may play important regulatory roles because they contain several conserved potential phosphorylation sites for protein kinase C, as well as sequences shown to be involved in ubiquitination and degradation of ENaC (12).

ENaC is localized to the apical surface of salt-reabsorbing epithelia, notably the cortical collecting ducts of the kidney, distal colon, and respiratory airways (13), where it constitutes the rate-limiting step in Na⁺ absorption. It thus plays a major part in salt homeostasis and control of blood pressure, as indicated by the existence of genetic hypertensive and hypotensive diseases associated with ENaC gene mutations such as Liddle's syndrome (17) and pseudohypoaldosteronism type 1 (18). In the lung, ENaC is important in controlling the amount of liquid in the lung air space. The critical role of ENaC function in clearance of lung fluid at the time of birth has been demonstrated in both pharmacologic and genetic studies. Instillation of the ENaC blocker amiloride in newborn guinea pig lungs led to hypoxemia, respiratory distress, and a failure to clear air-space fluid in the subject animals (19). Targeted "knockout" of the gene for ENaC in mice gave rise to newborns that were unable to clear their fetal lung liquid and died within 2 d of birth, despite morphologically normal lungs (20).

In addition to tissue specificity, expression of ENaC is regulated developmentally such that transcripts are upregulated during late gestation and in the neonatal period (21). Recent evidence from Dagenais and colleagues (23) has indicated that gestation-dependent alternative splicing of ENaC mRNA occurs in mouse colon. Hormonal effects include regulation by glucocorticoids (24- 26) and by female gender hormones (27) in lung, whereas aldosterone (28, 29) and arginine vasopressin (AVP) (30) regulate ENaC expression in colon and kidney. Reports of the effect of tri-iodothyronine (thyroid hormone, T3) on ENaC expression have not all been in agreement (see DISCUSSION). In addition, ENaC mRNAs have recently been shown to be positively regulated by increased concentrations of oxygen in adult rat lung (31) and cultured fetal (32) and adult (33) distal lung epithelial cells.

To understand the factors that mediate regulation of ENaC mRNA levels, we have isolated and characterized genomic clones containing rat ENaC (rENaC), and determined the transcriptional start site of the gene in adult kidney and lung, fetal lung, and cultured fetal distal lung epithelial cells (FDLE). In addition, we have sequenced approximately 1.5 kb of 5' flanking DNA and analyzed its ability to promote transcription of a reporter gene in transiently transfected cells.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Cell Culture

FDLE were prepared for culture from Wistar rat fetuses of 20 d gestation (term = 22 d) as described previously (32). SV40-transformed monkey kidney cells (COS7) and human lung carcinoma cells (A549) were obtained from American Type Culture Collection (Rockville, MD) and maintained in Dulbecco's modified Eagle's medium (DMEM) with 10% fetal bovine serum (FBS), 100 U/ml penicillin G, and 100 µg/ml streptomycin sulfate.

RNA Preparation and 5' Rapid Amplification of cDNA Ends

Rat RNA was prepared from cultured FDLE, fetal lung (Day 20), and adult lung and kidney using TRIzol reagent (Life Technologies, Burlington, ON, Canada) following the manufacturer's instructions. rENaC-specific primers used in rapid amplification of cDNA ends (RACE) were AHSP1: 5'-ACAGCACCGCCCAGAAG, corresponding to nucleotides 427-411 of Genbank sequence X70497, for the reverse transcription step; and AHSP2: 5'-CGCGGATCCGTCTTCATGCGGTTGTGTTT (nucleotides 411-388) and AHSP3: 5'-AGCGGTGGAAT TCAATCAGTGC (nucleotides 316-295) for nested polymerase chain reaction (PCR). 5' RACE was carried out using the 5' RACE System, version 2.0 (Life Technologies) following the manufacturer's instructions. Amplified products were cloned into pBluescript II KS (Stratagene, La Jolla, CA).

Library Screening

A Wistar rat genomic library in DASH II was obtained from Stratagene. Initially, 800,000 plaques were screened, following transfer to nylon membrane, in Expresshyb (Clontech, Palo Alto, CA) using a mixed probe consisting of [³²P]deoxycytidine triphosphate-labeled full-length cDNA for rENaC (1) and the subcloned 5' RACE product obtained from FDLE. Three clones (3.1, 3.2, and 4.1) were isolated and characterized. Screening of a further 800,000 plaques using a 660-base pair (bp) EcoRI-XbaI fragment of rENaC cDNA as probe resulted in the isolation of clones 1.2, 1.3, and 1.4. Growth of  phage, transfer to nylon membranes, and detection and isolation of positively hybridizing clones followed established procedures (34). Phage  DNA was isolated from purified clones by the QIAgen (Santa Clarita, CA) Lambda Midi kit. Genomic DNA was mapped using a combination of complete and partial restriction digests, Southern blotting, and hybridization to end-specific oligonucleotide probes (T7 and T3 for DASH II vector) (35). EcoRI fragments of  clones were shotgun-subcloned to pBluescript II KS for further analysis (34).

Primer Extension

Two antisense oligonucleotide primers were designed on the basis of the sequence of cloned RACE products. PE2 (5'-TCTGGTCTGGTCCAGCATCATTAG) and PE3 (5'-CCCTGGCCTCCAGCTCCGTGCTAC) were 5'-end labeled with [-³²P]adenosine triphosphate using T4 polynucleotide kinase. Labeled primers were hybridized as described (34) to 25 µg total RNA from rat lung or kidney for 90 min at 62°C, and extended using Superscript II reverse transcriptase (Life Technologies) for 1 h at 48°C. Extended products were analyzed on 8% denaturing polyacrylamide gels alongside a sequencing ladder generated using PE2 and PE3 to prime dideoxy sequencing reactions on a cloned rENaC gene fragment.

DNA Sequencing

All sequencing was carried out using the Pharmacia T7 sequencing kit (Pharmacia Biotech, Baie d'Urfé, PQ, Canada). Complete sequence of 5' RACE products was determined on both strands using subclones from naturally occurring restriction sites. Intron-exon junctions in genomic clones were determined using cDNA-specific primers (selected using OLIGO 4.1 software) on EcoRI fragments subcloned to pBluescript II. The 5' flanking gene sequence was determined completely on both strands using sets of nested deletions generated by exonuclease III/ mung bean nuclease digestion (34).

Long-Distance PCR

Intron size was estimated by amplifying intronic DNA using long-distance PCR on DNA isolated from rENaC clones with primers designed from cDNA sequence. Primers were 18 nucleotides in length with 60 to 70% GC content. Long-distance PCR was carried out using the Expand long template PCR kit (Boehringer Mannheim, Laval, PQ, Canada), following the manufacturer's instructions. Size of introns was estimated by agarose gel electrophoresis.

Reporter Constructs

Cloned genomic DNA fragments containing portions of the putative rENaC promoter were inserted upstream of the secreted alkaline phosphatase (SEAP) gene in the promoterless expression vector pSEAP2-basic (Clontech). A combination of naturally occurring restriction sites and fragments generated by exonuclease digestion were used to assemble reporter constructs.

Transfections

COS7 and A549 cells were plated in six-well tissue-culture dishes at 2.5 to 3 × 10⁵ cells per well 18 h before transfection. Cells were cotransfected with pSEAP2 constructs and a Rous sarcoma virus (RSV)-driven -galactosidase (gal)-expressing plasmid (RSVgal) as an internal control for transfection efficiency. Certain experiments involved cotransfection of expression vectors for human glucocorticoid and T3 receptors (36, 37) as indicated in figure captions. Transfections were carried out using Lipofectamine reagent (Life Technologies) according to the manufacturer's recommendations. Medium was collected 48 h after transfection for analysis of SEAP activity, using a Phospha-Light Chemiluminescent kit (Perkin-Elmer Applied Biosystems Division, Mississauga, ON, Canada). gal activity was determined via a colorimetric assay using o-nitrophenyl--D-galactopyranoside as described previously (38) on cell extracts prepared by lysis of cultured cells in 1% Triton X100/0.25 M Tris/HCl, pH 7.8.

Determination of rENaC mRNA Half-Life in FDLE

FDLE were seeded at 5 × 10⁵ cells/cm² in 25 cm² tissue-culture flasks in modified Eagle's medium, allowed to recover overnight, and cultured in the presence or absence of 5 µg/ml actinomycin D (time 0) (39). Cells were harvested and mRNA was prepared (40) at various time points up to 24 h after actinomycin addition. rENaC mRNA content was determined (relative to time 0) by Northern blot analysis as previously described (32).

Statistical Analysis

Reporter gene activities in transiently transfected cells are presented as means ± standard error, and statistical significances were calculated using an analysis of variance followed by Student-Neuman-Keuls multiple comparisons test (Instat software by Graphpad, Inc.; San Diego, CA). P < 0.05 was considered significant.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Transcription Start Site Analysis

As a prelude to defining the rENaC promoter, the exact 5' end of the mRNA was identified by the 5' RACE technique. This was initially studied in RNA from cultured rat FDLE (Figure 1A, lanes 1 and 2) using primers located approximately 300 bp downstream of the 5' end of the published rENaC cDNA sequence (1). 5' RACE gave a single band of approximately 800 bp from FDLE. The band amplified by 5' RACE was not found in control reactions from which the reverse transcriptase (RT) enzyme was omitted (lane 1), indicating that it did not arise from genomic DNA. This band was isolated from an agarose gel and subcloned, and three independently isolated clones were sequenced (Figure 1B). All clones were identical, ending at the same position and defining a 5' untranslated region (UTR) 518 bp in length. Sequence analysis of the cloned RACE product revealed identity to the previously reported cDNA sequence (Genbank entry X70497) from the primer up to 26 bp from the 5' terminus of the cDNA clone. These 26 bp could not be found anywhere within the 5' RACE product. Noting that this 26 bp possessed restriction sites for a large number of "rare cutting" restriction endonucleases and was present at the 5' terminus of reported cDNAs for rENaC and rENaC isolated from the same cDNA library (Genbank entries X77932 and X77933 [2]), it seems most likely that this is a cloning artifact, possibly part of a polylinker used in cDNA library construction. Becuase rENaC is known to be tissue-specifically regulated, we performed further 5' RACE reactions on RNA from fetal lung (Figure 1A, lanes 3 and 4), adult lung (lanes 5 and 6 ), and adult kidney (lanes 7 and 8). These tissues each gave two major bands, at approximately 800 and 300 bp, as well as faint minor bands. Cloning and sequencing of these products revealed that all were derived from rENaC. The 800-bp clones ended close to the position determined from FDLE (see Figure 1B), whereas the 300-bp clones ended 8 to 10 bp upstream of the end of homology to the published cDNA. Minor bands in fetal lung (lane 4) and adult kidney (lane 8) were identical in sequence to major bands but truncated at the 5' end. Thus, it appears that the rENaC gene may contain at least two transcriptional start sites, approximately 450 bp apart, generating 5' UTRs of different lengths. The additional 5' UTR sequence cloned here does not reveal any additional in-frame methionine codon upstream of the one previously proposed as the translational start site in the cDNA; therefore, no change is predicted in the protein sequence.

View larger version (54K): [in this window] [in a new window]   View larger version (29K): [in this window] [in a new window]   View larger version (75K): [in this window] [in a new window]   Figure 1.   Determination of transcription start site using 5' RACE and primer extension. (A) 5' RACE was carried out as described in the text on rat RNA from cultured FDLE (lanes 1 and 2), fetal lung (lanes 3 and 4), adult lung (lanes 5 and 6), and adult kidney (lanes 7 and 8). One-tenth of final PCR products was analyzed on a 1% agarose gel, stained with ethidium bromide. Lanes 1, 3, 5, and 7 contain negative controls from which RT was omitted. Sizes of marker fragments (1-kb ladder; Life Technologies) are indicated to the left in base pairs. (B) Nucleotide sequence of the 5' untranslated region of rENaC as determined from cloned 5' RACE products and primer extension. The vertical line indicates the start of homology to previously published cDNA sequence (Genbank no. X70497). The 5' ends of clones from various RNA sources are indicated as follows: § fetal lung,  FDLE, ¶ adult lung, * adult kidney. The EcoRI site used in subcloning is underlined. Primers PE2 and PE3 used in primer extension are double underlined. The locations of 5' termini indicated by primer extension are marked with vertical arrows. The translational start site is indicated by a bent arrow. This sequence has been submitted to Genbank under the accession number AF082073. (C) Primer extension was carried out on rat RNA from fetal lung (lanes 1 and 4), adult lung (lanes 2 and 5), and adult kidney (lanes 3 and 6 ) using primer PE3 (left panel ) and PE2 (right panel ). A sequencing ladder generated using PE2 and PE3 as primer on cloned genomic DNA was run alongside each primer extension product. Sequences surrounding the transcriptional start sites are indicated vertically to the right of each panel (uppercase indicates transcribed sequence).

Following isolation of genomic DNA clones for rENaC (see below), we wished to confirm the major transcriptional start sites defined by 5' RACE. We designed antisense oligonucleotide primers (PE2 and PE3; see Figure 1B) 80 to 100 bp downstream of the putative start sites. Primer extension products generated from these oligonucleotides were analyzed alongside a DNA sequencing ladder generated using these same primers on rENaC genomic clone DNA (Figure 1C). Using PE3, a single band was detected in all three mRNA preparations that comigrated with the sequence corresponding to the end of the UTR as defined by 5' RACE. Using PE2, major bands comigrating with the UTR end defined by 5' RACE were detected in fetal lung and adult kidney. Only a faint band was present at this position in adult lung. A minor band was also detected in fetal lung with primer PE2, suggesting that some transcripts may initiate 2 nucleotides further downstream.

Isolation of Genomic Clones

Initial screening of the rat genomic DNA library using a mixed probe covering the entire cDNA resulted in the isolation of three hybridizing phage clones (3.1, 3.2, and 4.1). Further screening with an EcoRI-XbaI fragment of the cDNA led to the isolation of clones 1.3, 1.2, and 1.4. A map of restriction endonuclease sites for EcoRI, BamHI, XbaI, and SalI on the rENaC gene, as determined from these six overlapping clones, is presented in Figure 2.

View larger version (19K): [in this window] [in a new window]   Figure 2.   Restriction map and intron-exon organization of the rENaC gene. The middle line shows restriction sites in the rENaC gene. Restriction endonucleases are abbreviated as: E, EcoRI; X, XbaI; B, BamHI; S, Sal I. Hatched boxes above the middle line indicate the location and extent of individual  clones. Thick lines at the bottom of the figure represent exons. The locations of translational initiation (ATG) and termination (TGA) codons are indicated by vertical arrows.

Characterization of rENaC Gene Intron-Exon Structure

On the basis of the known sequence of rENaC cDNA, a set of 18-mer oligonucleotide primers was designed and synthesized that would enable us to sequence through all coding regions of the gene to determine intron positions and splice junction sequences. Intron-exon boundaries were determined from the upstream transcriptional start site (exon 1A) through exon 12, which ends within the 3' UTR (Figure 2 and Table 1). Intron sizes (Table 1) were determined by complete sequencing or by long-distance PCR on cloned genomic DNA (data not shown). The gene extends over at least 21 kilobase pairs (kbp), and is broken into at least 12 exons, which vary in size from 56 bp in exon 10 to more than 1 kbp in exon 1A. The introns range in size from 74 bp to more than 11 kbp. The 5' splice donor and 3' splice acceptor sites of introns 1 through 11 agree well with published consensus sequences (41). The exception lies at the end of the final exon we characterized, exon 12. Here the homology to cDNA sequence ends 218 bp downstream of the translational stop codon, but the following nucleotide sequence, CAAGCA, does not resemble a 5' donor sequence. We did not investigate the genomic structure of the remaining 700 bp of 3' UTR.

The gene structure of one other ENaC gene, human ENac (hENaC), has been published (42). One notable characteristic of the hENaC gene was the existence of a large intron located within the 5' UTR. Sequence analysis of the rENaC genomic DNA indicates that the sequence present in the longer rENaC transcripts (i.e., the 518-bp 5' UTR), is contiguous with the genomic sequence, and that the first intron occurs 165 codons downstream of the translation initiation site. This indicates that, unlike the hENaC gene, the 5' UTR is not interrupted by an intron (the alternative transcripts must thus arise through use of alternative promoters rather than alternative splicing). The remainder of the gene structure, however, is highly conserved. When we aligned the amino acid sequences of rENaC and hENaC using the PILEUP program (Figure 3), all of the remaining introns previously identified in hENaC were present in rENaC, although the sizes of the introns were poorly conserved. rENaC contained one additional intron, intron 9 in Figure 3, which is absent from hENaC.

View larger version (41K): [in this window] [in a new window]   Figure 3.   Conservation of ENaC genomic organization. Alignment of rENaC and hENaC amino acid sequences. Alignment was produced using the PILEUP program of the GCG sequence analysis software. Positions of introns in hENaC determined by Thomas and associates (42) are indicated by triangles below the alignment. Positions of introns in rENaC are indicated above the alignment. Nucleotide numbering refers to the rat sequence.

Promoter Sequence

The nucleotide sequence of almost 1,500 bp of 5' flanking genomic DNA upstream of the start of exon 1A was analyzed using Genetic Computer Group (GCG) software (Figure 4). There were several consensus binding sequences for two widely expressed transcription factor elements, PEA3 and Sp1. A TATA-like sequence, T T TAA, was noted at 26 bp from the exon 1A start site, although it is not known whether this sequence is functional; no TATA element was present in the appropriate position upstream of the exon 1B start site. Several inducible transcription factor binding elements were noted between nucleotide positions 1,000 and 250. These potential regulatory sites include sites for C/EBP at 982, nuclear factor (NF)-B at 510, and activator protein (AP)-1 at 278. We also detected dyad response elements whose orientation and spacing were optimal for glucocorticoid receptor (GR) (at 791 and 800) and thyroid receptor (T3R) (at 599 and 609).

View larger version (42K): [in this window] [in a new window]   Figure 4.   Nucleotide sequence of the putative promoter region of rENaC. Potential transcription factor binding motifs are boxed. BamHI, XhoI, and HindIII restriction endonuclease sites used in the production of reporter constructs are indicated above the sequence. The transcription initiation sites as determined by 5' RACE and primer extension assay are indicated by bent arrows. The translational initiation codon is underlined. This sequence has been submitted to Genbank under the accession number AF081783.

Characterization of rENaC Promoter Activity

We employed a SEAP reporter assay system to test the ability of rENaC 5' flanking DNA to promote transcription in representative kidney (COS7) and lung (A549) cell lines. Although COS7 cells are not known to express ENaC, they are widely used for promoter analysis due to their ease of transfection. A549 cells have been previously shown to express amiloride-sensitive Na channels, and to contain proteins recognized by antibodies against ENaC subunits (43, 44). Regions of the putative promoter tested are illustrated in Figure 5. The constructs 297SEAP2, 548SEAP2, 1051SEAP2, and 1474SEAP2 contain various amounts of 5' flanking DNA, ranging from 302 to 1,477 bp upstream of the start site of exon 1A, but not including the second transcription start site at the start of exon 1B. The corresponding constructs 297d-, 548d-, 1051d-, and 1474dSEAP2 end at the same set of 5' nucleotide positions, but are extended downstream by 120 bp to include the second transcription start site but not the translation initiation codon.

View larger version (13K): [in this window] [in a new window]   Figure 5.   rENaC promoter-reporter constructs. Upper thick line represents the structural organization of the rENaC promoter (not to scale). Potential transcription factor binding motifs are indicated by boxes. Restriction endonuclease sites for BamHI, XhoI, and HindIII are indicated. Bent arrows indicate transcription initiation sites. Thin lines labeled on the right with the name of the promoter- reporter construct represent genomic DNA fragments corresponding to the region of the rENaC promoter directly above (thick line). Numbers to the immediate left of the thin line represent a nucleotide number of the 5' end of the fragment based on +1 being the "A" at the beginning of exon 1A. Numbers to the immediate right of the thin line denote the 3' end of the promoter fragment.

The ability of each promoter-reporter construct to induce reporter gene activity relative to a promoterless SEAP2 plasmid is illustrated in Figure 6. In COS7 cells, all eight rENaC-SEAP constructs produced an approximately 2-fold increase in activity over the negative control, promoterless SEAP2 (P < 0.001). There were no statistically significant differences between any of the rENaC promoter constructs although, surprisingly, there was a trend to lower activity when the promoter regions were extended to include the downstream transcription start site. In A549 cells, constructs 297-, 549-, 1051-, and 1474SEAP again gave low (2- to 3-fold, P < 0.001) increases in activity over the negative control. The activity of construct 1474SEAP2 was significantly lower than 548SEAP2 (P < 0.01), but not significantly different from 1051SEAP2. In A549, extension of the promoter region to include the downstream transcription start site, rather than increasing activity, significantly decreased activity of each construct (P < 0.001 for 297d relative to 297 and for 548d relative to 548; P < 0.01 for 1051d relative to 1051; P < 0.05 for 1474d relative to 1474).

View larger version (18K): [in this window] [in a new window]   View larger version (16K): [in this window] [in a new window]   Figure 6.   Activity of rENaC promoter-reporter constructs in kidney and lung cells. rENaC promoter-SEAP2 constructs were transiently transfected into COS7 (A, cells transfected with 1.25 µg SEAP2 and 0.05 µg RSVgal) or A549 (B, cells transfected with 0.85 µg SEAP2 and 0.4 µg RSVgal). The "dSEAP2" series of constructs contain further downstream sequences to include the exon 1B transcription start site. Data represent means of three separate transfection experiments, each experiment consisting of triplicate wells for each construct. Error bars are standard error of the mean. SEAP activity (normalized to gal activity) is expressed relative to the promoterless SEAP2-basic vector.

Results from the experiments described above suggest that the basal activity of the rENaC promoter is extremely low. We proceeded to examine whether corticosteroids, a well-known inducer of rENaC mRNA in a variety of animal models and tissue culture systems, could increase activity of the promoter-reporter constructs. COS7 cells were therefore transfected with promoter- reporter constructs 297-, 548-, 1051-, or 1474SEAP2 together with an expression vector directing the synthesis of the human GR. Treatment with 0.1 µM dexamethasone (DEX) for 24 h, initiated 24 h after transfection, induced a doubling in SEAP activity in constructs 1051- and 1474SEAP2 (Figure 7A). This suggests that the DEX- responsive element lies between 552 and 1,054, in agreement with the one identified between 790 and 805 by sequence analysis. This response is similar in magnitude to that obtained with a control reporter construct, GREtkSEAP2, which contains a single-dyad GR element in front of a thymidine kinase-driven reporter (data not shown). A similar pattern of DEX responsiveness was found using the 297d-, 548d-, 1051d-, and 1474dSEAP2 series of constructs, and in A549 cells (data not shown).

View larger version (11K): [in this window] [in a new window]   View larger version (27K): [in this window] [in a new window]   Figure 7.   Hormone responsiveness of rENaC promoter-reporter constructs. rENaC promoter-SEAP2 constructs were transiently transfected into COS7 cells. An RSV promoter-driven gal construct was cotransfected as an internal control. Data represent means of three separate transfection experiments, each experiment consisting of triplicate wells for each condition. Error bars are standard error of the mean. SEAP activity (normalized to gal activity) is expressed relative to the promoterless SEAP2-basic vector. In A, wells were cotransfected with 0.43 µg each of RSV-driven human glucocorticoid receptor (pRShGR), SEAP2 construct, and RSVgal. SEAP activity was assayed in the absence (open bars) and presence (hatched bars) of 0.1 µM DEX in DMEM plus 10% hormone-depleted FBS (36, 59). *P < 0.001 comparing presence and absence of DEX. In B, cells were cotransfected with 0.325 µg each pRShGR, RSVgal, pRShT3R, and 1051SEAP2. Cells were grown in DMEM plus 10% hormone-depleted FBS, or treated with 0.1 µM DEX, 0.1 µM T3, or both drugs together for 24 h before harvesting. *P < 0.05 compared with no hormone treatment. **P < 0.001 compared with no hormone treatment; P < 0.01 compared with DEX treatment alone.

Furthermore, because nuclear hormone receptors are known to function synergistically (45), we decided to investigate whether the consensus T3R element detected between 599 and 613 could confer T3 inducibility on the rENaC promoter, and whether the two hormones together could yield even higher levels of transcriptional activation. Following cotransfection of 1051SEAP2 with expression vectors for both the hGR and hT3R, we tested the effects of DEX and T3 (each at 0.1 mM), alone and combined, on SEAP expression. As shown in Figure 7B, T3 had no significant effect alone but potentiated DEX-mediated activation significantly when the two hormones were added together.

rENaC mRNA Stability

Our results indicate that the basal activity of the rENaC promoter is very low (Figure 6), in contrast with our observations of relatively high steady-state mRNA levels of rENaC in vivo (22). We therefore assessed rENaC mRNA stability by measuring the rate of degradation of rENaC mRNA in FDLE using actinomycin D, an inhibitor of RNA synthesis. To determine rENaC mRNA half-life (t_1/2), FDLE were cultured in the presence or absence of actinomycin D for up to 24 h. Cells were harvested at various times after addition of inhibitor for determination of rENaC mRNA content. Combined results from five experiments are shown in Figure 8. rENaC mRNA decay was log-linear, and had a t_1/2 of ~ 22 h. rENaC mRNA content in FDLE cultured in the absence of actinomycin D varied by < 20% from time 0 throughout the experiment.

View larger version (14K): [in this window] [in a new window]   Figure 8.   Rate of degradation of ENaC mRNA in actinomycin D-treated FDLE. Cells were cultured in the presence of 5 µg/ml actinomycin D from time 0. ENaC mRNA content was determined by Northern blot hybridization at various time points. Values are expressed as percentage of time 0 and are combined from five separate experiments.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Although considerable descriptive information exists in the literature surrounding the tissue specificity and developmental and hormonal regulation of ENaC mRNA expression in a variety of tissues, the mechanisms by which these controls are exerted have not previously been explored. The definition of the gene structure, transcription start sites, and 5' flanking DNA sequence provides the necessary framework for undertaking such studies.

We initiated our study by determining the 5' terminus of rENaC mRNA by 5' RACE, after noting that the reported size of rENaC mRNA in the literature was 3,500 to 3,700 bp (1, 22), whereas the reported cDNA sequence, which included a 982-bp 3' UTR with polyA tail and 82 bp of 5' UTR, was only 3,079 bp. As anticipated, our initial experiment employing oligonucleotides located ~ 300 bp downstream of the previously cloned rENaC cDNA end to prime reactions on RNA from cultured FDLE produced a 5' RACE product approximately 800 bp in length. In 5' RACE experiments using RNA from fetal and adult lung and adult kidney, we detected multiple RACE products, the main two being the 800-bp fragment previously detected in FDLE and a 300-bp product present in all three tissues. Although a single mRNA species for rENaC is most often seen by Northern analysis, we have occasionally observed a second hybridizing species in the region of 18S RNA (22), and two mRNAs for the human homologue of ENaC are routinely observed in kidney and lung, differing by approximately 500 bp (3, 4, 16). We confirmed the length and start sites of both 5' UTRs in kidney and lung mRNA using primer extension analysis.

Assuming that secondary structure effects are not impairing amplification efficiency of the alternative 5' UTRs, the relative intensities of different RACE products within a single reaction should reflect the relative abundance of each transcript within a particular tissue. The intensity of the bands obtained during 5' RACE suggests that the shorter 5' UTR is preferentially expressed in adult tissue (lung and kidney), whereas the longer UTR is more abundant in fetal lung and indeed was the only product detected in cultured FDLE. We note that the primer extension experiments showed only a weak band for the downstream start site in adult lung. Definitive confirmation of the relative abundances of the two transcripts will require the use of methods that do not require an RT step, such as nuclease protection assays.

Evidence for gestation-dependent expression of alternative ENaC transcripts has been previously published by others (23), who demonstrated that a 1.2-kb mRNA was the major mENaC transcript detected during gestation. During the first week after birth, it was gradually replaced by the 3.5-kb mRNA typically detected in lung and kidney. The 1.2-kb mRNA lacked sequences encoding the N-terminal cytoplasmic domain, first transmembrane domain, and part of the extracellular domain, and thus it was likely to result in a nonfunctional channel. Dagenais and associates (23), therefore, speculated a negative regulatory role for the 1.2-kb mRNA because it was replaced by full-length transcripts around the time of induction of amiloride-sensitive Na⁺ transport in suckling rats. Our results indicate that gestation-dependent alternative transcripts may also be important in control of ENaC activity in fetal lung. Further studies of the ontogeny of expression of the alternative transcripts in lung will be important in analyzing the potential role of the alternative 5' UTRs in developmental regulation of ENaC expression.

We have isolated six overlapping genomic DNA clones containing the complete 5' UTR and coding region of rENaC, plus a small portion of the 3' UTR. We deduced the exon-intron structure by a combination of sequencing, restriction-mapping, and long-distance PCR techniques. Comparison of this structure with one previously reported for the human gene for the related ENaC (42) revealed that the genomic structure was well conserved, with the exception that the intron found in the 5' UTR of hENaC is absent in rENaC, whereas rENaC possesses an intron following the codon for Glu535 which is absent in hENaC.

Alternatively spliced variants (ENaCa and ENaCb) have been reported for rENaC by Li and coworkers (46), initially isolated from rat taste tissue by RT-PCR but also detectable in lung and kidney. It is apparent from the genomic structure now in hand that the transcript termed ENaCa arises from the use of a cryptic 3' splice acceptor sequence within exon 8, whereas ENaCb derives from the splicing of exon 7 directly to exon 9.

The nucleotide sequence of the 5' flanking region of DNA in the rENaC gene revealed many potential binding sites for transcription factors. Alignment of this sequence with the previously reported sequence flanking the transcriptional start site of hENaC revealed no extensive regions of significant homology, although both contain multiple consensus motifs for SP1 and at least one potential binding site for PEA3. The PEA3/ETS family of transcription factors is a potential target for signal transduction, particularly the mitogen-activated protein kinase pathway (47). Like hENaC, exon 1B of rENaC lacks a TATA box, but a similar sequence (T T TAA), is found at the appropriate position immediately upstream of the exon 1A transcription start site.

Transient transfection of rENaC promoter-reporter constructs into both A549 and COS7 cells confirmed that a "minimal" promoter containing ~ 300 bp of 5' flanking DNA, the exon 1A transcription start site, and ~ 300 bp of exon 1A 5' UTR could induce about a 2-fold increase in reporter gene activity over a promoterless construct. However, the level of activity was not much affected by increasing the length of the promoter sequence to ~ 1,500 bp. The decrease in reporter gene activity seen when the promoter constructs were extended in the 3' direction to include the second transcription start site suggests the presence of a negative regulatory element. Such an element could be either transcriptional or translational in nature because it would be transcribed in mRNAs initiated at exon 1A.

The low level of activity conferred by promoter- reporter fusions suggests that the sites for "inducible" transcription factors identified within the putative promoter may be extremely important in generating significant levels of transcriptional activity. Because considerable literature exists describing the effects of corticosteroids on ENaC expression in a variety of animal and cell culture models, we investigated the effect of DEX on transcriptional activity. Our results confirm the location of a glucocorticoid-responsive element between nucleotides 552 and 1,054. Although other potential "half-site" consensus motifs for GR were detected by sequence analysis, the results suggest that only the inverted repeats at 791 and 800 contribute to glucocorticoid inducibility. Aldosterone and DEX preferentially bind to mineralocorticoid receptors (MRs) and GRs, respectively, which subsequently bind to the same cis-acting sequences within target genes. Thus, aldosterone would be expected to act through this element in its target tissues (colon and kidney) via the MR expressed there, and DEX through GR in the lung.

The effect of T3 on ENaC expression is controversial. In thyroidectomized fetal lambs, combined treatment with both thyroid hormones and corticosteroid was required to induce the normal activation of Na⁺ absorption in the lung in response to -agonists (48, 49). In the colon of hypothyroid rats (50), amiloride-sensitive short-circuit current is suppressed, even in the face of elevated aldosterone levels. This suppression is relieved by the administration of exogenous T3. Analysis of rENaC mRNA in FDLE from Day 21 rat fetuses revealed a significant increase in response to 0.1 µM DEX, which was further upregulated by the addition of 0.1 µM T3 (25). T3 alone produced no effect on the level of rENaC mRNA in this system. Conversely, treatment of pregnant rats (16 through 22 d gestational age) with thyroid-releasing hormone (TRH) and DEX, alone or in combination, showed that although DEX alone could increase fetal lung rENaC mRNA levels, TRH had no effect either alone or in combination with DEX (24). In these studies it was assumed, but not proven, that the hormones administered to the mothers altered the fetus' T3 levels. Similarly, treatment of cultured human fetal lung explants with DEX induced rENaC mRNA 2- to 3-fold, whereas treatment with T3 had no effect on rENaC mRNA levels in either the presence or absence of DEX (26). Differences in the experimental protocols and model systems used could explain the disagreement among published reports of T3 effects. Our results using transient transfection of rENaC promoter-reporter constructs support the studies that suggest that T3 potentiates the stimulatory effect of DEX on rENaC expression.

AVP is also known to regulate Na⁺ absorption in kidney in the short run, primarily by promoting translocation of ENaC from intracellular pools to the apical membrane (51). Long-term increases have been proposed to be mediated at the transcriptional level via cyclic adenosine monophosphate (cAMP)-responsive elements. Recent studies, however, have demonstrated an increase in steady-state mRNA levels for - and rENaC, but not rENaC, in a rat cortical collecting duct cell line after 5 to 6 h treatment with AVP (30). Our analysis of the putative rENaC promoter, which detected no consensus cAMP-responsive element motifs within 1,500 bp of the transcriptional start site, is consistent with this observation.

The role of other elements we have detected remains to be confirmed by experimental studies. Among the putative transcription factor binding sites identified within the rENaC promoter, AP-1 is redox-sensitive and NF-B activity is induced by reactive oxygen species, making both of these transcription factors candidates for mediators of oxygen inducibility of ENaC (31, 52). We have recently observed that the DNA-binding activity of NF-B, but not of AP-1, is induced in FDLE cultured under 21% oxygen relative to 3% oxygen (53).

Both Northern blot analysis and in situ hybridization indicate that ENaC is the most abundant ENaC mRNA at the cellular level (13, 16). We have recently determined ratios of 5:1 for : and 20:1 for : in human nasal turbinate epithelium by quantitative PCR (54). Although the basal level of transcription from the rENaC promoter is quite low, we have shown that the half-life of the mRNA is relatively long, greater than 20 h. Thus, the long half-life may account for the relatively high steady-state level of rENaC mRNA. Stoichiometry models suggest either equal numbers of , , and  subunits in ENaC or a 2:1 ratio of  to each of the other two (7, 8). To achieve these ratios, post-transcriptional mechanisms may attenuate the synthesis of rENaC protein from its relatively abundant mRNA. Translational regulation is a well-known phenomenon in development, wherein abundant mRNAs are held sequestered within the oocyte until signaled by fertilization to begin protein synthesis (55). The use of alternative transcription start sites to produce mRNAs with extended 5' UTRs has been shown to be a means of modulating translational efficiency in a tissue-specific manner (56). Careful analysis of the relative abundance of the alternative mRNAs in fetal and adult tissues, plus further deletion constructs within the 5' UTR of rENaC, will be required to investigate fully the potential role of translational control in the pathway leading to production of functional ENaCs. Transcriptional control is a rather slow process. In light of the importance of upregulation of Na⁺ absorption in the lung at the time of birth to clear the air spaces of fluid, and of the relative abundance of the longer mRNA in fetal lung and FDLE, it is intriguing to consider that the high level of rENaC mRNA in fetal lung just before birth might represent a pool of translationally silent messages, awaiting an as-yet-undefined signal at birth to release them for a rapid increase in channel production and activity to clear the lungs of fluid for the initiation of air breathing.