| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Departments of 1Internal Medicine and 4Physiology and the 2Graduate Program in Molecular Biology, University of Iowa College of Medicine and 5Veterans Affairs Medical Center, Iowa City, Iowa 52242; and 3Axcell Biosciences, Newtown, Pennsylvania 18940
Submitted 30 May 2003 ; accepted in final form 18 July 2003
 |
ABSTRACT |
|---|
 TOP ABSTRACT EXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
-, 
-, and 
-ENaC have strong affinity for WW domains 3 and 4 but not 1 and 2. These studies indicate that alternate transcripts of Nedd4–2 may interact with ENaC differently. Understanding the function of variant proteins will increase our knowledge of the role of hNedd4–2 in the regulation of ENaC and define protein domains important for Nedd4–2 function. ubiquitin ligase; epithelial sodium channel; sodium transport; Nedd4L
- or -subunit of ENaC results in unregulated Na+ reabsorption and a dominantly inherited form of severe hypertension called Liddle's syndrome (21, 24). It is now apparent that the PY motif in ENaC is a key domain involved in internalization and degradation of the multimeric channel complex and that, in its absence, assembled channels persist at the cell surface leading to an increase in Na+ transport. These studies clearly indicate that apical Na+ transport can be regulated, in part, by removal of assembled Na+ channels at the cell surface and underscore the importance of the "retrieval" pathway in ENaC function.
The discovery, by yeast two-hybrid analysis, that the PY motif of ENaC can bind to the WW domains of Nedd4 has led to a large body of work that indicates that ubiquitin ligases with multiple WW domains are involved in negatively regulating ENaC expression at the cell surface (40). Nedd4 has an NH2 terminal C2 domain and a COOH-terminal ubiquitin ligase domain; when coexpressed with ENaC in Xenopus oocytes, Na+ transport is reduced, in keeping with a role for Nedd4 in reducing ENaC surface expression (9, 11). Furthermore, lysine residues in the NH2 termini of -and -ENaC subunits are targets for ubiquitination, and mutation of these residues in ENaC increases cell surface expression and enhances Na+ transport (3, 41). These results are consistent with a model where ubiquitination of - and/or -ENaC marks the channel complex for endocytosis and possibly degradation. Mutations that disrupt the PY domain of ENaC, as is seen in Liddle's syndrome, increase the surface expression of ENaC, presumably by disrupting its interaction with Nedd4 or related proteins (1, 11).
Although it appears that Nedd4 is a protein that may regulate ENaC activity when heterologously reconstituted with ENaC, the true mediator of ENaC endocytosis in vivo has not been clearly identified. In certain tissues, the pattern of expression of Nedd4 does not strictly coincide with that of ENaC subunits. For example, the ENaC complex is expressed in surface epithelia of the colon, whereas Nedd4 is expressed in crypt cells, suggesting that there may be other ubiquitin ligases that may regulate ENaC (42). A homolog of Nedd4, named Nedd4–2, has been identified in mice and in humans, which physically associates with the endocytosis motif of ENaC subunits and substantially reduces ENaC-dependent Na+ transport in Xenopus oocytes, salivary duct cells, and in Fischer rat thyroid (FRT) cells (12, 18, 39). This molecule, like Nedd4, has a COOH-terminal ubiquitin ligase domain and several WW domains but, unlike Nedd4, appears to lack an NH2-terminal C2 domain. Interestingly, phosphorylation of Nedd4–2 by serum- and glucocorticoid-regulated kinase (sgk) inhibits Nedd4–2 function and may account for the ability of corticosteroids to increase ENaC function via activation of sgk1 (7, 39).
To begin to understand the structural organization of Nedd4–2 and to study its regulation, we have characterized the genomic structure of the Nedd4–2 gene. The human Nedd4–2 gene has a complex gene structure with multiple proximal exons where transcription is alternately initiated, varying polyadenylation signal sequences that create differing lengths of the 3'-untranslated region and internal exons where alternate splicing events occur. As a consequence of this organization, multiple transcripts are expressed in a tissue-specific manner, and these transcripts encode Nedd4–2 proteins with and without the NH2-terminal C2 domain and with a varying number of WW domains.
|
EXPERIMENTAL PROCEDURES |
|---|
|
TOPABSTRACTEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
-32P]UTP, [-32P]dCTP, and [-32P]ATP were obtained from New England Nuclear Life Science Products (Boston, MA). Cell culture media were obtained from Life Technologies (Gaithersburg, MD). DNA sequencing and synthesis were services provided by the University of Iowa DNA core facility. Adult human kidney and lung RNA and a panel of human cDNAs (Multiple tissue cDNA panel) were purchased from Clontech (Palo Alto, CA). Human placenta and fetal tissue were obtained from local sources, as approved by the University of Iowa Human Subjects Review Committee. The human lung epithelial cell lines H441 and A549 and FRT cells (gift from C. Zurzulo, Cornell University) were cultured as previously described (16, 36).
Southern blotting. Human genomic DNA (20 µg) was digested with Bgl II, Xba I, or Pvu II, run on an 0.8% agarose gel, and transferred to nylon membranes (Zetaprobe GT; Bio-Rad, Hercules, CA). The full-length Nedd4La transcript was cloned by RT-PCR from H441 cDNA using primers Nedd4–2 F3 and R3 (Table 1), and the product digested with Sac I and the 3' 200-bp fragment corresponding to exon 30 was gel purified and used as a probe. Briefly, cDNA was labeled by random primer extension with Klenow DNA polymerase and [-32P]dCTP (Decaprime II DNA labeling kit; Ambion, Austin, TX) and then hybridized with the transferred membrane, as previously described (44).
|
5'-Rapid amplification of cDNA ends. Two separate cDNA libraries were used for 5'-rapid amplification of cDNA ends (RACE). The first was a human lung cDNA library (Marathon-ready; Clontech) that has an adapter sequence 5'-CTAATACGACTCACTATAGGGCTCGAGCGGCCGCCCGGGCAGGT ligated to both ends. A reverse primer (Nedd4–2 R2) corresponding to sequence within exon 5, 5'-CCCAGGAAGTCGTCTCGTGT, was used with adapter-specific primer AP1 to obtain the initial set of 5'-RACE clones. Additional 5'-RACE reactions were then performed using exon-specific primers in standard or nested PCR reactions (Table 2). These reactions included the following primers: 5'-flank R1: 5'-AACAGCCTGGAGTGGGAGGT (KIAA) and 5'-flank R2: 5'-GCGCTGCTGGAAATCTACCTT (Nedd18).
|
The second cDNA library was made from H441 cell mRNA using an RNA ligase-dependent RACE kit (First Choice RLM-RACE; Ambion), essentially as previously described (17). Briefly, H441 RNA was first treated with calf intestinal phosphatase to remove 5'-phosphates from transcripts that did not include the 5'-cap. The RNA was then treated with tobacco acid phosphatase to remove the cap structure and an adaptor sequence 5'-GCUGAUGGCGAUGAAUGAACACUGCGUUUGCUGGCUUUGAUGAAA ligated to decapped RNA with T4 RNA ligase. Adaptor-ligated RNA was then reverse transcribed with oligo(dT) and Moloney murine leukemia virus (MMLV) reverse transcriptase (RT) in 1x RT buffer with four dNTPs and RNase inhibitor at 42°C for 60 s to synthesize first-strand cDNA. The synthesized cDNA was subjected to two rounds of PCR using adaptor-specific primers with exon 1f primers, 3rd NeddformR2: 5'-TGGCGAGATCAATTCCAGAAA and 3rd NeddformR3: 5'-TCTGAGAATACGGGACTCTCC.
RT-PCR for Nedd4–2 cDNA fragments. Total RNA was prepared from H441 and A549 cells with the RNeasy Mini Kit (Qiagen, Valencia, CA). Total RNA was prepared from human fetal tissues using Tri-Reagent (Molecular Research Center, Cincinnati, OH) according to the manufacturer's instructions. RT was performed using 1 µg total RNA with oligo(dT) or with random hexamers and MMLV RT at 42°C for 60 min. Fragments corresponding to the unique 5'-ends of various Nedd4–2 transcripts were obtained by amplifying cDNA from various tissues using a exon-specific forward primer (Table 3) and a common reverse primer in exon 5, Nedd4–2 R2: 5'-CCCAGGAAGTCGTCTCGTGT. To confirm the identity of the amplified fragments, the samples were resolved by agarose gel electrophoresis, transferred to nylon membranes, and hybridized with a labeled oligonucleotide probe corresponding to exon 2: 5'-GTCTGGACCAAAGCAAGTTCTC. To identify splice variants between exon 11 and 15, RT-PCR reactions were performed with two sets of primers, WWF and WWR or Nedd4–2 F6 and Nedd4–2R6 (Table 1). The PCR reactions were performed with Taq DNA polymerase (Amplitaq; Roche Biochemicals, Indianapolis, IN), 250 mM dNTPs, and 1.5 mM MgCl2 with annealing temperature between 53 and 59°C for 35 cycles.
|
RNase protection assay. To determine the prevalence of selected 5'-variant alternate transcripts, the 5'-Nedd4–2 cDNA fragments that included either exon 1a, 1n, or 1f and common downstream sequences (264 and 355 bp, respectively; see Table 3) were cloned into pCRXL-topo or pcDNA3 (Invitrogen, Carlsbad, CA), linearized, and used as templates for RNase protection assay (RPA). To determine the prevalence of WW variants, two cDNA templates were used for RPA. The first was a 494-bp Nedd4La fragment amplified by primers Nedd4–2 F1 and R1 (Table 1) from H441 cDNA and cloned into pCRXL-TOPO. The second was a 770-bp fragment of 
WW2,3 that was amplified by primers WWF and WWR (Table 1) from fetal liver and cloned into pcDNA3. To determine the downstream extent of the 3'-UTR, a 250-bp genomic fragment was amplified by PCR using primers Nedd4–2 3'-untranslated region (UTR) F3 and R3 (Table 1), cloned into pCRXL-topo and used as a template for RPA. Briefly, T7 or Sp6 polymerase was used to synthesize [32P]UTP-labeled antisense cRNA probes, hybridized overnight with RNA samples from various cell lines or tissues, and then treated with RNase A and T1, as previously described (25). An 18S rRNA template (Ambion) was used as a control for RNA loading, and RNA-RNA duplexes that were protected from nuclease digestion were resolved by PAGE and detected by autoradiography.
Expression in FRT epithelia. Full-length Nedd4–2B2 (Nedd4La), Nedd4–2C6 (NedL3), and Nedd4–2C2a were amplified by RT-PCR from kidney poly(A)+ RNA (Clontech) or H441 RNA using specific primers Nedd4La F and R, NedL3 F and R, and 1g_RT and Nedd4–2R3 (Table 1) and cloned into pMT3. Human -, -, and -ENaC in PMT3 was constructed as previously described (23). FRT epithelia were grown on permeable supports and transfected with 0.2 µg -, -, and -ENaC and varying concentrations of Nedd4–2 using TFX 50 (Promega, Madison, WI), as previously described (39). The total mass of DNA used per Millicell filter was kept constant at 1 µg by using appropriate amounts of an irrelevant plasmid expressing green fluorescent protein. After transfection (2–3 days), amiloride-sensitive Na+ transport was derived by measuring total and amiloride-inhibited short-circuit current (Isc) in Ussing chambers.
Peptide synthesis and WW domain-PY motif binding assay. The binding affinity of Nedd4–2 WW domains and PY motifs of ENaC were detected and quantified as previously described with some modifications (28). Briefly, Nedd4–2 WW domains 1, 2, 3, and 4 (Table 4) were amplified by PCR and subcloned into pGEX4T-2 (Pharmacia, Peapack, NJ) using Sal I and Not I sites downstream of and in frame with glutathione-S-transferase (GST). GST fusion proteins were expressed in Escherichia coli and purified from lysates following the manufacturer's instructions and quantified by the Bradford assay. Peptide sequences corresponding to the PY motif of -, -, and -ENaC and the Na+/H+ exchanger (NHE3; see Table 4) were synthesized with an NH2-terminal biotin-Lys-Lys-Lys-Gly sequence. Microtiter wells were coated with 0.6 µg GST alone or GST fusion protein in 100 mM NaHCO3, blocked overnight with Superblock TBS (Pierce, Rockford, IL), and then washed with PBS. Each biotinylated peptide was incubated with SA-AP (Streptavidin-Alkaline Phosphatase; Sigma) to generate peptide-streptavidin-alkaline phosphatase complexes, and these were incubated with GST fusion proteins for 1 h at room temperature and then washed with PBS containing 0.1% Tween 20. Bound complexes were detected by the addition of p-nitrophenyl phosphate (Kirkegard & Perry, Gaithersburg, MD) and quantified as absorbance units at 405 nm in a spectrophotometer.
|
|
RESULTS |
|---|
|
TOPABSTRACTEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
|
To begin to examine the genomic organization in more detail and before evaluating the differences at the 5'- and 3'-ends of these transcripts, we performed Southern blot analysis to confirm that these transcripts originated from the same gene. A 190-bp DNA fragment corresponding to exon 30 of hNedd4–2, which is perfectly conserved in these three transcripts, was used as a probe. A single hybridizing band of the appropriate size was seen in Bgl II, Pvu II, and Xba I digested lanes, a result consistent with the conclusion that all these transcripts originate from the same gene on chromosome 18 (Fig. 1B).
Heterogeneity at the 5'-end of hNedd4–2 transcripts. Structurally, the Xenopus, mouse, and hNedd4–2 proteins are highly homologous, with each consisting of four WW domains, two to three sgk1 phosphorylation sites (S221, T246, and S327 in Nedd4La), and a COOH-terminal Hect domain (7, 39). However, although xNedd4–2 has a C2 domain, this domain does not appear to be present in mNedd4–2 (Fig. 2 and Ref. 18). Interestingly, each of the hNedd4–2 transcripts had nucleotide sequences that, if translated, would have homology with the xNedd4–2 C2 domain; the first detected ATG in these transcripts was downstream of the predicted C2 domain. The data suggested that, with an appropriate 5'-end, the C2 domain could be translated in some Nedd4–2 transcripts. This was particularly relevant, because in a closely related gene, Nedd4, deletion of the C2 domain converts it into a more robust inhibitor of ENaC function (20, 38).
|
Given the uncertainty about the 5'-end of these transcripts, and the potential significance of the C2 domain, we performed 5'-RACE in adapter-ligated human lung cDNA using a reverse primer that corresponded to exon 5 of the hNedd4–2 gene. This exon is present in all Nedd4–2 forms in the database. A specific product was amplified by PCR, confirming that the hNedd4–2 gene is represented in the library. The amplified product was cloned into pCRXL-TOPO, and 20 clones were sequenced. Analysis of the sequenced clones confirmed that the cDNA library contained the 5'-ends of Nedd4La and KIAA0439, two of the previously detected clones in the database (Fig. 3A). To simplify the nomenclature of these transcripts and based on their common origin, we have named these forms hNedd4–2A and hNedd4–2B, respectively (Table 5). Two new forms, hNedd4–2A2 and hNedd4–2C, were also identified, which, like KIAA0439 and Nedd4La, diverged upstream of the common exon 2 (Fig. 3B). We then compared cDNA sequence information with the human genome database and confirmed that the 5'-sequences of each transcript arose from one or more distinct exons found within genomic contigs NT_028380.7, AC107896
[GenBank]
, AC090402
[GenBank]
, and AC015988
[GenBank]
. Nedd4–2A initiates in exon 1a and splices directly to exon 2 or through exon 1b to exon 2 (Fig. 3B). Exon 1c is 
50 kb upstream of exon 2 and splices to exon 2 to form KIAA0439. Using a similar 5'-RACE strategy, Dunn et al. (8) had identified the 5'-end of KIAA0439 in human kidney cDNA (isoform III) and another novel 5'-variant that they called isoform VI. Another cDNA form that we identified, Nedd4–2C1, is formed when exon 1d splices through 1e and 1f to exon 2. Interestingly, exons 1d and 1e are just 1 kb apart, but 11 Mb upstream of exon 1f!
|
|
Identification of additional 5'-variants. When we used a cDNA form that included exons 1e, 1f, and 2 as a probe for RPA, we identified an abundant RNA transcript that contained exon 1f and 2 but not 1e (data not shown). This raised the possibility that there were additional exons upstream of exon 1f or that exon 1f was an initiating exon. Because we had been unable to identify exons upstream of 1f (other than 1d and 1e) by 5'-RACE in adaptor-ligated human lung, we presumed that the lung library did not contain all variant 5'-sequences. This situation can arise, for instance, when there were structural constraints to efficient RT of some mRNA species during cDNA library construction. We therefore constructed a new cDNA library from H441 mRNA by the addition of a 5'-RACE adaptor to decapped mRNA with T4 RNA ligase, thus enriching the library with transcripts that include their authentic 5'-ends. We then performed 5'-RACE PCR reactions using reverse primers in exon 1f. Amplified products were ligated into pCR-XLTOPO, and several clones were sequenced, leading to the discovery of several new 5'-variants (Fig. 3A). When we compared the cDNA sequence with that in Genbank, the 5'-sequence in Nedd4–2C2, -C3, -C4, and -C5 appears to arise in exons 1g, 1h, 1k, and 1i, respectively. As this report was being prepared, two additional forms, NedL3 and Nedd4Lb (accession nos. AB071179 [GenBank] and AF385931 [GenBank] ), were also identified that appear to arise in upstream exons 1n and 1l, then splice to exon 1f and exon 2 (Table 5). Transcripts that arise in exon 1g either splice directly to 1f or through 1j to 1f, whereas other 5'-exons appear to splice exclusively to exon 1f. A search of Expressed Sequence Tag (EST) databases with the University of California, Santa Cruz human genome browser at http://genome.ucsc.edu identified two more cDNA clones (BF965237 [GenBank] and BF678906 [GenBank] ) that arise in a novel exon 16 kb upstream of exon 1f; this exon, named exon 1 by Dunn et al. (8), is shown as exon 1m in Fig. 3. Another EST clone (BU189167 [GenBank] ) arises in yet another novel exon, exon 1p and, like all other exons upstream of exon 1f, splices to exon 1f and then to exon 2 and beyond (Fig. 3). Of the 5'-Nedd4–2 variants identified to date, 11 transcripts appear to have unique 5'-exons, suggesting that they arise from distinct transcription initiation sites under the control of independent promoters. In addition to the generation of alternate transcripts by separate transcription initiation, further transcript diversity is achieved by the use of cryptic 5'-splice donor sites with exon 1h and exon 1i (Fig. 4, A and B). These give rise to exon lengths of 243, 351, and 355 nt for exon 1h and 53, 88, and 92 nt for exon 1i (Fig. 4C).
|
Tissue profile for Nedd4–2 5' variants. We identified a total of 12 exons upstream of the common exon 2 that, through an assortment of splicing patterns, gives rise to 14 5'-variant transcripts in H441 cells. To examine the tissue specificity of these 5'-variants, to determine the relative abundance of internal splice variants, and to identify other splicing patterns, we performed RTPCR in a number of adult tissues. We used a unique upstream primer in each 5'-variant exon and a common downstream primer in exon 5 to amplify each Nedd4–2 variant; as a control for RNA quality, we also amplified GAPDH in each tissue in separate but simultaneous reactions. Amplified products were analyzed by agarose gel electrophoresis, transferred to nylon membranes, and hybridized with a radiolabeled internal primer corresponding to exon 2 to enable the detection of additional splicing forms (Fig. 5A). With the exception of exons 1d and 1e, each of the transcripts was identified in pancreas and in H441 cells, with notable differences in the expression pattern in other tissues. Transcripts that included exon 1c were also detected in heart, 1g in liver, and 1h in kidney, although 1k was not detected in additional tissues. Transcripts that contain exon 1a were detected in all tissues tested (data not shown), although we could not detect spliced transcripts where exon 1a spliced to 1b before exon 2, nor did we find evidence for transcripts where exon 1g spliced to exon 1j, suggesting that these were extremely rare. Importantly, in no case did we find an initial exon of any of these transcripts as an internal exon for another transcript, which is further evidence that these initial exons arise independently under the control of different promoters. RT-PCR analysis also allowed us to look at the relative use of internal splice sites in exons 1h and 1i in various tissues. Exon 1h' was the predominant form in pancreas, whereas in kidney and in H441 cells both 1h and 1h' were equally abundant. A number of forward primers designed to correspond to exons 1d and 1e failed to yield an appropriate product by RT-PCR in any tissue.
|
We then selected two of the 5'-variant forms, Nedd4–2A (exon 1a to 2) and Nedd4–2C6 (exon 1n to 1f to 2), and used the 5'-ends of their cDNAs to perform RPA in selected tissues and in two lung cell lines, H441 and A549. This strategy allowed us to simultaneously detect two or more transcripts and thus better define the prevalence of these transcripts relative to all other transcripts. Transcripts containing 1a were less abundant than those without it in A549 cells and in kidney and fetal lung (Fig. 5B). In contrast, exon 1n-containing transcripts were at least as abundant as those without it in kidney and fetal lung (Fig. 5C).
Identified protein variants. Nedd4 and related proteins are characterized by a modular structure where a central core includes a variable number of WW domains and the COOH-terminus contains a Hect domain. Many but not all members of the Nedd4 family have an NH2-terminal C2 domain, which, in a Ca2+-dependent or -independent manner, can function as a protein-protein interaction domain (26, 31). We performed an open reading frame (ORF) analysis for each of the identified transcripts to determine if there were Nedd4–2 proteins that contained a C2 domain. Nedd4–2A (Nedd4La) and Nedd4–2B (KIAA0439) and most other transcripts have an in-frame upstream stop codon that would preclude translation of a C2 domain for these forms. For each of these forms, the longest ORF begins in exon 6 at a perfect Kozak consensus CCATGG (Fig. 6A). Transcripts that utilize exons 1g, 1h', and 1n utilize an upstream start codon within these exons to read through exons 1f to 6 and add a C2 domain to Nedd4–2. Transcripts that begin in exon 1m are also capable of encoding a C2 domain, although a polymorphism within this exon can activate a cryptic splice site disrupting the ORF (8). The C2 domain for each of the forms described is identical because the domain begins in exon 1f, common to all these forms. The differences between these forms, then, are produced by sequence diversity in the first translated exon, which is upstream of exon 1f (Fig. 6B). The nucleotide context in which translation begins in exon 1n is at a perfect Kozak consensus sequence, whereas that for other forms are not optimal, suggesting that there may be differences in translation efficiency between them.
|
Functional effect of NH2-variant Nedd4–2 proteins on Na+ transport. To begin to examine the role of the C2 domain in Nedd4–2 function, we compared the effects of two of the C2 domain forms, including the longest C2 domain form, Nedd4–2C6 (NedL3), with Nedd4La in FRT epithelia. FRT epithelia are a model polarized epithelial cell line, where functional amiloride-sensitive Na+ channels can be reconstituted by heterologous expression of ENaC (36). Although these cells lack endogenous Na+ transport, both Nedd4 and Nedd4–2 transcripts are detectable by RT-PCR (C. Estes, J. Steines, and P. M. Snyder, unpublished observation). Increasing amounts of Nedd4–2 isoforms were coexpressed with -, -, and -ENaC in FRT epithelia, and amiloride-sensitive Isc was measured 2–3 days later. As we have reported previously, Nedd4La was a potent inhibitor of ENaC-mediated Na+ currents (39). Both C2 domain forms also substantially reduced ENaC currents in a dose-dependent manner (Fig. 7). These results suggest that, with and without the C2 domain, Nedd4–2 can inhibit ENaC function when heterologously expressed in FRT epithelia and confirm similar studies previously reported (20). In these previously reported studies, a C2 domain was created by insertion of an upstream translation start codon in an hNedd4–2 cDNA, and the effect of the extended mutant form was compared with a wild-type hNedd4–2 that did not include a C2 domain. When coexpressed in Xenopus oocytes, the addition of a C2 domain did not appear to alter the ability of hNedd4–2 to inhibit ENaC function (20). Our studies with two distinct naturally occurring C2 domain forms of hNedd4–2 suggest that, unlike Nedd4, the C2 domain does not dramatically alter the potency of Nedd4–2 to affect ENaC function in heterologous epithelia (20, 38). It is possible, however, that, in native epithelia, cytosolic localization and association of Nedd4–2 with ENaC may be substantially different with a C2 domain.
|
Diversity within WW domains and putative phosphorylation sites. A comparison of the ORF of each of the three hNedd4–2 transcripts in the database revealed that differences in splicing patterns between exon 12 and 14 lead to differences in the encoded proteins (Fig. 8, A and B). Thus DKFZp434P2422 contains only three WW domains and one sgk1 phosphorylation site, whereas KIAA0439 and Nedd4La each have four WW domains and two or three sgk1 phosphorylation sites. This is interesting because Nedd4 and related proteins interact with the PY motif of ENaC via the WW domains. Given the potential significance of the WW domains and sgk1 phosphorylation sites to Nedd4–2 function, we asked if these and other variants might exist by performing RT-PCR in various tissues using primers that flank WW1 and WW4. We were able to identify Nedd4La and KIAA0439 in each of the tissues we analyzed. Although DKFZp234p2422, which has been amplified before from human testis, was not identified in our studies, we discovered two novel forms in fetal lung or liver or H441 cDNA. These forms arise from differences in splicing between exons 11 and 16 and, when translated, while retaining their reading frame, led to additional protein variants. Form 4 lacks both WW2 and WW3 (WW2,3) and two of the three sgk1 phosphorylation sites and form 5 lacks WW2 (WW2) and one of the phosphorylation sites (Fig. 8B).
|
To determine the relative prevalence of these forms, we performed RPA in certain cells and tissues using a Nedd4La probe that included exons 12 through 15. Our results show that Nedd4La and KIAA0439 (exon 12) are expressed in all tissues examined, and a third variant corresponding to a form where exon 15 is deleted is also seen in some tissues (Fig. 8C). There are tissue-specific differences in the relative proportion of these transcripts. For example, although KIAA0439 is as abundant as Nedd4La in fetal lung, it is substantially less than Nedd4La in human kidney. We did not detect DKFZp234p2422 or WW2 in any of the tissues examined. Because this probe would not detect WW2,3, we performed a second RPA using a WW2,3 probe that included exons 8–11, 16, and 17. Our results confirm that WW2,3 is expressed in all tissues examined. All other Nedd4–2 WW variants identified to date, including Nedd4La, are detected as two bands, 478 and 200 bp in size, and cannot be individually discriminated by this RPA (Fig. 8D).
WW domain and ENaC PY motif interaction. To determine the relative affinity of WW domains of Nedd4–2 for ENaC, we performed an in vitro binding assay using WW domain-GST fusion proteins as immobilized targets. None of the ENaC peptides (, , or ) bound to WW1 and WW2 but bound with high affinity to WW3 and WW4 (Fig. 9). In contrast, the PY motif of an unrelated ion transporter, NHE3, bound to WW2 and WW3 but not to WW1 or WW4. Although we have not performed a detailed characterization of the binding affinities of WW domains for ENaC PY motifs, our studies are in remarkable agreement with a recent study in which the binding kinetics between individual WW domains of Nedd4–2 and full-length ENaC subunits were determined (10). Together, these studies indicate that splice variants that do not contain WW3 and/or -4 may be limited in their ability to interact with each ENaC subunit.
|
Alternate polyadenylation sites to lead to varying lengths of the 3'-UTR. There are differences in the length of the 3'-UTR of the hNedd4–2 transcripts in the database, adding to the diversity of the transcripts identified. Nedd4La and DKFZp234p2422 have identical 3'-UTRs with an apparent length of 304 nt. Examination of the cDNA sequence and the corresponding genomic sequence reveals a polyadenylation signal sequence, AATAAA, 20 nt upstream of the poly(A) tail found in Nedd4La and DKFZp234P2422, confirming that this is the 3'-end of these processed transcripts. The 3'-UTR of KIAA0439 extends beyond that of Nedd4La; the poly(A) sequence at the 3'-end of this transcript is 1,889 nt downstream of the translation stop codon. However, this is not the authentic polyadenylate tail added to the 3'-end of the processed primary transcript by poly(A) polymerase but is contained in the primary transcript because it corresponds to a 37-nt oligo(A) sequence within the genomic sequence. The lack of a polyadenylation signal sequence just upstream of this region is in keeping with this schema. A search for additional polyadenylation signal sequences revealed a consensus sequence, AATAAA, 4,600 bp downstream of that used by Nedd4La. To determine if the alternate 3'-UTR could extend that far, a DNA fragment corresponding to a 250-bp sequence just upstream of this signal sequence was used to synthesize an antisense cRNA transcript, and RPA assay was performed. Our studies demonstrate that a 250-nt protected transcript was identified in all tissues examined (Fig. 10). These results confirm that the use of alternate polyadenylation signal sequences gives rise to processed transcripts that have a 3'-UTR of either 287 or 4,915 nt (Fig. 10).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
40 amino acids in length, contain two conserved tryptophan (W) residues, and appear to mediate the interaction of WW domain proteins with proteins that carry a PY motif. One such interacting partner of Nedd4 is the PY motif in the COOH-terminus of - and -ENaC (40). WW domains 2, 3, and 4 of Nedd4–1 are required for the direct interaction with ENaC subunits, and WW domain 1 may help to bring other proteins/cofactors required for ubiquitination in proximity to the ENaC multimer (9, 13, 22, 35, 38). Recently, a search for homologs of Nedd4 led to the identification of mouse and human Nedd4–2 (12, 18, 38). Nedd4–2 protein binds with higher affinity to ENaC subunits and more robustly inhibits ENaC function in Xenopus oocytes, suggesting that this ubiquitin ligase may be the authentic in vivo regulator of ENaC function (20). Nedd4–2 but not Nedd4 has a number of consensus sgk1 phosphorylation sites, and in heterologous systems Nedd4–2 is phosphorylated and negatively regulated by sgk1 (7, 39). Furthermore, in contrast to Nedd4, the identified and tested forms of mouse and hNedd4–2 did not appear to have a C2 domain (18, 20). Interestingly, Nedd4 is a more potent inhibitor of ENaC when the C2 domain is deleted, suggesting that the C2 domain might contribute to these differences (20, 38).
The identification of differences in function of Nedd4 and Nedd4–2 indicated that the C2 domain might alter the function of Nedd4–2. By BLAST analysis, the previously identified xNedd4 (accession no. AJ000085 [GenBank] ) appeared to be an ortholog of Nedd4–2 and in common with hNedd4–2 contained consensus phosphorylation sites for sgk1. Because xNedd4–2 appeared to have a C2 domain, we reasoned that the hNedd4–2 gene might also encode protein forms that include a C2 domain. We performed 5'-RACE analysis and found 10 different 5'-exons, spanning 250 kb, that appear to either splice to exon 2 or to exon 1f and then to exon 2 to form at least 13 5'-variant alternate transcripts. These transcripts are expressed in a tissue-specific manner, and many of these are expressed in kidney and in lung epithelia and may regulate ENaC in these sites. Three of the 5'-variants encode protein forms that are predicted to have a C2 domain. Importantly, five aspartate residues required for binding of Ca2+ to the C2 domain (34, 43) are perfectly conserved in each of the hNedd4–2 forms that include this domain. We tested two of the C2 domain forms (NedL3 and Nedd4–2C2a) in FRT epithelia and found that it robustly inhibits ENaC activity, similar to a form that lacks a C2 domain (Nedd4La). The function of the C2 domain in Nedd4–2 is not known, although, analogous to Nedd4, the C2 domain may dictate cytosolic localization and/or regulation by Ca2+, especially in native epithelia (29, 30). The potential interaction of hNedd4–2 with Ca2+ is particularly interesting because of the known effects of Ca2+ to inhibit epithelial Na+ transport (4, 6, 27).
By 5'-RACE, we found two 5'-exons, 1d and 1e, that appear to splice to exon 1f, 11 Mb downstream, to create another 5'-variant transcript. However, we were unable to identify Nedd4–2 transcripts that included 1d or 1e by RT-PCR or by RPA. In fact, these two exons appear to be the proximal exons of another transcript NM_138443 [GenBank] and its hypothetical protein BC014003 [GenBank] . Moreover, several discrete genes are present within the 11-Mb interval between exon 1e and exon 1n, and it remains unclear if exons 1d and 1e are really part of Nedd4–2 arising from the trans-splicing of two cotranscribed genes or if it came to be associated with Nedd4–2 because of a rare and illegitimate transcriptional event.
We also identified splice variants that predicted differences in the number of WW domains or sgk1 phosphorylation sites. Although Nedd4La and KIAA0439 (exon 12) were detected in all tissues that we examined, others have reported tissue-specific differences in expression of forms that include exon 12 vs. those that exclude exon 12 (5). We found at least two additional splice variants, of which one is predicted to exclude WW3 from its ORF. Data from a number of laboratories indicate that the individual WW domains of Nedd4 have varying affinities for ENaC and have different abilities to inhibit ENaC activity (13, 22, 38). Although WW1 has almost no detectable interaction with ENaC, WW3 interacts with ENaC at very high affinity and appears to be necessary and sufficient for the inhibition of ENaC activity. A similar theme appears to be emerging for Nedd4–2. Here we show that WW3 and -4 can bind the PY motifs of all three subunits of ENaC with strong affinity. Two other groups have reported that WW domains 3 and 4 are, together, critical for hNedd4–2 activity (10, 20). Kamynina et al. (18) demonstrated that, even after deletion of WW1 and WW2, Nedd4–2 potently reduced ENaC-mediated current in Xenopus oocytes, and Fotia et al. (10) showed that WW3 and WW4 are required for the Na+-dependent inhibition of ENaC in salivary duct cells. Fotia et al. (10) also examined the interaction of individual WW domains with full-length ENaC subunits by Far Western Analysis and surface plasmon resonance. In their studies, WW1 and WW2 had no affinity for full-length ENaC, whereas WW3 and WW4 reacted with ENaC with high affinity. Two recent studies measured the binding affinities between Nedd4–2 WW domains and the PY motifs of ENaC in vitro and demonstrated that WW3 had a greater affinity (3- to 6-fold) for ENaC compared with WW4 (2, 15).
The function of domains WW1 and WW2 in Nedd4–2 are not known. These domains may serve to bring other PY motif-bearing proteins in close proximity to the ENaC complex. This hypothesis has been strengthened by the observation that WW1 of Nedd4 is required for the interaction and processing of a yeast protein Sp23 with WW2 and WW3 (35). Alternatively, these domains may interact independently with a cohort of proteins that are unrelated to ENaC-mediated Na+ transport. In this regard, we show that the PY motif of NHE3 can bind to the WW2 domain of Nedd4–2 in an in vitro assay.
Unlike Nedd4, interaction with serine threonine kinases may regulate the ability of Nedd4–2 to process ENaC. Two recent reports indicate that sgk1 phosphorylates Nedd4–2 at one of three consensus sites to negatively regulate its function (7, 39). Although Nedd4La has all three phosphorylation sites, KIAA0439 has two and DKFZ043402422 has just one site. The new forms, WW2 and WW2,3, have one or two sgk1 phosphorylation sites. Whether this leads to differences in regulation by sgk1 or other kinases is not yet known.
Our data on the variable length of the 3'-UTR and the diversity in length of the alternate exons that initiate transcription predict that a large number of transcripts would be identifiable by Northern analysis. The expression pattern of hNedd4–2 has been examined previously by Northern analysis. With the use of a probe corresponding to the 3' 950-bp of Nedd4La, the only transcripts identified by Chen et al. (5) were 3.6, 3.4, and 3.2 kb in size even though the longest transcript in Genbank, KIAA0439 (AB007899 [GenBank] ), is 4,879 bp in length. Kamynina et al. (20), on the other hand, using a 500-bp probe corresponding to the 5'-portion of KIAA0439 detected 10.5-, 4.2-, and 4.0-kb transcripts, with the 10.5-kb transcript being particularly abundant in the kidney (20). Our data demonstrate that the extended form of the 3'-UTR alone is 4,915 nt in length, and the difference between the smallest and largest transcripts identified thus far could be explained, in large part, by the variation in size of the 3'-UTR.
There is tremendous complexity in the structural organization of the hNedd4–2 gene with transcription of alternate mRNA species, variable internal splicing, and the use of alternate polyadenylation signal sequences. This has major implications for tissue-specific expression of unique Nedd4–2 forms and for differences in the function of encoded protein variants. The significance of varying lengths of the 3'-UTR in hNedd4–2 is not known but may determine mRNA stability. In forms that do not encode a C2 domain, translation begins in exon 6 and ends in exon 30. In forms that encode a C2 domain, translation begins in exons 1g, 1h', 1m, or 1n, splices to exon 1f, and also exons 2 through 30. Within the coding region, at least four additional internally spliced forms are seen that alternately include or exclude exons 12 through 15, creating protein variants differing in the number of WW domains and in the number of sgk1 phosphorylation domains. Additional diversity of the Nedd4–2 proteome may arise from single nucleotide polymorphisms (SNPs) within the coding region. An analysis of the SNP database at http://www.ncbi.nlm.nih.gov/SNP confirms one SNP that has been identified to date in the coding regions of Nedd4–2. This SNP (rs646509), an A-to-C substitution within exon 2, results in an Ile to Leu substitution at position 66 in NedL3. A second SNP, reported by Dunn et al. (8) is a G-to-A change in the last nucleotide of exon 1m, which activates a cryptic splice site 11 bases downstream, resulting in a frameshift and premature stop within exon 1f (8).
Overall, up to 136 different transcripts may be generated from this gene, encoding at least 24 protein variants. Presumably, this complexity is required to dictate tissue or developmental or signal-specific expression of individual forms, with differing subcellular localizations and varying affinity for ENaC and other interacting partners. The studies described here provide the basis for future studies exploring the physiological relevance of multiple promoter usage and alternative splicing in regulating the function of Nedd4–2.
|
DISCLOSURES |
|---|
|
ACKNOWLEDGMENTS |
|---|
The nucleotide sequences reported in this paper will appear in DNA Data Bank of Japan, European Molecular Biology Laboratory, GenBank, and Genome Sequence Database Nucleotide Sequence Databases with accession numbers AY243313 [GenBank] to AY243322 [GenBank] and AY256662 [GenBank] .
|
FOOTNOTES |
|---|
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.
|
REFERENCES |
|---|
|
TOPABSTRACTEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
subunit of the epithelial sodium channel is an aldosterone-induced transcript in mammalian collecting ducts, and this transcriptional response is mediated via distinct cis-elements in the 5' flanking region of the gene. Mol Endocrinol 15: 575-588, 2001.ENaC subunit. Am J Respir Cell Mol Biol 27: 314-319, 2002.This article has been cited by other articles:
|
|
 |
|
  F. Luo, Y. Wang, X. Wang, K. Sun, X. Zhou, and R. Hui A Functional Variant of NEDD4L Is Associated With Hypertension, Antihypertensive Response, and Orthostatic Hypotension Hypertension, October 1, 2009; 54(4): 796 - 801. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 X. Huang, M. Beullens, J. Zhang, Y. Zhou, E. Nicolaescu, B. Lesage, Q. Hu, J. Wu, M. Bollen, and Y. Shi Structure and Function of the Two Tandem WW Domains of the Pre-mRNA Splicing Factor FBP21 (Formin-binding Protein 21) J. Biol. Chem., September 11, 2009; 284(37): 25375 - 25387. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. Usami, S. Popov, E. Popova, and H. G. Gottlinger Efficient and Specific Rescue of Human Immunodeficiency Virus Type 1 Budding Defects by a Nedd4-Like Ubiquitin Ligase J. Virol., May 15, 2008; 82(10): 4898 - 4907. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. S. Raikwar and C. P. Thomas Nedd4-2 isoforms ubiquitinate individual epithelial sodium channel subunits and reduce surface expression and function of the epithelial sodium channel Am J Physiol Renal Physiol, May 1, 2008; 294(5): F1157 - F1165. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Jespersen, M. Membrez, C. S. Nicolas, B. Pitard, O. Staub, S.-P. Olesen, I. Baro, and H. Abriel The KCNQ1 potassium channel is down-regulated by ubiquitylating enzymes of the Nedd4/Nedd4-like family Cardiovasc Res, April 1, 2007; 74(1): 64 - 74. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Sorkina, M. Miranda, K. R. Dionne, B. R. Hoover, N. R. Zahniser, and A. Sorkin RNA Interference Screen Reveals an Essential Role of Nedd4-2 in Dopamine Transporter Ubiquitination and Endocytosis J. Neurosci., August 2, 2006; 26(31): 8195 - 8205. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. Wang, H. Sun, F. Lang, and C. C. Yun Activation of NHE3 by dexamethasone requires phosphorylation of NHE3 at Ser663 by SGK1 Am J Physiol Cell Physiol, October 1, 2005; 289(4): C802 - C810. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. J. Russo, E. Melista, J. Cui, A. L. DeStefano, G. L. Bakris, A. J. Manolis, H. Gavras, and C. T. Baldwin Association of NEDD4L Ubiquitin Ligase With Essential Hypertension Hypertension, September 1, 2005; 46(3): 488 - 491. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
O. A. Itani, J. B. Stokes, and C. P. Thomas Nedd4-2 isoforms differentially associate with ENaC and regulate its activity Am J Physiol Renal Physiol, August 1, 2005; 289(2): F334 - F346. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J.-S. Rougier, M. X. van Bemmelen, M. C. Bruce, T. Jespersen, B. Gavillet, F. Apotheloz, S. Cordonier, O. Staub, D. Rotin, and H. Abriel Molecular determinants of voltage-gated sodium channel regulation by the Nedd4/Nedd4-like proteins Am J Physiol Cell Physiol, March 1, 2005; 288(3): C692 - C701. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. P. Thomas, R. W. Loftus, and K. Z. Liu AVP-induced VIT32 gene expression in collecting duct cells occurs via trans-activation of a CRE in the 5'-flanking region of the VIT32 gene Am J Physiol Renal Physiol, September 1, 2004; 287(3): F460 - F468. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. Fouladkou, R. Alikhani-Koopaei, B. Vogt, S. Y. Flores, L. Malbert-Colas, M.-C. Lecomte, J. Loffing, F. J. Frey, B. M. Frey, and O. Staub A naturally occurring human Nedd4-2 variant displays impaired ENaC regulation in Xenopus laevis oocytes Am J Physiol Renal Physiol, September 1, 2004; 287(3): F550 - F561. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Visit Other APS Journals Online |
