The epithelial Na+ channel (ENaC) is regulated by the ubiquitin-protein ligase Nedd4–2 via interaction with ENaC PY-motifs. These PY-motifs are mutated/deleted in Liddle's syndrome, resulting in elevated Na+ reabsorption and hypertension explained partly by impaired ENaC-Nedd4–2 interaction. We hypothesized that Nedd4–2 is a susceptibility gene for hypertension and screened 856 renal patients and healthy controls for mutations in a subset of exons of the human Nedd4–2 gene that are relevant for ENaC regulation by PCR/single-strand conformational polymorphism. Several variants were identified, and one nonsynonymous mutation (Nedd4–2-P355L) was further characterized. This mutation next to the 3′ donor site of exon 15 does not affect in vitro splicing of Nedd4–2 mRNA. However, in the Xenopus oocyte expression system, Nedd4–2-P355L-dependent ENaC inhibition was weaker compared with the wild type (Nedd4–2-WT), and this difference depended on the presence of intact PY-motifs on ENaC. This could not be explained by the amount of wild type or mutant Nedd4–2 coimmunoprecipitating with ENaC. When the phosphorylation level of human Nedd4–2 Ser448 (known to be phosphorylated by the Sgk1 kinase) was determined with a specific anti-pSer448 antibody, we observed stronger basal phosphorylation of Nedd4–2-P355L. Both the phosphorylation level and the accompanying amiloride-sensitive Na+ currents could be further enhanced to approximately the same levels by coexpressing Sgk1. In addition, the role of the two other putative Sgk1 phosphorylation sites (S342 and T367) appears also to be affected by the P355L mutation. The differential phosphorylation status between wild-type and mutant Nedd4–2 provides an explanation for the different potential to inhibit ENaC activity.
- genetic variation
- sodium ion homeostasis
the epithelial na+ channel (ENaC), composed of three homologous subunits (α, β, and γ), plays a fundamental role in the control of the whole body Na+ balance and consequently regulation of blood volume and pressure (26, 52). It is expressed in the apical membrane of segment-specific cells of the aldosterone-sensitive distal nephron (30), and facilitates the entry of Na+ in the cell. The regulation of ENaC is complex and involves the action of various hormones, including aldosterone, vasopressin, or insulin. The cell surface expression of ENaC appears to be tightly regulated, both by clathrin-dependant endocytosis (43, 48) and by ubiquitination via the ubiquitin-protein ligase Nedd4 (neuronal precursor cell expressed developmentally downregulated; also referred to as Nedd4–1; see Refs. 1 and 15), its closely related isoform Nedd4–2 (12, 18, 23, 25, 49), and/or WWP2 (WW-domain containing protein; see Ref. 35). Ubiquitination (also referred to as ubiquitylation) is a posttranslational modification in which ubiquitin, a polypeptide of 76 amino acids, is covalently linked to the ε-NH2 group of lysines in a target protein. This process involves an enzymatic cascade including the ubiquitin-activating enzyme E1, ubiquitin-conjugating (or ubiquitin-carrier) enzyme E2, and ubiquitin-protein ligase E3, the latter being involved in substrate recognition (14). In the case of ENaC, Nedd4–1, Nedd4–2, or WWP2 bind via WW-domains to the COOH-terminal PY-motifs on the ENaC subunits and reduce in a ubiquitination-dependent manner the cell surface expression of ENaC (1, 9, 15, 35, 50). Indeed ENaC is known to be ubiquitinated on its α- and γ-subunit and has been shown to be regulated via ubiquitination (51). Interestingly, the PY-motif of either β- or γ-ENaC is deleted/mutated in most forms of Liddle's syndrome, an inherited form of human hypertension (16, 17, 44). This disease is characterized by the early onset of salt-sensitive hypertension, hypokalemia, metabolic alkalosis, and low circulating aldosterone and renin levels. It is thought that this is the result of elevated ENaC activity, most likely because of reduced Nedd4–2/ENaC interaction, ENaC ubiquitination, and internalization (50, 51). Recently, Sgk1 (serum and glucocorticoid-induced kinase 1) has gained considerable interest with respect to ENaC regulation (24), since Sgk1 is one of the earliest proteins that are induced by aldosterone. It was also found to stimulate ENaC activity (6, 36) by increasing ENaC cell surface expression (3, 31) in Xenopus laevis oocytes and to promote transepithelial Na+ transport in renal epithelial cells (2, 11, 20); a Sgk1 knockout model shows defects in the proper handling of the Na+ balance when exposed to a low-Na+ diet (54). It was recently proposed that Nedd4–2 is a potential target of Sgk1, since it contains two or three putative phosphorylation sites for Sgk1 phosphorylation. Indeed, when Sgk1 is coexpressed with Nedd4–2 in X. laevis oocytes, it is able to phosphorylate X. laevis Nedd4–2 on Ser444 (corresponding to Ser448 in human Nedd4–2) and, though to a lesser extent, on S338 (S342 in human). Such phosphorylation interferes with ENaC interaction, hence with ubiquitination and internalization, and may therefore represent one of the mechanisms by which Sgk1 controls cell surface expression of ENaC (7, 47, 49).
Although Nedd4–2 (and Nedd4–1 or WWP2) are potential susceptibility genes for hypertensive or hypotensive disorders, no genetic linkage has been established so far. Nedd4–2 is encoded on chromosome 18q21, GenBank Locus ID 23327, a gene that spans ∼350 kb and comprises ∼34 exons (Ref. 10 and Fig. 1). There is suggestive linkage between this chromosomal region and several hypo- or hypertensive pathologies, including essential hypertension (4, 29), postural change in systolic blood pressure (39), and orthostatic hypotensive disorder (8).
In this study, we have searched for naturally occurring human Nedd4–2 variants in both controls and patients. Among several variants (see Table 2), we identified a nonsynonymous mutation changing proline-355 to leucine (Nedd4–2-P355L). We have characterized this mutant in X. laevis oocytes and found that it less potently inhibits ENaC compared with wild-type Nedd4–2. This difference is most likely related to altered phosphorylation of Nedd4–2, expected to interfere with the ENaC-Nedd4–2 interaction.
The subjects investigated have been described previously (32, 37, 38, 55). Briefly, the study comprised 362 Caucasian patients with end-stage renal disease (ESRD; 102 on dialysis and 260 after kidney transplantation) from our Division of Nephrology and Hypertension of the University Hospital in Bern, Switzerland, and 494 subjects without renal disorders. More than 70% of the patients with ESRD suffered from arterial hypertension. From the 494 subjects without renal diseases, 193 had hypertension, 42 patients had diabetes, and 259 were healthy controls. The female-to-male ratio of the subjects studied was 45 to 55 and the age ranged from 17 to 92 yr. Genetic analysis of the Nedd4–2 gene was performed in all 856 subjects. All participants gave informed consent to the study, which was approved by the local ethical committee of the University Hospital of Berne.
DNA preparation and PCR analysis.
Genomic DNA was isolated from peripheral blood using the Nucleon BACC3 DNA extraction kit (Amersham International, Buckinghamshire, UK). Mutation detection in the indicated exons of the Nedd4–2 gene was performed by single-strand conformational polymorphism (SSCP) of PCR products from exons 15, 16, 18, 19, 20, 21, 22, and 33, using the primers indicated in Table 1. All reactions were performed with 0.4 μM of each primer in a final volume of 25 μl containing 2 mM MgCl2, 0.5 mM of each dNTP, and 0.5 units of Ampli Taq Gold polymerase (PE Biosystems, Foster City, CA). The DNA was amplified for 35 cycles with denaturation at 94°C for 20 s, annealing from 64 to 70°C for 30 s, and extension at 72°C for 45 s. PCR products were analyzed on 12% acrylamide gels containing 7.25% glycerol using a two-buffer system (Fig. 2A). PCR sample buffer (4 μl) was loaded, and DNA was visualized by silver staining. Sequence changes were detected by double-band shifts on the gel. Variants were purified using a Qiaquick PCR purification column (Qiagen, Chatsworth, CA) according to the supplier's recommendations and sequenced by Microsynth (Balgach, Switzerland) using the same primers as for PCR (Fig. 2C). The obtained sequences were compared with the Nedd4–2 gene sequence (GenBank no. AC011331). Identified variants were further analyzed by restriction digestion of PCR products according to standard methods (Fig. 2B).
Human Nedd4–2c cDNA (GenBank no. AY312514; see Ref. 34) was cloned into the NotI/EcoR V site of pSDeasySB (41). The human Nedd4–2a (KIAA0439) construct was described previously (25). Mutations were created by site-directed mutagenesis (Quickchange; Stratagene), and each mutant was verified by sequencing (Synergene Biotech, Schlieren, Switzerland). Human Sgk1 construct was prepared from an EST clone (BM460788), and a myc-tag (AEEQKLISEEDLLRKRREQLKHKLEQLRNSCA) was added at the NH2-terminus. The cDNA was cloned into XbaI/EcoR V sites of pSDeasySB. α-, β-, and γ-subunits of human ENaC were provided by Dr. Pascal Barbry (Valbonne, France). The PY-motifs were mutated using site-directed mutagenesis, by changing tyrosine to alanine in the PY-motifs of α, β, and γ. A minigene human Nedd4–2 construct, containing exons 14–16 from either control or affected subjects, was made by PCR on genomic DNA using primes SPL-FWD and SPL-RWD (Table 1). After TA cloning (TOPO cloning kit; Invitrogen, Basel, Switzerland) and sequencing to identify the wild-type and mutant cDNA, the fragments were cut and cloned into XhoI/EcoR V of pcDNA3.1(−).
In vitro splicing assay.
The Nedd4–2 minigene construct was transiently transfected into SW620 cells (human, Caucasian, colon, adenocarcinoma, ATCC CCL 227) using the FuGENE transfection reagent (Roche). After 48 h incubation, total RNA was isolated using the RNAeasy kit (Qiagen) and reverse transcribed using Superscript III RT (Invitrogen, Basel, Switzerland). Reversed transcription was followed by PCR using EX-SP-FWD/EX-SP-RWD primers. The amplified products were analyzed on agarose gels.
Expression in X. oocytes and electrophysiological measurements.
Plasmid encoding ENaC, Sgk1, and Nedd4–2 proteins was linearized and in vitro transcribed, the cRNA was injected in X. laevis oocytes, and electrophysiological measurements were carried out after overnight incubation, as described previously (42). The following quantities of cRNA were injected: 3 ng for each ENaC subunit; 10 ng Sgk1 and Nedd4–2 as indicated legends for Figs. 1–11.
Generation and purification of anti-Nedd4–2 phosphopeptide antibodies.
The anti-Nedd4–2 phosphopeptide antibody was raised in rabbits using a keyhole limpet hemocyanin-coupled, synthetic phosphopetide (KPRSL-SPhos-SPTV) corresponding to amino acids 324–332 of mouse Nedd4–2 (25) as immunogen. The antiserum was further affinity-purified against the immunogenic peptide coupled to Actigel ALD (Sterogene, Carlsbad, CA) according to the manufacture's instructions. The resin was then packed in a Poly-Prep Chromatography Column (Bio-Rad) and washed with 10 bed volumes of PBS. The remainder of the procedure was performed at 4°C. Serum was loaded on the column two times, the column was washed with 10 bed volumes of PBS, and the antibodies were eluted with 10 bed volumes of 0.1 M glycine, pH 2.5, directly on a tube containing 1 bed volume of 2 M Tris·HCl, pH 8.0. The eluate was dialyzed overnight in PBS and subsequently concentrated in the dialysis tube with polyethylene glycol 20.000 (Sigma) until the volume was reduced to the initial volume of serum loaded on the column. The eluate was recovered, and albumin was added to a final 1% concentration. Before use, the antibody was mixed 1:1 (vol/vol) to a peptide resin slur containing the nonphosphorylated immunogenic peptide and rotated for 90 min at 4°C to eliminate antibodies that recognize nonphosphorylated mNedd4–2. The supernatant is the purified anti-Nedd4–2 phosphopeptide antibody.
Western blot analyses of X. laevis lysates were carried out as described previously (23), using either anti-c-myc (Santa Cruz Biotechnology), anti-Nedd4–2 (25), or anti-Nedd4–2 phosphopeptide antibody. Lysis of oocytes was carried out by washing oocytes one time with modified Barth's solution (42), followed by extraction in 25 μl/oocyte of Triton X-100 homogenization buffer (20 mM Tris·HCl, pH 7.4, 100 mM NaCl, 1% Triton X-100, 1 mM phenylmethylsulfonyl fluoride, 10 μg/ml leupeptin, 10 μg/ml pepstatin A, and 10 μg/ml aprotinin), and centrifuged for 10 min at 20,000 g (4°C). In the phosphorylation experiment (Fig. 8), one-half of the lysates was treated with 1/10 vol 2 M Tris·HCl, pH 9.0, and 10 μl calf intestinal alkaline phosphatase (CIP; Roche) and incubated 3 h at 37°C. The enzyme was inactivated by adding 5× sample buffer and heating for 5 min at 95°C. Coimmunoprecipitation analysis was performed as described by Kamynina et al. (23).
Quantification of gels and statistical analysis.
Fluorograms of Western blots were scanned using an Epson 1680 Pro scanner and quantified using the B10–1D software (Vilbert Lourmat). Data are expressed as mean values ± SE and were analyzed by two-tailed Student's t-test for unpaired data.
Identification of naturally occurring human Nedd4–2 mutants.
To search for genetic variants in the human Nedd4–2 gene, we analyzed 856 subjects with and without hypertension by PCR/SSCP (Fig. 2). As outlined above, the Nedd4–2 gene is complex and consists of ∼34 different exons, spanning a region of >350 kb (illustrated in Fig. 1 and described in Refs. 5, 10, and 22). Therefore, we focused in the present study on exons 15, 16, 18, 19, 20, 21, and 22, which encode regions of Nedd4–2 essential for the regulation of ENaC, including WW domains 3 and 4 that interact with ENaC (23) and two phosphorylation sites shown previously to be phosphorylated by Sgk1 (7). Whereas exons 15 and 18–22 were analyzed in all 856 subjects, exons 16 and 33, which account either for the putative phosphorylation site Thr367 or the catalytic HECT domain, respectively, were screened in 200 patients with ESRD. We found five variants (Table 2), two of which were nonsynonymous, one synonymous, and two located in introns. One of the nonsynonymous mutations was located in exon 19 and changed Ser497 to Arg in WW domain 3 (Table 2); the subject carrying this mutation was a healthy control without hypertension. The other mutation, located in exon 15, changed proline-355 to leucine (Fig. 2 and Table 2). The two nonrelated patients carrying this mutation suffered from ESRD and were subjected to renal transplantation. The underlying renal diseases in these two patients were autosomal dominant polycystic kidney disease and Senior-Loken syndrome (juvenile nephronophthisis). We further characterized the mutation Nedd4–2-P355L observed in these two patients.
Human Nedd4–2-WT and human Nedd4–2-P355L RNA are spliced in a similar fashion.
The mutation P355L is localized very close to the 3′-donor site of exon 15 and may therefore potentially influence the splicing at this site (Fig. 3). We therefore investigated this possibility by generating a construct in pcDNA3.1 that contained the genomic sequence, including exons 14–16, in the wild-type or mutant form (Fig. 3A). These constructs were transfected in SW620 cells. After 2 days, cells were lysed, RNA was extracted, and RT-PCR was carried out using two primers (EX-SP-FWD and EX-SP-RWD) covering the cloned sequence (Fig. 3A). Whereas PCR of the plasmid yielded a fragment of 5108 bp (Fig. 3B, lanes 2 and 3), one of ∼302 bp (spliced transcript) was amplified by RT-PCR after transfection in the SW620 cells (lanes 4 and 5), indicating that both constructs were properly spliced.
Human Nedd4–2-P355L is less efficient than wild-type human Nedd4–2 in downregulating ENaC in X. laevis oocytes.
One of the best characterized functions of Nedd4–2 is the regulation of ENaC via WW-domain-PY-motif interaction when coexpressed in X. laevis oocytes. This prompted us to investigate if the P355L mutation affected the ability of Nedd4–2 to control ENaC activity. We expressed wild-type or mutant Nedd4–2 together with human ENaC in Xenopus oocytes. In most experiments, a novel form of Nedd4–2, Nedd4–2c, was used, which we recently cloned from a human kidney cDNA library (GenBank accession no. AY312514; see Ref. 34). Nedd4–2c comprises a complete C2-domain resulting from alternative splicing at the 5′-end and exon 16, which is spliced out in some of the previously described forms of human Nedd4–2 (e.g., KIAA0439; see Ref. 25). This exon encodes a third putative phosphorylation site for Sgk1 (T367), described previously by Snyder and collaborators (47). We injected increasing amounts of Nedd4–2-WT or Nedd4–2-P355L (Nedd4–2m) cRNA together with ENaC cRNA in X. laevis oocytes and monitored amiloride-sensitive Na+ currents by the two-electrode voltage-clamp method, a measure of ENaC activity (Fig. 4C). In agreement with our previous reports (25), increasing amounts of Nedd4–2-WT lead to diminished ENaC activity, with >95% inhibition at maximal Nedd4–2 cRNA levels. In contrast, Nedd4–2-P355L consistently inhibited ENaC to a lesser extent. Similar expression of Nedd4–2-WT and Nedd4–2-P355L proteins was verified by Western blotting the lysates of the same oocytes using an anti Nedd4–2 antibody (Fig. 4A and Ref. 25) and quantification of the detected Nedd4–2 protein on the blot (Fig. 4B). To further support the notion that Nedd4–2-P355L inhibited ENaC less efficiently, we looked at the mean absolute currents of 86 different experiments that we carried out under various conditions [rat or human (h) ENaC, hNedd4–2a or hNedd4–2c, different concentrations of injected Nedd4–2 cRNA]. As can be seen in Fig. 4, D and E, this large number of experiments confirmed that the Nedd4–2 mutant had a smaller impact on ENaC than the wild-type protein. When various Nedd4–2-WT-to-Nedd4–2-P355L ratios were injected with ENaC, we observed that increasing proportions of Nedd4–2-P355L yielded higher ENaC currents, corroborating the idea that this mutant is a weaker inhibitor of ENaC compared with wild-type Nedd4–2 (Fig. 5).
It is well established that Nedd4–2 exerts its effect on ENaC via WW domain-PY-motif interaction (19, 23, 46, 50). To ensure that the difference between Nedd4–2-WT and Nedd4–2-P355L is also depending on this interaction, we expressed ENaC channels with all PY-motifs disrupted by mutating the tyrosine residues to alanine. As expected and reported previously (1), such channels displayed much higher activities compared with wild-type channels (Fig. 6, compare ENaC with ENaC-ΔPY), likely because of the loss of all binding sites for endogenous Nedd4–2. When Nedd4–2-WT or Nedd4–2-P355L (Nedd4–2m) was coexpressed with wild-type ENaC, the activity was significantly decreased, with Nedd4–2-WT being more efficient in inhibiting ENaC activity than Nedd4–2-P355L, as observed before. However, no effect either by Nedd4–2 or by Nedd4–2-P355L was observed with ENaC-ΔPY, suggesting that both forms of Nedd4–2 regulate ENaC via WW-domain-PY-motif interaction (Fig. 6).
Both Nedd4–2-WT and Nedd4–2-P355L interact with ENaC in X. laevis oocytes.
The reduced effect of Nedd4–2-P355L on ENaC activity could be the result of either 1) reduced enzymatic Nedd4–2 activity or 2) reduced interaction with ENaC. To date, no quantitative enzymatic assay is available for Nedd4/Nedd4-like ubiquitin-protein ligases. We therefore investigated if the interaction between ENaC and Nedd4–2 is changed in this mutant. We carried out coimmunoprecipitation experiments between ENaC and either Nedd4–2-WT or Nedd4–2-P355L in X. laevis oocytes, as described previously (7, 23). We coexpressed Flag-tagged rat ENaC subunits with human Nedd4–2 (wild type or mutant), metabolically labeled the oocytes overnight with [35S]methionine, solubilized a membrane fraction, and precipitated the ENaC subunits with anti-Flag antibodies. The precipitated proteins were visualized by SDS-PAGE autoradiography. As shown in Fig. 7, the ENaC subunits were precipitated each time ENaC was expressed (Fig. 7). When either Nedd4–2-WT or Nedd4–2-P355L was added, a supplementary band at ∼120 kDa was coimmunoprecipitated, corresponding to the predicted molecular weight of Nedd4–2 (Fig. 7), whereas no Nedd4–2 band was observed when ENaC was omitted. We observed no obvious difference in the amount of coimmunoprecipitated hNedd4–2-WT or hNedd4–2-P355L, suggesting that both forms of Nedd4–2 bind ENaC with comparable apparent affinity.
The level of Ser448 phosphorylation is different between Nedd4–2-WT and Nedd4–2-P355L.
We and others have shown that the aldosterone-induced Sgk1 kinase increases ENaC activity by phosphorylating Nedd4–2 on two sites, primarily on Ser444 and less efficiently on Ser338 in X. laevis Nedd4–2 (corresponding to Ser448 and Ser342 in human Nedd4–2; see Refs. 7 and 47), causing reduced interaction between ENaC and Nedd4–2. In addition, a third putative Sgk1-dependent phosphorylation site (Thr367) has been described by Snyder et al. (47). We therefore wished to know if the level of phosphorylation induced by Sgk1 varied between wild-type and mutant Nedd4–2. Because we had previously shown that Sgk1 phosphorylates primarily Nedd4–2 on Ser448, we focused on this residue and raised a phosphopeptide-specific antibody against this site. Figure 8A shows that this antibody is specific for phosphorylated Nedd4–2, since treatment with CIP completely removed the signal for Nedd4–2, whereas another anti Nedd4–2 antibody (25) still recognized the protein (Fig. 8A, compare lanes 2–5 with 8–11). The Nedd4–2 bands were quantified in three different experiments, and the ratio between pSer448-Nedd4–2 and total Nedd4–2 was calculated (Fig. 8B). Thereby, we normalized against the CIP-treated total Nedd4–2 pool (lanes 1–6, second panel), since our Nedd4–2 antibody appears to recognize the phosphorylated Nedd4–2 (in the presence of Sgk1) less well (Fig. 8A, second panel, compare Nedd4–2 in presence or absence of Sgk1 in −CIP; lanes 8–11). Remarkably, when Sgk1 was absent, Nedd4–2-P355L (Nedd4–2m) was stronger phosphorylated than Nedd4–2-WT (compare lane 8 with 9 in Fig. 8A and quantification in Fig. 8B), which likely is explained by an endogenous Sgk-like activity in X. laevis oocytes. When we coexpressed Sgk1, the phosph-Ser448 signal was increased for Nedd4–2-WT by ∼4.5-fold, whereas the signal for Nedd4–2-P355L increased only ∼2-fold (Fig. 8A and quantification in 8B), yielding the same phosphorylation levels as Nedd4–2-WT. The observed difference in basal phosphorylation may explain why the Nedd4–2 mutant is less efficient than Nedd4–2 wild type in inhibiting ENaC activity, since such phosphorylation is expected to interfere with the Nedd4–2-ENaC interaction. In parallel, we studied the functional effect of Sgk1 on ENaC activity in oocytes expressing either wild-type or mutant Nedd4–2 together with ENaC by measuring amiloride-sensitive Na+ currents (Fig. 9). As can be seen in Fig. 9A, Sgk1 increased, whereas Nedd4–2-WT and (to a lesser extent) Nedd4–2-P355L (Nedd4–2m) decreased, ENaC activity. When Sgk1 was coexpressed with Nedd4–2 or Nedd4–2-P355L, the inhibition was reversed, and stimulation of ENaC was observed, as reported previously (7). However, compatible with the phosphorylation experiments in Fig. 8, Sgk1 increased the currents in the presence of Nedd4–2-WT by nearly 20-fold (Fig. 9, A and B), whereas it affected the current with Nedd4–2-P355L less strongly (7.5-fold), yielding currents that were not distinguishable between wild-type and mutant Nedd4–2 (Fig. 9). If differential level of Ser448 phosphorylation is the explanation for the difference in ENaC inhibition between Nedd4–2 and Nedd4–2-P355L, one may expect that mutation of Ser448 will abolish such dissimilarity. Consequently, we mutated this site to alanine both in wild-type and mutant Nedd4–2 and monitored ENaC activity. We verified that both proteins were expressed at the same level (Fig. 10, A and B). As predicted, the difference between Nedd4–2 and Nedd4–2-P355L was smaller in the S448A background (Fig. 10C). Surprisingly, we now found an inverse relationship, i.e., Nedd4–2-P355L-S448A was slightly more potent in ENaC inhibition than Nedd4–2-S448A (Fig. 10C). Consistent with the important role of S448 in Sgk1 phosphorylation, coexpression of Sgk1 displayed a weaker increase of the ENaC currents compared with the condition in which S448 was not mutated (Fig. 10D; compare columns N4–2/N4–2m ± Sgk1 with N4–2 S448A/N4–2m S448A ± Sgk1).
P355L affects other putative Sgk1 phosphorylation sites as well.
The observation that mutation of S448 did not completely abolish but rather inversed the difference between wild type and Nedd4–2-P355L suggested that this mutation also affects other regions in Nedd4–2. Because P355 lies between the two other putative Sgk1 phosphorylation sites, S342 and T367, we tested if the mutation also had an impact on these sites and inactivated them together with S448A either in the wild-type or in the P355L background (Nedd4–2ΔP or Nedd4–2mΔP). As before, the effect of these mutations on ENaC currents was measured in oocytes. When Sgk1 was coexpressed with such Nedd4–2 mutants, no stimulation of ENaC activity was observed, consistent with the idea that Sgk1 stimulates ENaC via phosphorylation of Nedd4–2 on the three putative phosphorylation sites (S342, T367, and S448; Fig. 11, compare columns N42 ± Sgk1 with N42ΔP ± Sgk1). Moreover, in this background, the mutation P355L did not affect the ability of Nedd4–2 to regulate ENaC currents (compare columns N4–2ΔP with N4–2mΔP), suggesting that this mutation indeed also affects the other putative Sgk1 phosphorylation sites.
Our data reveal the existence of a naturally occurring human Nedd4–2 variant (P355L) that affects the property to regulate the epithelial Na+ channel ENaC when expressed in X. laevis oocytes. This mutant, found in 2 out of 856 subjects, inhibited ENaC less efficiently compared with wild-type Nedd4–2. This difference can be explained by the fact that, in Nedd4–2-P355L, Ser448 is stronger phosphorylated by an endogenous kinase and therefore expected to display weaker interaction with ENaC. In addition, this mutation also has an impact on the other Sgk1-dependant phosphorylation sites, S342 and T367.
One of the major difficulties when one analyzes the Nedd4–2 gene is its complexity. As shown in Fig. 1 and reported previously (5, 10, 22), the gene contains ∼34 exons spanning over a region of ∼350 kb. The gene encodes a large number of alternative transcripts, varying not only at the 5′-end but also generating proteins that either do or do not include the C2 domains (5, 10, 22). Other splice variants vary within the protein and either lack WW domains or a putative phosphorylation site for Sgk1 (12, 25, 47). We therefore limited our search for mutations in the regions with potential relevance for the regulation of ENaC, i.e., the WW domains 3 and 4, shown to interact with ENaC (23), or the regions close to or including the Sgk1-dependent phosphorylation sites (7). These regions are encoded by exons 15, 18, 19, 20, 21, and 22. Genomic DNA of some patients was also screened for the catalytic HECT domain (exon 33) or in exon 16 of Nedd4–2. Overall, we have identified five variants, two of which were nonsynonymous changes, one synonymous, and two mutations located in introns (see Table 2). This number of variants seems to be low compared with the study by Dunn et al. (10) in which 38 variants were found in 48 Caucasians. However, their mutations were mostly located either in 5′-noncoding regions or within introns.
Here, we have focused on the properties of Nedd4–2-P355L, one of the two nonsynonymous mutants, and analyzed it with respect to ENaC regulation in X. laevis oocytes. The oocyte system has proven to be a powerful and reliable tool for the study of ion channels and plasma membrane proteins in general (for a review, see Ref. 45). Our analysis showed that the Nedd4–2 mutant was less efficient in downregulating ENaC activity, which we tend to explain by stronger phosphorylation of Ser448 in the mutant. We do not know the nature of the endogenous kinase that is involved in this phosphorylation, but it may involve Sgk1, or one of its isoforms, Sgk2 and -3, since they all have similar consensus sequences for phosphorylation (27, 28, 40) and are activated by the phosphatidylinositol 3-kinase pathway, known to sustain ENaC activity in X. laevis oocytes (53). Moreover, both Sgk2 and -3 have been shown to increase ENaC activity in X. laevis oocytes (13). Notably, the two other putative Sgk1-dependent phosphorylation sites (S342 and T367), which are situated in close proximity to P355, appear to be affected as well by the P355L mutation, although to a lesser extent. The observation that Nedd4–2-P355L-S448A is more efficient in ENaC inhibition than Nedd4–2-S448A, and the fact that additional mutation of S342 and T367 eliminates this difference, suggests that one or both of these sites are less phosphorylated in the presence of P355L. This remains to be investigated further.
In our previous studies, we found that Sgk1-dependent phosphorylation of S448 impairs the interaction of ENaC with Nedd4–2 (7). Although Nedd4–2-P355L is stronger phosphorylated on Ser448 than wild-type Nedd4–2, we are not able to detect any differences between mutant and wild-type Nedd4–2 in our coimmunoprecipitation experiments (Fig. 7). This might be explained by the small change in impact on ENaC activity and phosphorylation state of S448 that we observed between wild-type and P355L mutant. Alternatively, we cannot exclude that the enzymatic activity of Nedd4–2 is also affected by phosphorylation.
It may come as a surprise that the mutation of Pro355 to leucine does affect the phosphorylation level of Ser448 to such an extent. However, mutations involving a proline may significantly change the overall conformation of the enzyme. Moreover, we do not know how Sgk1 is interacting with Nedd4–2. This may be via PY-motif-WW domain interaction, as proposed previously (7, 47), or by other mechanisms (21) and may be affected by this mutation.
What is the biological relevance of such a Nedd4–2 mutation? The two patients with the Nedd4–2-P355L mutation could not be investigated with respect to abnormal renal sodium and potassium handling and blood pressure-regulating hormones. Both patients had been on dialysis and had been transplanted before the mutations were found. The underlying clinical entities in the native kidneys, autosomal polycystic kidney disease and juvenile nephronophthisis (Senior-Loken syndrome), affect by an inherited mechanism the tubular interstitial area of the kidney. These are two complex renal diseases without apparent link to dysfunction of ENaC. Moreover, it has to be noted that this is a relatively rare mutation that, when expressed in X. laevis oocytes, displays subtle effects on Nedd4–2 function (i.e., weaker downregulation of ENaC resulting from elevated phosphorylation level of Ser448). Admittedly, these differences may be inherent to the oocyte system and will have to be demonstrated in an epithelium. Therefore, it is too early to speculate about a pathophysiological role for this mutation, and the potential role for the induction of or predisposition to hypertension remains to be established.
In summary, we have shown that the naturally occurring P355L mutant of human Nedd4–2 inhibits ENaC less efficiently in X. laevis oocytes. This appears to be caused by enhanced phosphorylation on Ser448 by an endogenous yet undefined kinase expected to interfere with ENaC-Nedd4–2 interaction. The function of the other putative Sgk1-dependent phosphorylation sites appears also to be affected. To our knowledge, these data represent the first report on functional aspects of a human Nedd4–2 variant and help to elucidate the structure-function relationship of the concerted action of Nedd4–2, ENaC, and Sgk1.
This work was supported by Swiss National Science Foundation grants to O. Staub (31–64052.00), F. J. Frey and B. M. Frey (31–61505.00), and J. Loffing (32–061742.00), Leenaards Foundation (O. Staub), the Roche Research Foundation (O. Staub), and the EMDO foundation (J. Loffing).
We thank Dr. Pascal Barbry for providing human ENaC cDNA. We are grateful to Drs. Laurent Schild, Dmitri Firsov, Hugues Abriel, Phil Shaw, Jean-Daniel Horisberger, and Laura Stanasila for critically reading the manuscript.
↵* B. M. Frey and O. Staub contributed equally to this work.
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.
- Copyright © 2004 the American Physiological Society