Identification of Murr1 as a Regulator of the Human δ Epithelial Sodium Channel*

The human δ epithelial sodium channel (δENaC) subunit is related to the α-, β-, and γENaC subunits that control salt homeostasis. δENaC forms an amiloride-sensitive Na+ channel with the β and γ subunits. However, the in vivo function of δENaC is not known. To gain insight into the function of δENaC, a yeast two-hybrid screen of a human brain cDNA library was carried out using the C- and N-terminal domains of δENaC. A novel δENaC-interacting protein called Murr1 (mouse U2af1-rs1 region) was isolated in the C-terminal domain screen. Murr1 is a 21-kDa protein mutated in Bedlington terriers suffering from copper toxicosis. The interaction of Murr1 and δENaC was confirmed by glutathione S-transferase pulldown assay and coimmunoprecipitation. To test the functional significance of the interaction, Murr1 was coexpressed with δβγENaC in Xenopus oocytes. Murr1 inhibited amiloride-sensitive sodium current in a dose-dependent manner. In addition, deletion of the last 59 amino acids of δENaC abolished the inhibition. Murr1 also bound to the β- and γENaC subunits and inhibited αβγENaC sodium current. Therefore, these results suggest that Murr1 is a novel regulator of ENaC.

The human ␦ epithelial sodium channel (␦ENaC) subunit is related to the ␣-, ␤-, and ␥ENaC subunits that control salt homeostasis. ␦ENaC forms an amiloride-sensitive Na ؉ channel with the ␤ and ␥ subunits. However, the in vivo function of ␦ENaC is not known. To gain insight into the function of ␦ENaC, a yeast two-hybrid screen of a human brain cDNA library was carried out using the C-and N-terminal domains of ␦ENaC. A novel ␦ENaC-interacting protein called Murr1 (mouse U2af1-rs1 region) was isolated in the C-terminal domain screen. Murr1 is a 21-kDa protein mutated in Bedlington terriers suffering from copper toxicosis. The interaction of Murr1 and ␦ENaC was confirmed by glutathione Stransferase pulldown assay and coimmunoprecipitation. To test the functional significance of the interaction, Murr1 was coexpressed with ␦␤␥ENaC in Xenopus oocytes. Murr1 inhibited amiloride-sensitive sodium current in a dose-dependent manner. In addition, deletion of the last 59 amino acids of ␦ENaC abolished the inhibition. Murr1 also bound to the ␤and ␥ENaC subunits and inhibited ␣␤␥ENaC sodium current. Therefore, these results suggest that Murr1 is a novel regulator of ENaC.
An additional human sodium channel subunit, ␦ENaC, was reported in 1995 (9). Among ENaC family members ␦ENaC has the highest amino acid identity (ϳ37%) to ␣ENaC and to a recently described ⑀-subunit from Xenopus laevis (10). A ␦ENaC gene appears to be present in chimpanzee (GenBank TM accession number O46547) and in rabbit (11), but there is no evidence for a rat or mouse ␦ENaC gene. Similar to ␣ENaC, when ␦ENaC is expressed alone in Xenopus oocytes a small, amiloride-sensitive Na ϩ current is induced (9). This current is increased 50-fold by coexpression with the ␤and ␥ENaC subunits, and the properties of the ␦ and ␦␤␥ channels were iden-tical (9). The highest expression levels of ␦ENaC mRNA were detected in brain, testis, ovary, and pancreas, indicating that the primary function of ␦ENaC may not be in epithelia (9).
ENaC subunits are members of the degenerin/ENaC gene family. Other family members such as brain Na ϩ channel 1, acid-sensing ion channel, and dorsal root acid-sensing ion channel are expressed in neurons of the central and peripheral nervous systems (12). These channels are stimulated by acidic pH and have been implicated in touch sensation, synaptic plasticity, and pain perception (12). A brain channel matching the properties of the ␦ENaC channel has not been reported.
␦ENaC shares a common predicted topology to the ␣-, ␤-, and ␥ENaC subunits: a large extracellular loop separated by two membrane-spanning domains leaving the short N-and C-terminal domains located inside the cell (13). The C-terminal domains of the ␣␤␥ENaC subunits provide binding sites for the Nedd4 family of ubiquitin ligases. Nedd4 decreases the surface expression of ENaC, probably by mediating ubiquitination and internalization of the channel, thus controlling sodium movement across epithelia (14,15). Nedd4 binding to the ␣␤␥ENaC subunits is mediated by WW domains in Nedd4 and a conserved PY motif (PPPXY) in the C-terminal domain of the ␣␤␥ENaC subunits. However, this motif is not conserved in ␦ENaC (9), suggesting that the ␦ENaC subunit may be regulated by binding proteins other than Nedd4 family members. Analysis of the amino acid sequence of the C-and N-terminal domains of ␦ENaC shows that the N-terminal domain is particularly proline-rich (Fig. 1A, bottom). Short proline-rich sequences are known to bind interaction domains such as the WW or SH3 domains (16). The C-terminal domain also contains prolines that might provide a binding site for these domains. Alternatively, ␦ENaC may contain novel binding motifs, and identification of proteins binding to novel motifs might provide information on ␦ENaC function. Therefore, we screened a human brain cDNA library with the N-and C-terminal domains of ␦ENaC to identify interacting proteins. Here we report that Murr1, a protein implicated in copper transport (17), interacts with the C-terminal domain of ␦ENaC. In addition, we found that Murr1 binds to ␤and ␥ENaC but not to ␣ENaC. Coexpression of Murr1 with either the ␦␤␥ or ␣␤␥ENaC subunits in Xenopus oocytes resulted in sodium current inhibition, and the inhibition of ␦␤␥ENaC was abolished by C-terminal truncation of ␦ENaC.

EXPERIMENTAL PROCEDURES
DNA Constructs-Full-length and truncated ␦ENaC constructs (all containing the FLAG epitope tag DYKDDDDK) were cloned into pMT3 after PCR using the primers described below (all primers from 5Ј to 3Ј). For full-length ␦ENaC: 5Ј primer CCATCGATATGGCTGAGCACCG-AAGC and 3Ј primer GGAATTCTCACTTGTCATCGTCGTCCTTGTA-GTCGGTGTCCAGAGTCTCAAG. The C-terminal truncations were all constructed using the same 5Ј primer, ACGCGTCGACGCCACCATGG-ACTACAAGGACGACGATGACAAGGCTGAGCACCGAAGCATG, and the following 3Ј primers, CGGAATTCTCATGGAAGCATCACCCGTGG for ␦C-23 (last 23 amino acids were deleted), CGGAATTCTCAATCTG-ACTGGCCTCTGG for ␦C-47, and CGGAATTCTCATGAGGCAGGGCT-GGCTCT for ␦C-59 (␦ T ). The N-terminal truncations were constructed using the same 3Ј primer as for full-length ␦ENaC and the 5Ј primers ACGCGTCGACGCCACCATGGAGCCCCCCAGGCCGGGGCCA for ␦N-27 (first 27 amino acids deleted) and ACGCGTCGACGCCACCAT-GGAGAAGGAGGGGCACCAGGAG for ␦N-41. Full-length ␦ENaC was also cloned into the vector pHM6 (Roche Applied Science) in-frame with the HA epitope tag using the 5Ј primer GGAATTCGATGGCTGAGCA-CCGAAGC and the 3Ј primer GGAATTCCGGTGTCCAGAGTCTCAAG.
Because sequence analysis showed that the isolated Murr1 clones were missing the first 15 base pairs, these missing base pairs were added as part of a PCR primer in a subsequent PCR reaction. Murr1-FLAG was cloned by PCR from a yeast two-hybrid library plasmid using the 5Ј primer CCATCGATGCCACCATGGAGGGCGAGCTTGAGGGT-GGCAAACCCCTG and the 3Ј primer CGGAATTCTCACTTGTCATCG-TCGTCCTTGTAGTCGTTAGGCTGGCTGATCAG and ligated into the vector pMT3. For the production of Murr1⅐GST fusion protein, Murr1 was cloned in-frame with GST into the vector pGEX-KG (AP Biotech) using the 5Ј primer CGGGATCCCACGAGGGCGAGCTTGAGGGTGG-CAAACCCCTG and the 3Ј primer GCTCTAGATCAGTTAGGCTGGCT-GATC.
Cell Culture and Transient Transfection-COS7 cells were obtained from the American Type Culture Collection and were grown in low bicarbonate Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum, 10 units/ml penicillin, and 10 g/ml streptomycin. Cells were maintained at 37°C and 5% CO 2 . The day before transfection, COS7 cells were plated at a density of 3 ϫ 10 5 cells in 35-mm plates. Cells were transfected with 1.5 g of each cDNA construct using FuGENE 6 (Roche Applied Science) and the manufacturer's protocol.
Yeast Two-hybrid Screen-All materials used for the screen were derived from the MATCHMAKER GAL4 two-hybrid system 3 (Clontech). A pretransformed human brain cDNA library (Clontech) was screened with the C-and N-terminal cytosolic regions of ␦ENaC. The baits were constructed by cloning the ␦ENaC C-terminal (amino acids 562-638) and N-terminal (amino acids 1-88) cytosolic domains inframe with the GAL4 binding domain into the EcoRI-SalI sites of pGBKT7. Correct insertion of the cytosolic domains was verified by DNA sequencing (Centre for Gene Research, University of Otago). The inability of the baits to autonomously activate the reporter genes in yeast strain AH109 was confirmed prior to the library screen. Also, untransformed and mated AH109-and Y187-strain cells did not activate the reporter genes. The cDNA library and the bait were grown overnight, and 5 ϫ 10 8 cells of each strain were mixed and plated on YPDA (yeast peptone dextrose adenine media; 20 g/liter peptone, 10 g/liter yeast extract, 2% dextrose, 0.003% adenine) medium to allow mating for 6 h. The cells were then resuspended in sterile water and plated onto SC (synthetic complete medium)-His-Leu-Trp. Colonies were picked 3-6 days after plating and regrown on SC-His-Leu-Trp. Regrown colonies were replica-plated on SC-Ade-Leu-Trp to check for expression of a second reporter gene, on S.D. (synthetic-deficient medium) to check for wild-type yeast, and on SC-His-Leu-Trp to show regrowth after replica plating. Colonies growing on both selection media, but not on S.D., were further cultivated and analyzed for ␤-galactosidase activity by performing filter assays. For further analysis, cDNA inserts were isolated from histidine and adenine prototrophic and ␤-galactosidase-expressing clones by PCR using the 5Ј primer CTATTCGATGATGAAGATACCCCACCAAACCC and the 3Ј primer GTGAACTTGCGGGGTTTTTCAGTATCTACGAT. Plasmids were restored by gap repair cloning (19) into the vector pGADT7, transformation into AH109, and selection on SC-Leu plates. False positives were excluded by performing autoactivation screens (on SC-Leu-Ade and SC-Leu-His) and retesting the interaction with the original (␦N/Cterminal domain) and an unrelated bait (lamininC). All remaining clones were again amplified by PCR and digested with HaeIII to determine the cDNA size and to exclude identical clones. Selected clones were further analyzed by DNA sequencing.
GST Pulldown Assay-pGEX-KG containing the cDNA for Murr1 or the empty vector were transformed into Escherichia coli strain BL21, and expression of the Murr1⅐GST fusion protein or GST alone was induced with 0.5 mM isopropyl-␤-D-1-thiogalactopyranoside for 3 h at 30°C. Bacteria were lysed using Bugbuster® (Novagen), and the GST fusion proteins were purified on glutathione-Sepharose beads (AP Biotech).
COS7 cells were transfected with different cDNA constructs. After 24 h cells were lysed in T-TBS (150 mM NaCl, 50 mM Tris, pH 7.4, 10 g/ml phenylmethylsulfonyl fluoride, 2 g/ml aprotinin ϩ 1% Triton X-100). Insoluble material was pelleted by centrifugation (5 min, 16,000 ϫ g), and the lysates were precleared with GST bound to glutathione-Sepharose beads. The precleared lysates were then incubated with Murr1⅐GST fusion protein or GST alone, attached to glutathione-Sepharose beads for 3 h at 4°C, washed extensively in T-TBS, and analyzed by Western blotting using anti-FLAG or anti-HA antibody (Sigma).
Coimmunoprecipitation-COS7 cells were cotransfected with ␦ENaC-HA and Murr1-FLAG in 35-mm dishes, and three plates were pooled for one experiment. Cells were lysed with T-TBS 24 h after transfection, insoluble material was removed by centrifugation (5 min, 16,000 ϫ g), and the supernatant was incubated with 25 g/ml anti-FLAG antibody for 3 h at 4°C. Then 25 l of protein G-Sepharose beads was added and incubation continued for 1 h at 4°C. The beads were washed four times in lysis buffer and resuspended in 2ϫ SDS sample buffer. The samples were analyzed by Western blotting using anti-HA antibody (Sigma).
Expression in Xenopus Oocytes and Immunoprecipitation-␦␤␥ENaC, ␦ T FLAG␤␥ENaC, ␦␤ T ␥ T ENaC, ␦ T FLAG␤ T ␥ T ENaC, or ␣␤␥ENaC (237 pg of each) together with secreted alkaline phosphatase (SEAP, 237 pg) as a negative control or Murr1-FLAG (35 or 70 pg) were injected into the nuclei of manually defolliculated stage V-VI oocytes. Injected oocytes were incubated in low Na ϩ MBS (5 mM NaCl, 83 mM NMDG, 1 mM KCl, 2.4 mM NaHCO 3 , 15 mM HEPES, 0.33 mM Ca(NO 3 ) 2 , 0.41 mM CaCl 2 , 0.82 mM MgSO 4 , 50 mg/l gentamycin, pH 7.6) at 18°C and studied 2 or 3 days after injection. Amiloride-sensitive Na ϩ current was measured by the two-electrode voltage-clamp technique. Oocytes were bathed in Ringer's solution (116 mM NaCl, 2 mM KCl, 0.4 mM CaCl 2 , 1 mM MgCl 2 , 5 mM HEPES, pH 7.4), and inward currents were recorded at a holding potential of Ϫ60 mV using the oocyte clamp 725C (Warner Instruments) and pCLAMP 8.2 software (Axon instruments). Amiloride-sensitive current was determined by subtracting the current obtained at Ϫ60 mV in the presence of 100 M amiloride from the current obtained without amiloride. Data were sampled at 1 kHz. Data were collected and pooled from at least two preparations of oocytes isolated on different days from different animals and are presented as mean Ϯ S.E. Statistical significance was calculated using analysis of variance and MS Excel software.
Expressing eggs were collected and lysed with T-TBS by rotating for 1 h at 4°C. The lysates were centrifuged for 20 min at 20,000 ϫ g, and the supernatants were incubated with 25 g/ml anti-FLAG antibody overnight to precipitate Murr1-FLAG or ␦ T FLAG. Immunocomplexes were isolated with protein G-Sepharose (1 h, 4°C) and analyzed by Western blotting using anti-FLAG antibody.

RESULTS
Yeast Two-hybrid Screen-To identify proteins binding to ␦ENaC, we performed a yeast two-hybrid screen with the intracellular C-and N-terminal domains of ␦ENaC. Because of the high expression of ␦ENaC in brain (9), we chose to screen a human brain cDNA library. The ␦ENaC C-terminal domain screen yielded 40 putative positive clones showing histidine and adenine prototrophy and ␤-galactosidase activity; however, the N-terminal domain screen did not yield any positive clones. The ␦ENaC C-terminal domain-interacting clones that activated the reporter genes with the original bait (␦ENaC C-terminal domain) were not capable of inducing expression of any reporter gene by themselves and did not interact with an unrelated bait (lamininC) (data not shown). Interacting clones were sequenced and subjected to a BLAST search. Three of the clones were identified as the human Murr1 gene. The Murr1 gene was first described by Nabetani et al. (20), who discovered that the imprinted gene U2af1-rs1 lies in an intron of the Murr1 gene, although the Murr1 gene itself does not appear to be imprinted. Thus, Murr1 was named as a gene that locates in the mouse U2af1-rs1 region. Murr1 has recently been implicated in copper metabolism because a mutated form of Murr1 is associated with copper toxicosis in Bedlington terriers (17). However, the function of Murr1 in copper metabolism is not known.
Murr1 appears to be widely expressed in human tissues, including tissues that express ␦ENaC (17). Therefore we investigated the possibility of an interaction between ␦ENaC and Murr1. The Murr1 gene codes for a 21-kDa protein that does not contain any known interaction domains, suggesting the interaction with ␦ENaC might involve novel binding motifs. Murr1 Interacts Specifically with the C-terminal Domain of ␦ENaC-Expression of ␦ENaC and its truncations in COS7 cells resulted in the appearance of two or more specific bands, most likely because of glycosylation of the extracellular domain. This observation is consistent with the multiple bands, observed upon expression of wild type and epitope-tagged ␣ENaC, ␤ENaC, and ␥ENaC subunits (1, 18); some of these bands represent glycosylated subunits. The interaction between Murr1 and ␦ENaC was confirmed by GST pulldown assay (Fig. 2). Initially, ␦ENaC-FLAG expressed in COS7 cells was bound with a Murr1⅐GST fusion protein. Fig. 2 shows that Murr1⅐GST bound ␦ENaC, but GST alone did not. To confirm that Murr1 binds to the C-and not to the N-terminal domain of ␦ENaC as predicted by the yeast two-hybrid screen, we used a series of N-and C-terminal ␦ENaC truncations (Fig. 1A) in a subsequent GST pulldown assay. Truncations from the N terminus had no effect on the interaction between Murr1 and ␦ENaC (Fig. 2). A series of ␦ENaC C-terminal truncations showed that removal of the last 23 amino acids from ␦ENaC had no effect on binding to Murr1 (Fig. 2). However, deletion of either the last 47 or the last 59 amino acids abolished the interaction between Murr1 and ␦ENaC. These results show that Murr1 specifically binds to the C-terminal domain of ␦ENaC and that the probable binding site for Murr1 is located between amino acids 592 and 615 of ␦ENaC (Fig. 1A).
To investigate an interaction between Murr1 and ␦ENaC in intact cells we used coimmunoprecipitation. COS7 cells were cotransfected with constructs encoding Murr1-FLAG and ␦ENaC-HA. Complexes between Murr1-FLAG and ␦ENaC-HA were purified with anti-FLAG antibody followed by protein G-Sepharose beads and analyzed by Western blotting using anti-HA antibody. Fig. 3 (lane 1) shows that Murr1 and ␦ENaC interact when coexpressed in mammalian cells.
Murr1 Inhibits Na ϩ Channel Current in Xenopus Oocytes-Because Murr1 and ␦ENaC interact when coexpressed in intact cells, we investigated whether the interaction between ␦ENaC and Murr1 is functionally significant. Initially the ␦-, ␤-, and ␥ENaC subunits were coexpressed with an equimolar amount of Murr1 in Xenopus oocytes. Murr1 significantly decreased the Na ϩ current by 88 Ϯ 3% (n ϭ 17) compared with the control (Fig. 4). To test whether the decrease in current was dose-dependent, oocytes were injected with ␦␤␥ENaC and half the amount of Murr1 used in the initial experiment. The Na ϩ current was decreased by 52 Ϯ 6%, (n ϭ 26) and differed significantly from both the control and currents recorded using an equimolar amount of Murr1. Thus, Murr1 inhibits ␦␤␥ENaC current in a dose-dependent manner.
The results of the in vitro interaction experiments suggested that Murr1 interacts specifically with the C-terminal domain of ␦ENaC. To test this hypothesis in Xenopus oocytes and to investigate whether the ␤ and ␥ subunits contribute to the inhibition of sodium current via Murr1, C-terminal truncation construct combinations were coexpressed with Murr1. First, full-length ␦ENaC was coexpressed with ␤ and ␥ subunits missing their C-terminal domains (␤ T and ␥ T ) in the presence or absence of Murr1 (Fig. 5A). When Murr1 was expressed with ␦␤ T ␥ T the Na ϩ current was reduced by 78 Ϯ 3%, (n ϭ 23) compared with the control (␦␤ T ␥ T ϩ SEAP), showing Murr1 could still inhibit the channel when the ␤and ␥ENaC Cterminal domains were removed. However, the inhibition was significantly less than that observed with ␦␤␥ENaC ϩ Murr1, suggesting that Murr1 may exert its inhibition in part via the ␤and/or ␥ENaC C-terminal domains. Next, we coexpressed a truncated ␦ENaC subunit missing the last 59 amino acids (␦ T , which did not bind to Murr1 in vitro, Fig. 2) with fulllength ␤and ␥ENaC. Coexpression of Murr1 with the ␦ T ␤␥ENaC channel did not result in significant downregulation of the sodium channel (96 Ϯ 12%, n ϭ 27 of ␦ T ␤␥ ϩ SEAP). This result suggests that Murr1 binds to the C-terminal domain of ␦ENaC to inhibit channel activity and that the contribution of the ␤ and ␥ subunits to Murr1 inhibition is minimal.
To confirm that both ␦ T ENaC and Murr1 were expressed in this experiment, sodium channel-expressing oocytes were collected and pooled after current measurement. Murr1 and ␦ T E-NaC were isolated by immunoprecipitation using anti-FLAG antibody. Fig. 5B shows that both proteins were expressed in the oocytes. Therefore, Murr1 binds to the C-terminal region of ␦ENaC to inhibit sodium current. Murr1 Interacts with ␤and ␥ENaC in Vitro-Although Murr1 still inhibited current when the C-terminal domains of the ␤and ␥ENaC subunits were removed, (Fig. 5, ␦␤ T ␥ T ENaC ϩ Murr1), the decrease was significantly less than that observed with Murr1 inhibition of ␦␤␥ENaC. Therefore, we investigated whether the ␣-, ␤or ␥ENaC subunits also interact with Murr1, using a GST pulldown assay. ␣-, ␤-, or ␥ENaC, either HA-or FLAG-tagged, were expressed in COS7 cells for 24 h, and the lysates were incubated with Murr1⅐GST or GST alone. Fig. 6A shows that ␤and ␥ENaC, but not ␣ENaC, bind to Murr1⅐GST in vitro.
Because the ␤and ␥ENaC subunits bind Murr1, we tested whether Murr1 could regulate a channel formed by the ␣-, ␤-, and ␥ENaC subunits. Fig. 6B shows that Murr1 inhibited ␣␤␥ENaC current by 53 Ϯ 7% (n ϭ 18). Therefore, Murr1 is able to regulate sodium channels formed by either ␦␤␥ or ␣␤␥ ENaC subunits. DISCUSSION In contrast to the relatively well-characterized ␣␤␥ENaC subunits, the ␦ENaC subunit has been poorly investigated. The most apparent difference between ␦ENaC and ␣␤␥ENaC is tissue distribution. Whereas the ␣, ␤, and ␥ subunits are predominantly expressed in epithelial tissues such as kidney, colon, or lung (1, 2), ␦ENaC expression is highest in brain, testis, ovary, and pancreas (9). Although ␦ENaC is expressed at low levels in kidney tissue (9), it is unlikely that ␦ENaC can compensate for ␣ENaC, because loss of function mutations in ␣ENaC cause the salt-wasting disorder, pseudohypoaldosteronism type 1 (3). This observation suggests that in the kidney, ␦ENaC is not expressed at a high enough level to compensate for ␣ENaC, that ␣ENaC and ␦ENaC are expressed in different cell types, or that channels formed by ␣ENaC and ␦ENaC are functionally unique and perhaps regulated by different cellular pathways.
To begin to investigate the function and regulation of ␦ENaC we performed a yeast two-hybrid screen with the N-and Cterminal domains of ␦ENaC to isolate ␦ENaC binding partners. Such binding partners might represent novel regulators or subunits of a ␦ENaC channel. Murr1, a gene implicated in copper metabolism (17), was identified as a novel ␦ENaC binding protein.
The Murr1 gene is mutated in inherited autosomal recessive copper toxicosis in Bedlington terriers. Affected dogs showed a homozygous deletion of exon 2 of the Murr1 gene (17). Copper toxicosis is characterized by inefficient excretion of copper into the bile, resulting in accumulation of copper in the liver (21). In affected dogs excess copper remains in the lysosomes (22), perhaps because of a defect in lysosomal vesicle trafficking to the bile canalicular membrane. These findings suggest that Murr1 might be involved in ion transport and/or in vesicle trafficking. In humans, the majority of copper toxicosis patients have Wilson's disease. The genetic defect for this disease has been localized to the ATP7B gene, which encodes a P-type ATPase (21). Recently Tao et al. (23) demonstrated that Murr1 interacts directly with the intracellular N-terminal domain of the Wilson disease gene ATP7B, indicating a link between Murr1 and ATP7B. However, investigation of 23 patients with non-Wilsonian hepatic copper toxicosis did not identify any mutations or polymorphisms in the Murr1 gene (24). The function of Murr1 remains unknown because the protein shows no homology to other proteins and does not contain any identifiable amino acid motifs or domains.
Using immunoprecipitation and pulldown data, we have demonstrated that Murr1 interacts with the C-terminal domain of ␦ENaC. Using N-and C-terminal truncations of ␦ENaC, the binding site for Murr1 was located between amino acids 592 and 615 of ␦ENaC. The C-terminal domain of ␦ENaC is only 25% identical to that of ␣ENaC (Fig. 1B), and the PY motif found in the ␣␤␥ENaC subunits that mediates interaction with Nedd4 family members is not conserved in ␦ENaC. Murr1 does not contain any recognizable protein interaction motifs; thus further work will be required to identify the exact amino acids involved in the interaction of ␦ENaC and Murr1.
Two-electrode voltage-clamp studies in Xenopus oocytes revealed that Murr1 is a potent inhibitor of ␦␤␥ENaC sodium current. This effect might be because Murr1 alters the trafficking of subunits to or from the plasma membrane, it might have a direct effect on channel gating, or it might be an adaptor linking the channel to another regulatory protein. ENaC subunit trafficking is known to be influenced by syntaxins (25,26), which decrease surface expression of ENaC presumably by interfering with its insertion into the membrane (27), and Nedd4 family members, which facilitate ubiquitination and internalization of ENaC (28).
When the C-terminal domains of ␤and ␥ENaC were deleted, Murr1 still inhibited the current produced by the ␦␤ T ␥ T ENaC channel, though this current was significantly larger than currents obtained with ␦␤␥ENaC ϩ Murr1. These results suggest that Murr1 exerts its inhibitory effect primarily via the ␦ENaC subunit. This hypothesis was confirmed by deleting the C-terminal domain of ␦ENaC; Murr1 was unable to inhibit ␦ T ␤␥ channel function. However, the ␤and ␥ENaC subunits might also be involved in Murr1-meditated ␦␤␥ENaC channel regulation because Murr1 bound both subunits in a pulldown assay.
Murr1 also reduced ␣␤␥ENaC sodium current. The effect of Murr1 on the ␣␤␥ENaC current was less than that observed for ␦␤␥ENaC, possibly because Murr1 does not bind to ␣ENaC. These results suggest that Murr1 may be a general regulator of ENaC family members and potentially of other channels or transporters.
A link between sodium and copper transport has been demonstrated in fish gill and intestinal epithelia, including evidence for copper leak through ENaC (29). Although neither Murr1 nor ␦ENaC appears to contain consensus copper binding sites (30,31), it is possible that Murr1 links the transport of sodium and copper across epithelia. Alternatively, Murr1 might be important for correct trafficking of both sodium channels and copper transporters. Both the ATP7A and ATP7B copper-transporting proteins are localized to the trans-Golgi network in low copper concentrations. Increasing the copper concentration causes a shift in the location of ATP7A and ATP7B to the plasma membrane or a vesicular population (1,32).
Further experiments such as cellular colocalization, effect on ␦ENaC trafficking, and single-channel patch clamp analysis will be necessary to elucidate the mechanism by which Murr1 inhibits both ␦␤␥ENaC and ␣␤␥ENaC channel function.