NFAT5 up-regulates expression of the kidney-specific ubiquitin ligase gene Rnf183 under hypertonic conditions in inner-medullary collecting duct cells

We previously reported that among the 37 RING finger protein (RNF) family members, RNF183 mRNA is specifically expressed in the kidney under normal conditions. However, the mechanism supporting its kidney-specific expression pattern remains unclear. In this study, we elucidated the mechanism of the transcriptional activation of murine Rnf183 in inner-medullary collecting duct cells. Experiments with anti-RNF183 antibody revealed that RNF183 is predominantly expressed in the renal medulla. Among the 37 RNF family members, Rnf183 mRNA expression was specifically increased in hypertonic conditions, a hallmark of the renal medulla. RNF183 up-regulation was consistent with the activation of nuclear factor of activated T cells 5 (NFAT5), a transcription factor essential for adaptation to hypertonic conditions. Accordingly, siRNA-mediated knockdown of NFAT5 down-regulated RNF183 expression. Furthermore, the −3,466 to −3,136-bp region upstream of the mouse Rnf183 promoter containing the NFAT5-binding motif is conserved among mammals. A luciferase-based reporter vector containing the NFAT5-binding site was activated in response to hypertonic stress, but was inhibited by a mutation at the NFAT5-binding site. ChIP assays revealed that the binding of NFAT5 to this DNA site is enhanced by hypertonic stress. Of note, siRNA-mediated RNF183 knockdown increased hypertonicity-induced caspase-3 activation and decreased viability of mIMCD-3 cells. These results indicate that (i) RNF183 is predominantly expressed in the normal renal medulla, (ii) NFAT5 stimulates transcriptional activation of Rnf183 by binding to its cognate binding motif in the Rnf183 promoter, and (iii) RNF183 protects renal medullary cells from hypertonicity-induced apoptosis.

We previously reported that among the 37 RING finger protein (RNF) family members, RNF183 mRNA is specifically expressed in the kidney under normal conditions. However, the mechanism supporting its kidney-specific expression pattern remains unclear. In this study, we elucidated the mechanism of the transcriptional activation of murine Rnf183 in inner-medullary collecting duct cells. Experiments with anti-RNF183 antibody revealed that RNF183 is predominantly expressed in the renal medulla. Among the 37 RNF family members, Rnf183 mRNA expression was specifically increased in hypertonic conditions, a hallmark of the renal medulla. RNF183 up-regulation was consistent with the activation of nuclear factor of activated T cells 5 (NFAT5), a transcription factor essential for adaptation to hypertonic conditions. Accordingly, siRNA-mediated knockdown of NFAT5 down-regulated RNF183 expression. Furthermore, the ؊3,466 to ؊3,136-bp region upstream of the mouse Rnf183 promoter containing the NFAT5-binding motif is conserved among mammals. A luciferase-based reporter vector containing the NFAT5-binding site was activated in response to hypertonic stress, but was inhibited by a mutation at the NFAT5-binding site. ChIP assays revealed that the binding of NFAT5 to this DNA site is enhanced by hypertonic stress. Of note, siRNA-mediated RNF183 knockdown increased hypertonicity-induced caspase-3 activation and decreased viability of mIMCD-3 cells. These results indicate that (i) RNF183 is predominantly expressed in the normal renal medulla, (ii) NFAT5 stimulates transcriptional activation of Rnf183 by binding to its cognate binding motif in the Rnf183 promoter, and (iii) RNF183 protects renal medullary cells from hypertonicity-induced apoptosis.
The renal medulla is the only tissue that is constantly under hypertonic conditions to concentrate urine through water reabsorption in mammals (1). The driving forces for water reabsorption are elevated concentrations of sodium chloride and urea, which create exceptionally hypertonic conditions (600 and 2,000 mosmol/kg H 2 O under conditions of diuresis and antidiuresis in rodents, respectively) (2). Renal cells, particularly renal medullary cells, partly adapt to the hypertonic conditions by accumulating intracellular organic osmolytes that reduce intracellular ionic strength. Nuclear factor of activated T cells 5 (NFAT5) 3 /tonicityresponsive enhancer binding protein plays a key role in the adaptation to hypertonic conditions by binding to the consensus DNAbinding element RNRNNTTTCCA (where N is any nucleotide, and R is any purine) (3,4). NFAT5 stimulates the transcription of aldose reductase (AR/AKR1B1) (5,6), sodium-and chloridedependent betaine transporter (BGT1/SLC6A13) (4,6), sodium myo-inositol co-transporter (SMIT/SLC5A3) (3,6), and heat shock protein 70 (HSP70/HSPA1B) (7,8), which mediate the intracellular accumulation of sorbitol, betaine, and myo-inositol and function as an intracellular molecular chaperone, respectively. In addition to these genes, many downstream genes of NFAT5 are known (9 -15). However, a ubiquitin ligase that is directly activated by NFAT5 has not yet been reported.
Ubiquitin ligases are involved in a wide variety of cellular events through ubiquitination (16,17). RING finger protein 183 (RNF183) is a member of the RNF family, which functions as a ubiquitin ligase (18). Previous studies have demonstrated that RNF183 mRNA expression in the colon of patients with inflam-matory bowel disease (IBD) and colorectal cancers was ϳ5and 2-fold higher than that in control subjects; in these patients, RNF183 promotes intestinal inflammation (19) and proliferation and metastasis of cancers (20), respectively. On the contrary, we previously demonstrated that RNF183 was specifically expressed in human and mouse kidney, and that mouse Rnf183 mRNA expression in the kidney was ϳ324fold higher than in the colon (21). To date, however, the reason why Rnf183 mRNA is selectively expressed in the kidney remains unclear.
In this study, we demonstrated that RNF183 is dominantly expressed in the renal medulla and that NFAT5 regulates Rnf183 transcription in mouse inner-medullary collecting duct (mIMCD-3) cells.

RNF183 is predominantly expressed in the renal medulla
RNF183 has been described as a ubiquitin ligase, which is expressed in normal colonic epithelial cells and colorectal cancer cells (19,20). In addition, we previously reported that Rnf183 mRNA expression in the kidney is ϳ324-fold higher than in the colon (21); however, RNF183 protein expression in the kidney remains unclear. To detect RNF183 protein expres-sion, we generated an affinity-purified antibody using recombinant deletion mutant RNF183 (amino acids 61-158) lacking a RING finger domain at its N terminus and a transmembrane domain at its C terminus (Fig. 1A). The antibody specifically recognized GFP-tagged RNF183 at ϳ50 kDa, which was transiently transfected into human embryonic kidney 293 (HEK293) cells (Fig. 1B). Immunofluorescence staining using HeLa cells transiently transfected with GFP-RNF183 showed that punctate signals of GFP-RNF183 were recognized by the anti-RNF183 antibody (Fig. 1C). GFP signals were localized in the late endosome/lysosome (Lamp1) and early endosome (EEA1), with additional weak signal in the recycling endosome (transferrin receptor) and Golgi complex (GM130) but not in the endoplasmic reticulum (calnexin) or mitochondria (COX IV) ( Fig. 1D and Fig. S1). To evaluate endogenous RNF183 protein expression, we performed RT-PCR and Western blot analysis using tissue extracts in 4-week-old mice. Western blotting revealed that endogenous RNF183 protein was expressed markedly in the kidney, particularly in the renal medulla, and in the thymus (Fig. 1, F and H), consistent with Rnf183 mRNA expression (Fig. 1, E and G). Megalin and AKR1B1 were used as positive controls for the renal cortex and medulla, respectively. , and GFP-RNF183 (bottom) used in this study. Anti-RNF183 antibody was generated using deletion mutant RNF183. The number on the right indicates the peptide length. B, Western blotting of HEK293 cells transfected with GFP-RNF183. The anti-RNF183 and GFP antibody recognized a GFP-RNF183 at ϳ50 kDa. C, confocal images of immunofluorescence with anti-RNF183 antibody (top, red) and GFP signals (middle, green) in HeLa cells transfected with GFP-RNF183. Bar, 10 m. D, confocal images of immunofluorescence with GFP signals (middle panels, green) and various antibodies for organelle markers (EEA1 and Lamp1; top panels, red) in HeLa cells transfected with GFP-RNF183. Bars, 10 m. E, RT-PCR analysis of Rnf183 mRNA in 10 tissues from mice. F, tissue immunoblot analysis of RNF183. Tissue lysates containing equal amounts of total protein were analyzed by Western blotting. G and H, expression patterns of RNF183 mRNA and protein in the renal cortex and medulla. Mouse tissue lysates were analyzed by RT-PCR (G) and Western blotting (H). Liver was used as a negative control. Megalin and AKR1B1 were used as positive controls for the renal cortex and medulla, respectively. Arrowhead, full-length megalin (H, middle). The results shown are representative of at least three replicates in independent observations.

NFAT5 regulates Rnf183 transcription RNF183 expression is up-regulated in response to hypertonic stress
Renal medullary cells are constantly exposed to extreme hypertonic stress to concentrate the urine (1). Therefore, using quantitative real-time RT-PCR (qRT-PCR), we examined whether RNF183 expression and the other transmembrane RNF family members, those containing RING finger (C3H2C3 or C3HC4) and transmembrane domains, are up-regulated in response to hypertonic stress. Mouse IMCD-3 cells were incubated with isotonic medium or NaCl-or sucrose-supplemented medium. The extracellular tonicity was increased by ϳ150 mosmol/kg H 2 O (final osmolality: 450 mosmol/kg H 2 O) by adding 75 mM NaCl or 150 mM sucrose. Akr1b1 (5, 6) and Hapa1b (7,8) were used as positive controls for tonicity dependence. We found that Rnf183 mRNA expression was markedly up-regulated in a tonicity-dependent manner in both hypertonic NaCl-and sucrose-treated cells compared with that in isotonic control cells ( Fig. 2A), which was consistent with the Akr1b1 and Hapa1b up-regulation patterns ( Fig. 2A). Rnf128 and Rnf19B were modestly up-regulated in a tonicity-dependent manner. Further, Rnf24, Rnf13, Rnf167, Rnf43, Syvn1, Amfr, Rnf145, Rnf5, Rnf185, and Rnf170 were slightly up-regulated in cells treated with only 75 mM NaCl; the other RNF family members were not up-regulated ( Fig. 2A). Western blot analysis showed that RNF183 and AKR1B1 protein expression was up-regulated in a tonicity-dependent manner (Fig. 2, B-E). However, Rnf183 mRNA was not up-regulated under hypoxic conditions (oxygen concentration, 1 and 0.3%) (Fig. S2), which is another characteristic of the renal medulla. These results suggest that hypertonic conditions play a more important role . RNF183 is up-regulated in response to hypertonic stress. A, the hypertonic response of Rnf183 (top) and the other transmembrane RNF family member (middle, 18 of the C3H2C3 type; bottom, 17 of the C3HC4 type) mRNA levels in mIMCD-3 cells. Akr1b1 and Hspa1b were used as tonicity-dependent positive controls (top). Cells were treated with isotonic or the indicated NaCl-or sucrose-supplemented medium for 12 h and analyzed by qRT-PCR (n ϭ 5). B and D, the tonicity-dependent induction of RNF183 and AKR1B1 protein in mIMCD-3 cells. Cells were cultured under hypertonic conditions by adding the indicated NaCl (B) or sucrose (D) supplementation for 12 h and analyzed by Western blotting. C and E, quantitative analysis of RNF183 (left) and AKR1B1 (right) protein expression in NaCl (B) and sucrose (D) (n ϭ 5). F, the up-regulation patterns of RNF183 and AKR1B1 proteins in four renal cell lines in response to hypertonic stress. NRK-52E, NRK-49F, mIMCD-3, and HEK293 cells were treated with isotonic or 75 mM NaCl-supplemented medium for 12 h and analyzed by Western blotting. HEK293 cells transfected with mouse RNF183 were used as a positive control. Data were analyzed by one-way ANOVA, followed by post hoc tests using t tests with Bonferroni correction. Values represent mean Ϯ S.D. (error bars). *, p Ͻ 0.05; **, p Ͻ 0.01; ***, p Ͻ 0.001 (versus isotonic control).

NFAT5 regulates Rnf183 transcription
in RNF183 expression than hypoxic conditions. Next, we examined whether RNF183 up-regulation was different between mIMCD-3 cells and other renal cell lines. Normal rat kidney (NRK)-52E (a rat kidney tubular epithelial cell line), NRK-49F (a rat kidney interstitial fibroblast cell line), mIMCD-3, and HEK293 cells were used. HEK293 cells transiently transfected with mouse RNF183 were used as a positive control. We found that RNF183 expression increased markedly in mIMCD-3 cells treated with hypertonic NaCl and increased slightly in NRK-52E cells, whereas no expression was detected in NRK-49F and HEK293 cells (Fig. 2F). Interestingly, the difference in RNF183 expression among renal cell lines was similar to that in AKR1B1 expression. The similarity between RNF183 and AKR1B1 upregulation in response to hypertonic stress suggests that NFAT5 regulates RNF183 expression as well as AKR1B1 expression.

RNF183 expression is up-regulated concurrently with NFAT5 activation
The NFAT5 transcription factor is the master regulator for hypertonic stress in mammals (22). In response to hypertonic stress, NFAT5 activation is achieved by a combination of NFAT5 induction and localization into the nucleus (4, 23, 24). Thus, we evaluated the effects of hypertonicity on NFAT5 acti-vation in mIMCD-3 cells. We performed immunofluorescence staining of NFAT5 and plotted the fluorescence intensity along a line drawn through the nucleus. These analyses showed that NFAT5 was present in both the cytoplasm and nucleus under isotonic conditions (Fig. 3A). Hypertonic stimulation for 3 h reduced NFAT5 cytoplasmic expression and slightly enhanced nuclear expression, and stimulation for 12 h drastically enhanced nuclear expression (Fig. 3A). Western blot analysis demonstrated that the total abundance of NFAT5 protein significantly increased after 3 h of hypertonic stimulation and peaked at 9 h (Fig. 3, B and C), consistent with NFAT5 localization (Fig. 3A). Moreover, qRT-PCR analysis revealed that the expression of Rnf183 and NFAT5 downstream genes, Akr1b1 (5, 6) and Hspa1b (7,8), concurrently increased after 3 h of hypertonic stimulation (Fig. 3D). The RNF183 and AKR1B1 protein levels significantly increased after 6 and 9 h of hypertonic stimulation (Fig. 3, E and F), respectively. In contrast to the downstream genes of NFAT5, the Tnfa mRNA level, which is that of NF-B, peaked after 3 h of hypertonic stimulation and decreased after 6 h (Fig. S3). These results demonstrated that RNF183 expression was up-regulated under hypertonic conditions and was accompanied by NFAT5 activation and the up-reg-

NFAT5 regulates Rnf183 transcription
ulation of NFAT5 downstream genes, suggesting that RNF183 expression is directly mediated by NFAT5.

NFAT5 knockdown and p38/MAPK inhibitor SB203580 attenuate hypertonicity-induced RNF183 expression
To determine the effect of NFAT5 inhibition on RNF183 expression, mIMCD-3 cells were transfected with siRNA oligonucleotides targeting NFAT5 (NFAT5 siRNA 1) or negative control siRNA (NC siRNA). After 12 h of hypertonic stimulation, the NFAT5 protein levels were significantly reduced in mIMCD-3 cells transfected with NFAT5 siRNA 1 (Fig. 4, A and B). NFAT5 knockdown significantly inhibited the hypertonicity-induced RNF183 protein and mRNA expression of both Rnf183 and Akr1b1 (Fig. 4, C-E). Another NFAT5 siRNA (NFAT5 siRNA 2) produced similar results (Fig. S4, A-C). The hypertonic stress signal is mediated through p38/MAPK, which regulates NFAT5 activation in mammals (12,13,25,26). Thus, to determine the effect of p38/MAPK inhibition on RNF183 expression, mIMCD-3 cells were treated with 75 mM NaClsupplemented medium in the presence of the indicated concentration of SB203580. The hypertonicity-induced protein and mRNA expression of RNF183 and NFAT5 downstream genes Akr1b1 and Sgk1 were concurrently inhibited by SB203580 in a dose-dependent manner (Fig. 4, F and G). These results demonstrated that RNF183 expression depends on NFAT5.

Identification of the Rnf183 enhancer region that responds to hypertonic stress
To identify the putative enhancer region in the Rnf183 promoter, we searched for an evolutionarily conserved region (ECR) within 5,000 bp upstream of the human RNF183 gene using the ECR browser (27). The 319-bp region (Ϫ4,367 to Ϫ4,049-bp region; numbers indicate nucleotide positions relative to the transcription start site) upstream of the human RNF183 gene was identified as conserved in mammals (blue; Fig. 5A). Although its actual position from the transcription start site varies among the species (Fig. 5B), it shows high sequence identity with the corresponding region in the following mammals: Mus musculus, 71%; Rattus norvegicus, 73%; Canis familiaris, 81%; Bos taurus, 80%; Macaca mulatta, 94%; and Pan troglodytes, 98%. We next searched for potential binding sites for various transcription factors within the mammalian conserved region of RNF183 genes (Ϫ3,466 to Ϫ3,136, mouse (Table S1); Ϫ1,978 to Ϫ1,664, rat (top right); Ϫ4,367 to Ϫ4,049, human (bottom right)) using the JASPAR database (28). This analysis identified one binding site for NFAT proteins conserved among mouse, rat, and human (Table 1 and Table  S1). The sequences (GTACTTTTCCA (Ϫ3,342 to Ϫ3,332; numbers indicate the nucleotide position relative to the transcription start site of mouse Rnf183), GTACTTTTCCA (Ϫ1,870 to Ϫ1,860; numbers indicate the nucleotide position   Table 1 Conserved putative transcription factor binding sites in the RNF183 enhancer region Transcription factor-binding sites predicted using the JASPAR software are shown in the mammalian conserved region of RNF183 genes (mouse, Ϫ3,466 to Ϫ3,136; rat, Ϫ1,978 to Ϫ1,664; human, Ϫ4,367 to Ϫ4,049). Conserved binding sites from mice, rats, and humans were selected from the potential binding sites with a Ͼ95% chance (relative score) of binding to any of the listed transcription factors (Table S1). The conserved NFAT5 DNA-binding site is indicated in boldface type. NFIC, nuclear factor 1 C-type; ZNF354C, zinc finger protein 354C; NFE2L1-MafG, nuclear factor erythroid 2-related factor 1 and V-maf avian musculoaponeurotic fibrosarcoma oncogene homolog G complex; NFAT, nuclear factor of activated T cells; Atoh1, atonal homolog1; Prrx2, paired-related homeobox 2.  5B). Furthermore, the sequences around the binding site were highly conserved among mouse, rat, and human (blue background). Between the conserved region and the transcription start site of the mouse Rnf183 gene (Ϫ3,135 to Ϫ1), we found another site (AGGTTTTTCCA; Ϫ538 to Ϫ528) that was completely homologous with the NFAT5 consensus sequence, but not conserved among the species. To determine whether NFAT5 acts on the Rnf183 promoter containing these sites to activate Rnf183 transcription, we performed luciferase reporter assays using the promoter region in mIMCD-3 cells. A 3.5-kbp (containing both NFAT5 DNA-binding sites) or 2.0-kbp (containing the nonconserved NFAT5 DNA-binding site) promoter region of the mouse Rnf183 gene (pGL4 -3.5 kbp-Luc and pGL4 -2.0 kbp-Luc, respectively; Fig. 5C) was used. The reporter activity of pGL4 -3.5 kbp-Luc, but not that of pGL4 -2.0 kbp-Luc or pGL4-Luc, was considerably induced by hypertonic stress (Fig. 5D). These results demonstrated that the putative Rnf183 enhancer region has responsiveness to hypertonic conditions.

Rnf183 enhancer region depends on NFAT5
To determine whether the putative Rnf183 enhancer region depended on NFAT5, we examined the effects of NFAT5 knockdown, p38/MAPK inhibition, and the dominant negative (DN) form of NFAT5 on luciferase activities. NFAT5 knockdown markedly reduced the reporter activity of pGL4 -3.5 kbp-Luc ( Fig. 6A and Fig. S4D), whereas FLAG-NFAT5 overexpression fully rescued the down-regulated reporter activity ( Fig. 6B and Fig. S5). p38/MAPK inhibitor SB203580 significantly inhibited the hypertonicity-induced Rnf183 luciferase activities (Fig. 6C), which is consistent with mRNA and protein expression of RNF183 (Fig. 4, F and G). Moreover, Rnf183 luciferase activities were inhibited by overexpression of DN-NFAT5, whereas they were up-regulated by that of the full-length NFAT5 (Fig. 6D). These results suggest that the putative Rnf183 enhancer region depends on NFAT5.

Rnf183 is a direct target of NFAT5
We next performed luciferase reporter assays using pGL4 -3.5 kbp-Luc of WT and mutant having a mutation in the conserved NFAT5 DNA-binding site (Fig. 7A). In mIMCD-3 cells transfected with the mutant, reporter activities were markedly reduced even in hypertonic conditions (Fig. 7B). Consistent with this, NFAT5 overexpression markedly induced reporter activity of WT, which was inhibited by the mutation (Fig. 7C). To confirm whether NFAT5 directly binds to the conserved binding site in the Rnf183 promoter, we performed ChIP assays using mIMCD-3 cells under hypertonic conditions (Fig. 7D) and detected a high level of NFAT5 binding to the conserved binding site in the endogenous Rnf183 promoter (Fig. 7, E and  F). These results indicated that NFAT5 directly acts on the conserved NFAT5 DNA-binding site within the Rnf183 promoter and facilitates its transcription in mIMCD-3 cells.

RNF183 expression provides protection against hypertonicityinduced apoptosis
Stress induced by an elevated NaCl concentration mediates apoptosis in mIMCD3 cells when osmolality is acutely raised to Ͼ600 mosmol/kg (29 -31). Moreover, several NFAT5 downstream genes show a protective effect against hypertonicityinduced apoptosis (9,32,33). Thus, to determine whether RNF183 expression affects the osmotic tolerance of mIMCD-3 cells to high NaCl levels, mIMCD-3 cells were transfected with siRNA oligonucleotides targeting RNF183 (RNF183 siRNA) or NC siRNA. RNF183 siRNA effectively reduced the Rnf183 mRNA levels compared with NC siRNA (Fig. 8A). After 4 h of exposure to 200 mM NaCl-supplemented medium, cleaved caspase-3 protein levels and cleaved caspase-3-positive cells (which are indexes of apoptosis) significantly increased in mIMCD-3 cells transfected with RNF183 siRNA (Fig. 8, B and  C). Crystal violet assays revealed that RNF183 knockdown significantly reduces cell viability under hypertonic conditions (Fig. 8D). In contrast, RNF183 knockdown did not promote caspase activation upon treatment of mIMCD-3 cells with the apoptosis inducer staurosporine (Fig. S6). These results dem-

Rescue of RNF183 knockdown attenuates hypertonicityinduced apoptosis
To eliminate the possibility of siRNA off-target effects, mIMCD-3 cells transfected with RNF183 siRNA or NC siRNA were rescued with siRNA-resistant 3ϫFLAG-RNF183, which was expressed in mIMCD-3 cells under RNF183 knockdown (Fig. 9A). After 4 h of exposure to 200 mM NaCl-supplemented medium, cleaved caspase-3 protein levels significantly decreased in mIMCD-3 cells rescued with 3ϫFLAG-RNF183 (Fig.  9B). Crystal violet assays showed that the rescue of RNF183 knockdown significantly improved cell viability under hypertonic conditions (Fig. 9C). These results confirm that RNF183 protects inner-medullary cells from hypertonicity-induced apoptosis.

Discussion
We identified the conserved NFAT5 DNA-binding site in the Rnf183 enhancer region and found that this site induced RFN183 reporter activity under hypertonic conditions, whereas mutation at the site resulted in almost complete inhibition of Rnf183 reporter activity. In addition, ChIP assays revealed a high level of NFAT5 binding to the site. Thus, these results demonstrated a mechanism of RNF183 up-regulation under hypertonic conditions by which NFAT5 directly binds to the conserved NFAT5 DNA-binding site in the Rnf183 enhancer region and leads to its transcriptional activation. This is the first study to demonstrate that ubiquitin ligase can be directly regulated by NFAT5. However, Yu et al. (19) demonstrated that microRNA-7 (miR-7) directly bound to the 3Ј-UTR of RNF183 mRNA, leading to RNF183 degradation and trans-lational inhibition, and that down-regulation of miR-7 induced RNF183 elevation in inflamed colon tissues of IBD patients and colitic mice. Of the two regulators of RNF183 expression, we considered NFAT5 to be more important than miR-7 in a normal kidney because the tissue distribution pattern of RNF183 is different from that of miR-7 (34). The expression of miR-7 is restricted to specific tissues of the brain, colon, and thymus, whereas it is extremely low in the kidney, heart, liver, lung, spleen, and stomach (34), which is not consistent with our results on RNF183 expression (Fig. 1, E and F). Whereas NFAT5 is broadly expressed as an essential factor during exposure to the hypertonic environment (4), systemic or immature thymocyte-specific NFAT5 knockout mice display profound defects in the renal medulla and reduced thymocyte compartment in the thymus, respectively (6,35,36). This is because NFAT5 is activated by hypertonicity-dependent and -independent pathways in the renal medulla (6,37,38) and thymus (36), respectively. Consistent with these findings, RNF183 was highly expressed in the renal medulla and was also expressed in the thymus (Fig. 1, E-H). Therefore, we hypothesized a mechanism by which high RNF183 expression in the normal kidney and thymus is predominantly regulated by the NFAT5 activation, but not miR-7.
We observed an ϳ2-3-fold increase in the induction of Rnf183 reporter activity following hypertonic stimulation. One possible explanation for this mild induction of the Rnf183 promoter is that only a short sequence of the Rnf183 promoter was investigated. Consistent with our results, Hasler et al. (39) and Chen et al. (12) have demonstrated that the activity of the AQP2 and SGK1 promoters (each containing one NFAT5-binding element) increased ϳ1.5and 5-fold, respectively, following hypertonic stimulation. However, at least five NFAT5-binding

NFAT5 regulates Rnf183 transcription
elements spread over 50-kb pairs, which participate in NFAT5mediated transcriptional stimulation, were revealed in the SMIT promoter (3). Similarly, three NFAT5-binding sequences that were identified in the AKR1B1 promoter may collectively participate in the hyperosmotic response (5). Therefore, other NFAT5-binding elements may be present in regions of the Rnf183 promoter that are further upstream or downstream. Moreover, we demonstrated that Rnf183 reporter activity and Rnf183 and Akr1b1 mRNA expression were concurrently reduced by NFAT5 knockdown under isotonic conditions. Several previous studies have shown that, to some extent, NFAT5 knockdown or dominant negative NFAT5 expression inhibited mRNA expression or promotor activity of its downstream genes under isotonic conditions (7,9,39,40). In contrast, two studies have shown that NFAT5 inhibition did not affect the basal expression of its downstream genes under isotonic conditions (12,41). This difference in NFAT5 inhibition may result from variation in the basal NFAT5 protein levels because basal AKR1B1 protein expression levels were high in mIMCD-3 cells compared with those in other renal cell lines (Fig. 2F).
We further examined whether RNF183 was up-regulated by hypoxia in mIMCD-3 cells because the renal medulla is charac-teristically hypoxic (42). Although vascular endothelial growth factor A (Vegfa) mRNA, which is downstream of the hypoxic master regulator hypoxia-inducible factor 1 (HIF-1) (43), was significantly up-regulated under hypoxia (oxygen concentration, 1 and 0.3%), Rnf183 was not up-regulated (Fig. S2). Moreover, the HIF-1 DNA-binding site (ACGTG) was not observed in the Rnf183 enhancer region (Table S1). These results suggest that the high RNF183 expression in the renal medulla does not result from hypoxia.
Our results demonstrated that Rnf183 mRNA was markedly up-regulated compared with the other transmembrane RNF family members under hypertonic conditions and that the siRNA-induced knockdown of RNF183 exacerbates apoptosis and reduces cell viability in mIMCD-3 cells at ϳ700 mosmol/ kg. Although the effects of RNF183 on adaptation to hypertonicity appear to be mild, we consider RNF183 to be as important as other NFAT5-regulated genes. Lee et al. (44) demonstrated up to a 1.5-fold increase in the hypertonicity-induced LDH release in MEF cells transfected with siRNA targeting SMIT, BGT1, AKR1B1, HSPA1B, etc. Shim et al. (33) have also demonstrated an ϳ20 -30% decrease in cell viability following 100 mM NaCl supplementation in HSPA1B-knockout MEF

NFAT5 regulates Rnf183 transcription
cells. Their effects on adaptation to hypertonicity appear to be similar to that of RNF183 in the present study. Therefore, RNF183 expression is induced by hypertonic stress and plays an important role in the kidney's osmoprotective function under hypertonic conditions. However, we could not clarify the substrate of RNF183 under hypertonic conditions. To date, three previous studies have demonstrated the different functions of RNF183 in cancers and the colon as follows. (i) In tumors, RNF183 interacts with fetal and adult testis-expressed 1 (FATE1), which is a cancer/testis antigen, and negatively regulates the apoptotic effector Bcl-2-interacting killer (BIK), a component of FATE1 complex, leading to increased tumor cell viability (45). (ii) In the colon, RNF183 degrades IB␣, thereby inducing NF-B activation, which might contribute the pathogenesis of IBD (19). (iii) In the colorectal cancers, RNF183 induces the activation of NF-B and expression of its downstream gene, IL8, resulting in increased proliferation and metastasis of colorectal cancer cells (20). However, FATE1 expression is restricted to cancer or testis and is absent in the other tissues, such as the kidney (45). Moreover, NF-B activation by hypertonicity is transient and starts decreasing after 3 h of hypertonic stimulation in cortical collecting duct cells due to the restoration of IB␣ protein levels (46,47). Consistent with this, the mRNA level of Tnfa, an NF-B downstream gene, peaked after 3 h of hypertonic stimulation and decreased after 6 h at the time of RNF183 protein level elevation (Fig. S3). Hence, these results suggest that RNF183 ubiquitylates other substrates under hypertonic conditions as there is a contradic-tion on whether the NFAT5-RNF183 axis regulates BIK and IB␣ expression.
RNF183 expression in response to hypertonicity is mediated by NFAT5, which is essential for long-term adaptation to hypertonic conditions (48). Short-term adaptation involves rapid influx of electrolytes and amino acids through various transporters, such as Na ϩ -K ϩ -Cl Ϫ cotransporters, Na ϩ /H ϩ exchangers, Cl Ϫ /HCO 3 Ϫ exchangers (49), and sodium-coupled neutral amino acid transporter 2 (50). Long-term adaptation involves delayed replacement of ionic osmolytes with uncharged small organic osmolytes via expression of various enzyme and transporters, such as AKR1B1 (5, 6), BGT1 (4, 6), and SMIT (3,6), mediated by NFAT5. However, in long-term adaptation, the mechanism by which the dispensable transporters of ionic osmolytes are down-regulated remains unclear. Here, the punctate signals of RNF183 were predominantly localized to endosomes and lysosomes (Fig. 1D). Early or late endosome-localized ubiquitin ligases can sort internalized cell-surface receptors from the recycling pathway to lysosomal degradation through ubiquitination (51)(52)(53). Therefore, we speculated that RNF183 plays a role in sorting internalized dispensable transporters to lysosomes for long-term adaptation to hypertonic conditions.
In summary, NFAT5 directly binds to the conserved DNA binding site in the Rnf183 enhancer region and leads to Rnf183 transcriptional activation, thereby inducing its high expression in the renal medulla. Among the transmembrane RNF family members, Rnf183 is specifically up-regulated by hypertonic stress and protects the cells from hypertonicity-induced apoptosis. However, our study did not include identification of substrate for RNF183 and validation by animal experiments under hypertonic conditions. Further study is warranted to elucidate the localization and function of RNF183 in the kidney.

NFAT5 regulates Rnf183 transcription
N terminus of the insert sequences. pFLAG-NFAT5, the expression vector of NFAT5, was gifted to us by Dr. B. C. Ko (23) and Dr. Takashi Ito (9). The pFLAG-DN-NFAT5 was a generous gift from Dr. Takashi Ito (9).

Animals
Four-week-old C57BL/6 mice were used in this study. All animal experiments were performed in accordance with the National Institutes of Health guidelines for the care and use of laboratory animals and were approved by the Committee of Animal Experimentation, Hiroshima University. The mice were maintained in a room at 23°C and a constant day-night cycle and were provided food and water ad libitum.

Analysis of mRNA levels in culture cells
Total RNA from mIMCD-3 cells was extracted using ISOGEN. Reverse transcription was performed with ReverTra Ace. The reverse-transcribed cDNA was measured by TaqManbased real-time PCR assay using the ⌬⌬Ct method.

ChIP assay
Two 10-cm plates of ϳ80% confluent mIMCD-3 cells were cultured for 24 h and treated with 75 mM NaCl for 4 h prior to cross-linking. Cells were treated with 1% (v/v) formaldehyde for 15 min at 37°C, followed by an additional 5 min in 150 mM glycine. Subsequently, cells were washed with cold PBS and lysed in nuclear lysis buffer (50 mM Tris-HCl (pH 8.0), 10 mM EDTA, 1% SDS, and protease inhibitor mixture) for 10 min on ice. Nuclear lysates were sonicated to a final average fragment length of 1,000 bp using a Bioruptor II (BM Equipment, Tokyo, NFAT5 regulates Rnf183 transcription Japan). Sonicated chromatin was cleared by centrifugation at 13,000 rpm for 10 min and then diluted 10-fold with dilution buffer (16.7 mM Tris-HCl (pH 8.0), 1.2 mM EDTA, 0.01% SDS, 1.1% Triton X-100, 167 mM NaCl, and protease inhibitor mixture). The soluble chromatin was incubated with 2 g of normal IgG rabbit (catalog no. 2729, Cell Signaling Technology) or anti-NFAT5 antibody (catalog no. ab3446, Abcam) on a rotating platform at 4°C overnight, followed by incubation with 20 l of Magna ChIP Protein A ϩ G magnetic beads (catalog no. 16-663, Millipore) for 1 h. Subsequently, the beads were washed with the following four buffers: low-salt buffer (20 mM Tris-HCl (pH 8.0), 2 mM EDTA, 0.1% SDS, 1% Triton X-100, 150 mM NaCl), high-salt buffer (same as low-salt buffer with 500 mM NaCl), LiCl buffer (10 mM Tris-HCl (pH 8.0), 1 mM EDTA, 0.25 M LiCl, 1% Nonidet P-40, 1% sodium deoxycholate), and TE buffer (10 mM Tris-HCl (pH 8.0), 1 mM EDTA). Precipitated complexes were eluted twice from the beads in fresh elution buffer (1% SDS, 0.1 M NaHCO 3 ) for 30 min each at room temperature and subsequently heated at 65°C overnight to reverse formaldehyde-induced cross-linking. The DNA was purified by phenol-chloroform extraction and ethanol precipitation. The purified DNA was amplified by PCR in a total volume of 20 l containing a 0.45 M concentration of each primer, 0.4 mM dNTPs, 1 unit of KOD FX DNA polymerase, and 2ϫ PCR buffer (catalog no. KFX-101, Toyobo). The following primer set was used, yielding a 303-bp product: 5Ј-GTTTTACCCTTTTGC-CTGTTTCCTT-3Ј (forward) and 5Ј-CAGGCTAAGAAGTC-CTTGGTAAACA-3Ј (reverse). The PCR conditions were as follows: 94°C for 5 min, 33 cycles at 94°C for 15 s, 60°C for 30 s, 68°C for 20 s, and 68°C for 3 min. The PCR products were resolved by electrophoresis on a 4.8% acrylamide gel. The density of each band was quantified using the Adobe Photoshop Elements version 2.0 program (Adobe Systems Inc.).

Cell viability assay
mIMCD-3 cells were transfected with RNF183 or NC siRNA in 6-well plates (12.5 pmol/well). After 24 h, the cells were reseeded in a 24-well plate and cultured in 5% CO 2 at 37°C for an additional 24 h. Subsequently, cells were treated with isotonic or 200 mM NaCl-supplemented medium in 5% CO 2 at 37°C for 12 h and washed with PBS and then stained with crystal violet. The wells were washed with water and air-dried, and the dye was eluted with water-containing 0.5% SDS. The absorbance was measured at 590 nm using a Bio-Rad SmartSpec 3000 spectrophotometer, according to the method reported by Omura et al. (56).

Electroporation
For RNF183 rescue experiments using electroporation, mIMCD-3 cells were transfected with RNF183 or NC siRNA in 10-cm dishes (62.5 pmol/dish). After 12 h, the cells were trypsinized and resuspended in DMEM/F-12. They were then pelleted at 1,000 rpm for 5 min and resuspended in Opti-MEM at a working concentration of ϳ1.0 ϫ 10 7 cells/ml. One hundred microliters of cell suspension was added to 10 g of empty or 3ϫFLAG-RNF183-pcDNA vectors in electroporation cuvettes (catalog no. SE-202, BEX Co., Ltd., Tokyo, Japan) and then transfected by one direct-current 150-V electrical pulse and five direct-current 20-V electrical pulses at intervals of 50 ms using an electroporator (CUY 21 Vitro-EX, BEX Co.). After 48 h, cells were treated with 200 mM NaCl in DMEM/F-12 for the indicated time and analyzed using Western blotting or a crystal violet assay.

Statistical analysis
Results are expressed as the mean Ϯ S.D. Statistical evaluation was performed using JMP statistical software (version Pro 12). Comparisons between two groups were analyzed using the two-tailed t test. For multiple-group comparisons, one-way analysis of variance (ANOVA), followed by two-tailed Student's t test with Bonferroni correction were applied.