Abstract

The E3 ubiquitin (Ub)-protein ligases (E3s) play a role as regulators of protein trafficking and degradation. We aimed to integrate the profile of E3s in rat kidney and examine the changes in protein abundance of the selected E3s in response to 1-deamino-8-d-arginine vasopressin (dDAVP) stimulation/withdrawal. Sprague-Dawley rats were infused with vehicle (n = 13), dDAVP for 5 days (n = 13), or dDAVP was withdrawn for periods (15 min, 30 min, 1, 3, 6, 12, or 24 h) after 5-day infusion (n = 46). Total RNA was isolated from the inner medulla (IM) for transcriptome analysis. Plasma membrane (PM)- or intracellular vesicle (ICV)-enriched fractions of whole kidney were immunoisolated for liquid chromatography-tandem mass spectrometry analysis. dDAVP infusion for 5 days (D5d) significantly increased urine osmolality, which was maintained during 3-h withdrawal of dDAVP after 5-day infusion (D5d-3h). Consistent with this, aquaporin-2 (AQP2) expression in the PM fractions of D5d and D5d-3h increased, whereas AQP2 expression in the ICV fractions of D5d-3h was further increased, indicating internalization of AQP2. Transcriptome analysis revealed 86 genes of E3s and LC-MS/MS analysis demonstrated 16 proteins of E3s. Among these, seven E3s (BRCA1, UBR4, BRE1B, UHRF1, NEDD4, CUL5, and FBX6) were shared. RT-PCR demonstrated mRNA expressions of the seven identified E3s in the kidney, and immunoblotting demonstrated changes in protein abundance of the selected E3s (BRE1B, NEDD4, and CUL5) in response to dDAVP stimulation/withdrawal or lithium-induced nephrogenic diabetes insipidus. The rate of AQP2 degradation was retarded in mpkCCDc14 cells with small interfering RNA-mediated knockdown of NEDD4 or CUL5. Taken together, identified E3s could be involved in the degradation of proteins associated with vasopressin-induced urine concentration.

  • proteomic analysis
  • siRNA
  • transcriptome analysis
  • ubiquitination

urine concentration is critical for maintaining body water homeostasis. Regulation of aquaporin-2 (AQP2) induced by vasopressin stimulation in the collecting duct principal cells plays a central role in this process. AQP2 has a pivotal role in water reabsorption by both short-term regulated translocation of intracellular AQP2-expressing vesicles to the apical plasma membrane (29, 37, 40, 53) and long-term adaptation of AQP2 protein abundance (41, 52). In contrast to the relatively well-established pathways involved in vasopressin-regulated urine concentration mediated by AQP2 trafficking and de novo synthesis of AQP2, intracellular degradation of the proteins which is regulated in response to vasopressin-induced urine concentration is poorly understood.

Previous studies demonstrated that ubiquitination plays a major role in the endocytosis and lysosomal or proteasomal degradation of proteins (10, 12, 20, 51). Ubiquitin (Ub) is a 76-amino acid protein, which is covalently attached to the lysine residue of the substrate proteins (13, 18). Activation and attachment of Ub to a target protein are known to be mediated by the action of three enzymes (i.e., E1, E2, and E3) (9, 16, 42, 44). The E1 (Ub-activating enzyme) activates the C terminus of Ub in an ATP-dependent manner, and both E2 (Ub-conjugating enzyme) and E3 (Ub-protein ligase) are involved in the attachment of Ub to a target protein through the ε-amino group of a lysine residue (17). In particular, E3 Ub-protein ligases (E3s) are known to have the substrate specificity, and hence they play an important role in determining the selectivity of Ub-mediated protein degradation (54). The degradation pathways, therefore, balance the abundance of proteins which play an important role in body water homeostasis.

In the present study, we aimed to identify a profile of E3 genes and proteins in the rat kidney which could be involved in the intracellular degradation of proteins associated with vasopressin-induced urine concentration. We hypothesized that identified E3s play a role in the degradation of vasopressin-induced proteins in the kidney. In particular, to address the potential role of selected E3s in the regulation of AQP2 protein abundance, small interfering RNA (siRNA)-mediated gene silencing of the selected E3s was done in mouse cortical collecting duct cells (mpkCCDc14 cells) (7), and the altered degradation rate of AQP2 abundance was examined. In addition, semiquantitative immunoblotting was done to examine the changes in protein abundance of the selected E3s in response to 1-deamino-8-d-arginine vasopressin (dDAVP) stimulation/withdrawal or lithium-induced nephrogenic diabetes insipidus (Li-NDI).

METHODS

Experimental protocols for long-term and short-term dDAVP withdrawal after 5 day-dDAVP infusion (protocols 1 and 2).

Pathogen-free male Sprague-Dawley rats (200- 250 g) were obtained from Charles River (Orient Bio, Seongnam, Korea). The animal protocols were approved by the Animal Care and Use Committee of Kyungpook National University, and all animal experiments were conducted according to the guidelines of Kyungpook National University.

For protocol 1 (long-term dDAVP withdrawal after 5-day dDAVP infusion), vehicle-treated control rats (n = 7) and dDAVP-treated rats (n = 29) were maintained in metabolic cages. All animals were maintained on a fixed amount of standard rodent diet (20 g·220 g body wt−1·day−1; 2918C, HarlanTeklad, Madison, WI) with free access to water intake. For vehicle (saline, 1 μl/h sc for 5 days, n = 7) or dDAVP (40 ng/h sc for 5 days, n = 29, V1005, Sigma) infusion, osmotic minipumps (model 2001, Durect, Cupertino, CA) were implanted subcutaneously in the neck of each rat (28). In the dDAVP-treated group (n = 29), rats were divided into five subgroups: dDAVP infusion for 5 days (D5d; n = 7) and 5-day dDAVP infusion followed by dDAVP withdrawal for 3 h (D5d-3h; n = 7), 6 h (D5d-6h; n = 7), 12 h (D5d-12h; n = 4), or 24 h (D5d-24h, n = 4). For dDAVP withdrawal after 5-day dDAVP infusion, osmotic minipumps were removed from rats under light enflurane inhalation anesthesia and the animals were awakened, returned to metabolic cages, and kept for periods to collect urine output.

For protocol 2 (short-term dDAVP withdrawal after 5-day dDAVP infusion), vehicle-treated control rats (n = 6) and dDAVP-treated rats (n = 30) were maintained in metabolic cages. This protocol was identical to protocol 1 except for the durations of dDAVP withdrawal after 5-day DAVP infusion. In the dDAVP-treated group (n = 30), rats were divided into five subgroups: dDAVP infusion for 5 days (D5d, n = 6) and 5-day dDAVP infusion followed by dDAVP withdrawal for 15 min (D5d-15m; n = 6), 30 min (D5d-30m; n = 6), 1 h (D5d-60m; n = 6), or 3 h (D5d-180m; n = 6).

Experimental protocols for Li-NDI.

Experiments were performed using kidney samples from previously studied Li-NDI rats (19). Male Sprague-Dawley rats [200–250 g, control rats (n = 6) and Li-treated rats (n = 12)] received daily food rations of a food mixture consisting of 15 g ground rat chow (2018S, HarlanTeklad,) and 20 ml tap water, as previously described (19, 27, 39). The Li-treated rats (n = 12) received 0.6 mmol LiCl (L 4408, Sigma) added to the food mixture per day for 7 days, and thereafter rats received either 0.6 mmol LiCl (low-dose Li group, n = 6) or 0.9 mmol LiCl (high-dose Li group, n = 6) added to the food mixture per day for the following 16 days. Rats had free access to water intake.

Membrane fractionation and immunoisolation.

Rats were anesthetized under enflurane inhalation, and kidneys were rapidly removed. Whole kidneys were homogenized in 10 ml of dissection buffer (0.3 M sucrose, 25 mM imidazole, 1 mM EDTA, 8.5 μM leupeptin, 1 mM phenylmethylsulfonyl fluoride, pH 7.2) containing iodoacetamide (IAA; 20 mM, catalog no. I-1149, Sigma). IAA was used as an inhibitor of deubiquitinating enzymes (DUB), since it alkylates the cysteine residues at the DUB active site. Membrane fractions enriched for either the plasma membrane (PM) or intracellular vesicle (ICV) from rat whole kidney were prepared by differential centrifugation, as demonstrated previously (30, 36). Immunoisolated samples were made by either anti-AQP2 (H7661AP, ∼2 μg/107 beads) or anti-Ub antibody (P4D1, ∼2 μg/107 beads, Cell Signaling), as demonstrated previously (30).

Electrophoresis and immunoblotting of AQP2.

SDS-PAGE was performed on 9–12% polyacrylamide gels, as previously described (28, 30, 31). Primary antibodies used were anti-AQP2 (H7661AP)(24, 38), anti-Ub (P4D1, 3936, Cell Signaling), anti-BRE1B (ab62528, Abcam), anti-CUL5 (42–2200, Zymed), and anti-NEDD4 (2740, Cell Signaling).

Immunohistochemistry of AQP2.

Left kidneys were fixed by retrograde perfusion via the aorta with 3% paraformaldehyde in PBS, pH 7.4. Immunolabeling was performed on sections from paraffin-embedded preparation (2-μm thickness) using the methods described previously (38).

Transcriptome analysis of E3 ligases in the inner medulla.

Gene expression profiles (Table 1) were analyzed on a GeneChip Rat Gene 1.0 ST array (Affymetrix, Santa Clara, CA) according to the manufacturer's instructions. Total RNA from rat kidney inner medulla was extracted by using TRIzol reagent. RNA that had a high RNA integrity number (RIN >9.0) and A260/A280 absorbance ratio ranging from 1.8 to 2.1 was used for cDNA synthesis. The amplification cycle of RNA to cDNA, cDNA to cRNA and cRNA to single-stranded cDNA was performed according to the manufacturer's instructions using the GeneChip Poly-A RNA control kit and the GeneChip WT cDNA Synthesis and Amplification Kit (Affymetrix). The single-stranded cDNA was then purified, enzymatically fragmented, biotinylated, and the labeled fragmented cDNA was hybridized to the GeneChip at 45°C for 17 h. The hybridized probe array was stained, washed, and the stained GeneChip probe array was scanned by a GeneChip Scanner 3000 (Affymetrix) 7G at 570 nm (GEO accession no. GSE27449). The signal intensity of the gene expression level was calculated by Expression Console software, version 1.1 (Affymetrix).

View this table:
Table 1.

E3 Ub-protein ligases identified by transcriptome analysis

Liquid chromatography-tandem mass spectrometry analysis for identification of E3 ligases and quantification.

The immunoisolated samples from three groups (control, D5d, and D5d-3h) were reduced with 10 mM DTT at 56°C for 20 min, and the protein samples were digested by trypsin (trypsin:protein = 1:50) overnight at 37°C. The digested peptides were eluted from beads, loaded on a C18 trapping column, and desalted using 0.1% formic acid in water. Separation was performed using a C18 nanocolumn (75-μm inner diameter, 150-mm length, 1.7-μm-sized particles) and then analyzed by nanoUPLC Q-Tof Premier (Waters). Accurate mass liquid chromatography-mass spectrometry (LC-MS) data were collected in a data-independent, alternate scanning mode of acquisition (MSE mode) and processed using ProteinLynx Global Server v2.25 (Waters) (11). The processed spectrum was converted to PKL files, which were searched against the IPI rat database using the MASCOT search engine v2.0 (Matrix Science, www.matrixscience.com). Search parameters were as follows: trypsin digest (2 missed cleavage), MS tolerance: ±0.6 Da, tandem MS (MS/MS) tolerance; ±0.1 Da, and modifications: methionine oxidation and carbamidomethyl-cysteine. Positive protein identification was based on standard MASCOT criteria for ESI-MS/MS data (33). Peptides and proteins were considered identified if their MASCOT individual ion score was higher than the MASCOT identity scores (P < 0.05, protein score with ≥31). Acceptable criteria for protein identification were set to a requirement of more than two peptides per protein. All the false discovery rates that were measured by a MASCOT decoy search against a randomized database for each analysis were under 3% (8). For protein quantification, label-free quantification was used, requiring exact mass and retention time (EMRT) among samples runs (49). Clustering in the deconvoluted mass is typically conducted with a precision of better than 10 ppm and in the retention time better than 0.25 min. Collected EMRT pairs of all same peptides identified at more than two groups were used to measure ion intensities of peptides of each E3 protein. The average intensity ratios of peptides corresponding to each protein from two biological replicates were calculated (Tables 2 and 3).

View this table:
Table 2.

E3 Ub-protein ligases identified by LC-MS/MS analysis in the AQP2-immunoisolated PM- and ICV-enriched fractions

View this table:
Table 3.

E3 Ub-protein ligases identified by LC-MS/MS analysis in Ub-immunoisolated PM- and ICV-enriched fractions

RT-PCR of E3 Ub-protein ligases.

Total RNA from rat whole kidney or inner medulla was used, and one-step RT-PCR was performed using the Access RT-PCR system (Promega, Madison, WI). Primer sequences for RT-PCR are demonstrated (Table 4).

View this table:
Table 4.

Primer sequences for RT-PCR

Cycloheximide studies.

To examine the role of selected E3s in the degradation of AQP2, mpkCCDc14 cells (passages 32–36) were cultured in a semipermeable filter of the Transwell system (0.4-μm pore size, Transwell Permeable Supports, catalog no. 3460, Corning) for 1 day. Thereafter, cells were transiently transfected with nontargeting siRNA or E3-directed siRNA for 48 h in the presence of dDAVP (10−9 M, both at the apical and basolateral side) stimulation for the last 30 h. For transient transfection of CUL5 or NEDD4-directed siRNAs to mpkCCDc14 cells, siGENOME SMARTpool mouse CUL5-siRNA, NEDD4-siRNA, or nontargeting siRNA (Dharmacon, Chicago, IL) was mixed with Lipofectamine RNAi MAX (Invitrogen, Carlsbad, CA) in opti-MEM, respectively. Each siRNA-Lipofectamine RNAi MAX complex was incubated with mpkCCDc14 cells, respectively, for 48 h in a 37°C incubator. The nontargeting siRNA-transfected mpkCCDc14 cells were used as a control.

Then, cells were treated by cycloheximide (10−4 M) for periods (0, 2, 4, and 8 h) to block the protein synthesis at the translational level. Protein samples prepared at various time points were subjected to AQP2 immunoblotting.

In addition, to examine whether proteosomal and/or lysosomal degradation of AQP2 could play a critical role in regulation of AQP2 abundance, inner medullary collecting duct (IMCD) cells from male Sprague-Dawley rats (200–250 g) were primary cultured as described previously (31). The IMCD cell suspension was then seeded in rat fibronectin-coated filter plates in the Transwell system (catalog no. 3460, Corning). At day 3, cells were treated with cycloheximide (10−4 M), MG-132 (10−3 M, Calbiochem), or chloroquine (10−3 M, Sigma) for 8 h and were subjected to immunoblotting with AQP2 or Ub antibodies.

Statistical analyses.

Values are presented as means ± SE. Data were analyzed by one-way ANOVA followed by Tukey's honestly significant difference multiple comparisons test. Multiple comparisons tests were only applied when a significant difference was determined in the ANOVA, P < 0.05.

RESULTS

Changes in urine output, urine osmolality, and renal AQP2 protein abundance in response to 5-day dDAVP infusion or dDAVP withdrawal after 5-day dDAVP infusion in vivo.

In protocol 1, dDAVP infusion for 5 days (D5d) significantly decreased urine output and increased urine osmolality, compared with vehicle-treated controls (Fig. 1, A and B). Increased urine osmolality was maintained in rats during 3-h dDAVP withdrawal after 5-day dDAVP infusion (D5d-3h), whereas it returned to the control level at 6-h dDAVP withdrawal (D5d-6h, Fig. 1B). Whole kidney AQP2 levels were increased in response to D5d and upregulated AQP2 abundance was maintained during 12-h dDAVP withdrawal (Fig. 1, C and F). Importantly, consistent with the observed changes in urine osmolality (Fig. 1B), AQP2 abundance in the PM-enriched fractions of both D5d and D5d-3h was significantly increased, and thereafter AQP2 abundance returned to the control level (Fig. 1, D and G). In contrast, AQP2 level in the ICV-enriched fractions was significantly increased in response to D5d, and it was further increased at 3-h and 6-h dDAVP withdrawal after 5-day dDAVP infusion (Fig. 1, E and H). This finding indicates that dDAVP-induced AQP2 abundance in the PM is retrieved into the ICV when dDAVP stimulation is withdrawn, along with a gradual reduction in whole kidney AQP2 abundance to the control level. Immunofluorescence microscopy further demonstrated the changes in subcellular AQP2 localization of the PM and ICV at D5d and D5d-3h (Fig. 1, IK).

Fig. 1.

Changes in urine output, urine osmolality, and aquaporin-2 (AQP2) abundance in response to 5-day dDAVP infusion (D5d) or dDAVP withdrawal after D5d in vivo (protocol 1). A: changes in urine output. D5d significantly decreased urine output. B: changes in urine osmolality. Long-term dDAVP infusion-induced increase in urine osmolality was maintained in rats despite the removal of dDAVP stimulation for 3 h after D5d (D5d-3h). Thereafter, it returned to the control level. CH: semiquantitative immunoblotting of AQP2 in whole kidney (WK; C and F), plasma membrane (PM)-enriched fractions (D and G), and intracellular vesicle (ICV)-enriched fractions (E and H). Changes in AQP2 levels in response to vehicle infusion (control; Con), D5d, or 3-, 6-, 12-, and 24-h dDAVP withdrawal after D5d (D5d-3h, D5d-6h, D5d-12h, and D5d-24h) are shown. *P < 0.05 compared with control. #P < 0.05 compared with D5d; n, no. of rats. IK: immunofluorescence microscopy of AQP2 in the inner medulla of vehicle-treated control (I), D5d (J), and D5d-3h (K) rats. Increased AQP2 intensity in the apical PM (arrows in J) was seen at D5d. In contrast, the AQP2 labeling intensity in the apical PM was decreased at D5d-3h, along with an increase in the intracellular AQP2 labeling (K).

The changes in urine output, urine osmolality, and AQP2 abundance were also examined in a rat model of short-term withdrawal of dDAVP (protocol 2). D5d significantly decreased urine output and increased urine osmolality (Fig. 2, A and B), which were accompanied by upregulation of whole kidney AQP2 levels (Fig. 2, C and F). D5d increased AQP2 level in the PM-enriched fractions, which was maintained during 3-h dDAVP withdrawal (Fig. 2, D and G). Importantly, AQP2 abundance in the ICV-enriched fractions was unchanged within 1-h dDAVP withdrawal, but it was significantly elevated at 3-h dDAVP withdrawal (Fig. 2, E and H). This finding indicates that long-term dDAVP-induced AQP2 abundance in the PM is retrieved into the ICV when dDAVP stimulation is withdrawn for 3 h. This finding led us to select kidney samples collected from the time point of D5d-3h, in addition to the samples of vehicle-treated control and D5d to identify a profile of E3s which could play a role in the degradation of vasopressin-induced proteins in the kidney.

Fig. 2.

Changes in urine output, urine osmolality, and AQP2 expression (protocol 2). A and B: changes in urine output and urine osmolality. D5d significantly decreased urine output and increased urine osmolality. B: urine osmolality was measured in the urine collected in the morning of day 5 at the same time instead of at various time points during the dDAVP withdrawal period after long-term stimulation. CH: semiquantitative immunoblotting of AQP2 in WK (C and F), PM-enriched fractions (D and G), and ICV-enriched fractions (E and H). Changes in AQP2 levels were examined in response to vehicle infusion (contol), D5d, or 15-, 30-, 60-, and 180-min dDAVP withdrawal after D5d (D5d-15m, D5d-30m, D5d-60m, and D5d-180m). *P < 0.05 compared with control. #P < 0.05 compared with D5d; n, no. of rats.

Changes in the prevalence of AQP2 among ubiquitinated proteins during dDAVP withdrawal after 5-day dDAVP stimulation.

PM- or ICV-enriched fractions of rat whole kidneys were immunoisolated by ubiquitin antibody for isolating the ubiquitinated proteins. Then, the immunoisolated proteins were subjected to AQP2 immunoblotting to examine the changes of AQP2 prevalence among the ubiquitinated proteins during dDAVP withdrawal after long-term stimulation. In protocol 1, AQP2 abundance in the ubiquitinated proteins from PM-enriched fractions was unchanged (Fig. 3, A and C), whereas its abundance in the ubiquitinated proteins from ICV-enriched fractions was significantly increased at D5d-3h and the increased level was maintained up to 24-h dDAVP withdrawal (Fig. 3, B and D). Moreover, in protocol 2, the AQP2 prevalence among the ubiquitinated proteins from PM-enriched fractions was unchanged (Fig. 3, E and G), whereas it was markedly increased in the ICV-enriched fractions at D5d and D5d-180 min (Fig. 3, F and H). This finding indicates that AQP2 per se, significantly retrieved into the ICV during dDAVP withdrawal (Figs. 1 and 2), could be highly ubiquitinated during withdrawal of dDAVP stimulation and/or was highly associated with other ubiquitinated proteins. Again, this observation also led us to select the kidney samples of D5d-3h, in addition to the samples from vehicle-treated control and D5d, for identification of the E3s associated with proteins involved in vasopressin-induced urine concentration.

Fig. 3.

Semiquantitative immunoblotting of AQP2 in the immunoisolated ubiquitinated proteins from PM- or ICV-enriched fractions of rat WK. A and B: changes in AQP2 levels in response to vehicle treatment (control), D5d, or D5d-3h, D5d-6h, D5d-12h, and D5d-24h (protocol 1). C and D: changes in AQP2 levels in response to vehicle treatment (control), D5d, or D5d-15m, D5d-30m, D5d-60m, and D5d-180m (protocol 2). *P < 0.05 compared with control. #P < 0.05 compared with D5d; n, no. of rats.

Identification of E3s by transcriptome analysis in rat kidney inner medulla.

Kidney inner medullas collected from control (n = 2), D5d (n = 2), and D5d-3h (n = 2) groups were subjected to transcriptome analysis (Affymetrix GeneChip Rat Gene 1.0 ST array containing 27,342 gene-level probe sets) to profile the gene expression of E3s at the mRNA level in rat kidney (GEO accession no. GSE27449). Eighty-six genes of E3s were identified in rat kidney inner medullas (Table 1). The mRNA signals relative to the levels of vehicle-treated control are demonstrated (Table 1).

Identification of E3s by LC-MS/MS analysis in AQP2- or Ub-immunoisolated PM- and ICV-enriched fractions.

LC-MS/MS analysis was carried out in AQP2- or Ub-immunoisolated PM- and ICV-enriched fractions of rat whole kidney (Tables 2 and 3). Twelve E3s (BRCA1, BRE1B, MIB2, UBR4, UBR5, UHRF1, NEDD4, CUL5, DDB1, FBX6, FBX43, and SOCS2) in the AQP2-immunoisolated PM fractions and four E3s (BRCA1, BRE1B, MIB2, and UBR4) in the AQP2-immunoisolated ICV fractions were identified (Table 2). Moreover, in the Ub-immunoisolated fractions, 13 E3s (BIRC4, BRCA1, BRE1B, CBL-B, MARCH10, UBR4, UBR5, UHRF1, NEDD4, CUL5, DDB1, FBX6, and FBX43) in the PM fractions and 7 E3s (BRCA1, BRE1B, CBL-B, UBR4, UHRF1, CUL5, and DDB1) in the ICV fractions were identified (Table 3). Identified proteins were quantitated by label-free quantification based on the peak intensity of each peptide (Tables 2 and 3).

Expression of E3s (BRE1B, NEDD4, and CUL5) in rat kidney.

Among the E3s identified by LC-MS/MS analysis, seven E3s (BRE1B, BRCA1, UHRF1, UBR4, NEDD4, CUL5, and FBX6) were also identified by transcriptome analysis. RT-PCR of rat whole kidney and IM revealed the mRNA expression of these seven E3s (Fig. 4). Among them, BRE1B, NEDD4, and CUL5 were selected for immunoblotting analysis, since 1) label-free quantification revealed that NEDD4 and CUL5 levels were the most highly increased in response to 5-day dDAVP infusion compared with other identified E3s (Table 2, AQP2-immunoisolated PM-enriched fractions of rat whole kidney); and 2) antibodies for these three E3s were available for immunoblot analysis.

Fig. 4.

RT-PCR analysis of 7 E3s (BRE1B, BRCA1, UHRF1, UBR4, NEDD4, CUL5, and FBX6; see definitions below) in rat WK and inner medulla (IM). Products of the predicted sizes were generated by RT-PCR of BRE1B (175 bp), BRCA1 (212 bp), UHRF1 (407 bp), UBR4 (220 bp), NEDD4 (149 bp), CUL5 (140 bp), FBX6 (183 bp), and β-actin (101 bp) in WK and IM. Negative controls included the omission of reverse transcriptase (− RT; −; RT negative). +, RT positive; M, molecular weight marker; BRCA1, breast cancer type 1 susceptibility protein; BRE1B, ring-finger protein 40; CUL5, cullin-5; FBX6, F-box only protein 6; NEDD4, neural precursor cell expressed, developmentally downregulated 4; UBR4, ubiquitin protein ligase E3 component n-recognin 4; UHRF1, ubiquitin-like PHD and ring-finger domains 1.

Protein expression was examined in the kidney cortex, outer medulla, and inner medulla. AQP2 (nonglycosylated 29 kDa and glycosylated ∼35–50 kDa) and NEDD4 (∼115 kDa) expression were the most abundant in the inner medulla, whereas BRE1B (∼105 kDa) and CUL5 (∼80 kDa) were abundantly expressed in all the kidney zones (Fig. 5A). Moreover, protein expression was examined in the whole kidney homogenates, PM-enriched (17,000-g pellet), ICV-enriched (200,000-g pellet), or cytosolic fractions (200,000-g supernatant) (Fig. 5B). AQP2 expression was abundantly seen in both PM and ICV fractions of whole kidney, whereas no expression was seen in the supernatant of 200,000-g spin (cytosolic fractions) of whole kidney. BRE1B expression (indicated an arrow in Fig. 5B) was similar to that of AQP2, while NEDD4 and CUL5 expression was more abundantly expressed in the cytosolic fractions of whole kidney than PM-enriched or ICV-enriched fractions (Fig. 5B).

Fig. 5.

Immunoblotting of AQP2, BRE1B, CUL5, and NEDD4. A: protein expression of AQP2, BRE1B, CUL5, and NEDD4 in the cortex, outer medulla (OM), and IM of rat kidney. Six micrograms of protein was loaded onto each lane. B: protein expression of AQP2, BRE1B, CUL5, and NEDD4 in the WK, PM- and ICV-enriched fractions, and cytosolic fractions (SUP). Six micrograms of protein was loaded onto each lane. An arrow indicates the ∼105-kDa band of BRE1B expression.

Altered protein abundance of BRE1B, NEDD4, and CUL5 in vivo in response to 5-day dDAVP infusion, dDAVP withdrawal after 5-day dDAVP infusion, or in Li-NDI.

AQP2 abundance in rat whole kidney was significantly increased at D5d compared with control, and it was decreased at D5d-6h compared with D5d (Fig. 6D). In contrast, protein abundance of NEDD4 and CUL5 at D5d-6h was markedly increased compared with D5d (Fig. 6, A and C), whereas BRE1B abundance was unchanged (Fig. 6B). Immunoperoxidase microscopy also demonstrated that NEDD4 labeling intensity in the cytoplasm of the collecting duct cells increased during dDAVP withdrawal after long-term infusion (Fig. 6, EH). Moreover, in rats with Li-NDI where AQP2 abundance was markedly decreased, NEDD4 and BRE1B levels were significantly increased, whereas CUL5 abundance was unchanged (Fig. 7, AE).

Fig. 6.

Semiquantitative immunoblotting of NEDD4 (A), BRE1B (B), CUL5 (C), and AQP2 (D) in rat WK of vehicle-treated (control), D5d, D5d-3h, or D5d-6h rats. *P < 0.05 compared with control. #P < 0.05 compared with D5d. ^P < 0.05 compared with D5d-3h. EH: immunoperoxidase microscopy of NEDD4 in the IM of vehicle-treated (control; E), D5d (F), D5d-3h (G), or D5d-6h (H) rats. CD, collecting duct; n, no. of rats.

Fig. 7.

Semiquantitative immunoblotting of AQP2 (A and B), NEDD4 (A and C), BRE1B (A and D), and CUL5 (A and E) in the WK of rats with lithium (Li)-induced nephrogenic diabetes insipidus (NDI). Low-Li, low-dose Li group; High-Li, high-dose Li group. *P < 0.05 compared with control; n, no. of rats.

Reduced rate of AQP2 degradation in mpkCCDc14 cells with NEDD4 or CUL5 knockdown.

To address the potential role of the markedly upregulated NEDD4 and CUL5 expression observed at D5d-6h (Fig. 6), we examined the effects of siRNA-mediated gene silencing of either NEDD4 or CUL5 on the rate of AQP2 degradation in mpkCCDc14 cells. The cells were cultured on a semipermeable filter of the Transwell system in the presence of dDAVP stimulation. Protein synthesis was blocked at the translational level with cycloheximide, and protein samples prepared at various time points were subjected to AQP2 immunoblotting. The rate of AQP2 degradation in mpkCCDc14 cells with NEDD4 knockdown (n = 8) was significantly decreased from that of control siRNA-transfected mpkCCDc14 cells (n = 8, Fig. 8) at 2 and 8 h after cycloheximide treatment. Moreover, the rate of AQP2 degradation was also retarded in mpkCCDc14 cells with CUL5 knockdown (n = 8) compared with control siRNA-transfected mpkCCDc14 cells (n = 8) at 2 h after cycloheximide treatment (Fig. 8).

Fig. 8.

AQP2 degradation rate in mpkCCDc14 cells transfected with CUL5 (Cul5-siRNA), NEDD4 (Nedd4-siRNA), or nontargeting siRNA (Con-siRNA). siRNA, small interfering RNA. AC: protein synthesis was blocked at the translational level with cycloheximide, and protein samples prepared at various time points (0, 2, 4, and 8 h after cycloheximide treatment) were subjected to AQP2 immunoblotting. D: time course of changes in AQP2 expression. The rate of AQP2 degradation in mpkCCDc14 cells with NEDD4 knockdown was significantly decreased compared with the control siRNA-transfected mpkCCDc14 cells at 2 and 8 h after cycloheximide treatment. The rate of AQP2 degradation was also retarded in mpkCCDc14 cells with CUL5 knockdown at 2 h after cycloheximide treatment; 1.5 μg of protein was loaded onto each lane. *P < 0.05 compared with control; n, no. of cell lysate preparations from 4 independent experiments.

DISCUSSION

We demonstrated that dDAVP-induced AQP2 in the PM fractions was retrieved into the ICV fractions when dDAVP stimulation was removed. Three-hour dDAVP withdrawal after 5-day dDAVP infusion greatly enhanced AQP2 endocytosis. Moreover, the prevalence of AQP2 among the ubiquitinated proteins from ICV fractions was significantly increased during dDAVP withdrawal. This suggests that internalized AQP2 per se is highly ubiquitinated or internalized AQP2 interacts with other ubiquitinated proteins during dDAVP withdrawal. A previous study revealed that AQP2 is ubiquitinated with one UbLys63-linked poly-Ub chain at K270 of AQP2, and lysosomal degradation was extensive for ubiquitinated AQP2 (20). It is well known that UbLys48-linked poly-Ub chains represent a signal for proteasomal degradation of substrate proteins (15). In contrast, UbLys63 has a more extended conformation compared with UbLys48, and thus it likely has distinct targets and functions in the cell (13). Consistent with this, either MG132 (a specific proteasome inhibitor) or chloroquine (a blocker of the lysosomal pathway of protein degradation) treatment in primary cultured IMCD cells significantly reduced AQP2 degradation (Fig. 9), indicating that ubiquitination and subsequent proteosomal and/or lysosomal degradation of AQP2 could play a critical role in regulation of AQP2 abundance.

Fig. 9.

Effects of MG-132 or chloroquine treatment in primary cultured inner medullary collecting duct (IMCD) cells on AQP2 degradation. IMCD cells of rat kidney were primary cultured in rat fibronectin-coated filter plates of the Transwell system. Cycloheximide (10−4 M), MG-132 (10−3 M), or chloroquine (10−3 M) were treated for 8 h. Either MG132 (a specific proteasome inhibitor) or chloroquine (a blocker of the lysosomal pathway of protein degradation) treatment in primary cultured IMCD cells significantly reduced AQP2 degradation.

To identify the profile of E3s genes and proteins in rat kidney which could be involved in the intracellular degradation of proteins associated with vasopressin-induced urine concentration, transcriptome analysis and LC-MS/MS analysis were performed. In particular, proteomic analysis was done using kidney samples at D5d-3h where AQP2 was significantly internalized to the ICV, in addition to the samples from vehicle-treated control and D5d where AQP2 was abundantly expressed at the apical plasma membrane domain. Among the identified E3s, BRE1B (single ring-finger type), NEDD4 (HECT type), and CUL5 (multisubunit ring-finger type) were selected for further analysis, since label-free quantification on LC-MS/MS analysis revealed that NEDD4 and CUL5 levels were the most highly increased in response to 5-day dDAVP stimulation. Both in vivo and in vitro results suggest that the selected three E3s, BRE1B, NEDD4, and CUL5, could play a potential role in urine concentration: e.g., 1) semiquantitative immunoblotting revealed an increase in NEDD4 and CUL5 during dDAVP withdrawal after long-term stimulation or an increase in NEDD4 and BRE1B in rats with Li-NDI where AQP2 abundance was significantly decreased; and 2) siRNA-mediated gene silencing of NEDD4 or CUL5 in mpkCCDc14 cells significantly decreased the rate of AQP2 degradation.

The NEDD4/Rsp5p family belongs to the HECT-domain superfamily of E3s (14, 26). NEDD4 proteins are composed of an N-terminal C2 domain, 2-4 WW domains that bind PY motifs, and a C-terminal HECT domain (22, 23, 43, 50). Mammals possess nine NEDD4 members, with NEDD4 (Nedd4-1) and NEDD4L (Nedd4-2) being most closely related to each other (21). We observed a significant increase in NEDD4 abundance along with decreased AQP2 abundance in response to the dDAVP withdrawal after long-term dDAVP stimulation and Li-NDI, suggesting its involvement in the intracellular degradation of proteins induced by long-term vasopressin stimulation for urine concentration. Consistent with this, the observed significantly decreased degradation rate of AQP2 in mpkCCDc14 cells with NEDD4 knockdown suggests that NEDD4 is likely to play a direct role in regulating AQP2 abundance by ubiquitination or it regulates other substrate proteins importantly involved in the AQP2 regulation. For example, Persaud et al. (43) demonstrated Akt1, Akt3, MAP2, and PKC are the substrates of rat NEDD4, all of which contain the PY motif, a binding site of WW domain of NEDD4, and are known to regulate AQP2 expression (24, 45), whereas rat AQP2 per se does not contain PY motifs (LPxY or PPxY). Interestingly, a recent study using Nedd4-1 gene knockout mice suggested that Nedd4-1 positively regulates cell surface expression of insulin-like growth factor (IGF-1) (3). Thus the role of NEDD4 in the regulation of cell surface AQP2 expression is required to be examined in the future study using NEDD4-deficient mice. In addition, Nedd4-2 gene knockout mice confirmed that Nedd4-2 is the critical regulator of ENaC activity/expression and blood pressure (48, 50).

CUL5, a vasopressin-activated calcium-mobilizing (VACM-1) receptor (2), is a member of the cullin gene family of scaffold proteins of the E3 complex (47). CUL5 appears to be an important E3 when cells are exposed to sustained physical stress, as demonstrated by the upregulated CUL5 mRNA level in rat cerebral cortex in response to water deprivation (4). Consistent with this, in the present study CUL5 protein abundance was increased in the kidney when dDAVP stimulation was removed after 5-day dDAVP stimulation. This finding is consistent with a previous study showing an increased renal CUL5 mRNA expression during a rehydration period after 48-h water deprivation (4). Thus the observed upregulation of CUL5 during dDAVP withdrawal could play a role in the intracellular degradation of proteins induced by long-term vasopressin stimulation. Moreover, reduced degradation of AQP2 was also seen in the mpkCCDc14 cells with CUL5 knockdown, although it was seen only at one data point (Fig. 8). This suggests that CUL5 may play a role in the attachment of Ub to AQP2, resulting in an internalization of AQP2 and reduction of AQP2 abundance, presumably via lysosomal and/or proteosomal degradation. However, we could not perform a direct ubiquitination assay for the confirmation, since the components of the whole CUL5 E3 complex attaching Ub to AQP2 have not yet been identified, especially the parts for recognizing AQP2. Moreover, biotinylation of cell surface AQP2 in the collecting duct cells of CUL5-deficient mice is required to examine the changes in AQP2 endocytosis in future studies. In addition, CUL5 could also affect AQP2 stability indirectly through the ubiquitination of unknown substrate proteins which are importantly involved in AQP2 regulation and urine concentration. For example, CUL5 is known to inhibit adenylyl cyclase activity through regulating inositol 1,4,5-trisphosphate abundance in stably CUL5-transfected cells (1).

BRE1B, a single ring-finger type E3, is known to mediate the degradation of syntaxin 1, which is dominantly expressed in the brain for the neurotransmitter release machinery (5). We observed a significant increase in BRE1B abundance in Li-NDI where AQP2 abundance and trafficking are markedly decreased (27), whereas its abundance was unchanged in response to dDAVP stimulation and withdrawal. In the kidney, syntaxin-1, -2, -3, and -4 are expressed (35) and in particular, syntaxin-3, syntaxin-4, and SNAP23 are involved in the AQP2-vesicle fusion into the apical membrane (34, 46). In contrast, syntaxin-1A is known to play a role in the targeted exocytosis of the H+-ATPase to the apical plasma membrane of intercalated cells (32). Thus the substrate proteins of BRE1B in the principal cells are unclear and need to be identified. Previously, we demonstrated that chronic Li treatment is associated with a major cellular reorganization of the collecting duct cells with a marked increase in the fraction of intercalated cells and a significant decrease in the fraction of principal cells (6). Moreover, H+-ATPase abundance was markedly increased (25), suggesting that upregulated BRE1B might be involved in the regulation of syntaxin-1 abundance, hence resulting in the reduction of H+-ATPase exocytosis in the intercalated cells of Li-NDI (25).

In summary, we identified a number of endogenously expressed E3s in the kidney. The significant increase in the abundance of selected E3s in response to dDAVP stimulation/withdrawal or in Li-NDI may suggest that they are likely to play a role in the ubiquitination of substrate proteins induced by vasopression stimulation for urine concentration. Further studies, including cross-linking assay, are needed to examine the specificity of the identified E3s to the unknown substrate proteins. Detailed knowledge of the molecular mechanisms for intracellular degradation of proteins induced by urinary concentration could provide new therapeutic targets for water-balance disorders in patients.

GRANTS

This study was supported by National Research Foundation grant (2010-0008225 and 2010-0019393) funded by the Ministry of Education, Science and Technology (MEST), Korea, and the Korea Healthcare Technology R&D Project, Ministry of Health and Welfare, Korea (A100617). The Water and Salt Research Centre at the University of Aarhus is established and supported by the Danish National Research Foundation (Danmarks Grundforskningsfond).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

REFERENCES

  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.
  40. 40.
  41. 41.
  42. 42.
  43. 43.
  44. 44.
  45. 45.
  46. 46.
  47. 47.
  48. 48.
  49. 49.
  50. 50.
  51. 51.
  52. 52.
  53. 53.
  54. 54.
View Abstract