Calcium-sensing Receptor Ubiquitination and Degradation Mediated by the E3 Ubiquitin Ligase Dorfin*

Calcium-sensing receptors (CaR) contribute to regulation of systemic calcium homeostasis by activation of Gq- and Gi-linked signaling pathways in the parathyroids, kidney, and intestine. Little is known about the mechanisms regulating CaR synthesis and degradation. Screening of a human kidney yeast two-hybrid library identified the E3 ubiquitin ligase dorfin as a binding partner for the intracellular carboxyl terminus of CaR. Interaction between CaR and dorfin was confirmed by coimmunoprecipitation from HEK293 cells. Ubiquitination of CaR was observed in the presence of the proteasomal inhibitor MG132; mutation of all putative intracellular loop and carboxyl-terminal lysine residues abolished ubiquitination of CaR. Coexpression with dorfin decreased the amount of total CaR protein and increased CaR ubiquitination, whereas a dominant negative fragment of dorfin had opposite effects. The AAA-ATPase p97/valosin-containing protein associates with both CaR and dorfin in HEK293 cells. Treatment with tunicamycin, an inhibitor of N-linked glycosylation, induced the appearance of the unglycosylated 115-kDa CaR form, which was further increased by exposure to MG132, or upon transfection with a dorfin dominant negative construct, suggesting that dorfin-mediated proteasomal degradation of immature CaR occurs from the endoplasmic reticulum. Because endogenous CaR in Madin-Darby canine kidney cells is also subject to degradation from the endoplasmic reticulum, dorfin-mediated ubiquitination may contribute to a general mechanism for CaR quality control during biosynthesis.

The calcium-sensing receptor (CaR) 2 contributes to maintenance of systemic Ca 2ϩ homeostasis, regulating parathyroid hormone secretion and absorption/resorption of Ca 2ϩ by the intestine and kidney and may also have effects on bone (1). CaR belongs to family C of the GPCR superfamily, having structural similarities to metabotropic glutamate receptors, ␥-aminobutyric acid, type B receptors, and some putative pheromone/taste receptors (1,2). Common to all members of family C is a large extracellular domain of more than 600 amino acids containing the agonist binding site, a heptahelical transmembrane domain, and a large intracellular carboxyl-terminal tail of more than 200 amino acids (1,2). Members of Family C, including CaR, function as dimers, stabilized either by a disulfide bond or non-covalent interactions (3)(4)(5). The CaR extracellular domain contains N-linked glycosylation sites (2,6,7) and is stabilized by multiple intramolecular disulfide bonds, as well as one intermolecular disulfide bond between monomers in the dimer (8 -10). Upon agonist stimulation, CaR activates diverse signaling pathways leading to changes in hormone secretion, cell proliferation, differentiation, and/or apoptosis (1). Although considerable progress has been made in understanding the structure, activation, and signaling of CaR, the biosynthesis, trafficking, targeting, and turnover mechanisms regulating CaR remain largely unexplored.
To identify novel proteins that might regulate trafficking and/or targeting of CaR, the intracellular carboxyl terminus of CaR was used as bait in a yeast two-hybrid (Y2H) screen of a human kidney cDNA library. One of the proteins identified in the screen was the E3 ubiquitin ligase dorfin (double-RING finger protein) (11). Dorfin was originally cloned from human spinal cord and is expressed in many organs, including kidney, liver, intestine, and the central nervous system (11). Dorfin localizes to a region near the centrosome in an aggresome-like structure in cultured cells (11). In the nervous system, dorfin ubiquitinates superoxide dismutase-1 (12) and synphilin-1 (13) and is a component of Lewy bodies observed in Parkinson's and other neurodegenerative diseases (12)(13)(14). Ubiquitination results in the attachment of ubiquitin, a highly conserved 76-amino acid polypeptide, to the ⑀-amino group of lysine residues of target proteins, and requires the sequential actions of three enzymes. Final transfer of activated ubiquitin to target proteins is coordinated by the E3 ligase, which specifically interacts with both E2-ubiquitin and the target protein (15). Dorfin contains two RING domains at its amino terminus, which function as recruiting motifs for specific E2s (11). The carboxyl terminus of dorfin has no identifiable motifs but has been shown to confer specificity of binding to synphilin-1 (13). Dorfin interacts directly with VCP (valosin-containing protein, also called p97 or Cdc48 homologue) (14), an AAA-ATPase proposed to have a role in endoplasmic reticulum-associated protein degradation (ERAD). VCP assists in translocation of ubiquitinated proteins from the ER and acts as a chaperone, targeting ubiquitinated proteins to the proteasome for degradation (16 -19).
In this report, we characterize the functional interactions between CaR and dorfin in HEK293 cells and demonstrate that dorfin mediates CaR ubiquitination, leading to degradation by the proteasome. Both dorfin and CaR interact with VCP in HEK293 cells, and a dominant negative fragment of dorfin protects immature forms of CaR from degradation. Finally, endogenous CaR in MDCK cells is also subject to ER-associated degradation, suggesting a common mechanism may regulate quality control of both exogenously and endogenously expressed

EXPERIMENTAL PROCEDURES
Materials-Human kidney cDNA library and all the materials for the Y2H screen were purchased from Clontech. HEK293 and MDCK cells were from the American Tissue Culture Collection and used through laboratory passage number 30. Restriction enzymes were from New England Biolabs and Promega. The EGFP-dorfin plasmid (dorfin chimera with EGFP at the amino terminus) and rabbit polyclonal antibody against dorfin (D-30) were generated as described previously (11). The HA-ubiquitin plasmid was generously provided by Dr. Richard J. H. Wojcikiewicz (SUNY Upstate Medical University, Syracuse, NY). Monoclonal antibodies were from various sources (anti-FLAG M2 and anti-actin antibodies, Sigma; anti-HA antibody, Roche Diagnostics; anti-VCP antibody, Research Diagnostics; and anti-GFP antibody, Molecular Probes). Rabbit polyclonal antibody against CaR (LRG) was generated as described (20). Anti-phospho-ERK1/2 (p42/44) polyclonal antibody was from Cell Signaling Technology. ECL anti-mouse and anti-rabbit, horseradish peroxidase-conjugated secondary antibodies were purchased from Amersham Biosciences. MG132 and tunicamycin were purchased from Sigma.
Plasmid Construction-CaR with an amino-terminal FLAG epitope (FLAG-CaR) was generated as described (21). CaR carboxyl-terminal (CaR-CT) fragments in the Y2H bait vector pGBKT7 were prepared by PCR using primers containing NdeI and SalI sites, using FLAG-CaR as template. PCR products were digested with NdeI/SalI and subcloned into NdeI/SalI-digested pGBKT7. Dorfin carboxyl-terminal fragments in the Y2H prey vector pACT2 were prepared similarly, using NcoI and XhoI sites. The dominant negative construct of dorfin was generated as a chimera with EGFP at the carboxyl terminus (DCT-EGFP). The DCT fragment containing dorfin residues 561-838 was prepared by PCR using primers containing BglII and SalI sites. The PCR product was digested with BglII/SalI and subcloned into BglII/SalI-digested PEGFP-N1 vector (Clontech). Another version of dominant negative dorfin (DCT-c-myc) was prepared similarly by PCR the DCT fragment using primers containing XbaI and HindIII sites. The PCR product was digested with XbaI/ HindIII and subcloned into XbaI/HindIII-digested pcDNA3.1A(Ϫ) vector (Clontech). Another construct of dorfin (DNT-EGFP) containing the N-terminal RING-finger domains of dorfin from residue 1-367 was prepared by PCR using primers containing SacI and KpnI sites. The PCR product was digested with SacI/KpnI and subcloned into SacI/ KpnI-digested pEGFP-N1 vector (Clontech). Point mutations were generated by a modified inverse PCR mutagenesis method (22). All PCR reactions used Pfu DNA polymerase (Stratagene). All constructs were verified by dideoxy-DNA sequencing (DNA Sequencing Facility, Cornell University, Ithaca, NY). Sequences of primers will be provided upon request.
Y2H Assay Screening-Gal4-based Y2H library screening was performed by yeast mating as recommended in the manufacturer's instructions (Clontech). Yeast Saccharomyces cerevisiae MATa strain AH109 was transformed with bait plasmid containing CaR-CT (amino acids 866 -1078) and incubated with yeast S. cerevisiae MAT␣ strain Y187 pretransformed with human kidney cDNA library (Clontech) in 2xYPDA/Kan at 30°C for 24 h. The mixture was plated on SD/-Ade/ -His/-Leu/-Trp plates to screen for ADE2 ϩ and HIS3 ϩ clones. Plates were incubated at 30°C until colonies appeared. Colonies were restreaked on SD/-Ade/-His/-Leu/-Trp plus X-␣-gal plates to screen for MEL1 ϩ clones. Clones that activated three reporter genes, ADE2, HIS3, and MEL1, were considered positive and identified by purifying plasmids and sequencing inserts.
Directed Y2H studies were performed by cotransformation using the lithium acetate method (23). Bait plasmids containing CaR-CT fragments and prey plasmids containing dorfin carboxyl-terminal fragments were cotransformed into AH109 and plated on SD/-Ade/-His/ -Leu/-Trp plus X-␣-gal plates. An interaction was considered positive when three reporter genes (ADE2, HIS3, and MEL1) were activated.
Cell Culture and Transfection-HEK293 and MDCK cells were grown in high glucose Dulbecco's modified Eagle's medium, supplemented with 10% heat-inactivated fetal calf serum, 50 units/ml penicillin, and 50 g/ml streptomycin (37°C, 5% CO 2 ). HEK293 cells were transiently transfected with Novafector (Venn Nova LLC, Pompano, FL) according to the manufacturer's instructions; experiments were done 72 h after transfection. For inhibitor studies, cells were treated with MG132 and/or tunicamycin (solubilized in Me 2 SO) for 12 h prior to lysis. Comparable levels of Me 2 SO had no effect on measured parameters.
Immunoprecipitation-Transfected HEK293 cells were washed with PBS and lysed as described. After sonication on ice, samples were agitated for 30 min at 4°C and incubated with 10 l of protein G-agarose (Invitrogen) to minimize nonspecific binding. Samples were centrifuged, and supernatants were incubated with antibody for 3 h at 4°C. 15 l of protein G-agarose was then added, and samples were rotated overnight (4°C). Precipitates were incubated in loading buffer for 30 min at 25°C and run on SDS-polyacrylamide gels as described. To detect specific ubiquitination, a two-step immunoprecipitation was performed. After the first immunoprecipitation as described, pellets were washed three times with lysis buffer and incubated with 1% SDS/PBS for 30 min (25°C) to disrupt non-covalent interactions. Samples were centrifuged, and supernatants were diluted with lysis buffer, followed by a second immunoprecipitation. The resultant precipitates were incubated in loading buffer and separated on SDS-polyacrylamide gels as described.
Densitometry and Statistics-Blots were digitized using an Epson Expression 800 Photo scanner and quantified by densitometry using AlphaEaseFC StandAlone Software (San Leandro, CA). Results were mean Ϯ S.D. of at least three independent experiments; graphs were generated using Sigma Plot 2000. Student's t test (two comparisons) was performed, and p Ͻ 0.05 was considered significant.

Y2H Screen Identifies the E3 Ubiquitin Ligase Dorfin as a CaR-binding
Protein-The cytoplasmic carboxyl terminus of CaR (CT, residues 866 -1078) was used as bait to screen a human kidney cDNA library (Fig. 1A). Positive clones were selected by activation of three reporter genes, ADE2, HIS3, and MEL1. Blast searches of the NCBI data base with positive clones identified in the screen yielded a fragment corresponding to the carboxyl terminus (residues 561-838) of the E3 ubiquitin ligase dorfin, an 838-amino acid protein ( Fig. 1B) (11).
To localize the dorfin interaction site on the CaR carboxyl terminus, various truncations of the CaR carboxyl terminus (in pGBKT7), illustrated in Fig. 1A, were screened against the carboxyl terminus of dorfin (residues 561-838, in pACT2) after cotransformation in the AH109 strain. The full CaR-CT (residues 866 -1078) interacted with dorfin, confirming the results of the initial screen. Truncations of the carboxyl end of CaR-CT were well tolerated, displaying positive interactions with dorfin. Only the smallest fragment, containing residues 866 -886, did not interact with dorfin. Truncations from the amino terminus of the CaR-CT narrowed the region for interaction with dorfin to residues 880 -900 (Fig. 1A).
Complementary directed Y2H screens were performed to localize the CaR-CT interaction site on the carboxyl terminus of dorfin (residues 561-838). Any truncations from the carboxyl terminus of dorfin inhibited interactions with CaR-CT, whereas interaction with CaR-CT was retained upon removal of up to 100 residues from the amino terminus of the dorfin fragment (residues 561-660). The minimal fragment of the dorfin carboxyl terminus required for interaction with CaR-CT contains residues from 660 -838 (Fig. 1B).
Coimmunoprecipitation of CaR and Dorfin from HEK293 Cells Confirms Their Interaction-Human CaR having an amino-terminal FLAG epitope (FLAG-CaR) was transfected into HEK293 cells, and FLAG-CaR was immunoprecipitated with anti-FLAG antibody. The immunoprecipitate was separated on a 4 -15% SDS-PAGE reducing gel, blotted to nitrocellulose, and probed with anti-dorfin antibody D-30. Endogenous dorfin was observed as a doublet with molecular mass of ϳ100 kDa in the presence (lane 1) but not in the absence (lane 2) of transfected FLAG-CaR ( Fig. 2A, top panel). The two lower panels in Fig. 2 illustrate To confirm that the dorfin fragment identified in the Y2H studies is required for the interaction of dorfin and CaR in mammalian cells, FLAG-CaR and full-length dorfin (EGFP-dorfin) or dorfin truncations (DNT-EGFP or DCT-EGFP) were tested for coimmunoprecipitation from HEK293 cells. DNT-EGFP contains the amino-terminal RINGfinger domains of dorfin from residues 1-367; DCT-EGFP contains the carboxyl-terminal domain of dorfin from residues 561-838. Anti-FLAG antibody was used to immunoprecipitate FLAG-CaR, and blots were probed with anti-GFP antibody to detect the dorfin species. EGFPdorfin (130 kDa) and DCT-EGFP (60 kDa) coprecipitated with FLAG-CaR (Fig. 2B, top panel), but DNT-EGFP (68 kDa) did not. The middle and bottom panels of Fig. 2B illustrate expression of the indicated constructs in cell lysates when probed with anti-GFP (middle) or anti-CaR LRG (bottom) antibodies. These results confirm that the domain of dorfin mediating the interaction with CaR in yeast, i.e. the carboxyl terminus, is also required for the interaction of dorfin with CaR in mammalian cells.
CaR Is Ubiquitinated-Interaction between CaR and dorfin suggests that CaR may be ubiquitinated. To test this possibility, FLAG-CaR and amino-terminal HA-tagged ubiquitin (HA-Ub) were cotransfected into HEK293 cells, and a two-step, denaturing protocol was used to immunoprecipitate FLAG-CaR. Briefly, anti-FLAG antibody was used to immunoprecipitate CaR, followed by treatment of the pellet with 1% SDS/PBS to disrupt non-covalent interactions between CaR and its associated proteins. The supernatant was diluted with lysis buffer and subjected to a second round of immunoprecipitation with anti-FLAG antibody. The Western blot was probed with anti-HA antibody to detect ubiquitinated species. Ubiquitination of CaR was observed in the presence of the proteasomal inhibitor MG132, appearing in the range from 150 to Ͼ250 kDa, while ubiquitination was barely detectable in the absence of MG132 (Fig. 3A, upper panel, lanes 1 and 2). MG132 increased the amount of CaR protein (Fig. 3A, lower panel, lanes 1 and  2), suggesting that ubiquitination followed by proteasomal degradation contributes to regulation of CaR.
Ubiquitin ligases covalently conjugate ubiquitin to lysine residues of target proteins. To confirm that CaR is ubiquitinated, we mutated intracellular lysine residues to arginine. There are nine lysine residues within the CaR carboxyl terminus (residues 863, 882, 897, 917, 931, 963, 965, 984, and 1002) and seven lysine residues within the three intracellular loops (residues 636, 644, 709, 717, 793, 796, and 805). When each lysine residue was mutated to arginine individually, the resulting mutants were still heavily ubiquitinated (data not shown). Because single point mutations did not abolish ubiquitination, CaR must be ubiquitinated at more than one lysine residue. All 16 lysine residues were therefore mutated to arginine simultaneously, and ubiquitination of the mutant, termed FLAG-CaR(0K), determined in the absence or presence of MG132 (Fig. 3A, upper panel, lanes 3 and 4). FLAG-CaR(0K) was not  2 and 4) 10 M proteasomal inhibitor MG132 for 12 h prior to lysis. Cell lysates were immunoprecipitated (IP) with anti-FLAG antibody, and the precipitated pellets were treated with 1% SDS/PBS to disrupt non-covalent interactions. Supernatants were diluted with lysis buffer, followed by a second immunoprecipitation using anti-FLAG antibody. The blot was probed with anti-HA antibody (upper panel). The same blot was then stripped and probed with anti-CaR antibody LRG (lower panel). B, FLAG-CaR cDNA or FLAG-CaR(0K) cDNA was transfected into HEK293 cells. Cells were preincubated with 0.5 mM Ca 2ϩ overnight prior to exposure to either 0.5 or 5 mM Ca 2ϩ for 10 min (37°C), followed by immunoblotting (IB) of lysates with anti-phospho-ERK1/2 antibody. C, FLAG-CaR cDNA or FLAG-CaR(0K) cDNA was transfected into HEK293 cells. Anti-FLAG antibody was used to immunoprecipitate samples, and blots were probed with antidorfin D-30 antibody (top blot). Lysates were probed with anti-dorfin D-30 antibody (middle blot) or anti-CaR LRG antibody (bottom blot).

FIGURE 4. Regulation of CaR ubiquitination by dorfin.
A, dorfin mediates ubiquitination of CaR. FLAG-CaR (3 g) and HA-Ub (2 g) cDNAs were cotransfected into HEK293 cells, without (Ϫ) or with (ϩ) EGFP-dorfin cDNA (6 g). Cells were incubated with 10 M MG132 for 12 h prior to lysis. Cell lysates were immunoprecipitated (IP) with anti-FLAG antibody, and blots were probed with anti-HA antibody. CaR ubiquitination was quantified and normalized to basal ubiquitination in the absence of exogenous dorfin; significance at p Ͻ 0.05 (*). B, dominant negative dorfin fragment DCT inhibits ubiquitination of CaR. FLAG-CaR cDNA (3 g) and HA-ubiquitin cDNA (2 g) were cotransfected into HEK293 cells, without (Ϫ) or with (ϩ) DCT-EGFP cDNA (6 g). Methods were as described in A. IB, immunoblot. ubiquitinated to a significant extent even in the presence of MG132. In addition, the amount of FLAG-CaR(0K) was not significantly changed upon addition of MG132 (100% in the absence versus 107.6 Ϯ 10.7% in the presence of MG132) (Fig. 3A, lower panel, lanes 3 and 4), whereas wild-type FLAG-CaR was sensitive to MG132 treatment (100% in the absence versus 148.4 Ϯ 4.7% in the presence of MG132) (Fig. 3A, lower  panel, lanes 1 and 2). Both FLAG-CaR and FLAG-CaR(0K) achieved mature glycosylation consistent with plasma membrane localization (Fig. 3A, lower panel) (6, 7) and had a comparable ability to stimulate ERK1/2 phosphorylation upon exposure of cells to 5 mM Ca 2ϩ (Fig. 3B). The absence of FLAG-CaR(0K) ubiquitination is not the result of an inability to associate with dorfin, because immunoprecipitation with anti-FLAG antibody of either FLAG-CaR or FLAG-CaR(0K) results in coprecipitation of comparable levels of endogenous dorfin (Fig. 3C, top  panel). Also illustrated in Fig. 3C is the presence of endogenous dorfin (middle panel) or transfected FLAG-CaR (bottom panel) in cell lysates. These results confirm that CaR is ubiquitinated at multiple lysine residues and degraded by the proteasome.
Dorfin Mediates CaR Ubiquitination-Dorfin is an E3 ubiquitin ligase and interacts with CaR and, therefore, likely mediates CaR ubiquitination. To test this possibility, FLAG-CaR and HA-Ub cDNAs were cotransfected into HEK293 cells, without or with EGFP-dorfin cDNA. Anti-FLAG antibody was used to immunoprecipitate CaR; Western blots were probed with anti-HA antibody to detect ubiquitinated species. In the absence of dorfin, a low level of CaR ubiquitination was detected; cotransfection with dorfin dramatically increased CaR ubiq-uitination (Fig. 4A). HEK293 cells express endogenous dorfin (data not shown, but see, e.g. Figs. 2A or 3C), and thus basal CaR ubiquitination (in the absence of cotransfected dorfin) might be catalyzed by endogenous dorfin. To test this possibility, we used DCT-EGFP as a dominant negative to interfere with ubiquitination mediated by endogenous dorfin. This construct cannot catalyze ubiquitination of substrates, because it does not contain the amino-terminal RING domains, which are essential for interaction with ubiquitin-conjugating enzymes (11). Cotransfection of HEK293 cells with FLAG-CaR, HA-Ub, and DCT-EGFP resulted in a reduction in CaR ubiquitination, compared with ubiquitination mediated by endogenous dorfin (-DCT) (Fig. 4B). These results suggest that the E3 ubiquitin ligase dorfin mediates ubiquitination of CaR.
Dorfin Regulates the Amount of CaR Protein in HEK293 Cells-Because dorfin mediates CaR ubiquitination, it must contribute to regulation of total cellular CaR protein. FLAG-CaR cDNA was transfected into HEK293 cells with increasing amounts of EGFP-dorfin cDNA (Fig.  5A). Total cDNA was kept constant with pcDNA3.1. The expression of CaR and dorfin were characterized by immunoblotting lysates from HEK293 cells with either anti-CaR LRG antibody (Fig. 5A, top blot) or anti-GFP antibody (Fig. 5A, bottom blot). Actin was used as a loading control (Fig. 5A, middle blot). When dorfin was increased, the amount of CaR decreased in a dose-dependent manner (Fig. 5A). The graph in Fig. 5A shows averaged results for three independent experiments and demonstrates a significant decrease in CaR protein as dorfin protein is increased. When CaR cDNA was transfected into HEK293 cells with increasing amounts of the dominant negative DCT-EGFP cDNA, CaR protein in cell lysates increased as a function of DCT protein (Fig. 5B, top blot). The graph illustrates averaged results from three independent experiments (Fig.  5B). To confirm that EGFP made no contributions to the observed responses, a DCT construct containing a carboxyl-terminal c-myc epitope was also generated. DCT-c-myc protected CaR against dorfin-mediated degradation in a manner comparable to DCT-EGFP (data not shown), indicating that the dorfin fragment specifically competed with endogenous dorfin to protect against CaR degradation. These results suggest that endogenous dorfin regulates CaR degradation.
To test if dorfin mediates degradation of CaR in a proteasome-dependent manner, the effect of the proteasomal inhibitor MG132 on dorfin-mediated CaR degradation was examined. When cells cotransfected with FLAG-CaR and EGFP-dorfin were treated with MG132, the dorfin-dependent decrease of CaR protein level was abrogated (Fig. 5C), suggesting that a proteasome-dependent mechanism underlies dorfinmediated degradation of CaR.
VCP Interacts with Both CaR and Dorfin-Dorfin interacts directly with VCP, an AAA-ATPase proposed to have a role in ERAD of proteins (14). In HEK293 cells transiently transfected with FLAG-CaR, an antibody against endogenous VCP immunoprecipitated both FLAG-CaR and endogenous dorfin (Fig. 6, lane 1). In the absence of FLAG-CaR, the anti-VCP antibody pulled down endogenous dorfin (Fig. 6, lane 2), as has previously been shown (14). Lower blots of Fig. 6 indicate endogenous expression of VCP and dorfin, as well as FLAG-CaR in lysates of transfected cells. Because both dorfin and CaR can interact with VCP, it is likely that dorfin-mediated ubiquitination and degradation of CaR is occurring at the ER via a VCP-facilitated ERAD pathway.
All Forms of CaR Are Degraded via the Proteasome-If CaR interacts with dorfin and VCP at the ER, it is possible that misfolded or unfolded, immature CaR is targeted for proteasomal degradation by dorfin. To test this possibility, we examined the effects of tunicamycin, an inhibitor of glucosaminyl 1-phosphate transferase, for the amounts and molecular weights of CaR in the absence or presence of MG132. Treatment with tunicamycin induced the appearance of unglycosylated CaR, with a molecular mass of 115 kDa (Fig. 7A). Two differentially glycosylated forms of CaR were also observed, i.e. bands at molecular masses of 130  Lysates were immunoprecipitated (IP) with anti-FLAG antibody and probed with anti-CaR LRG antibody. B, MG132 increases the amounts of immature forms of CaR. HEK293 cells were transfected with FLAG-CaR and HA-Ub cDNAs, and incubated with tunicamycin (5 g/ml) without or with MG132 (10 M), for 12 h prior to lysis. Lysates were immunoprecipitated with anti-FLAG antibody and probed with anti-CaR LRG antibody. Three forms of CaR are evident in tunicamycin: 1, fully glycosylated CaR (150 kDa); 2, ER-resident high mannose CaR (130 kDa); and 3, unglycosylated CaR (115 kDa). The amount of CaR protein in each form was quantified in the absence or presence of MG132, and average results for three independent experiments are illustrated (each band normalized to the amount observed in the absence of MG132), significance was at p Ͻ 0.05 (*). C, DCT increases the amounts of immature forms of CaR. HEK293 cells were transfected with FLAG-CaR without or with DCT-EGFP. Cells were incubated with tunicamycin (5 mg/ml) for 12 h prior to lysis. Experiments were as in A. Graph indicates amounts of three forms of CaR protein in the absence or presence of DCT-EGFP, normalized to the amount in the absence of DCT-EGFP. Significance was at p Ͻ 0.05 (*). D, MG132 increases the amounts of immature CaR in MDCK cells. MDCK cells were treated with tunicamycin (5 g/ml) without or with MG132 (10 M) for 12 h prior to lysis. Lysates were immunoblotted (IB) with anti-CaR LRG antibody. and 150 kDa (Fig. 7A), respectively. Previous studies have shown that the 130-kDa form of CaR is the ER-localized high mannose-modified receptor, and the 150-kDa CaR is the mature receptor at the plasma membrane (6,7). Addition of MG132 dramatically increased the amount of the 115-kDa form of CaR (185.3 Ϯ 29.6% normalized to the amount in the absence of MG132), whereas the two glycosylated forms of CaR were increased to a lesser extent (130-kDa form, 159.2 Ϯ 12.6%; 150-kDa form, 131 Ϯ 10.3%) (Fig. 7B). All forms of CaR are therefore sensitive to ubiquitination and degradation, although the immature forms of CaR (115 kDa), which localize to the ER, represent the strongest ubiquitination targets.
To determine whether dorfin-mediated ubiquitination is responsible for proteasomal degradation of the variously processed forms of CaR observed in the presence of tunicamycin, we cotransfected HEK293 cells with FLAG-CaR with or without the dorfin dominant negative fragment, DCT, treated cells overnight with tunicamycin, and quantified the abundance of CaR forms on Western blots (Fig. 7C). As illustrated in the blot and associated graph (average of three independent experiments), the presence of DCT had an effect on CaR abundance comparable to addition of MG132 (compare Fig. 7, B and C), suggesting that ubiquitination leading to proteasomal degradation of all forms of CaR is mediated by dorfin.
To determine whether endogenous CaR is subjected to proteasomedependent degradation in the ER, we tested whether MG132 could alter CaR protein levels in MDCK cells, which express endogenous CaR (24,25) and endogenous dorfin (data not shown). MG132 significantly increased the tunicamycin-induced immature 115-kDa form of CaR in MDCK cells (Fig. 7D), suggesting that endogenous CaR is regulated by ERAD, presumably via a dorfin-mediated pathway.

DISCUSSION
The molecular mechanisms underlying the trafficking, targeting, and turnover of CaR remain largely unknown. In this study we demonstrate that the E3 ubiquitin ligase dorfin, identified as a binding partner to the intracellular carboxyl terminus of CaR by Y2H screening of a human kidney library, interacts with CaR in HEK293 cells and regulates CaR abundance.
Dorfin contains two RING-finger domains and an in-between RINGfinger domain at its amino terminus, through which it interacts specifically with the ubiquitin-conjugating enzymes Ubc7 and Ubc8 (11). In common with other RING domain E3 ubiquitin ligases, the dorfin carboxyl terminus confers specificity for substrate proteins, including mutant superoxide dismutase-1 (12) and synphilin-1 (13). The region identified in the initial Y2H screen for CaR carboxyl-terminal binding partners corresponds to the distal carboxyl terminus of dorfin, from residues 561 to 838, further narrowed by directed Y2H screens to residues 660 -838, implicating the distal carboxyl terminus of dorfin in binding to a range of its target proteins. Dorfin also interacts with VCP through its carboxyl terminus (14). It remains to be determined whether specific interaction motifs can be identified within the dorfin carboxyl terminus, which may aid in identifying additional dorfin targets.
Dorfin has been suggested to have role in the pathology of a variety of neurodegenerative diseases. Dorfin localizes to Lewy bodies in Parkinson's disease and dementia (12,14,26) and in Lewy body-like inclusions in amyotrophic lateral sclerosis (13,14). CaR is expressed in tissues that contribute to maintenance of systemic Ca 2ϩ homeostasis, including the parathyroids, kidney, intestine, and bones (1), as well as other cell types where its one or more roles are less well defined (27). Our results therefore suggest that dorfin may have more general roles as an E3 ligase in a variety of cells and tissues. Our data support a model in which dorfin mediates ubiquitination and proteasomal degradation of CaR via the ERAD pathway, which suggests that dorfin plays a crucial role in regulating the post-translational level of CaR. It will be of interest to determine whether alterations in CaR protein levels in CaR-related diseases are mediated by dorfin, given that endogenous CaR in MDCK cells is subject to ERAD.
The interactions between CaR and dorfin led us to investigate CaR ubiquitination. Ubiquitinated CaR is not observed in the absence of MG132, suggesting that ubiquitinated CaR is rapidly de-ubiquitinated or degraded. MG132 increased the overall amount of CaR protein, an effect mimicked by a dominant negative fragment of dorfin, suggesting that dorfin-mediated ubiquitination and proteasomal degradation play a significant role in regulating CaR protein. Ubiquitinated CaR immunoreactivity is observed over a wide range of molecular masses, 150 to Ͼ250 kDa, typical of polyubiquitination, a potent signal for degradation (28,29). In addition, single point mutations of intracellular lysine residues to arginine did not significantly reduce CaR ubiquitination, suggesting that CaR is ubiquitinated at multiple lysine residues. Overall, our results demonstrate that CaR is multi/polyubiquitinated, and a target for proteasomal degradation.
Agonist-independent ubiquitination has been observed for ␦-opioid receptors (37,38), rhodopsin (39,40), thyrotropin-releasing hormone receptors (41), and as described in the current report, CaR. Agonistindependent ubiquitination results from ERAD, i.e. misfolded, unfolded, or abnormal proteins from the ER, are retrotranslocated into the cytoplasm, deglycosylated, and degraded by the proteasome (42). VCP is an AAA-ATPase that plays crucial roles in multiple aspects of the ERAD pathway in conjunction with its cofactors, Ufd1 and Ndl4 (16 -19, 42). Studies indicate that retrotranslocation of ERAD substrates from the ER and delivery to the proteasome is catalyzed by the VCP-Ufd1-Ndl4 complex, which binds first to polypeptide backbone and then the polyubiquitin chains on ERAD substrates. ATP hydrolysis catalyzed by VCP is required to complete retrotranslocation; VCP also chaperones polyubiquitinated proteins to the proteasome for degradation (42). Dorfin interacts with VCP and has been colocalized with VCP in perinuclear aggresomes, i.e. membrane-free, cytoplasmic inclusions containing misfolded, ubiquitinated proteins (14), strongly suggesting the involvement of dorfin in the ERAD pathway. Immunoprecipitation of VCP with both CaR and dorfin from HEK293 cells supports the notion that CaR and dorfin interact at the ER. ERAD-mediated degradation of CaR is further supported by the observation that the tunicamycin-stabilized, unglycosylated form of CaR is most sensitive to MG132 treatment. The ability of the dominant negative fragment DCT to increase the abundance of all CaR forms demonstrates that proteasomal degradation of CaR is initiated by dorfin-mediated ubiquitination.
CaR is a disulfide-linked dimer (1, 5, 8 -10). CaR monomers have a large extracellular domain (ECD) of Ͼ600 amino acids (1). The ECD harbors 11 potential N-linked glycosylation sites, 8 of which are utilized (1, 7). The CaR ECD also contains 19 cysteine residues; mutations at any of 14 cysteine residues abolish or dramatically reduce cell-surface expression and/or function (1,9). N-Linked glycosylation and disulfide bond formation occurs at the ER during membrane protein biosynthesis (43). Given the complexity of the structure of CaR, it is likely that some fraction of newly synthesized receptors are retained intracellularly for quality control purposes, as has been observed for both wt and mutant forms of CaR by Western blotting (5-10); only properly glycosylated, dimerized, and folded CaR are transported to Golgi complex and the plasma membrane (5)(6)(7)(8)(9)(10). Dorfin may represent a critical step in the quality control mechanism, ensuring that only properly folded CaR exits the ER.
In summary, dorfin regulates ubiquitination and degradation of CaR via a VCP-mediated ERAD pathway. The molecular mechanisms underlying differential sorting of CaR to either the Golgi complex or the ERAD pathway remain to be explored.