Site-specific polyubiquitination differentially regulates parathyroid hormone receptor–initiated MAPK signaling and cell proliferation

G protein–coupled receptor (GPCR) signaling and trafficking are essential for cellular function and regulated by phosphorylation, β-arrestin, and ubiquitination. The GPCR parathyroid hormone receptor (PTHR) exhibits time-dependent reversible ubiquitination. The exact ubiquitination sites in PTHR are unknown, but they extend upstream of its intracellular tail. Here, using tandem MS, we identified Lys388 in the third loop and Lys484 in the C-terminal tail as primary ubiquitination sites in PTHR. We found that PTHR ubiquitination requires β-arrestin and does not display a preference for β-arrestin1 or -2. PTH stimulated PTHR phosphorylation at Thr387/Thr392 and within the Ser489–Ser493 region. Such phosphorylation events may recruit β-arrestin, and we observed that chemically or genetically blocking PTHR phosphorylation inhibits its ubiquitination. Specifically, Ala replacement at Thr387/Thr392 suppressed β-arrestin binding and inhibited PTHR ubiquitination, suggesting that PTHR phosphorylation and ubiquitination are interdependent. Of note, Lys-deficient PTHR mutants promoted normal cAMP formation, but exhibited differential mitogen-activated protein kinase (MAPK) signaling. Lys-deficient PTHR triggered early onset and delayed ERK1/2 signaling compared with wildtype PTHR. Moreover, ubiquitination of Lys388 and Lys484 in wildtype PTHR strongly decreased p38 signaling, whereas Lys-deficient PTHR retained signaling comparable to unstimulated wildtype PTHR. Lys-deficient, ubiquitination-refractory PTHR reduced cell proliferation and increased apoptosis. However, elimination of all 11 Lys residues in PTHR did not affect its internalization and recycling. These results pinpoint the ubiquitinated Lys residues in PTHR controlling MAPK signaling and cell proliferation and survival. Our findings suggest new opportunities for targeting PTHR ubiquitination to regulate MAPK signaling or manage PTHR-related disorders.


Expression and signaling of wildtype and Lys-deficient PTHR
To identify the intracellular PTHR domains possessing candidate Lys residues that could be targeted for ubiquitination upon PTH stimulation, we generated four mutant receptor constructs by replacing cytosolic Lys (K) residues with Arg (R) as follows (Fig. 1A): 2K-PTHR (loop 2 Lys intact with all other intracellular Lys mutated to Arg); 3K-PTHR (loop 3 Lys protected; others replaced with Arg); CTK-PTHR (C-terminal Lys present; others converted to Arg); and 0K-PTHR (Lys-deficient construct), where all 11 Lys residues were replaced by Arg. To verify that these Lys mutant receptors were functional, the capacity of the activated receptor to promote G s -stimulated cAMP production was examined. WT-PTHR and the various Lys mutant constructs were transiently introduced into human embryonic kidney 293 (HEK-293) cells stably expressing the GloSensor cAMP reporter (22). As shown in Fig. 1B, all mutant constructs elicited cAMP formation comparable with WT-PTHR upon PTH stimulation. Interestingly, the 0K-PTHR displayed significantly increased signaling compared with WT-PTHR or partial Lys-deficient forms of the receptor. Such a finding is in keeping with previous results, pointing to an inhibitory domain within the intracellular PTHR tail that reduces cAMP formation, which upon truncation enhances cAMP accumulation (23,24).

Trafficking of Lys-deficient PTHR
Ubiquitination alters the trafficking of many GPCRs (25,26). To determine the functional significance of PTHR ubiquitination, we first applied confocal microscopy to assess the effect of ligand-stimulated ubiquitination on receptor internalization. HEK-293S GnTI Ϫ cells stably expressing TAP(HA)-WT-PTHR or TAP(HA)-0K-PTHR were labeled for 1 h at room temperature with an anti-HA primary antibody directly conjugated to DyLight 488, which recognizes the HA tag in the N-terminal domain of PTHR. WT-PTHR and 0K-PTHR rapidly internalized after a 5-min treatment with 100 nM PTH , as indicated by the disappearance of plasma membrane PTHR and the formation of intracellular fluorescent puncta (Fig. 1C). Residual, non-internalized PTHRs were detected by labeling with a goat anti-mouse IgG secondary antibody conjugated to Alexa 594 that recognizes the mouse IgG heavy and light chains of the primary antibody. The fraction of internalized receptor can then be quantified by comparing the intensity of Alexa 594 staining in control and PTH-treated cells (Fig. 1D). The results show comparable reduction of WT-PTHR (33%) and 0K-PTHR (25%) after PTH stimulation (Fig. 1D). Thus, PTHR ubiquitination is not required for activity-dependent receptor internalization.
Of note, HEK-293S GnTI Ϫ cells lack N-acetylglucosaminyltransferase I (GnTI), which is required for the processing of complex N-glycans. GnTI Ϫ cells are a convenient tool for overexpressing membrane proteins for biochemical and related analyses (27). PTHR possesses four N-glycosylation sites, and the absence of GnTI restricts PTHR glycosylation to a single Man 5 GlcNAc 2 . Thus, the absence of complex N-glycans might alter PTHR localization and function. Notably, the residual glycosylation is sufficient for PTHR function, as demonstrated here by normal receptor trafficking (Fig. 1C) and signaling, as reported previously (28 -30).
We complemented the imaging studies by measuring cell surface biotinylation to quantify the effect of ubiquitination on recycling of sequestered PTHR. The extent of recycled receptor is determined as the difference between internalized receptor and non-recycled receptor (Fig. 1E). The results indicated that over 50% of internalized WT-PTHR and 0K-PTHR recycled to the cell membrane within 30 min of PTH treatment (Fig. 1F). Mutant 0K-PTHR was as efficiently internalized and recycled as WT-PTHR. Thus, ubiquitination does not interfere with efficient PTHR endocytosis and is not required for PTHR recycling. Establishing that the Lys mutant receptors displayed normal signaling and trafficking permitted investigation of the consequences of PTH-induced ubiquitination independent of these biological activities.

Agonist-promoted PTHR ubiquitination and degradation
Previous work established that PTH(1-34) induced receptor ubiquitination followed by partial deubiquitination and recycling, with negligible degradation in the absence of cycloheximide to prevent de novo protein synthesis. In contrast, PTH(7-34) evoked receptor ubiquitination accompanied by degradation (19). Here, in GnTI Ϫ cells stably expressing PTHR, PTH(1-34) stimulated stable PTHR ubiquitination for at least 30 min (Fig. 2, A and C). After 45 min, ubiquitination decreased dramatically ( Fig. 2A) primarily due to degradation of ubiquitinated receptors (Fig. 2B), resulting in diminished PTHR expression. After pretreatment with cycloheximide followed by 45-min PTH stimulation, WT-PTHR abundance decreased by 40% (Fig. 2, B and C), whereas the mutant 0K-PTHR was essentially refractory to metabolic degradation, with expression unchanged from control levels (Fig. 2, B and C). PTHR depletion as observed here is probably replenished by equivalent newly synthesized receptor protein because in the absence of cycloheximide, PTHR expression was stable upon PTH(1-34) stimulation (19). Thus, the time course of PTHR ubiquitination  (43). Loops 1-3 (L1-L3) and the C terminus (CT) are marked in red. Intracellular Lys are highlighted in green and listed above. B, cAMP signaling in response to 100 nM PTH(1-34) by HEK-293 cells stably expressing luciferase-based GloSensor TM cAMP reporter and transiently expressing WT-PTHR or mutant PTHR. Error bars were omitted for clarity (**, p Ͻ 0.01, ANOVA with post hoc multiple comparison using the Bonferroni procedure (Prism). C, internalization of WT-PTHR and Lys-deficient 0K-PTHR visualized by confocal microscopy. HEK-293S GnTI Ϫ cells stably expressing TAP(HA)-WT-PTHR or TAP(HA)-0K-PTHR were incubated with anti-HA antibody conjugated to DyLight 488 to label surface receptors as detailed under "Experimental Procedures." After washing, cells were incubated with vehicle (Control) or 100 nM PTH . After 5 min, cells were fixed, and remaining, non-internalized, surface receptors were detected with an anti-mouse secondary antibody conjugated to Alexa 594. Scale bars, 10 M. D, quantitative analysis of WT-PTHR and 0K-PTHR from confocal images shown in C. Internalization was calculated from the ratio of post-treatment surface receptor expression (Alexa 594) to initial surface-expressed receptor (Dylight 488). Each bar represents the mean Ϯ S.D. (error bars) (n ϭ 11-14); *, p Ͻ 0.05. E, recycling of PTH receptors measured by cell-surface biotinylation. HEK-293S GnTI Ϫ cells stably expressing TAP(HA)-WT-PTHR or TAP(HA)-0K-PTHR were labeled with biotin as described under "Experimental procedures." After quenching unreacted biotin and washing, cells were treated with 100 nM PTH(1-34) for 15 min at 37°C to internalize the PTHR. The samples were then chilled, stripped, and allowed to recover at 37°C for 0, 15, or 30 min to permit the internalized PTHR to recycle. For each time point, one dish was stripped after recycling, and one paired dish was not. These dishes represent the non-recycled and total internalized PTHR, respectively. Biotinylated receptors were immunoprecipitated with streptavidin and detected with an anti-HA antibody. Molecular weights are indicated at the left. Recycled receptor equals total internalized receptor minus non-recycled receptor. F, quantification of receptor recycling shown in E expressed as a percentage of the total receptors at the indicated time. Results are means Ϯ S.D. (error bars) (n ϭ 3). ANOVA indicated equivalent recycling of WT-PTHR and 0K-PTHR.   was added for 5, 15, 30, 45, or 60 min. Cells were harvested and lysed as described under "Experimental procedures," incubated overnight with streptavidin-agarose beads at 4°C. The eluted protein using SDS sample buffer was incubated at 37°C for 30 min. A higher-percentage gel (12%) was run for a shorter time to produce sharper bands on the immunoblot. Anti-ubiquitin P4D1 and HA antibodies ( Table 3) were used to detect ubiquitination and PTHR. ␤-Actin was used as a loading control. Molecular weights are indicated at the left. The illustrated immunoblots are representative of three replicates. Blocking kinase activity with staurosporine virtually abolished receptor ubiquitination (lane 1). B, time-dependent degradation of WT-PTHR and 0K-PTHR. HEK-293S GnTI Ϫ cells stably expressing TAP(HA)WT-PTHR or TAP(HA)0K-PTHR were transfected with Mycubiquitin. After 36 h, cells were seeded on 6-cm dishes and grown overnight at 37°C, serum-starved for 2 h, and incubated with 10 g/ml cycloheximide for 1 h. 100 nM PTH(1-34) was added for the indicated time. Protein lysates were resolved by SDS-PAGE. WT-PTHR and 0K-PTHR were detected with an anti-HA antibody. Molecular weights are indicated at the left. C, quantification of receptor degradation. WT-PTHR and 0K-PTHR abundance was calculated from the respective band intensity and was normalized to ␤-actin. Results are reported as means Ϯ S.D. (error bars) (n ϭ 3; ***, p Ͻ 0.001, two-way ANOVA).

PTHR ubiquitination and MAPK signaling
and degradation were highly correlated. Moreover, preventing receptor ubiquitination abolished degradation, as exemplified by the 0K-PTHR. These results imply that PTHR metabolism essentially proceeds in a ubiquitination-sensitive manner. Interestingly, when cells were treated with 5 M staurosporine, a broad-spectrum protein kinase inhibitor, PTHR ubiquitination was virtually abolished ( Fig. 2A, lane 1), suggesting a role for receptor phosphorylation in the ubiquitination process and implying signaling crosstalk between receptor phosphorylation and ubiquitination.

Sites of PTHR ubiquitination
We next sought to identify intracellular Lys residues ubiquitinated upon PTH treatment. GnTI Ϫ cells stably expressing WT-PTHR or the various Lys-deficient mutant PTHRs were transfected with Myc-ubiquitin and treated with PTH(1-34) for 30 min. Similar amounts of receptor protein were pulled down by streptavidin beads and detected by immunoblotting (Fig. 3A). The characteristic smearing of ubiquitinated WT-PTHR was absent in 2K-PTHR and 0K-PTHR (Fig. 3A). In contrast, 3K-PTHR and CTK-PTHR constructs, where loop 3 Lys and C terminus Lys are present, displayed ubiquitination comparable that of WT-PTHR, suggesting that loop 3 and the carboxyl PTHR tail harbor the principal sites of Lys ubiquitination.
We applied mass spectrometry of purified receptor protein to identify Lys residues targeted for ubiquitination. After a 30-min challenge with PTH(1-34), Lys 388 in the third loop and Lys 484 in the C terminus exhibited the characteristic covalently linked di-Gly modifications ( Fig. 3 B and C and Table 1) indicative of ubiquitination. Loop 3 harbors lysines at Lys 388 , Lys 405 , and Lys 408 . The latter two sites are not covered by trypsin digestion as needed for mass spectrometry. Therefore, to determine whether these residues in the Loop 3 construct are ubiquitinated, we separately reverted each individual mutated Arg to Lys (R388K, R405K, or R408K) and transiently transfected them together with Myc-ubiquitin in GnTI Ϫ cells. As shown in Fig. 3D, 0K-PTHR displayed no ubiquitination as expected, and R388K-PTHR showed a high-molecular weight ubiquitination signal similar to WT-PTHR, confirming the mass spectrometry results for this Lys residue. In contrast, neither R405K nor R408K exhibited detectable ubiquitination (Fig. 3D). Collectively, these results show that only Lys 388 in the third loop and Lys 484 in the intracellular PTHR C terminus are ubiquitinated upon PTH treatment.
Following PTH stimulation, the PTHR is phosphorylated at Thr 387 and Thr 392 , Ser 489 , Ser 491 , Ser 492 , and Ser 493 ( Fig. 5A and Table 2). These sites of Ser phosphorylation are consistent with a recent report (34). To determine which of these phosphorylation sites is involved in or responsible for ␤-arrestin recruitment, Thr at Thr 387 /Thr 392 and Ser in the 489 SGSSS 493 cluster (Fig. 5B), alone or together, were mutated to Ala (Ala 387 /Ala 392 , 489 AGAAA 493 ). These constructs were then used to characterize the binding to ␤-arrestin1 or ␤-arrestin2. As shown in Fig. 5 (C and D), replacing Thr 387 /Thr 392 decreased PTHR binding both to ␤-arrestin1 and ␤-arrestin2. In contrast, mutating the 489 SGSSS 493 motif failed to disrupt binding to ␤-arrestin1 or ␤-arrestin2. Combined mutation of Ala 387 /Ala 392 with 489 AGAAA 49 had no greater effect than that of Ala 387 /Ala 392 alone on binding to ␤-arrestin1 or ␤-arrestin2.
␤-Arrestin binding to phosphorylated receptors initiates desensitization. As shown here for the PTHR, phosphorylation sites in the Thr 387 /Thr 392 region are critical to ␤-arrestin recruitment (Fig. 5). Interestingly, ubiquitinated Lys 388 is located within the 387 TKLRET 392 region. Overlap of phosphorylation and ubiquitination positions suggests that receptor phosphorylation and ubiquitination are at least correlated and probably interdependent. Indeed, abrogation of phosphorylation at positions Thr 387 /Thr 392 dramatically decreased PTHR ubiquitination (Fig. 5E). However, preventing phosphorylation within the 489 SGSSS 493 cluster did not markedly interfere with PTHR binding to ␤-arrestin (Fig. 5, C and D) and did not affect receptor ubiquitination (Fig. 5E), suggesting that ␤-arrestin binding to the receptor site-specifically phosphorylated at 387 TKLRET 392 is required for subsequent PTHR ubiquitination.

Ubiquitinated PTHR differentially regulates PTHR-mediated MAPK signaling
We further inquired into the cellular and signaling consequences and biological function of ubiquitinated PTHR. Early evidence revealed that ␤-arrestins play an important role not only in GPCR desensitization (35) and trafficking (36), but also in G-protein-independent MAPK signaling (37). We reasoned that if ␤-arrestin-dependent receptor ubiquitination affects PTHR-mediated downstream MAPK signaling, then G-protein-independent signaling of WT-PTHR or 0K-PTHR stimulated by PTH(7-34) should be impaired. MAPK signaling pathways include ERK1 and ERK2 (ERK1/2), JNK, and p38 (38). Because ␤-arrestins can scaffold MAPKs (39,40), we investigated the effects of ␤-arrestindependent PTHR ubiquitination on downstream MAPK signaling. The PDZ protein NHERF1 (Na ϩ /H ϩ exchanger regulatory factor 1) can inhibit PTH-induced ERK signaling (24). Therefore, we used GnTI Ϫ cells, where endogenous NHERF1 expression is negligible (data not shown) and does not affect receptor trafficking (Fig. 1, C and E), to examine the influence of PTHR ubiquitination on downstream MAPK signaling. Further, to exclude the possibility of G q /protein kinase C-induced ERK activation (41,42), we used PTH , which triggers PTHR ubiquitination and MAPK signaling without activating PKC. Here, PTH(7-34) induced Lys 48 /Lys 63 -linked PTHR polyubiquitination in a PTHR ubiquitination and MAPK signaling time-dependent fashion (Fig. 6A). Lys 48 ubiquitination is the dominant form of modification and targets PTHR for degradation (19). As shown here, polyubiquitination can also be Lys 63 -linked in the presence of either PTH  or PTH(7-34) (Fig.  6C). The observed discrete bands characteristic of Lys 63 -linked ubiquitination are lower than combined Lys 63 plus Lys 48 high-PTHR ubiquitination and MAPK signaling molecular-weight smears detected by the P4D1 ubiquitin antibody (Fig. 6A). This difference probably arises from the absence of Lys 48 -linked polyubiquitin. A similar pattern of antibody-dependent size differences has been described (44,45). Lys 63linked ubiquitination has been implicated in a variety of cellular events, including ERK signal transduction (46).
MAPK signaling was analyzed from 0 to 30 min in the absence of proteasome inhibitor MG132 because PTHR ubiquitination ( Fig. 6A) and degradation (Fig. 6B) were stable for the first 30 min following PTH . PTH(7-34) stimulated ERK1/2 phosphorylation in a time-dependent manner in cells expressing WT-PTHR. Phosphorylation of ERK1/2 reached a maximum at 5 min, after which the level declined (Fig. 7, A and C). In marked contrast, the 0K-PTHR exhibited a small early ERK1/2 increase at 2 min but a markedly lower peak at 5 min and a modestly delayed response at 10 min (Fig.  7, B and C).
p38 displayed a conspicuously different pattern of ubiquitination-sensitive activation compared with ERK1/2. Upon PTH(7-34) treatment, p38 signaling of WT-PTHR dramatically decreased at 5 min compared with 0K-PTHR (Fig. 8, A-C), where there was no measurable change. The substantial difference in p38 signaling between WT-PTHR and 0K-PTHR in response to PTH  suggested that PTH-induced ubiquitination of discrete Lys residues is required to inhibit p38 signaling. We used the PTHR mutants described earlier (Figs. 1A and 3D) to test this hypothesis. As shown in Fig. 8 (D and E), refined analysis of p38 signaling by the various PTHR Lys mutants revealed a pattern of differential responses. 3K-PTHR, CTK-PTHR, and 388K-PTHR exhibited decreased p38 signaling similar to WT-PTHR, whereas 2K-PTHR and 0K-PTHR were refractory to PTH(7-34) stimulation, thus implicating Lys 388 and Lys 484 as critical for PTH-induced receptor ubiquitination.
The differential response of ␤-arrestin-dependent MAPK signaling triggered by PTH  suggests that ubiquitination initiates distinct MAPK signaling signatures. MAPK signaling is importantly involved in cell proliferation and apoptosis (47). We therefore tested the effect of PTH on these biological processes in cells expressing WT-PTHR or 0K-PTHR. PTH  increased the rate of proliferation in cells expressing WT-PTHR but not in cells expressing 0K-PTHR (Fig. 9A). Enhanced proliferation was accompanied by reduced apoptosis in cells expressing WT-PTHR, whereas cells expressing 0K-PTHR exhibited increased apoptosis (Fig. 9B).

Discussion
Cyclical receptor desensitization and down-regulation protect cells against persistent agonist-induced overstimulation
Notably, PTHR ubiquitination requires antecedent phosphorylation and ␤-arrestin recruitment and causes differential activation of MAPK signaling upon PTH treatment. Thus, a possible scenario is that phosphorylated receptors recruit ␤-arrestin, which subsequently forms a multicomponent complex that includes the E3 ubiquitin ligase that is responsible for PTHR ubiquitination, as in the case of ␤2AR (31) and V2R (49). Ubiquitination commonly targets GPCRs for degradation in lysosomes or proteasomes (50,51). Internalized PTHR are targeted to endosomes, as demonstrated by colocalization with EEA1 (52). PTHR then undergoes recycling or degradation (53). Post-translational ubiquitin modification also has been implicated in regulating receptor endocytosis and cell signaling (54). Early evidence established that internalization of the yeast GPCR Ste2 depends on auto-ubiquitination (55,56). Internalization of most mammalian GPCRs, however, has proven to be independent of ubiquitination (57). Interestingly, upon agonist stimulation, PAR1 displays a hybrid pattern of ubiquitination-dependent and -independent internalization (58). The Lys-deficient 0K-PTHR mutant lacking all intracellular Lys and hence refractory to ubiquitination internalized and recycled comparably with WT-PTHR, implying that ubiquitination is not required for PTHR internalization or recycling and does not affect the rate of receptor endocytosis.
Several lines of investigation reveal an interplay between receptor phosphorylation and ubiquitination (59). The earliest evidence again came from yeast Ste2p, where phosphorylation of the cytoplasmic receptor tail facilitated ubiquitination of a vicinal Lys (60). This phenomenon was subsequently extended to ␤2AR, the prototype mammalian GPCR (31). Ligand-induced activation results in ␤2AR and PTHR phosphorylation by GRK2 (12,(61)(62)(63). Because isotype-specific GRK inhibitors are unavailable to determine whether phosphorylation affects PTHR ubiquitination, we used staurosporine, a Ser/Thr kinase inhibitor that at high concentrations nonspecifically blocks a broad gamut of protein kinases, including GRKs (64). The results show clearly that preventing PTHR phosphorylation virtually abolished receptor ubiquitination ( Fig. 2A). Similar findings have been reported for the platelet-derived growth factor receptor-␤ (65). Consistent with this observation, we found that as opposed to kinase inhibition, blocking phosphorylation by site-specific Ala substitution at Thr 387 /Thr 392 , but not at Ser 489 -Ser 493 , reduced ␤-arrestin binding to PTHR by 80% (Fig.  5, C and D) with attendant reduction of PTHR ubiquitination (Fig. 5E). The partial reduction of ubiquitination by Thr mutation but not protein kinase inhibition implies the presence of additional phosphorylation sites besides Thr 387 /Thr 392 that contribute to PTHR binding to ␤-arrestin. Indeed, a recent report showed that the cluster Ser 501 -Thr 506 contributes to the interaction of PTHR with ␤-arrestin (34). The remaining input may stem from Thr 387 /Thr 392 , which was not characterized therein.
We initially expected ubiquitination to affect PTHR internalization or recycling, which turned out not to be the case. The distinct requirement for phosphorylation, on the one hand, and striking biased effects on MAPK signaling of Lys-deficient forms of the PTHR, on the other, suggested that cell proliferation and apoptosis may rely on or be influenced by ubiquitination. Indeed, Lys-deficient 0K-PTHR ubiquitination displayed reduced cell proliferation and increased apoptosis. ERK signaling has been shown to increase cell proliferation (47), consistent with our observation that WT-PTHR exhibits greater ERK activity than 0K-PTHR upon PTH(7-34) treatment. Increased cell death for 0K-PTHR cells is probably associated with the loss of regulated p38 signaling (72). The biased PTHR signaling upon ubiquitination could be connected with dysfunctional regulation and disease. For instance, PTH  accumulates to high levels in end-stage kidney disease (5). Further, ubiquitin ligase RNF146 has been implicated in cleidocranial dysplasia, an autosomal dominant form of abnormal bone disease (73). Moreover, PTH-induced osteoblast proliferation requires direct regulation of the ubiquitin-specific-processing protease 2 gene, USP2 (74), which we demonstrated is required for PTHR deubiquitination (19).
In summary, the present report identifies dual sites of PTHR ubiquitination and illustrates how receptor ubiquitination differentially affects MAPK signaling and cell proliferation. The divergent MAPK signaling responses activated by PTH(7-34)mediated ubiquitination could be potential targets for pharmacological intervention against diseases such as chronic kidney disease-mineral and bone disorder involving PTHR dysfunction. Additional work will be needed to ascertain the origin of

PTHR ubiquitination and MAPK signaling
the biased agonism by which ubiquitination triggers differential MAPK signaling. MAPK is a prospective therapeutic target (75)(76)(77). Our work suggests that novel compounds could potentially target specific elements of MAPK signaling of or signaling of PTHR-related skeletal disorders or chronic kidney disease (78 -80).

DNA constructs
Human TAP-PTHR in pIRES-puro-SS-GLUE (a gift from Drs. Jean-Luc Parent (University of Sherbrooke) and Terence E. Hébert (McGill University, Montreal, Canada)) was generated by PCR. TAP contains calmodulin-binding protein, an HA epitope, a tobacco etch virus cleavage site, and streptavidinbinding protein tags (83) and was inserted at the N terminus of PTHR. Mutant PTHRs were engineered by changing Lys to Arg, Arg to Lys, or Ser/Thr to Ala using the QuikChange (Agilent) kit following the manufacturer's instructions. All constructs were confirmed by DNA sequencing.

Cell culture and transfection
Murine embryonic fibroblasts (MEF) (33) and HEK-293S GnTI Ϫ cells were grown at 37°C in high-glucose Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% FBS and 100 units/ml penicillin-streptomycin in a humidified atmosphere containing 5% CO 2 . Plasmid transfections were performed using FuGENE 6 (Promega), Effectene (Qiagen), or Lipofectamine 3000 (Thermo Fisher Scientific) according to the manufacturer's instructions, unless otherwise specified. For transient transfection, cells were treated as above, and experiments were performed 48 h thereafter. Stably transfected cells expressing epitope-tagged receptors were generated by selecting for neomycin or puromycin resistance using 500 g/ml G418 or 1 g/ml puromycin. Positive colonies were isolated, and clones were further selected based on protein expression levels as assessed by immunoblot.

Immunostaining and confocal microscopy
HEK-293S GnTI Ϫ cells stably expressing TAP(HA)-WT-PTHR or TAP(HA)-0K-PTHR were plated on coverslips coated with poly-D-Lys. Cells were labeled for 1 h at room temperature with an anti-HA primary antibody directly conjugated to DyLight 488, which recognizes the HA tag in the N-terminal domain of PTHR. After washout, the cells were incubated with vehicle or 100 nM PTH(1-34) for 5 min at 37°C. Cells were fixed in 3.6% formaldehyde for 15 min and quenched in 0.1 M glycine for 5 min. Nonspecific binding was blocked by incubation in 10% goat serum. Non-internalized PTHR remaining on the cell surface were detected by labeling with a goat antimouse IgG secondary antibody conjugated to Alexa 594 that recognizes the mouse IgG heavy and light chains of the primary antibody. Coverslips were mounted on glass slides with Fluoromount TM (Diagnostic BioSystems, Pleasanton, CA). Single-plane confocal images were captured using a Nikon Ti-E microscope with A1 confocal unit and ϫ60/1.45 numeric aperture objective. Fluorescent proteins were excited using 488-nm (DyLight488) and 560-nm (Alexa594) lasers. Laser intensity and microscope gain settings were maintained for all image acquisitions. Image acquisition was performed with NIS Elements software, and analysis of fluorescence intensity was executed using ImageJ software (84). The fraction of internalized receptor was calculated by com-

PTHR ubiquitination and MAPK signaling
paring the intensity of Alexa 594 staining in control and in cells treated with PTH.

PTHR biotinylation and receptor internalization and recycling
Stably transfected HEK-293S GnTI Ϫ cells expressing either WT-PTHR or 0K-PTHR were grown to Ն80% confluence on 10-cm dishes and serum-starved overnight. Cells were then chilled, washed with ice-cold PBS, and immediately incubated with 0.5 mg/ml disulfide-cleavable EZ-Link sulfo-NHS-S-S-biotin (Thermo Fisher Scientific, 21331) in a buffer containing 10 mM HEPES, pH 7.6, 154 mM NaCl, 3 mM KCl, 10 mM MgCl 2 , 0.1 mM CaCl 2 , and 10 mM glucose for 45 min at 4°C. Unreacted biotin was quenched by ice-cold PBS supplemented with 100 mM glycine. Cells were subsequently washed twice with icecold PBS and incubated with prewarmed, serum-free DMEM containing 100 nM PTH(1-34) for 15 min to internalize the PTHR. Cells were then chilled on ice, and remaining cell-surface biotinylated receptors were scavenged by incubating at 4°C with two changes of 2-mercaptoethanesulfonate stripping buffer (50 mM Tris-HCl, pH 8.6, 50 mM Na-2-mercaptoethanesulfonate, 100 mM NaCl, 1 nM MgCl 2 , and 0.1 mM CaCl 2 ) for 20 min each. Cells were washed twice with ice-cold PBS and incubated with prewarmed DMEM at 37°C for 0, 15, or 30 min to permit PTHR recycling. At each time point, the first dish of each pair was stripped as above to cleave newly appearing surface biotin from recycled PTHR. The remaining PTHRs detected by immunoblot from this fraction represent non-recycled receptors. The second dish without the additional round of stripping represents the total internalized receptor. Cells were washed with ice-cold PBS and lysed in 50 mM Tris-HCl, pH 8.0, 150 mM NaCl, and 0.5% DDM. After centrifugation at 14,000 rpm for 30 min, supernatants containing equal amounts of total protein were digested with tobacco etch virus protease overnight to remove streptavidin-binding peptide in the PTHR TAP tag, followed by incubation with streptavidin-agarose beads overnight to capture biotinylated proteins. Beads were extensively washed. Protein was eluted with SDS sample buffer, resolvedbySDS-PAGE,andtransferredtopolyvinylidenedifluoride membranes and detected by immunoblotting. The recycled receptor fraction was calculated as the difference between total internalized receptor and non-recycled receptor for each pair.

Purification of TAP-PTHR
GnTI Ϫ cells stably expressing PTHR were grown on two 15-cm tissue culture dishes and transferred to 200 ml of Gibco FreeStyle 293 expression medium (Invitrogen). After a 60-h incubation at 37°C at 8% CO 2 , cells were serum-starved for 1 h and treated with 10 M MG132 for a second hour and then challenged with 100 nM PTH for 30 min. Cells were harvested by centrifugation at 1,500 rpm for 3 min and washed with cold PBS. Cells were lysed in a buffer containing 50 mM Tris-HCl, pH 8.0, 150 mM NaCl, 10 mM N-ethylmaleimide, 0.5% DDM, and Protease Inhibitor Mixture Set I (Calbiochem, 539131). After overnight incubation at 4°C, cell extracts were clarified by centrifugation at 16,000 rpm for 45 min. The resulting supernatant was mixed with pre-equilibrated streptavidin-conjugated agarose beads in the lysis buffer and incubated at 4°C for 4 h. Beads were extensively washed with a buffer containing 50 mM Tris-HCl, pH 8.0, 150 mM NaCl, 10 mM N-ethylmaleimide, 2 mM CaCl 2 , and 0.05% DDM. TAP-PTHR was eluted with a buffer consisting of 50 mM Tris-HCl, pH 8.0, 150 mM NaCl, 10 mM biotin, 2 mM CaCl 2 , and 0.05% DDM. The protein samples were further incubated with pre-equilibrated calmodulin-conjugated agarose beads in a buffer of 50 mM Tris-HCl, pH 8.0, 150 mM NaCl, 2 mM CaCl 2 , and 0.05% DDM. Proteins were eluted with a buffer consisting of 50 mM Tris-HCl, pH 8.0, 150 mM NaCl, 10 mM EGTA, and 0.05% DDM. Protein concentrations were measured using the Bradford assay (Bio-Rad) (85).

Detection of ubiquitinated receptors
For transiently expressed receptors, GnTI Ϫ cells were grown on 10-cm dishes and transiently transfected with 2 g each of DNA encoding WT-PTHR or the indicated mutant PTHR and Myc-ubiquitin. GnTI Ϫ cells stably expressing WT-TAP(HA)-PTHR, 0K-TAP(HA)-PTHR, or the indicated mutant were transiently transfected with Myc-ubiquitin. After 48 h, cells were washed with cold PBS, serum-staved for 2 h, and incubated with 10 M MG132 for 30 min. Cells were then stimulated with 100 nM PTH(1-34) for 30 min. Cells were harvested, lysed, and incubated with streptavidin affinity beads as described above. After extensive washing, protein was eluted in SDS sample buffer and incubated at 37°C for 30 min. Following electrophoresis, proteins were electroblotted onto a polyvinylidene difluoride membrane and detected by immunoblotting. To reprobe with other antibodies, the membrane was stripped in a PTHR ubiquitination and MAPK signaling denaturation buffer (62.5 mM Tris-HCl, pH 6.8, 100 mM ␤-mercaptoethanol, and 2% SDS) for 30 min at 50°C. Proteins were detected by ECL Western blotting (GE Healthcare), unless otherwise stated, and the signals were quantified with ImageJ software (84).

PTHR degradation
HEK-293S GnTI Ϫ cells stably expressing either WT-PTHR or 0K-PTHR were transfected with Myc-ubiquitin. 36 h later, cells were seeded at equal density on 6-cm dishes and grown overnight at 37°C. Cells were serum-starved for 2 h and incubated with 10 g/ml cycloheximide for 1 h to arrest newly synthesized receptor. Cells were subsequently incubated at 37°C with or without 100 nM PTH(1-34) for the times indicated, washed with cold PBS, and lysed in a buffer containing 50 mM Tris-HCl, pH 8.0, 150 mM NaCl, and 0.5% DDM supplemented with protease inhibitors (Protease Inhibitor Mixture Set I). Cell lysates were collected and normalized based on total protein concentration measured by the Bradford method (85). Equivalent amounts of cell lysates were analyzed by immunoblotting.

In-gel digestion and mass spectrometry
Purified PTHR protein was resolved by 7% SDS-PAGE and stained with Coomassie Brilliant Blue. After destaining, the gel band containing receptor protein was excised and cut into small pieces. The gel was further destained overnight in 50% acetonitrile containing 25 mM NH 4 HCO 3 and dehydrated in 100% acetonitrile. The in-gel protein was reduced with 10 mM DTT and alkylated by 55 mM iodoacetamide. The gel was washed with 25 mM NH 4 HCO 3 and dehydrated with 100% acetonitrile. Sufficient trypsin was added to the dried gel pieces to perform overnight in-gel digestion at 37°C. The digested peptides were desalted using Pierce C18 spin columns and eluted in a buffer containing 75% acetonitrile and 0.1% TFA. The eluates were lyophilized and reconstituted in 0.1% formic acid. 3 l of the peptide solution was injected and analyzed by LC-MS/MS using an LTQ Orbitrap Velos mass spectrometer (Thermo Fisher Scientific) as described (20). Collected raw data were analyzed for both ubiquitination and phosphorylation using the Sequest algorithm (86). For ubiquitination, a dynamic modification of 114.0429 Da was used for Lys residues to match the di-Gly signature. For phosphorylation, a dynamic modification of 79.9663 Da was used for Ser, Thr, and Tyr residues.

Cell proliferation and apoptosis
A total of 2 ϫ 10 4 stably transfected GnTI Ϫ cells expressing WT-PTHR or 0K-PTHR were seeded in triplicate on 96-well plates (Costar, 07-200-95) and cultured at 37°C for the time indicated. Cell proliferation was evaluated by measuring bromodeoxyuridine incorporation using the cell proliferation ELISA kit (catalog no. 11647229001) from Roche Applied Science. Apoptosis was assessed with the ELISAPLUS TM cell death detection kit (Roche Applied Science, 11544675001).

Data analysis
Results were analyzed using Prism version 7 software (GraphPad, La Jolla, CA). Results represent the mean Ϯ S.D. of n Ն 3 independent experiments and were compared by analysis of variance with post hoc testing using the Bonferonni procedure. p values Ͻ 0.05 were considered statistically significant.