The Role of Ubiquitination and Hepatocyte Growth Factor-Regulated Tyrosine Kinase Substrate in the Degradation of the Adrenomedullin Type I Receptor

Calcitonin receptor-like receptor (CLR) and the receptor activity-modifying protein 2 (RAMP2) comprise a receptor for adrenomedullin (AM). Although it is known that AM induces internalization of CLR•RAMP2, little is known about the molecular mechanisms that regulate the trafficking of CLR•RAMP2. Using HEK and HMEC-1 cells, we observed that AM-induced activation of CLR•RAMP2 promoted ubiquitination of CLR. A mutant (CLRΔ9KR), lacking all intracellular lysine residues was functional and trafficked similar to the wild-type receptor, but was not ubiquitinated. Degradation of CLR•RAMP2 and CLRΔ9KR•RAMP2 was not dependent on the duration of AM stimulation or ubiquitination and occurred via a mechanism that was partially prevented by peptidase inhibitors. Degradation of CLR•RAMP2 was sensitive to overexpression of hepatocyte growth factor-regulated tyrosine kinase substrate (HRS), but not to HRS knockdown, whereas CLRΔ9KR•RAMP2 degradation was unaffected. Overexpression, but not knockdown of HRS, promoted hyperubiquitination of CLR under basal conditions. Thus, we propose a role for ubiquitin and HRS in the regulation of AM-induced degradation of CLR•RAMP2.

. AM induces ubiquitination of CLR but not RAMP2. (A,B,D) HEK-CLR•RAMP2 or HMEC-1-CLR•RAMP2 cells were not treated (NT) or incubated with AM (100 nM, 0-60 min). CLR was immunoprecipitated (IP) using an antibody to CLR (RK11, HEK; rat-HA, HMEC-1) and RAMP2 was immunoprecipitated using an antibody to the extracellular epitope tag of RAMP2 (rabbit-Myc, rMyc). Immunoprecipitates were analyzed by Western blotting (WB) and probed using antibodies to ubiquitin and the extracellular epitope tags of CLR (rabbit-HA, rHA) and RAMP2 (rMyc).  •RAMP2 or HMEC-1-CLRΔ9KR•RAMP2 cells were untreated or incubated with AM (100 nM, 0-60 min). CLRΔ9KR was immunoprecipitated (IP) using an antibody to CLR (HEK, RK11; HMEC-1, ratHA). Immunoprecipitates were analyzed by Western blotting (WB) and probed using antibodies to ubiquitin and CLR (HEK, rabbit-HA, rHA; HMEC-1, mouse-HA, mHA). AM did not induce ubiquitination of CLRΔ9KR, whereas CLRΔ9KR was readily detected by the HA antibody. There were no signals in vector control cells, confirming specificity of antibodies. n = 3-4; HEK-vc = HEK-vector control. (C) Serum-starved HEK-CLR•RAMP2 and HEK-CLRΔ9KR•RAMP2 cells were not treated (NT) or incubated with AM (100 nM, 0-10 min), washed and incubated in AM-free medium (0-120 min). Lysates were then analyzed for levels of phosphorylated (p) ERK2 (pERK2) and ERK2 by Western blotting. In untreated cells, levels of pERK2 were similarly low in both HEK-CLR•RAMP2 and HEK-CLRΔ9KR•RAMP2 cells. AM-induced a prompt increase in levels of pERK2 and returned towards basal levels after removal of agonist. The magnitude and duration of The ubiquitination of eukaryotic GPCRs can function to facilitate internalization 27 , target the GPCR to lysosomes [28][29][30] or regulate GPCR trafficking through the endocytic network and hence their rate of their proteolytic destruction 31,32 . In this study we aimed to determine if transient stimulation with AM promotes CLR•RAMP2 recycling and the molecular mechanism regulating the post-endocytic sorting of CLR•RAMP2.

Results
AM Induces ubiquitination of CLR, but not RAMP2. It has previously been reported that CLR•RAMP2 traffics to acidic vesicles following stimulation with AM, however the molecular mechanism regulating this trafficking is unknown 25 . Attachment of ubiquitin moieties to lysine residues is required for the targeting of some, but not all GPCRs to lysosomes [28][29][30]33 . To determine whether AM caused ubiquitination of CLR or RAMP2, we incubated HEK-CLR•RAMP2 cells with AM, immunoprecipitated CLR and RAMP2 and probed Western blots for CLR, RAMP2 and ubiquitin. At all time points both CLR and RAMP2 were readily detected, with no signals present in cells not expressing receptor components (Fig. 1A,B). No signals for ubiquitinated RAMP2 were detected (Fig. 1B). However, ubiquitinated CLR could readily be detected 60 min post-stimulation (Fig. 1A,C).
As AM plays an important role in the maintenance of the endothelial cell barrier 7 , we investigated whether AM also induces ubiquitination of CLR when expressed in microvascular endothelial cells. HMEC-1 cells naturally express functional CLR•RAMP2 but to facilitate biochemical studies, we virally transduced HMEC-1 cells with CLR and RAMP2 (Supplementary Fig. 1A-D). HMEC-1-CLR•RAMP2 cells were challenged with AM (0-60 min), CLR immunoprecipitated and Western blots probed for CLR and ubiquitin. CLR was readily detected, with no signals present in cells not expressing receptor components, indicating antibody specificity (Fig. 1D). AM induced ubiquitination of CLR after 60 min (Fig. 1D,E). Thus, sustained activation of CLR•RAMP2 with AM causes ubiquitination of CLR in HEK and HMEC-1 cells.

Mutation of the intracellular lysines of CLR does not affect CLR•RAMP2 expression and internalization.
As ubiquitin molecules are mainly attached to lysine residues of target proteins 34,35 , we mutated all predicted intracellular lysine residues of CLR to arginine and examined the expression, trafficking, ubiquitination and mitogenic signaling of CLRΔ9KR•RAMP2. First, we determined if mutation of the lysine residues affected expression of the receptor. We expressed CLRΔ9KR•RAMP2 in HEK cells and compared the expression with the wild-type receptor (CLR•RAMP2). An antibody to the extracellular epitope (HA) of CLR recognized single proteins of approximately the same size (~85 kDa) in both HEK-CLR•RAMP2 and HEK-CLRΔ9KR•RAMP2 cells (Supplementary Fig. 2A,B). The predicted molecular mass of CLR is 54 kDa, although there are six sites for potential N-linked glycosylation, which probably accounts for the larger apparent mass. An antibody to the extracellular epitope (Myc) of RAMP2 recognized two proteins in both HEK-CLR•RAMP2 and HEK-CLRΔ9KR•RAMP2 cells with approximate molecular masses of 32 kDa and 37 kDa (Supplementary Fig. 2A,B). The predicted molecular mass of RAMP2 is 19 kDa, thus we predict that the observed immunoreactive proteins represent unglycosylated and glycosylated dimers of RAMP2, respectively. No signals were detected in HEK-vc cells, except for β-actin, which acted as a loading control, indicating specificity of detection (Supplementary Fig. 2A,B). A requirement for a functional receptor is expression at the cell-surface. To determine if mutation of lysine residues affected cell-surface expression we simultaneously localized CLR or CLRΔ9KR and RAMP2 using antibodies to the extracellular epitope tags of the receptor components. In unstimulated cells, CLR, CLRΔ9KR and RAMP2 were present at the cell-surface. AM similarly induced trafficking of CLR with RAMP2 and CLRΔ9KR with RAMP2 to the same intracellular vesicles ( Supplementary Fig. 2C,D). Together these data suggest that mutation of the intracellular facing lysine residues has no effect on the expression and internalization of CLR•RAMP2.

Mutation of the intracellular lysines of CLR prevents ubiquitination of CLR. Next we determined
if mutation of the intracellular lysines residues affects ubiquitination of CLR. HEK-CLRΔ9KR•RAMP2 cells were stimulated with AM (0-60 min), CLR immunoprecipitated and Western blots probed for ubiquitin and CLRΔ9KR. In contrast, to CLR (Fig. 1A), no ubiquitination of CLRΔ9KR could be detected ( Fig. 2A). We performed a similar experiment in HMEC-1 cells expressing CLRΔ9KR and similarly did not detect ubiquitination of CLRΔ9KR (Fig. 2B). Thus, the intracellular lysine residues are required for AM-induced ubiquitination of CLR.
Ubiquitination has no effect on AM-induced ERK Activation. Ubiquitination of receptors and associated proteins has been shown to influence the signaling properties of activated receptors 36-38 , therefore we examined if ubiquitination of CLR regulates AM-induced ERK1/2 activation. Serum-starved HEK-CLR•RAMP2 and HEK-CLRΔ9KR•RAMP2 cells were incubated with AM (10 min), incubated in AM-free medium (0-110 min) and phosphorylated (p) and total ERK2 analyzed by Western blotting. In untreated cells, levels of pERK2 were low (Fig. 2C,D). In HEK-CLR•RAMP2 and HEK-CLRΔ9KR•RAMP2 cells, AM induced phosphorylation of ERK2 that peaked at 10 minutes and before returning back to basal. Lack of CLR ubiquitination had no effect on the magnitude of the initial and sustained activation of ERK2 (Fig. 2C,D; Supplementary Fig. 3). Thus, we conclude that ubiquitination of CLR has no role in the generation of AM-induced ERK signaling. GPCRs is required for lysosomal targeting and degradation [28][29][30] , therefore determined ubiquitination of CLR affects trafficking. First, we determined if CLRΔ9KR traffics to early endosomes, similar to that reported for the wild-type receptor 25 . HEK-CLR•RAMP2 and HEK-CLRΔ9KR•RAMP2 cells were incubated with AM (30 min) and CLR and a marker for early endosomes, early endosomal antigen-1 (EEA1) localized by immunofluorescence and confocal microscopy. In unstimulated cells, both CLR and CLRΔ9KR were present at the cell-surface and EEA1 was present in intracellular vesicles (Fig. 3A,B). AM similarly induced trafficking of both CLR and CLRΔ9KR from the cell-surface to colocalize with EEA1, indicating trafficking to early endosomes ( Fig. 3A,B,E). We next determined if ubiquitination of CLR is required for lysosomal targeting. HEK-CLR•RAMP2 and HEK-CLRΔ9KR•RAMP2 cells were incubated with lysosomal peptidase inhibitors, incubated with AM for 4 h and CLR and LAMP1 (a marker for lysosomes) were localized. In unstimulated cells, CLR and CLRΔ9KR were present at the cell-surface and LAMP1 was detected in intracellular vesicles (Fig. 3C,D). Incubation with AM caused internalization of CLR and CLRΔ9KR to colocalize with LAMP1, indicating trafficking to lysosomes (Fig. 3C,D,F). Thus, ubiquitination of CLR is not required for the lysosomal targeting of CLR.
The duration of the AM challenge does not affect the trafficking or fate of activated CLR or CLRΔ9KR. It has previously been shown that the trafficking and fate of the CGRP receptor (CLR•RAMP1) is dependent on the duration of stimulation 14 . To determine whether the duration of stimulation with AM similarly affects the trafficking of CLR•RAMP2, we examined levels of CLR, CLRΔ9KR and RAMP2 following either continuous or transient stimulation with AM. To examine degradation of cell-surface CLR and RAMP2, we employed a cell-surface biotinylation assay. HEK-CLR•RAMP2 and HEK-CLRΔ9KR•RAMP2 cells were challenged with AM (30 min), washed and incubated in AM-free medium for 16 h or alternatively, cells were incubated continuously with AM (16 h), biotinylated proteins immunoprecipitated and levels of CLR, CLRΔ9KR, RAMP2 and transferrin receptor (loading control) determined by Western blotting. In untreated HEK-CLR•RAMP2 and HEK-CLRΔ9KR•RAMP2 cells, CLR, CLRΔ9KR and RAMP2 were readily detected ( Fig. 4A-D, Supplementary  Fig. 4A,B). In contrast, to CLR•RAMP1, which only recycles following transient stimulation 14 , both transient and continuous stimulation of CLR•RAMP2 with AM caused significant degradation of CLR, CLRΔ9KR and RAMP2 ( Fig. 4A-D, Supplementary Fig. 4A,B). As significantly less CLR, CLRΔ9KR and RAMP2 were degraded following a transient stimulation compared to a sustained stimulation (Fig. 4E,F), we investigated whether this was due , fixed, permeabilized and CLR and a marker for early endosomes (EEA1) localized by immunofluorescence and confocal microscopy. In unstimulated cells, CLR and CLRΔ9KR were present at the cell-surface (arrowheads) and EEA1 was detected in intracellular vesicles. AM induced trafficking of CLR and CLRΔ9KR to colocalize with EEA1 in early endosomes (arrows). (C,D) HEK-CLR•RAMP2 and HEK-CLRΔ9KR•RAMP2 cells were incubated with lysosomal protease inhibitors, challenged with vehicle (control) or AM (100 nM, 4 h), fixed, permeabilized and CLR and a marker for lysosomes (LAMP1) localized by immunofluorescence and confocal microscopy. In unstimulated cells, CLR and CLRΔ9KR were present at the cell-surface (arrowheads) and LAMP1 was detected in intracellular vesicles. AM induced trafficking of CLR and CLRΔ9KR to colocalize with LAMP1 in lysosomes (arrows). (E,F) Quantification of colocalization of CLR with EEA1 and LAMP1, respectively. n = 3, Scale bar, 10 µm.   Fig. 6A,B). Thus, we conclude that once activated, CLR•RAMP2 is removed from the cell-surface, does not recycle to the cell-surface, traffics to LAMP1-positive vesicles and that CLR•RAMP2 and CLRΔ9KR•RAMP2 are lost from the cell-surface to similar degrees following a 30 min exposure to AM.

Lysosomal peptidases participate in the degradation of CLR•RAMP2 and CLRΔ9KR•RAMP2.
Lysosomal and proteasomal peptidases have both been implicated in the degradation of GPCRs 39,40 . Therefore, we investigated the peptidases responsible for the degradation of CLR, CLRΔ9KR and RAMP2. HEK-CLR•RAMP2 and HEK-CLRΔ9KR•RAMP2 cells were preincubated with lysosomal peptidase inhibitors or vehicle (control) and stimulated with AM. We then determined levels of CLR, CLRΔ9KR, RAMP2 and TfR by Western blotting. In vehicle-treated cells CLR, CLRΔ9KR and RAMP2 were degraded to similar levels ( Fig. 5A-D). Incubation with lysosomal inhibitors partially prevented degradation of CLR, CLRΔ9KR and RAMP2 by similar amounts (Fig. 5A-D). Thus, lysosomal peptidases partially regulate the degradative process, irrespective of whether CLR is ubiquitinated or not.
To further determine if lysosomal peptidases are involved in the degradation of CLR and RAMP2, we determined the effect of the serine protease inhibitor, leupeptin an inhibitor that has been used to prevent the degradation of other GPCRs 41 . HEK-CLR•RAMP2 cells were treated as above and degradation of CLR, RAMP2 and TfR determined by Western blotting. Incubation of cells with leupeptin did not prevent the degradation of CLR and RAMP2 (Supplementary Fig. 7A-C). Although the degradation of GPCRs by the proteasome is unusual 39,40 , we investigated the role of the proteasome in the degradation of CLR•RAMP2. We quantified degradation of CLR, CLRΔ9KR and RAMP2 in the presence of the highly specific proteasome inhibitor, epoxomicin. In vehicle-and epoxomicin-treated cells, CLR, CLRΔ9KR and RAMP2 were degraded to similar levels ( Supplementary Fig. 8A-D). Thus, the proteasome plays no role in the degradation of CLR, CLRΔ9KR and RAMP2.

Kinetics of degradation of CLR•RAMP2 are not regulated by AM-induced ubiquitination.
Ubiquitination of GPCRs has also been shown to regulate the degradation rate by altering the trafficking through the endocytic system (25,26). Therefore we determined if ubiquitination altered the rate of degradation of CLR, CLRΔ9KR and RAMP2 by examining degradation following stimulation with AM at a shorter time points. HEK-CLR•RAMP2 and HEK-CLRΔ9KR•RAMP2 cells were stimulated with AM (0-4 h) and we then determined levels of CLR, CLRΔ9KR, RAMP2 and TfR. In HEK-CLR•RAMP2 and HEK-CLRΔ9KR•RAMP2 cells, CLR, CLRΔ9KR and RAMP2 were degraded to similar levels ( Fig. 6A-D). We also performed the same experiments in HMEC-1 cells and observed similar results (Supplementary Fig. 9A-C). Thus, AM-induced ubiquitination of CLR does not affect the kinetics of degradation.

HRS regulates the degradation of ubiquitinated CLR•RAMP2, but not CLRΔ9KR•RAMP2. HRS
is an endosomal sorting protein that has previously been shown to play a role in the trafficking and degradation of CLR•RAMP1 19 . Therefore we investigated if HRS also regulates the degradation of CLR•RAMP2. We quantified the overexpression of HRS in HEK cells expressing HRS and either CLR•RAMP2 or CLRΔ9KR•RAMP2 (Fig. 7A,B). The same cells were incubated with AM (0-4 h) and levels of CLR, CLRΔ9KR and RAMP2 determined. Expression of HRS prevented the degradation of both CLR and RAMP2 in cells expressing CLR•RAMP2 (Fig. 7C-F). In contrast, expression of HRS did not affect the degradation of CLRΔ9KR and RAMP2 in CLRΔ9KR•RAMP2 cells ( Fig. 7C-F). Thus, HRS is not required for the degradation of CLRΔ9KR.

Ubiquitin-interaction motif of HRS does not regulate degradation of ubiquitinated CLR. HRS
concentrates ubiquitinated receptors by direct interaction with its ubiquitin-interaction motif (UIM) 42 . Therefore we investigated if the UIM of HRS was essential for its regulation of CLR degradation. Overexpression of HRSΔUIM in HEK cells expressing HRS and either CLR•RAMP2 or CLRΔ9KR•RAMP2 was quantified by Western blotting (Supplementary Fig. 8A,B). The same cells were incubated with AM (0-4 h) and levels of CLR, CLRΔ9KR and RAMP2 quantified ( Supplementary Fig. 10C-F). In HEK cells expressing HRSΔUIM and either CLR•RAMP2 or CLRΔ9KR•RAMP2 cells, we observed similar degradation profiles for CLR, CLRΔ9KR and RAMP2 as when they are co-expressed with HRS ( Supplementary Fig. 10C-F). Thus, the UIM of HRS is not involved in the HRS-dependent regulation of CLR degradation.

Knockdown of HRS does not affect AM-induced degradation of CLR•RAMP2.
To further examine the role of HRS in the degradation of CLR•RAMP2, we used two separate siRNAs to knockdown expression of HRS. Quantification of HRS knockdown was determined by Western blotting (Fig. 8A,B). HEK cells transiently expressing CLR•RAMP2 that had been treated with transfection reagent alone (Mock), non-targeting pool of siRNA (Scr) or two single siRNAs against HRS were exposed to AM (0-4 h) and degradation of CLR and RAMP2 examined by Western blotting. In Mock and Scr cells, both CLR and RAMP2 were degraded (Fig. 8A,C,D, Supplementary Fig. 11A,B). Surprisingly, knockdown of HRS had no effect of the degradation of CLR and RAMP2 (Fig. 8A,C,D). CLR•RAMP2 constitutively traffics to HRS-positive vesicles. HRS regulates the ubiqutination and deubiquitination of the epidermal growth factor receptor 21 . We hypothesized that HRS may also regulate the degradation of CLR•RAMP2 by regulating the basal ubiquitination pattern of CLR. As a first step, we investigated if CLR•RAMP2 constitutively trafficked to HRS-positive vesicles. An antibody to the N-terminal epitope tag of RAMP2 was added to HEK cells expressing CLR•RAMP2 and cells immediately fixed or incubated for 2 h to allow constitutive trafficking to occur. We observed that CLR•RAMP2 constitutively internalizes and traffics to HRS-positive vesicles, presumably endosomes (Fig. 9A). Next, we determined the effect of HRS overexpression on the basal localization of CLR•RAMP2. In HEK-CLR•RAMP2 cells expressing empty vector (control), CLR was present at the cell-surface and HRS was present in intracellular vesicles (Fig. 9B). However, in HEK-CLR•RAMP2 cells overexpressing HRS, CLR was present at both the cell-surface and in HRS-positive intracellular vesicles (Fig. 9B). We also investigated the effect of HRS knockdown on the localization of CLR•RAMP2. We observed that for both siRNAs, CLR was present at the cell-surface, similar to that observed in Mock and Scr cells (Fig. 9C).
HRS regulates the basal ubiquitination of CLR. As CLR was observed to co-localize with both endogenous and overexpressed HRS, we determined the effect of overexpressing and knockdown of HRS on the basal ubiquitination levels of CLR using immunoprecipitation and Western blotting. Quantification of HRS expression in each experimental condition was determined by Western blotting (Fig. 9D,E). In the control cells (empty vector, Mock, Scr) a comparable level of basal CLR ubiquitination could be detected (Fig. 9D,F). Similar basal levels of ubiquitination were also detected in cells lacking HRS (Fig. 9D,F). In contrast, basal CLR ubiquitination was significantly enhanced in cells overexpressing HRS (Fig. 9D,F). Therefore, CLR constitutively traffics to HRS-positive vesicles and HRS regulates the basal level of CLR ubiquitination. CLR•RAMP2 and CLRΔ9KR•RAMP2 traffic to HRS-and LAMP1-positive vesicles in HRS overexpressing cells. Next we determined the effect of increased HRS expression on AM-induced trafficking of CLR and CLRΔ9KR. We overexpressed HRS, live labeled CLR at the cell-surface and used immunofluorescence and confocal microscopy to localize CLR, CLRΔ9KR, HRS and LAMP1 following treatment with AM (0-4 h). In unstimulated HEK-CLR•RAMP2 and HEK-CLRΔ9KR•RAMP2 cells, CLR was present at the cell-surface, HRS in enlarged endosomes and LAMP1 intracellular vesicles (Fig. 10A,B). Following stimulation with AM, CLR and CLRΔ9KR similarly trafficked to enlarged HRS-positive vesicles (Fig. 10A-C). However, both CLR and CLRΔ9KR could also be similarly detected in LAMP1-positive vesicles (Fig. 10A,C,D). Of note, we could also detect LAMP1 in HRS-positive vesicles ( Supplementary Fig. S12A,B).  (Fig. 11A,B). We then treated the same cells with AM (0-16 h) and levels of CLR, RAMP2 and TfR quantified by Western blotting (Fig. 11A,C,D). In control cells, AM induced significant degradation of both CLR and RAMP2 compared to the not treated control and AM-treated cells overexpressing HRS. In contrast, AM induced significant degradation of RAMP2, but not CLR in cells overexpressing HRS (Fig. 11A,C,D). Thus, overexpression of HRS does not prevent, but only impedes the degradation of CLR•RAMP2.

Discussion
In this study, we used HEK cells to detail the molecular mechanisms regulating the downregulation of the AM 1 receptor. Our results show that AM induces trafficking of CLR•RAMP2 to early endosomes and then onto lysosomes. Furthermore, CLR is trafficked to lysosomes following both transient and sustained stimulation with AM. We also show that AM induces ubiquitination of CLR on intracellular lysines residues and that although ubiquitin does not target CLR•RAMP2 for degradation, a basal hyperubiquitination mediated by HRS overexpression does prevent degradation.
Ubiquitination has been shown to play a major role in the regulation of many GPCRs (reviewed in refs 43,44 ). However, ubiquitination of CLR expressed with RAMP2 was unexpected, given that CGRP does not induce ubiquitination of CLR when expressed with RAMP1 14 . The kinetics of agonist-induced CLR ubiquitination was unexpected. Most previously reported agonist-induced ubiquitination of GPCRs occurs within minutes 28,30 . However, we only detected ubiquitinated CLR, 1 h post-stimulation. To our knowledge, the only other GPCR that undergoes delayed ubiquitination is the neurokinin 1 receptor. However, these studies were performed in a cell line that does not naturally express that receptor 45 . Our finding that AM induced ubiquitination of CLR in an epithelial (HEK) and endothelial cell (HMEC-1) suggests that this may be a universal mechanism that regulates AM 1 receptors.
We generated a mutant CLR lacking all intracellular facing lysine residues to study the role of CLR ubiquitination. We observed that AM did not induce ubiquitination of CLRΔ9KR, but this lack of ubiquitination did not affect the endocytic trafficking to lysosomes. This finding was not unexpected as although ubiquitination does affect the trafficking of some GPCRs 28,30 , it does not regulate the lysosomal targeting of others 14,31,33 . We also observed a basal level of CLR ubiquitination, indicating that perhaps CLR is constitutively ubiquitinated. Again, this is not unusual as PAR1 is constitutively ubiquitinated in order to prevent its internalization 46 . The duration of agonist exposure in known to influence ultimate fate of GPCRs. Transient exposure of CLR•RAMP1 to CGRP results in efficient recycling 14,15 . However, continued CGRP exposure promotes trafficking to lysosomes 14 . The trafficking of the neurokinin 1 receptor can be similarly modified 45,47 . CLR•RAMP3, the other receptor for AM, has been shown to internalise and recycle back to the cell-surface following agonist exposure 24 . Our results indicate that once activated CLR•RAMP2 is degraded. This was unexpected as we had expected CLR•RAMP2 to be similarly regulated by endothelin-converting enzyme 1, especially as this endosomal peptidase cleaves AM in a pH-dependent manner, similar to that observed with CGRP (Cottrell, unpublished observation). However, the lack of CLR•RAMP2 recycling is in agreement with a previous study where it has been reported that following a transient stimulation with AM (30 min), CLR•RAMP2 is internalised and less than 10% of CLR•RAMP2 recycles back to cell-surface after 2 h 25 . Thus, it appears that in addition to modifying the pharmacology of CLR 13 , RAMPs play an important role in the post-endocytic sorting of CLR following agonist stimulation and can change the fate of CLR•RAMP complex.
As CLR colocalized with the lysosomal marker LAMP-1, we expected that degradation of CLR•RAMP2 would be prevented by inhibitors of lysosomal peptidases. However, the lysosomal protease inhibitors cocktail that completely prevented CGRP-induced degradation of CLR•RAMP1 14 , did not affect AM-induced degradation of CLR•RAMP2. Furthermore, leupeptin that has been shown to prevent degradation of other GPCRs 41 , did not prevent degradation of the AM 1 receptor. This result suggests that GPCRs may be degraded by different repertoires of peptidases. Studies have shown that δ-opioid receptor (δ-OR) degradation is prevented by MG-132, which is often used to inhibit the proteolytic activity of the proteasome 33,39 . However, MG-132 has been found to inhibit certain cathepsins e.g. B, L and K found in lysosomes [48][49][50] . Our data using the highly specific proteasome inhibitor, epoxomicin, suggests the proteasome has no role in the degradation of CLR•RAMP2.
Ubiquitin-dependent alterations in mitogenic signaling have been observed for other cell-surface receptors 36, 38 . Ubiquitination of insulin-like growth factor 1 receptor is an absolute requirement for insulin-like growth factor-induced ERK activation 36 . We were unable to detect any changes in AM-induced mitogenic signaling, but that does not rule out the fact that other AM-induced signaling cascades may be altered. Ubiquitination has been shown to participate in many different phases of GPCR regulation, such as promoting 27 or preventing internalisation 46 , targeting to lysosomes for degradation 28,30 and increasing the rate of degradation 31 . We examined if AM promoted degradation of CLR•RAMP2 and CLRΔ9KR•RAMP2 with different kinetics by examining degradation at early time points. However, we did not observe any significant differences in the degradation of CLR, CLRΔ9KR or RAMP2 in HEK cells or HMEC-1 cells. This finding was unexpected as ubiquitination usually either regulates internalization, lysosomal targeting or degradation kinetics 27,28,30,31,46 .
HRS overexpression regulates the degradation of the δ-OR 20 , protease-activated receptor 2, the CGRP receptor (CLR•RAMP1) 19 and the recycling of the β 2 -adrenoceptor 18 . The pronounced difference in the sensitivity of CLR•RAMP2 and CLRΔ9KR•RAMP2 to HRS overexpression indicates that ubiquitination plays a key role in the movement of AM 1 receptors through the endocytic machinery. However, as the lysine defective mutant is still degraded, an alternative pathway must be open to the receptor if it is not ubiquitinated. One possibility is that it could follow a similar pathway to PAR1 which has been shown to be degraded in lysosomes in an HRSand ubiquitin-independent but ESCRT-III-dependent mechanism 51,52 , although both CLR and RAMP2 lacks the Figure 9. Colocalization of CLR and RAMP2 with HRS and effect of HRS expression on the basal levels of CLR ubiqutination. (A) HEK cells expressing CLR•RAMP2 were live labeled with an antibody to the extracellular epitope tag of RAMP2 (rabbit-Myc, rMyc) and fixed (control) or left for 2 h at 37 °C (2 h Recovery) before fixation. RAMP2 and HRS were localized using immunofluorescence. In control cells, RAMP2 was detected at the cell-surface (arrowheads). Following 2 h incubation, RAMP2 was also at the cell-surface (arrowheads), but could be detected in intracellular HRS-positive vesicles. (B) HEK cells expressing CLR•RAMP2 and empty vector (pcDNA3, control) or Myc-HRS were fixed permeablized and CLR (RK11) and HRS localized using immunofluorescence and confocal microscopy. In control cells CLR was detected at the cell-surface (arrowheads). In cells overexpressing HRS-myc, CLR was detected at the cell-surface, but also in enlarged HRS-positive endosomes. (C) HEK cells expressing CLR•RAMP2 and either mock transfected (control) or transfected a non-targeting pool siRNA (Scrambled) or siRNA to HRS (siRNA #1 and #2) were fixed and CLR and HRS localized by immunofluorescence and confocal microscopy. In all conditions CLR was detected at the cell-surface and HRS was detected in intracellular vesicles, with the exception of siRNA#1-and #2-treated cells were limited detection of HRS was possible due to knockdown. n = 3. Scale bar, 10 µm. (D) HEK cells expressing CLR•RAMP2 and either empty vector (pcDNA3, control, +vc), Myc-HRS, mock transfected (Mock, control) or transfected with a non-targeting pool siRNA (Scr) or siRNA to HRS (siRNA #1 and #2) were lysed, CLR immunoprecipitated (mouse HA, mHA) and Western blots probed using antibodies to ubiquitin and CLR (mHA). There were no signals for CLR in vector control cells, confirming antibody specificity. In Scr, siRNA#1 and siRNA#2 cells the basal levels of ubiquitination were similar to control (Mock). In contrast, in cells overexpressing HRS, the basal level of ubiquitinated CLR was significantly increased over control (+vc). (E) Quantification of HRS expression. (F) Quantification of ubiquitinated CLR. Data were examined using ANOVA and Student-Newman-Keuls post-hoc test, n = 4. *p < 0.05; **p < 0.01; ***p < 0.001. HEK-vc = HEK-vector control.
YPX(3)L motif necessary for the interaction with ALIX (an ESCRT-III-interacting protein). However, this still reinforces the idea that similar to the protease-activated receptors the AM 1 receptor is a one-shot receptor because irrespective of whether the receptor is ubiquitinated or not, it is destined for degradation. Deletion of the UIM of HRS did not rescue CLR•RAMP2 degradation indicating HRS is not the endosomal sorting protein responsible for the recognition of ubiquitinated CLR. Numerous endosomal sorting proteins contain UIMs including EPS15, epsins, the ESCRT-0 component, STAM, the ESCRT-I component TSG101, the ESCRT-II component VPS36, as well as EPS15b and GGA3 (reviewed in ref. 53 ). Thus, it is plausible that anyone of these proteins is responsible for the recognition and sorting of ubiquitinated AM 1 receptors.
Considering that degradation of other GPCRs and their lysine-deficient mutants is inhibited by HRS knockdown 19,20 , it was surprising that knockdown of HRS had no effect on AM-induced degradation of CLR•RAMP2 or CLRΔ9KR•RAMP2. Furthermore, CLR•RAMP2 also localized to enlarged HRS-positive vesicles following HRS overexpression, similar to protease-activated receptor 2, CLR•RAMP1 and δ-OR 19,20 . Investigation into the effect of HRS overexpression on AM-induced trafficking revealed that both CLR and CLRΔ9KR accumulate in HRS-and LAMP1-positive vesicles, although it was clear that both receptors predominantly colocalized with HRS. It was surprising that there was no difference in the colocalization of CLR and CLRΔ9KR with LAMP1 and that the colocalization of CLRΔ9KR with LAMP1 was far less than in the absence of HRS overexpression given the HRS overexpression did not affect AM-induced degradation of CLRΔ9KR. It should also be noted that we used antibody-tagged receptors in our immunofluorescence experiments (to circumvent the high level of CLR colocalization with HRS under basal conditions) and so we visualized the antibodies used for tagging and not the receptors themselves. Therefore, we cannot rule out that the antibodies using for tagging the receptors are degraded at similar rates, whereas the receptors themselves are not. Furthermore, in our biochemical experiments, we cannot rule out pre-lysosomal proteolysis of the epitope tag used to detect both CLR and CLRΔ9KR.
Unlike other receptors that have been shown to be regulated by HRS overexpression it is clear from our results that HRS overexpression does not prevent degradation of CLR•RAMP2, but only delays degradation. Thus, it is possible that degradation of other GPCRs such as the δ-OR 20 and PAR2 19 is only retarded, as in those investigations longer time points (>4 h) were not investigated.
Our discovery that CLR•RAMP2 constitutively traffics to HRS-positive endosomes and HRS overexpression promotes hyperubiquitination of CLR under basal conditions could provide an explanation for the differences observed in the degradation of CLR•RAMP2 and CLRΔ9KR•RAMP2 (Fig. 12). If HRS recruits an E3 ligase, such as UBE4B 21 that ubiquitinates CLR, then the hyperubiquitination that occurs during the constitutive trafficking of CLR•RAMP2 could perturb the endocytic transport of CLR•RAMP2 to the lysosome following activation. A similar effect on the ubiquitination state and degradation of PAR2 is observed following overexpression of a dominant-negative mutant of the deubiquitinating enzyme, associated molecule with the SH3 domain of STAM (AMSH) 54 . In contrast, trafficking of non-ubiquitinated CLRΔ9KR•RAMP2 is compromised by disruption of the endocytic network by HRS overexpression, but degradation can still occur by incorporation of LAMP1-positive vesicles into the MVB. HRS knockdown on the other hand, would not result in increased ubiquitination and wild-type receptors would either traffic as normal or via the route of non-ubiquitinated mutant CLRΔ9KR•RAMP2. Our data shows that basal ubiquitination of CLR•RAMP2 can be regulated by HRS expression and that basal ubiquitination levels regulate agonist-induced degradation. Plasmids. cDNA encoding rat CLR has been described 14,55 . cDNA encoding rat RAMP2 was obtained by RT-PCR from rat heart using Trizol ® (Invitrogen) and Taqman Reverse Transcription reagents (Applied Biosystems, Warrington, UK) according to the manufacturer's guidelines. An N-terminal Myc tag was added to RAMP2 by PCR. A single vector (pcDNA5/FRT) expressing both rat CLR and RAMP2 was created as described for CLR and RAMP1 14 . A CLR mutant in which all predicted intracellular lysines were replaced with arginines (designated CLRΔ9KR) was generated using a mega-primer PCR method. Lentiviral vectors expressing either CLR or CLRΔ9KR and RAMP2 were generated by releasing expression cassettes from existing plasmids by restriction digest, followed by cloning into a circularized pLenti6.3V5 (Invitrogen). All PCR amplified constructs were sequenced to verify integrity. Primer sequences are available on request. pcDNA3.0-Myc-HRS and Myc-HRSΔUIM were a kind gift from Professor M. von Zastrow (University of California, San Francisco, CA).

Lentivirus Production and Transduction.
Immunofluorescence and Confocal Microscopy. Following drug treatments cells were washed in 100 mm PBS, pH 7.4, and fixed in PBS containing 4% paraformaldehyde, pH 7.4 (20 min, 4 °C). Cells were washed with PBS containing 0.1% saponin and 2% normal horse serum for 30 min. Proteins were localized using the primary antibodies CLR (RK11, 1:2000), EEA1 (1:500), LAMP1 (1:1000), rabbit or mouse anti-HRS (1:100) (overnight, 4 °C). Cells were washed and incubated with secondary antibodies coupled to fluorescein isothiocyanate, rhodamine Red-X, or Cy5 (1:500, 2 h, room temperature). To examine trafficking of CLR•RAMP2 from the cell-surface, CLR, CLRΔ9KR and RAMP2 were labeled by incubating cells with antibodies to the extracellular epitope tags (CLR, rat anti-HA, 1:1000 or goat anti-HA.11, 1:100; RAMP2, rabbit anti-c-Myc, 1:1000; 30 min, 37 °C). Cells were washed with PBS, stimulated with AM, fixed and incubated with appropriate secondary antibodies. Cells were observed with a Zeiss laser-scanning confocal microscope (LSM Meta 510 or LSM 510) using a Plan-Apochromat 63x/1.4 Oil DIC objective. Alternatively, cells were observed with a Nikon Eclipse Ti laser-scanning confocal microscope using a 100x/1.45 Oil DIC N2 objective. Images were collected at a zoom of 1-2 and an iris of <3 μm, and at least five optical sections were taken at intervals of 0.5 μm. Single sections are shown. Images were processed using ImageJ and Adobe Photoshop software. Colocalization of proteins was determined using the NIS-Elements AR software.

Statistics.
Results are expressed as mean ± S.E.M. of n ≥ 3 experiments and were compared by Student's t test. Differences among multiple groups were examined using ANOVA and Student-Newman-Keuls post-hoc