c-Cbl-Dependent Monoubiquitination and Lysosomal Degradation of gp130

Abstract

Interleukin 6 (IL-6), a pleiotropic cytokine, functions in cells through its interaction with its receptor complex, which consists of two ligand-binding α subunits and two signal-transducing subunits known as gp130. There is a wealth of studies on signals mediated by gp130, but its downregulation is less well understood. Here we found that IL-6 stimulation induced lysosome-dependent degradation of gp130, which correlated with an increase in the K63-linked polyubiquitination of gp130. The stimulation-dependent ubiquitination of gp130 was mediated by c-Cbl, an E3 ligase, which was recruited to gp130 in a tyrosine-phosphorylated SHP2-dependent manner. We also found that IL-6 induced a rapid translocation of gp130 from the cell surface to endosomal compartments. Furthermore, the vesicular sorting molecule Hrs contributed to the lysosomal degradation of gp130 by directly recognizing its ubiquitinated form. Deficiency of either Hrs or c-Cbl suppressed gp130 degradation, which leads to a prolonged and amplified IL-6 signal. Thus, our present report provides the first evidence for involvement of a c-Cbl/SHP2 complex in ubiquitination and lysosomal degradation of gp130 upon IL-6 stimulation. The lysosomal degradation of gp130 is critical for cessation of IL-6-mediated signaling.

Interleukin 6 (IL-6) is a typical cytokine and mediates a variety of biological activities, including cell proliferation or differentiation, acute-phase reaction, and inflammation (3, 5, 35, 38, 55), in a wide range of tissues. The IL-6 receptor system is composed of two ligand-binding alpha receptor subunits, IL-6Rα, and two gp130 subunits; gp130 is a type I transmembrane glycoprotein and functions in the receptor complexes of other IL-6 family cytokines, including IL-11, LIF, OSM, CNTF, and CT1 (20, 21, 37, 57, 73). While it is soluble, the extracellular portion of IL-6Rα (soluble IL-6R [sIL-6R]) binds IL-6, and gp130 transduces the signal (25). The binding of IL-6 to its receptor induces homodimerization of gp130, which results in the transphosphorylation and activation of associated Janus tyrosine kinases (Jaks) and the phosphorylation of gp130 at multiple tyrosine residues (50, 67). Tyrosine phosphorylation creates special docking sites for signaling molecules containing SH2 and/or TKB (tyrosine kinase-binding) domains. gp130-mediated signals activate two major signaling cascades, the STAT3 and the MAP kinase pathways (22, 42). Among the six tyrosine residues within the cytoplasmic region of gp130, the five phospohotyrosines most proximal to the C terminus bind STAT3. STAT3 forms a homodimer upon phosphorylation and eventually translocates to the nucleus (9, 66). Thus, proliferation- and/or differentiation-promoting signals from gp130 are mediated by the four most C-terminal tyrosine residues. In contrast, the other tyrosine residue, tyrosine 759 (Y759, corresponding to Y757), which is the second-closest tyrosine to the membrane, mediates proliferation- and/or differentiation-inhibiting signals (53, 54) through the binding of a tyrosine phosphatase, SHP2. The SHP2 signal eventually activates the MAP kinases, ERK1/2, and SOCS3, which is transcriptionally induced by IL-6, LIF, and other factors that activate STAT3 and has been shown to regulate the strength of cytokine signals by inhibiting the Jaks (63, 68). Despite these findings, the details of how Y759 mediates the inhibitory signal through SHP2 activity have remained unclear.

The duration and strength of receptor-mediated signals are tightly regulated. This is accomplished by terminal signal inactivation via the endocytosis and degradation of activated receptors and of their associated signaling proteins. Following ligand binding, growth factor receptors, such as the epidermal growth factor receptor (EGFR), are rapidly internalized from the cell surface via several mechanisms, including clathrin-dependent endocytosis (31, 40). Internalized receptors are initially delivered to early endosomes, which mature into late endosomes and multivesicular bodies (MVBs). EGFRs are sorted into intraluminal vesicles within the MVBs as MVB cargoes. An alteration in their topology allows the cargoes to be eventually destroyed by the lysosomes (19, 23). These processes are essential to avoid constitutive signaling that would lead to tumorigenesis.

For many receptors, ubiquitination plays a key role in endosomal trafficking (33, 49). Ubiquitin (Ub) is an evolutionarily conserved protein consisting of 76 amino acids that can be covalently attached to a target protein via a process known as ubiquitination (32, 61). Attachment of a single Ub to a target lysine residue is called monoubiquitination; the formation of Ub chains is called polyubiquitination. The type of Ub modification largely determines the fate of the ubiquitinated proteins. Monoubiquitination is associated with the endocytosis and lysosomal sorting of plasma membrane proteins (27), but the formation of lysine 48 (K48)-linked polyubiquitin chains is the primary signal that targets the protein to the 26S proteasome for degradation (49, 71). K63-linked polyubiquitin chains also serve nonproteolytic functions, including acting as a signal for endocytosis, DNA repair, and kinase activation, although these processes may themselves be involved in proteolytic functions. Regardless of these complexities, it is clear that these two forms of ubiquitination regulate different cellular processes that may be indirectly involved with proteolytic degradation (34).

Specificity in the ubiquitination process is mediated by E3 ligases (15). The Cbl family of Ub ligases plays pivotal roles in the polyubiquitination of EGFR (36). c-Cbl binds directly to phosphorylated EGFRs via its TKB domain, while the RING finger domain of Cbl recruits Ub-conjugating enzymes (E2, Ubc) and mediates the transfer of Ub to the receptor (26, 62). However, the role(s) of ubiquitination and E3s in receptor trafficking and in the sorting of cytokine receptors has been poorly understood.

In studies to find molecules involved in cytokine/growth factor signaling, we identified Hrs as a signaling molecule related to IL-2 (4). Hrs is an evolutionarily conserved vesicular sorting protein belonging to the endosomal-sorting proteins required for transport (ESCRT) (72). Hrs possesses a FYVE domain, which enables the molecule to anchor to the internal surface of the lipid bilayer (44, 45), and a UIM domain, through which Hrs binds monoubiquitinated receptors (59, 65). By associating with ubiquitinated receptors such as the MET, E-cadherin, and EGFRs, Hrs plays a critical role in their degradation through the endosomal/lysosomal pathway (41, 47, 69). Given that ligand-activated EGFRs accumulate within the early endosomes of cells deprived of Hrs, Hrs appears to be involved in ubiquitinated receptor sorting (58, 64, 69). These findings provided a strong rationale for exploring whether the monoubiquitination of cytokine receptors would be followed by endosomal sorting.

Here, we show that gp130 is modified mainly with K63-linked polyubiquitin chains upon stimulus by IL-6. The polyubiquitination of gp130 is triggered by phosphorylation at Y759, resulting in an association with SHP2 and then c-Cbl. Following the ubiquitination, gp130 undergoes Hrs-dependent endosomal sorting and lysosomal degradation, which is essential for termination of IL-6 signaling.

MATERIALS AND METHODS

Reagents and antibodies.

Recombinant human IL-6 (300 ng/ml) and sIL-6R (500 ng/ml) were purchased from Peprotech. Lys48-linked diubiquitin and Lys63-linked diubiquitin were from Boston Biochem. Antibodies used for immunoprecipitation and Western blotting were as follows. Antibodies purchased from Santa Cruz Biotechnology were rabbit anti-gp130 (M-20), anti-STAT3 (C-20), anti-SHP2 (C-18), anti-SOCS3 (H-103), anti-early endosome antigen 1 (anti-EEA1) (N-19), and anti-CD63 (MX-49) antisera and anti-Ub monoclonal antibody (MAb; P4D1). Anti-Myc, antihemagglutinin (anti-HA), and anti-phospho-STAT3 MAbs were from Cell Signaling Technology. Anti-V5 MAb was from Invitrogen. Anti-FLAG-M2 MAb and mouse anti-α-tubulin antisera were from Sigma. A rat anti-Hrs MAb was described elsewhere (4).

Plasmids.

A wild-type gp130 expression vector, pEF4-Myc-gp130^WT, was generated by inserting mouse gp130 cDNA into pEF4/Myc-HisA (Invitrogen). The mutant gp130 constructs, in which each of the six cytoplasmic tyrosine residues (Y) was replaced with a phenylalanine (F), were generated using a QuikChange site-directed mutagenesis kit (Stratagene) according to the manufacturer's instructions. The resulting constructs were named pEF4-Myc-gp130^Y757F, pEF4-Myc-gp130^Y765F, pEF4-Myc-gp130^Y812F, pEF4-Myc-gp130^Y859F, pEF4-Myc-gp130^Y904F, and pEF4-Myc-gp130^Y914F, with the superscripted number giving the amino acid position of the mutated tyrosine. p3xFLAG-Cbl-b was generated by a PCR-based procedure using p3xFLAG-CMV-10 (Sigma). pcDNA3.1-c-Cbl was a generous gift from Yosef Yarden (Weizmann Institute of Science). p3xFLAG-c-Cbl^C381A and p3xFLAG-c-Cbl^dTKB, for expression of a catalytically inactive and a TKB domain-deleted c-Cbl, respectively, were generated using site-directed mutagenesis and a PCR-based method. For the wild-type human SHP2 expression vector, pcDNA3.1-V5-SHP2 was constructed using pcDNA3.1-V5-HisA (Invitrogen). Vector pcDNA3.1-V5-SHP2^dSH2x2 encodes a mutant of human SHP2 lacking both SH2 domains. pcDNA3.1-HA-Ub (WT-Ub) and pcDNA3.1-HA-monoUb were, respectively, an expression vector for an HA-tagged wild-type Ub and a mutant Ub (Mono-Ub) in which all seven lysine (K) residues, K6, K11, K27, K29, K33, K48, and K63, were replaced by arginines to prevent the formation of polyubiquitin chains. pcDNA3-HA-Ub^K48R (K48R-Ub) and pcDNA-HA-Ub^K63R (K63R-Ub) expressed HA-tagged Ub mutants carrying a lysine-to-arginine mutation (K48R, K63R). pME18S/Myc-NT/Hakai expression vector was a gift from S. Higashiyama (Ehime University, Japan). DNA sequences for the newly developed constructs were verified using a DNA sequencer (ABI3100; Applied Bioscience) and a BigDye Terminator kit (Applied Biosciences).

Cell culture and transient transfections.

The cell lines used here were HeLa and 293T cells (American Type Culture Collection) and two MEF cell lines (gp130^H/H and gp130^F759/F759) derived, respectively, from knock-in mice expressing human wild-type gp130 (gp130^H/H) or a mutant gp130 (gp130^F759/F759) (6), a c-Cbl-deficient (c-Cbl^−/−) MEF, a wild-type c-Cbl-tranfected c-Cbl^−/− (c-Cbl^WT) MEF (17), and a retroviral packaging cell line, Phoenix-Ampho (Orbigen), all of which were maintained in Dulbecco's modified Eagle's medium containing 10% (vol/vol) fetal calf serum (FCS) and antibiotics at 37°C in 7% CO₂ in humidified incubators. BAF/gp130 (a gift from T. Taga, Kumamoto University), a mouse BAF/B03 pro-B-cell line stably expressing human gp130, was maintained in RPMI 1640 medium containing 10% FCS and 10% (vol/vol) WEHI-231 culture supernatant. To transfect the cells with plasmids, FuGENE 6 (Roche Diagnostics) or Lipofectamine 2000 (Invitrogen) was used according to the manufacturer's instructions. For IL-6 stimulation experiments, HeLa cells were first pretreated with 200 μM E-64-d-20 μM pepstatin A (Pep) and/or 10 μM epoxomicin (Epx; Peptide Institute Inc.) along with cycloheximide (CHX; Sigma) in starvation medium (Dulbecco's modified Eagle's medium containing 0.1% bovine serum albumin) for 1 or 3 h and then stimulated with the combination of recombinant human IL-6 and sIL-6R.

Immunoprecipitation and immunoblotting.

Cells were lysed in NP-40 lysis buffer (1% Nonidet P-40, 40 mM Tris-HCl [pH 7.5], 150 mM NaCl, 2 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, 20 μg/ml aprotinin, 2 mM Na₃VO₄, 50 mM NaF, 30 mM Na₄P₂O₇, 2 mM dithiothreitol, 0.25 mM Na-p-tosyl-l-lysine chloromethyl ketone, 0.25 mM N-tosyl-l-phenylalanine chloromethyl ketone). For Ub experiments, the lysis buffer was also supplemented with 10 mM N-ethylmaleimide. The cell lysates were precleared of cellular debris by centrifugation (10,000 × g) for 30 min at 4°C and were then subjected to immunoprecipitation with antibodies immobilized on protein A-Sepharose beads (Amersham Biosciences) at 4°C overnight. The monoclonal and polyclonal antibodies are described above. After extensive washes with the lysis buffer, the immunoprecipitates were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred onto polyvinylidene difluoride membranes (Millipore). After being blocked with 5% nonfat milk in Tris-buffered saline containing 0.1% Tween 20, the membranes were probed with the indicated primary antibodies. After three washes, the membranes were incubated with horseradish peroxidase-conjugated secondary antibodies (horse anti-mouse immunoglobulin G [IgG] and goat anti-rabbit IgG; Cell Signaling Technology). Signals were visualized with SuperSignal West Pico chemiluminescent substrate (Pierce), and digital images were collected and analyzed with a Lumi-Imager F1 system (Roche). The signal density was analyzed with Image-J software, and the protein densities relative to the initial expression levels were determined.

Sample preparation for quantitative MS.

Absolute quantification of Ub and Ub linkages on gp130 was determined using iTRAQ methodology (60). K48- and K63-linked diubiquitins were used as internal controls. Concentrations of proteins were determined by scanning of SDS-PAGE gels after Coomassie blue staining and comparison to a dilution series by use of bovine serum albumin. Protein digestion in solution was performed as described previously (11), with minor modification. Immunopurified gp130 proteins (10 μg) were precipitated by six volumes of cold acetone (at −20°C for 2 h) and were resuspended in 22 μl of 0.5% RapiGest surfactant (Waters) reconstituted with 0.5 M triethylammonium bicarbonate (pH 8.5). After being boiled for 5 min, the proteins were reduced [4 mM Tris-(2-carboxyethyl) phosphine (TCEP) at 60°C for 1 h], alkylated (8 mM methyl methanethiosulfonate at 25°C for 15 min), and digested with trypsin (1:2 [wt/wt] at 37°C for 18 h). The resulting peptides were then labeled with the following multiplex iTRAQ reagents (iTRAQ reagent kit; Applied Biosystems): iTRAQ114 for K48-linked diubiquitin, iTRAQ115 for K63-linked diubiquitin, iTRAQ116 for untreated gp130, and iTRAQ117 for IL-6-treated gp130. After being labeled for 1 h at 25°C, the residual reagent was quenched by adding 200 μl of water. The samples were concentrated to 20 μl by use of a SpeedVac apparatus and were acidified with 20 μl of 1% trifluoroacetic acid (TFA) to degrade the RapiGest surfactant. The labeled peptides were further diluted to 0.25 μM with 0.5% TFA. Each sample (1 pmol) was analyzed on an ABI 4800 matrix-assisted laser desorption ionization-time of flight/time of flight (MALDI-TOF/TOF) proteomics analyzer (Applied Biosystems), and it was found that almost all peptides were fully labeled using iTRAQ at the N-terminal and lysine side chains without side reactions. Then, four iTRAQ-labeled samples were combined (5 pmol of gp130 and 2 pmol of Ubs) and the peptides were desalted using C18 Ziptips (Millipore) prior to liquid chromatography-tandem mass spectrometry (LC-MS/MS).

RNA interference (RNAi).

pSIREN-RetroQ-Hrs, which expresses a human Hrs-specific short hairpin RNA in a retrovirus-based system, was generated using pSIREN-RetroQ (BD Biosciences). The target sequence for the human Hrs cDNA spans bp 302 to 320 (5′-AGGTAAACGTCCGTAACAA-3′). For a control, pSIREN-RetroQ-Renilla, which targets a sequence within the Renilla luciferase gene (413 to 434 bp; 5′-GCAATAGTTCACGCTGAAAAG-3′), was used (69). Replication-defective retrovirus was harvested from the cell culture supernatants of Phoenix-Ampho packaging cells, and virus particles were concentrated 30-fold by centrifugation at 6,000 × g for 16 h. The titer of the concentrated virus was monitored using 293T cells, and an effective number of Pfu polymerases (1 Pfu/cell) was used to transduce the HeLa cells. Transduced cells were selected for an additional 5 days in 2 μg/ml puromycin (Sigma) and then used for the assays.

LC-MS/MS analysis of gp130 ubiquitination.

Peptide separation was performed on a direct nano-LC system equipped with a nano-LC/MALDI spotting device (DiNA MaP system; KYA Tech, Japan). The peptides were injected and captured onto a 0.8- by 3-mm trap column (HiQ sil C18-3; KYA Tech, Japan) and then eluted onto a 0.5- by 50-mm analytical column (HiQ sil C18-W3; KYA Tech, Japan) by use of an automated binary gradient (300 nl/min) from solvent A (0.1% TFA-2% acetonitrile [ACN]) to 50% solvent B (0.1% TFA-80% ACN) over 50 min and then to 50 to 100% solvent B for 10 min. Column effluent was mixed with matrix solution (4 mg/ml α-cyano-4-hydroxycinnamic acid, 2% [wt/wt] ammonium citrate, 0.1% TFA, and 70% ACN) at 1,300 nl/min and spotted onto the MALDI plate (30 s/spot for a total of 192 spots). MALDI plates were analyzed on an ABI4800 MALDI-TOF/TOF proteomics analyzer (Applied Biosystems). To obtain precursor masses, the mass spectrometer was set to perform data acquisition in the positive ion mode, with a selected mass range of m/z 600 to 4,000. MS/MS spectra were acquired in a data-dependent acquisition mode that automatically selected and fragmented the 30 most intense peaks from each MS spectrum by use of 4000 Series Explorer software (version 3.5; Applied Biosystems). For each MS/MS spectrum acquired, signature-ion peak areas at m/z 114.1, 115.1, 116.1, and 117.1 were extracted using ProteinPilot software version 2.0 (Applied Biosystems). We used a Ub peptide, TITLEVEPSDTIENVK (12 to 27 amino acids; referred to as TITLE), for calculation of total Ub levels. Precursor masses of iTRAQ-labeled TITLE, K48 linkages, and K63 linkages were detected at m/z 2,076.36, 1,749.12, and 2,532.55, respectively.

Immunofluorescence microscopy and immunolabeling.

HeLa cells were grown in 35-mm-diameter glass-bottomed dishes (Matsunami) at a density of 4 × 10⁴ cells/dish and transfected with Myc-gp130. After 48 h, the cells were washed twice with phosphate-buffered saline, fixed with 4% paraformaldehyde for 15 min, washed in wash buffer (0.1% Triton X-100-phosphate-buffered saline), and then blocked in the same buffer containing 10% FCS. For immunostaining, the fixed samples were incubated in wash buffer containing 5% FCS at 4°C overnight with the indicated primary antibodies. After three washes, the samples were further incubated with the secondary antibodies (anti-rabbit, -mouse, and -goat IgG conjugated with Alexa488, Alexa594, and Alexa633, respectively) (Molecular Probes) at room temperature for 1 h. Confocal images were obtained with a 510 META microscope and a 60/1.30 to 0.60 oil-immersion objective (Carl Zeiss).

Proliferation assay.

BAF/B03 cells transduced with pSIREN-RetroQ-Hrs or pSIREN-RetroQ-Renilla and named BAF/gp130^siHRS or BAF/gp130^siCont, respectively, were washed extensively with culture medium to remove the WEHI231 supernatant. Triplicate samples of equal numbers (5 × 10³) of cells were cultured in a 96-well microtiter plate with or without IL-6 for 72 h. [³H]thymidine (1 mCi) was added to each well 6 h prior to the termination of the incubation. The incorporated radioactivity level was determined with a 1405 microBeta liquid scintillation counter (Pharmacia).

Statistical analysis.

All results are presented as the means ± standard deviations of data from experiments performed in triplicate. Comparisons between groups were analyzed using Student's t test.

RESULTS

Degradation of gp130 is mediated by two distinct pathways.

We first investigated the degradation pathway of gp130. There are two protein degradation pathways, the proteasomal and the lysosomal (12). HeLa cells expressing gp130 were preincubated for 1 h in starvation medium. The cells were then treated with a proteasome inhibitor, Epx, and/or the lysosomal enzyme inhibitors E-64-d and Pep together with CHX and IL-6 for the times indicated. To stimulate the cells with IL-6 and thereby activate gp130-mediated signals, we used a combination of recombinant human IL-6 and sIL-6R. The CHX was added to the culture to stop de novo protein synthesis.

In the absence of IL-6 stimulation, 72% of the gp130 was degraded during 120 min of incubation. This degradation was almost completely inhibited by treatment with Epx but not by treatment with the lysosomal enzyme inhibitors (Fig. 1A). In contrast, after 120 min of IL-6 stimulation, 100%, 47%, 11%, or 4% of the gp130 was degraded when cells were treated, respectively, with dimethyl sulfoxide (DMSO), Epx, E-64-d plus Pep, or E-64-d plus Epx (Fig. 1B). These results suggest that gp130 is degraded through the proteasome-dependent pathway in the absence of IL-6 stimulation but that the lysosome-dependent pathway is mainly responsible for degrading gp130 following IL-6 stimulation.

FIG. 1.

Lysosome- and proteasome-dependent degradation of gp130. (A and B) Degradation of gp130 in the absence (A) or presence (B) of IL-6 stimulation. HeLa cells were pretreated for 1 h with media containing DMSO, 10 μM proteasome-specific inhibitor Epx, a combination of lysosome inhibitors E-64-d and Pep (200 μM each), or a combination of all the three inhibitors (Epx plus Pep plus E-64-d). At the beginning of the chase, CHX (25 μg/ml) was added to each culture samples. At 0, 30, 60, and 120 min time points, cells were harvested. Whole-cell lysates were prepared and analyzed by immunoblotting for gp130 and α-tubulin as indicated. Signals corresponding to two molecular-weight forms of gp130 were quantified using Image-J software. Relative amounts of gp130 were calculated after normalization by use of signals from α-tubulin and are expressed relative to the values from each initial amount (100%). For panel B, parallel cultures were stimulated by the combination of IL-6 (300 ng/ml) and sIL-6R (500 ng/ml) (IL-6/sIL-6R) in the presence of CHX.

K63-linked ubiquitination of gp130 upon IL-6 stimulation.

We examined ubiquitination of gp130 upon IL-6 stimulation. 293T cells were left untreated or were treated with IL-6/sIL-6R, and their lysates were immunoprecipitated with anti-gp130 antibody. The immunoprecipitates were then separated by SDS-PAGE and visualized by Coomassie blue staining or Western blotting with anti-Ub MAb (P4D1). Ubiquitinated gp130 was detected as smear bands by Western blotting, and levels significantly increased upon IL-6 stimulation. This result indicates that IL-6 stimulation promotes ubiquitination of gp130 (Fig. 2A).

FIG. 2.

Ub forms associated with gp130. (A) Detection of ubiquitinated gp130 following IL-6 stimulation. 293T cells were left untreated or were treated with IL-6/sIL-6R for 10 min at 37°C, and gp130 was purified from their lysates by immunoprecipitation with anti-gp130 antibody. Aliquots of the purified gp130 samples were subjected to SDS-PAGE analysis. The gels were stained with Coomassie blue or immunoblotted with anti-Ub antibody (P4D1). WB, Western blot. (B) Depiction of overall workflow for absolute quantification of Ub. The purified gp130 samples from non-IL-6-stimulated and IL-6-stimulated cells and control K48-di-Ub and K63-di-Ub cells were digested, labeled with multiplex iTRAQ, and subjected to LC-MS/MS analysis. The resulting peptides from each sample were labeled with iTRAQ114 to iTRAQ117. The labeled peptides corresponding to 2 pmol of Ubs and 5 pmol of gp130 were then mixed and analyzed by LC-MS/MS. Peptides from differently labeled samples are isobaric on LC and MS but yield specific reporter ions at m/z 114.1, 115.1, 116.1, or 117.1 following MS/MS. (C) The iTRAQ reporter ion regions from MS/MS spectra. The total Ub level was calculated using a Ub fragment (TITLE). Correct loading of two gp130 samples was confirmed with a gp130 peptide (832 to 847 residues termed SNQVLS). (D and E) Absolute amounts of total Ub and each Ub chain linkage (measured values are shown in panel E).

We next attempted to determine the ubiquitination pattern of gp130 by use of MS (56). Myc-tagged gp130 and HA-tagged Ub were cotransfected into 293T cells, and the cells were left untreated or were treated with IL-6/sIL-6R. Myc-tagged gp130 was immunopurified, denatured, and digested with trypsin. The resultant peptide samples were subjected to LC-MS/MS. More than 500 of the MS/MS spectra were acquired with high accuracy. Database searching revealed that the IL-6-stimulated sample contained predominantly gp130 (70.2% of sequence coverage) and Ub (80.3%). We observed that the IL-6-stimulated sample contained K63-linked Ub chains but not other Ub linkages (data not shown). To determine the amount of the K63-linked Ub chains associated with gp130, we next performed a quantitative MS analysis using iTRAQ multiplex reagents, including an isobaric tag (43, 48). Diubiquitins were used as controls (Fig. 2B). Total Ub levels were calculated with a linear Ub fragment (12 to 27 residues termed TITLE), since this fragment is efficiently trypsinized and is not utilized in K48 and K63 linkages. It was notable that each diubiquitin generated TITLE and a specific linkage at a 2:1 ratio. The Ub peptides were well resolved in both LC and MS. The iTRAQ reporter ions were monitored at MS/MS spectra and quantified by use of internal Ub peptides (Fig. 2C). Compared with unstimulated sample results, the total Ub level of gp130 from IL-6-stimulated cells was increased 3.76-fold (Fig. 2C to E). Approximately 19% of the total Ub amount was utilized for K63-linked chains, 4% for K48-linked chains, and remainder for mono/endcap Ub (Fig. 2E).

c-Cbl binds to SHP2 upon IL-6 stimulation and contributes to the ubiquitination of gp130.

We investigated which E3 ligase was responsible for ubiquitination of gp130. Since c-Cbl, an E3 ligase, is associated with SHP2 following stimulation by stromal cell-derived factor 1α (13) and since gp130 binds to SHP2 upon IL-6 stimulation (63, 68), we looked for a possible contribution of c-Cbl to gp130 ubiquitination in comparison with the results seen with other E3 ligases (Cbl-b and Hakai).

Transfection of c-Cbl apparently induced ubiquitination of gp130 in the presence or absence of IL-6/sIL-6R, whereas neither Cbl-b nor Hakai induced it (Fig. 3A). We next looked at the c-Cbl domains required for gp130 ubiquitination by use of the TKB domain- or RING domain-deleted c-Cbl mutant. Neither mutant induced ubiquitination of gp130 after IL-6 stimulation (Fig. 3B). These results suggest that gp130 specifically requires c-Cbl for its ubiquitination and that its RING and TKB domains are essential for this ligase activity.

FIG. 3.

c-Cbl, but not Cbl-b or Hakai, induces significant ubiquitination of gp130 upon IL-6 stimulation. (A) (Left panel) Catalytic function of c-Cbl and Cbl-b for the ubiquitination of gp130 with IL-6/sIL-6R. 293T cells were transfected with gp130^WT and Mono-Ub along with FLAG-tagged c-Cbl (c-Cbl^WT) and FLAG-tagged Cbl-b (Cbl-b^WT). V, vector control. At 48 h after the transfection, cells were stimulated with IL-6 (IL-6/sIL-6R) for 0 or 10 min. Cell lysates were subjected to immunoprecipitation (IP) for gp130 and were then immunoblotted with anti-HA MAb to detect ubiquitinated gp130. The same blot was analyzed with anti-FLAG MAb to detect c-Cbl^WT or Cbl-b^WT. Amounts of c-Cbl and Cbl-b in the cell lysates were also examined. (Center panel) Wild-type Myc-tagged Hakai ligase was examined. Experiments were performed as described for the left panel except that anti-Myc MAb was used to detect Hakai. (Right panel) Schematic structures of c-Cbl, Cbl-b, and Hakai. RF, RING-finger domain; Pro-rich, proline-rich domain; UBA, Ub-associated domain. Each number on the right indicates the amino acid length of the corresponding protein. (B) (Left panel) c-Cbl binds gp130 through its TKB domain and mediates ubiquitination of gp130. 293T cells were transfected with gp130 and Mono-Ub along with FLAG-tagged c-Cbl^WT, c-Cbl^C381A, or c-Cbl^dTKB or the vector control. At 48 h later, cells were stimulated with IL-6 for 0 or 10 min. Cell lysates were prepared and used for immunoprecipitation with anti-Myc MAb. Immunoblotting experiments were performed using anti-HA, anti-FLAG, and anti-Myc MAbs. Molecular masses of 250, 100, and 75 kDa are shown on the left. Data represent the results of at least three experiments. (Right panel) Schematic structures of c-Cbl^WT, c-Cbl^C381A, and c-Cbl^dTKB.

Next, we attempted to verify an interaction among gp130, SHP2, and c-Cbl. 293T cells were cotransfected with V5-tagged wild-type SHP2 (SHP2^WT) or SH2-deleted mutant SHP2 (SHP2^dSH2x2), FLAG-tagged wild-type c-Cbl (c-Cbl^WT) or c-Cbl mutant with amino acid substitutions (c-Cbl^Y700, c-Cbl⁷³¹, and c-Cbl^774F), and Myc-tagged wild-type gp130 (gp130^WT). After 10 min of stimulation with IL-6/sIL-6R, the cell lysates were immunoprecipitated with anti-Myc, anti-V5, or anti-FLAG antibody and then immunoblotted with anti-Myc, anti-V5, or anti-FLAG antibody. Either SHP2^WT or c-Cbl^WT was coimmunoprecipitated with gp130^WT, but c-Cbl^Y700, c-Cbl⁷³¹, c-Cbl^774F, and SHP2^dSH2x2 were not (Fig. 4A). Moreover, c-Cbl^WT was coimmunoprecipitated with SHP2^WT but SHP2^dSH2x2 was not, and a reverse precipitation procedure gave the same results (Fig. 4A). We confirmed that tyrosine phosphorylation of c-Cbl was detected upon IL-6 stimulation (Fig. 4A). These results suggest that gp130 binds directly to SHP2 upon IL-6 stimulation and that SHP2 also binds directly to c-Cbl upon IL-6 stimulation but that gp130 does not bind directly to c-Cbl.

FIG. 4.

IL-6 induces tyrosine-phosphorylation of c-Cbl and its association with gp130 through the tandem SH2 domains of SHP2. (A) Myc-tagged gp130^WT with FLAG-tagged wild-type c-Cbl (c-Cbl^WT) or c-Cbl with mutated phosphorylation sites (c-Cbl^Y700, c-Cbl⁷³¹, c-Cbl^774F) and V5-tagged wild-type SHP2 (SHP2^WT) or two SH2 domain-deleted mutant SHP2s (SHP2^dSH2x2) were introduced to 293T cells. At 48 h later, cells were starved for the serum for 3 h and then stimulated with IL-6/sIL-6R for 0 or 10 min. Cell lysates were subjected to immunoprecipitation (IP) by use of anti-FLAG or anti-V5 MAb and analyzed by immunoblotting with anti-Myc, anti-V5, or anti-FLAG MAb. “C” and “TCL” indicate lanes with parental cell control and the all-positive total cell lysate control, respectively. Data shown represent the results of at least three independent experiments. Schematic structures of the wild-type and deletion mutants of SHP2 and c-Cbl, namely, SHP2^WT, SHP2^dSH2x2, c-Cbl^WT, and c-^c-CblY700, c-Cbl⁷³¹, and c-Cbl^774F, are shown. (B) 293T cells were cotransfected with wild-type gp130 or various mutants of gp130 in the presence of c-Cbl^WT and Mono-Ub. At 48 h later, cells were stimulated with IL-6/sIL-6R for 0 or 10 min. Prepared cell lysates were used for immunoprecipitation of gp130. Results of monoubuiquitination of gp130 were examined by immunoblotting using anti-HA MAb. The same membranes were analyzed with anti-FLAG MAb and anti-SHP2 antiserum to monitor associated c-Cbl and (intrinsic) SHP2, respectively. Precipitated gp130 is represented with two Molecular mass markers. Amounts of intrinsic SHP2 and introduced c-Cbl in lysates were analyzed. Molecular mass markers of 250, 150, 100, and 75 kDa are shown on the left. Data represent the results of three sets of independent experiments. Schematic structures of wild-type gp130 and gp130 mutants with various mutations of tyrosine (Y) to phenylalanine (F) are shown in the lower panel. Wild-type gp130 (gp130^WT; lane 1), gp130^Y757F (lane 2), gp130^Y765F (lane 3), gp130^Y812F (lane 4), gp130^Y859F (lane 5), gp130^Y7904F (lane 6), and gp130^Y914F (lane 7) are shown. “EXT” and “TM” indicate extracellular and transmembrane regions of gp130, respectively.

To investigate further the contribution of c-Cbl to the ubiquitination of gp130, we used gp130 mutants in which each of six tyrosine residues was replaced with phenylalanine. All the mutants except gp130^Y757F bound to SHP2 10 min after IL-6 stimulation (Fig. 4B). The Y757F mutation caused a nearly complete loss of IL-6-induced ubiquitination of gp130 and was not coimmunoprecipitated with c-Cbl^WT (Fig. 4B, lane 2). The five other gp130 mutants showed ubiquitination patterns similar to those seen with gp130^WT and bound to c-Cbl^WT (Fig. 4B, lanes 1, 3, 4, 5, 6, 7). These results suggest that Tyr-757 of gp130 is primarily required for the interaction with SHP2 after IL-6 stimulation and that SHP2 then recruits c-Cbl for the ubiquitination of gp130.

Suppression of ubiquitination and degradation of gp130 in c-Cbl-knockout cells and gp130^F759/F759 knock-in cells.

As shown above, gp130 is degraded in lysosomes upon IL-6 stimulation, and Hrs is an endosomal sorting protein known to be required for lysosomal degradation of ubiquitinated receptor- tyrosine kinases (7, 39, 45, 47, 58). Hence, we investigated the relationship between c-Cbl-dependent ubiquination and lysosomal degradation of gp130. To examine the requirement of c-Cbl for gp130 ubiquitination and degradation, we used c-Cbl^−/− and c-Cbl^WT MEF cells. We confirmed that the c-Cbl^−/− MEF cells did not express c-Cbl but significantly expressed Cbl-b (data not shown). The two MEF cell lines were transfected with HA-tagged Mono-Ub, pretreated with CHX to prevent de novo synthesis of gp130, and stimulated with IL-6/sIL-6R for the indicated times after 1 h in starvation medium. Their lysates were immunoprecipitated with anti-gp130 antibody and immunoblotted with anti-HA, anti-SHP2, anti-c-Cbl, anti-Hrs, or anti-gp130. Following stimulation with IL-6, ubiquitination of gp130 was detectable, and SHP-2, c-Cbl, and Hrs were coimmunoprecipitated with gp130 in c-Cbl^WT MEF cells but not in c-Cbl^−/− MEF cells (Fig. 5A). The gp130 was significantly degraded upon IL-6 stimulation in c-Cbl^WT MEF cells but was scarcely degraded in c-Cbl^−/− MEF cells (Fig. 5A). IL-6-induced signals were also compared between c-Cbl^−/− and c-Cbl^WT MEF cells. The c-Cbl^−/− MEF cells showed prolongation of STAT3 phosphorylation and SOCS3 induction upon IL-6 stimulation (Fig. 5B). In similarity to the result described above, gp130^F759/F759 and gp130^H/H MEF cell lines transfected with HA-tagged Mono-Ub were examined for coimmunoprecipitation, ubiquitination, and degradation of gp130. Upon IL-6 stimulation, SHP2, c-Cbl, and Hrs were coimmunoprecipitated with gp130^H/H but not with gp130^F759/F759 MEF cells (Fig. 5C). The significant ubiquitination and degradation of gp130 were detectable in gp130^H/H but not in gp130^F759/F759 MEF cells, and the IL-6-induced signals were also prolonged in gp130^H/H MEF cells compared with gp130^F759/F759 MEF cells (Fig. 5D). These results suggest that c-Cbl association with SHP2 is essential for the IL-6-induced ubiquitination and degradation of gp130 that affects the IL-6 signal and that ubiquitinated gp130 binds to Hrs.

FIG. 5.

Impairment of gp130 ubiquitination, degradation, and prolongation of signaling in F759-knock-in and c-Cbl^−/− MEFs. (A) c-Cbl^−/− and c-Cbl^WT MEFs were transfected with Mono-Ub. At 48 h later, cells were stimulated with IL-6/sIL-6R in the presence of CHX for the indicated time periods. After the harvest, cell lysates were prepared and subjected to immunoprecipitation (IP) using anti-gp130 antiserum. Immunoblot analyses were performed using anti-HA MAb, anti-SHP2, anti-c-Cbl, anti-Hrs, anti-gp130, and antitubulin antisera. Relative gp130 protein amounts are plotted as percentages of initial amounts after normalization by use of α-tubulin. (B) Whole-cell lysates from the same samples were analyzed for SOCS3 induction and STAT3 activation by use of anti-SHP2, anti-Hrs, anti-STAT3, and anti-SOCS3 antisera and anti-phospho-STAT3 MAb (pSTAT3). Amounts of c-Cbl and STAT3 were monitored. Data represent the results of at least three independent experiments. (C) gp130^Y759F knock-in MEF cells manifest inefficient ubiquitination and degradation of gp130, as well as Hrs-binding defects. Two MEF cell lines derived from gp130 knock-in mice, gp130^H/H and gp130^F759/F759, were transfected with Mono-Ub. After 48 h, cells were first starved for the serum for 3 h and then stimulated with IL-6/sIL-6R. At each time point indicated, cells were harvested. Anti-gp130-directed precipitates were analyzed by immunoblotting with anti-HA MAb, anti-SHP2, anti-c-Cbl, anti-Hrs and anti-gp130 antisera. Each cell lysate was monitored for SOCS3 induction as well as Hrs and α-tubulin expression. Relative gp130 protein amounts were calculated after normalization by use of α-tubulin, and the results are indicated as percentages of the initial gp130 protein amount. (D) Whole-cell lysates were analyzed in a manner similar to that described for panel B.

Involvement of Hrs in IL-6-induced degradation of gp130.

Using retrovirus-mediated gene transfer, we knocked down Hrs expression in HeLa cells by use of Hrs RNAi and compared ubiquitination results for gp130. The cells pretreated with CHX were stimulated with IL-6/sIL-6R for the indicated times, and the expression levels of gp130 were determined by immunoblotting. Upon IL-6 stimulation, the amount of gp130 in control-RNAi-expressing HeLa cells gradually decreased and became undetectable 60 min after stimulation; in contrast, in the Hrs knockdown cells, gp130 degradation was significantly slower, and gp130 was still detectable at least 120 min after the stimulation (Fig. 6A). In further contrast, in the absence of IL-6 stimulation the amount of gp130 degradation was unchanged by the knockdown of Hrs expression (Fig. 6B), suggesting that Hrs is not required for the constitutive degradation of gp130. Furthermore, IL-6-induced STAT3 phosphorylation and SOCS3 expression were rapidly induced and significantly prolonged in the Hrs-RNAi HeLa cells compared with the control-RNAi HeLa cell results (Fig. 6A). Collectively, these results indicate that Hrs is required for the IL-6-induced degradation of gp130 and that Hrs-mediated degradation of gp130 contributes to suppression of the IL-6 signal.

FIG. 6.

Depletion of Hrs results in delay of degradation of gp130 upon IL-6 stimulation. (A and B) HeLa-HrsRNAi and HeLa-control cells were stimulated with IL-6/sIL-6R in the presence of CHX (A) or were left unstimulated in the presence of CHX (B). At the indicated time points after the addition of CHX plus IL-6/sIL-6R or CHX alone, cells were harvested and lysates were prepared. Immunoblotting analyses were carried out with anti-gp130, anti-phospho-STAT3, STAT3, SOCS3, Hrs, and α-tubulin antibodies. The signal density was measured, and the relative protein amounts (percentages of initial gp130 normalized by use of α-tubulin) are shown. (C and D) HeLa-HrsRNAi cells were analyzed for evidence of degradation of gp130 in a manner similar to that described for Fig. 1. Data represent the results of at least three independent experiments.

Next, we investigated whether mediated degradation of gp130 occurs in lysosomes or proteasomes by using each enzyme inhibitors. By use of a procedure similar to that represented by Fig. 1, Hrs-RNAi HeLa cells were left unstimulated or were stimulated with IL-6 and were then treated with a proteasome inhibitor (Epx) or lysosomal enzyme inhibitors (Pep and E-64-d). In the absence of IL-6 stimulation, degradation of gp130 was significantly suppressed by Epx treatment but not by treatment with the lysosomal enzyme inhibitors (Fig. 6C), suggesting that the constitutive degradation of gp130 is mainly mediated by proteasomes. On the other hand, although the lysosomal enzyme inhibitors (but not Epx) induced an inhibition of gp130 degradation upon IL-6 stimulation, as shown in Fig. 1, they showed no effect on the IL-6-induced degradation of gp130 in Hrs-RNAi HeLa cells (Fig. 6D). These results indicate that Hrs is involved in the lysosomal degradation of gp130 upon IL-6 stimulation but not in its proteasomal degradation, which is constitutively induced in the absence of IL-6 stimulation.

Hrs is required for the transport of gp130 from early endosomes to late endosomes.

Because the Hrs deficiency suppressed the IL-6-induced lysosomal degradation of gp130, we next examined the intracellular localization of gp130 and Hrs in HeLa cells after IL-6 stimulation by immunofluorescence confocal microscopy. After 1 h in starvation medium, control-RNAi HeLa cells expressing endogenous gp130 were stimulated with IL-6/sIL-6R for the indicated times and were stained for gp130, Hrs, EEA1, and CD63, a marker for late endosomes. After 10 min of stimulation, the cell surface gp130 was internalized and colocalized with EEA1 and Hrs in early endosomes and with CD63 in late endosomes; gp130 was no longer detectable after 60 min (Fig. 7A). A similar colocalization of gp130 and EEA1 was detected in the Hrs-RNAi HeLa cells, and gp130 remained colocalized with EEA1 for at least 60 min after stimulation (Fig. 7B). The EEA-1-positive vesicles containing gp130 were enlarged in the Hrs-RNAi HeLa cells and did not show CD63 expression (Fig. 7B, right panel). These results suggest that Hrs is involved in the transportation of internalized gp130 from early endosomes to late endosomes.

FIG. 7.

Depletion of Hrs results in the aberrant endosomal sorting of gp130. (A and B) HeLa-control cells (A) or HeLa-HrsRNAi cells (B) were cultured on glass-bottom dishes and transfected with Myc-tagged gp130. After 48 h, cells were starved for 1 h and stimulated with IL-6/sIL-6R in the presence of CHX for the indicated durations. (Left panels) After fixation, cells were immunolabeled with anti-Myc MAb (green), anti-Hrs (red), and anti-EEA (blue) antisera. (Right panels) In parallel experiments, fixed samples were immunolabeled with anti-Myc MAb (green), anti-Hrs (red), and anti-CD63 (blue). Arrowheads indicate colocalization of gp130 with EEA1 (left panel) or of gp130 with CD63 (right panel). Bars, 10 μm. Experiments were performed at least three times, and representative data from multiple cells are shown.

Biological significance of Hrs deficiency in IL-6 signaling.

We also examined the biological significance of Hrs deficiency in BAF/gp130 cells, which proliferate in response to IL-6 stimulation (52). BAF/gp130 cells were infected with retroviruses containing Hrs-RNAi or control RNAi and were selected with puromycin. They were stimulated with various concentrations of IL-6 for 6 h, and their growth curves were measured by [³H]thymidine incorporation. IL-6-dependent proliferation was significantly enhanced in Hrs-RNAi BAF/gp130 cells compared with the control RNAi BAF/gp130 cell results (Fig. 8A). This enhancement was similar to that observed in BAF/gp130^Y757F cells (which express gp130^Y757F). We then examined the phosphorylation status of STAT3 and the induction of SOCS3 in these cells. IL-6-induced STAT3 activation (i.e., phosphorylation) and SOCS3 expression were clearly prolonged in Hrs-RNAi BAF/gp130 cells compared with the control-RNAi BAF/gp130 cells (Fig. 8B). Prolonged STAT3 phosphorylation was also observed when BAF/gp130^Y757F cells were stimulated with IL-6. These results are compatible with those obtained with the HeLa cell lines (Fig. 6), suggesting that Hrs contributes to the negative regulation of cellular responses to IL-6 through the degradation of gp130.

FIG. 8.

Effects of Hrs deficiency on IL-6 signaling. (A) Accelerated IL-6-dependent proliferation of BAF/gp130 cells depleted of Hrs. BAF/gp130 cells transduced with retrovirus expressing Hrs-RNAi (BAF/gp130^HrsRNAi) or control-RNAi (BAF/gp130^control) were first starved for the starvation medium for 1 h. Cells were then cultured in the presence of various concentrations of IL-6/sIL-6R (0 to 500 ng/ml). After 18 h of culture in a 96-well microplate, cells were pulse-labeled with [³H]thymidine. Averages of the results of triplicate measurements of incorporated radioactivity are shown with standard deviations (vertical bars). (B) Hrs depletion induces prolonged STAT3 phosphorylation as well as enhanced SOCS3 induction in response to IL-6 stimulation. BAF/gp130^control and BAF/gp130^HrsRNAi cells were first starved as described for panel A and were then stimulated with IL-6/sIL-6R for the indicated time periods. Cells were harvested, and the resultant cell lysates were subjected to immunoblot analyses using anti-phospho-STAT3, anti-STAT3, anti-SOCS3, anti-Hrs, and anti-α-tubulin antisera. Data shown represent the results of at least three independent experiments.

DISCUSSION

Here we demonstrated that c-Cbl, an E3 ligase, is involved in the IL-6-dependent ubiquitination of gp130 and that IL-6- and c-Cbl-mediated ubiquitination of gp130 is required for lysosomal degradation of gp130. Although gp130 has been reported to be ubiquitinated regardless of ligand stimulation (12), the present study revealed that the c-Cbl-mediated ubiquitination of gp130 was detectable only when the cells were stimulated with IL-6/sIL-6R. In the absence of IL-6 stimulation, degradation of gp130 was also detected and was found to be inhibited by Epx, a proteasome inhibitor, but not by the lysosomal inhibitors, suggesting that the constitutive degradation of gp130 without IL-6 stimulation is mediated mainly by proteosomes whereas the IL-6-dependent degradation of gp130 was totally inhibited by the lysosomal inbibitors.

EGF and platelet-derived growth factor receptors are directly associated with c-Cbl after ligand binding, which induces their monoubiquitination at multiple lysine residues (28) and K63-linked polyubiquitination (36). We observed also that gp130 contains K63-linked Ub chains but not other Ub linkages. In the case of gp130, our coimmunoprecipitation study revealed that gp130 is not directly associated with c-Cbl. Given that gp130 is known to be associated with SHP2 (46, 53, 63) and that SHP2 is associated with c-Cbl in stromal cell-derived factor 1α-stimulated cells (13), we looked for an association of gp130 with c-Cbl via SHP2. We found that SHP2 binds c-Cbl upon IL-6 stimulation and that the gp130^F759/F759 mutant, which lacks the binding site for SHP2, was not ubiquitinated. Furthermore, a c-Cbl mutant that lacked its SHP2-binding site did not ubiquitinate gp130 upon IL-6 stimulation. These data indicate that SHP2 contributes to the recruitment of c-Cbl to gp130 for its ubiquitination upon IL-6 stimulation. Other E3 ligases (Cbl-b and Hakai) contain domains similar to those found with c-Cbl, including the Ub-associated, SH2, RING, zinc-finger, and proline-rich domains, and are ubiquitously expressed (14, 18). However, these ligases did not cause ubiquitination of gp130, suggesting that gp130 is ubiquitinated specifically by c-Cbl. Furthermore, c-Cbl-depleted cells showed impaired IL-6-induced degradation of gp130, indicating that c-Cbl-mediated ubiquitination of gp130 is important for its degradation following IL-6 stimulation.

We and others have reported that Hrs binds to ubiquitinated cargo proteins such as EGFR, E-cadherin, Met, hepatocyte growth factor receptor, and fibroblast growth factor receptor and is involved in regulatory mechanisms for their degradation (2, 8, 41, 69). The present study revealed that Hrs binds directly to ubiquitinated gp130 upon IL-6 stimulation. The confocal microscopic analyses showed colocalization between Hrs and gp130 in EEA1-positive early endosomes and CD63-positive late endosomes in an IL-6-dependent manner. In Hrs-RNAi HeLa cells, gp130 was detected in the early endosomes but not in the CD63-positive late endosomes, and the lysosomal degradation of gp130 was significantly impaired. Similar results were obtained with Hrs-knockout MEF cells (data not shown). These data indicate that Hrs plays a critical role in the transport of the ubiquitinated gp130 from the early endosome to the late endosome, where gp130 is degraded.

We also investigated whether the ubiquitination of gp130 and its association with Hrs are essential for internalization of gp130. Hrs-depleted cells showed significant internalization of gp130, indicating that Hrs is dispensable for this function, despite being required for the transport of gp130 from the early endosome to the late endosome. Furthermore, we showed that the internalization of gp130 is clathrin dependent, because the IL-6-induced internalization of gp130 was significantly suppressed in clathrin-deficient cells, and gp130 was not detectable in GM-1-positive detergent-resistant membranes, which do not contain the clathrin/AP-2 complex (data not shown). Several other receptors, such as EGF, growth hormone, leptin, and transferrin receptors, are known to be internalized through a clathrin-dependent endocytosis pathway (10, 23, 30, 70), and their internalization also requires their ubiquitination (10, 70). They contain conserved cytoplasmic tyrosine, serine, and lysine residues that are essential for their ubiquitination and subsequent internalization (33, 49). However, gp130 internalization was not affected, even when all 14 of the conserved cytoplasmic lysines were replaced with arginines (data not shown). Therefore, it seems that the ubiquitination of gp130 is not essential for its internalization. We also found that c-Cbl depletion had little effect on gp130 internalization, as is consistent with ubiquitination being dispensable for gp130 internalization. These data suggest that gp130 has some internalization signal other than ubiquitinated lysine. Such internalization signals could be located in a dileucine-based endocytosis motif in the cytoplasmic portion of gp130 (16). Such motifs are involved in the endocytosis of growth hormone and granulocyte colony-stimulating factor receptors (1, 24).

The degradation of effector proteins is an important means of terminating cytokine signals. Hrs knockdown by Hrs-RNAi retards the hepatocyte growth factor-dependent degradation of MET in the lysosome and prolongs its phosphorylation (29). Therefore, the Hrs-dependent lysosomal degradation pathway appears to be important in the negative regulation of cytokine signaling. In IL-6 signaling, SOCS3, which is rapidly induced by IL-6 stimulation, is a known negative regulator (63, 68). However, IL-6-mediated STAT3 activation and cell proliferation were significantly enhanced in Hrs-deficient HeLa cells and BAF/gp130 cells, respectively, although SOCS3 expression was somewhat increased in these cells. This unexpected phenomenon may be explained by the fact that, because the lysosomal degradation of gp130 was impaired in the Hrs-deficient cells, the gp130 signal was enhanced. Hence, we propose that termination of IL-6 signaling is mediated primarily by the lysosomal degradation of gp130. Similarly, MEF cells expressing the gp130^F759/F759 mutant, which cannot be ubiquitinated by c-Cbl or associated with Hrs, showed enhanced IL-6 signaling. This finding is compatible with the impaired lysosomal gp130 degradation seen with Hrs-deficient and c-Cbl-depleted cells showing significant increases in IL-6-mediated STAT3 activation, SOCS3 induction, and cell proliferation (Fig. 5 and 8).

c-Cbl-deficient mice produce thymic cells that show enhanced MAP kinase activity and enhanced tyrosine phosphorylation on multiple cellular proteins, and the mice show an impairment in the positive selection of CD4⁺ thymocytes (51). We reported that Hrs acts as an oncogenic factor contributing to the maintenance of malignant cell phenotypes, including anchorage-independent growth, in vivo tumorigenesis, and metastasis. The malignant cell phenotypes associated with Hrs expression are mediated by the Hrs-dependent lysosomal degradation of E-cadherin, which results in the upregulation and promotion of β-catenin signaling (69).

Excessive gp130-mediated signals cause pathological phenotypes such as gastrointestinal disorders, autoimmune diseases, and chronic inflammatory proliferative diseases. For example, gp130^F759/F759 knock-in mice display splenomegaly, lymphadenopathy, and the development of gastric cancer, all of which are consistent with the deficient SHP2 binding of this mutant (54). Our present findings indicate that c-Cbl and Hrs also contribute to the extinction of the gp130-mediated signals. Since neither c-Cbl nor Hrs seems to influence gp130-mediated functions, it is unlikely that mice deficient for either of these genes would show the gp130^F759/F759-like phenotypes. Nevertheless, it is intriguing to speculate about the existence of a pathological condition in which E3 function and/or the Ub-mediated sorting machinery is defective. In such conditions, activated receptors that escape proper degradation might induce neoplasms, inflammation, and/or apoptosis. Our next steps will be to create such animal models and to scrutinize human diseases, such as rheumatoid arthritis and various cancers, for evidence of dysfunction in ubiquitination or in lysosomal degradation of ubiquitinated proteins. In sum, we report here a novel mechanism for regulating cytokine signaling that uses ubiquitination and endosomal sorting to determine the longevity of the activated cytokine receptor.