Growth hormone receptor ubiquitination coincides with recruitment to clathrin-coated membrane domains.

Endocytosis of the growth hormone receptor (GHR) depends on a functional ubiquitin conjugation system. A 10-amino acid residue motif within the GHR cytosolic tail (the ubiquitin-dependent endocytosis motif) is involved in both GHR ubiquitination and endocytosis. As shown previously, ubiquitination of the receptor itself is not required. In this paper ubiquitination of the GHR was used as a tool to address the question of at which stage the ubiquitin conjugation system acts in the process of GHR endocytosis. If potassium depletion was used to interfere with early stages of coated pit formation, both GHR endocytosis and ubiquitination were inhibited. Treatment of cells with methyl-beta-cyclodextrin inhibited endocytosis at the stage of coated vesicle formation. Growth hormone addition to methyl-beta-cyclodextrin-treated cells resulted in an accumulation of ubiquitinated GHR at the cell surface. Using immunoelectron microscopy, the GHR was localized in flattened clathrin-coated membranes. In addition, when clathrin-mediated endocytosis was inhibited in HeLa cells expressing a temperature-sensitive dynamin mutant, ubiquitinated GHR accumulated at the cell surface. Together, these data show that the GHR is ubiquitinated at the plasma membrane, before endocytosis occurs, and indicate that the resident time of the GHR at the cell surface is regulated by the ubiquitin conjugation system together with the endocytic machinery.

Endocytosis of the growth hormone receptor (GHR) depends on a functional ubiquitin conjugation system. A 10-amino acid residue motif within the GHR cytosolic tail (the ubiquitin-dependent endocytosis motif) is involved in both GHR ubiquitination and endocytosis. As shown previously, ubiquitination of the receptor itself is not required. In this paper ubiquitination of the GHR was used as a tool to address the question of at which stage the ubiquitin conjugation system acts in the process of GHR endocytosis. If potassium depletion was used to interfere with early stages of coated pit formation, both GHR endocytosis and ubiquitination were inhibited. Treatment of cells with methyl-␤-cyclodextrin inhibited endocytosis at the stage of coated vesicle formation. Growth hormone addition to methyl-␤-cyclodextrin-treated cells resulted in an accumulation of ubiquitinated GHR at the cell surface. Using immunoelectron microscopy, the GHR was localized in flattened clathrin-coated membranes. In addition, when clathrin-mediated endocytosis was inhibited in HeLa cells expressing a temperature-sensitive dynamin mutant, ubiquitinated GHR accumulated at the cell surface. Together, these data show that the GHR is ubiquitinated at the plasma membrane, before endocytosis occurs, and indicate that the resident time of the GHR at the cell surface is regulated by the ubiquitin conjugation system together with the endocytic machinery.
Clathrin-mediated endocytosis involves the formation of clathrin-coated vesicles from coated pits at the plasma membrane. Recruitment of membrane proteins into clathrin-coated pits is mediated by specific amino acid sequences within their cytoplasmic domain (for review, see Refs. 1,2). The best-defined coated pit localization signals are the tyrosine-based motifs NPXY as described for, e.g. the low-density lipoprotein receptor (3), and YXX⌽ (where ⌽ is an amino acid with a bulky hydrophobic group) found in, e.g. the transferrin receptor (4). Alternatively, internalization of the insulin and ␤2-adrenergic receptor is mediated by a dileucine-containing motif (5,6). Many receptors, including the low-density lipoprotein receptor and the transferrin receptor, are clustered in coated pits and internalized constitutively, independent of ligand occupancy.
The heterotetrameric adaptor complex AP-2 binds directly to the tyrosine-based motif and nucleates assembly of clathrin triskelions onto the plasma membrane (for review, see Ref. 7). Invagination of the plasma membrane results in the formation of constricted coated pits, followed by the dynamin-dependent detachment of coated vesicles from the plasma membrane. More complex situations exist when plasma membrane proteins enter cells on stimuli such as hormone binding or specific signal transduction events. In this case the internalization signal is only recognized on ligand binding, or the stimulus induces the addition of a signal, which results in the subsequent recruitment of the receptor into the coated pit. The agonist-induced phosphorylation of the ␤2-adrenergic receptor resulting in the binding of ␤-arrestin, a specialized adaptor that binds to clathrin, is an example of a protein modification that regulates internalization (8). Recently, it was shown that the attachment of ubiquitin moieties is involved in the internalization of several plasma membrane proteins (for review, see Refs. 9, 10). In mammalian cells, the ubiquitin conjugation system regulates the endocytosis of the epithelial sodium channel (11) and the growth hormone receptor (GHR 1 ; Ref. 12).
The GHR was initially found to be ubiquitinated on amino acid sequencing of the receptor from rabbit liver (13). The ubiquitin conjugation system is involved in GHR internalization and degradation (12,14). In particular, a 10-amino acid motif within the GHR cytosolic tail (the ubiquitin-dependent endocytosis motif, DSWVEFIELD) is involved in both receptor ubiquitination and endocytosis (15). We have recently shown that the proteasome is also involved in GHR internalization (16). Ligand-induced internalization of the GHR is blocked in the presence of specific proteasomal inhibitors such as carbobenzoxy-L-leucyl-L-leucyl-L-leucinal and ␤-lactone, the more membrane-permeable analogue of lactacystin. Disruption of clathrin-mediated endocytosis by cellular potassium depletion (17), hypertonic medium treatment (18), or cellular cytosol acidification (19) inhibits internalization and ubiquitination of the GHR (14). Although ubiquitination of the GHR itself is not required for endocytosis (15), GHR internalization requires the activity of the ubiquitin conjugation system, which acts together with the endocytic machinery in targeting the receptor into the coated pit.
In this study the question was addressed of at which stage the ubiquitin conjugation system acts in the process of GHR endocytosis. GHR ubiquitination was taken as a biochemical marker for the location of the activity of the ubiquitin conjuga-tion system. Conditions in which coated pit formation was prevented were compared with conditions that allowed coated pit formation but prevented coated vesicle formation.

EXPERIMENTAL PROCEDURES
Materials and Antibodies-Antibody mAb5 recognizing the lumenal part of the GHR was from AGEN Inc. (Parsippany, NJ). Polyclonal antibodies against amino acid residues 271-318 of the cytosolic tail of the GHR (Anti-T; Ref. 16) and against human growth hormone (GH) were raised in rabbits. Antiserum specific for protein-ubiquitin conjugates was a generous gift from Dr. A. Ciechanover (Technion-Israel Institute of Technology, Haifa, Israel). Antibody 4G10 (anti-pY), recognizing phosphotyrosine residues, was obtained from Upstate Biotechnologies Inc. (Lake Placid, NY). Anti-biotin was from Rockland (Gilbertsville, PA); monoclonal anti-clathrin was from Transduction Laboratories (Lexington, NY); anti-mouse IgG was from Nordic Immunological Laboratories (Tilburg, The Netherlands); and monoclonal anti-hemagglutinin (HA) antibody 12CA5 was from Babco (Richmond, CA). Human GH was a kind gift from Lilly; methyl-␤-cyclodextrin (M␤CD) was from Sigma; LipofectAMINE was from Life Technologies, Inc.; and carbobenzoxy-Lleucyl-L-leucyl-L-leucinal was from Calbiochem.
Plasmids, Cell Culture, and Transfection-Wild-type rabbit GHR cDNA was cloned into the cytomegalovirus-Neo expression plasmid pcDNA3.1 (Invitrogen BV/Novex) and used for transient transfections.
The internalization-deficient mutant GHR(F327A) was constructed by site-directed mutagenesis and cloned into pcDNA3.1 as described (20). The Chinese hamster cell line ts20, stably transfected with a pCB6 construct containing the rabbit GHR cDNA sequence, was used in this study (12). Cells were grown at 30°C in MEM␣ supplemented with 10% fetal calf serum, 4.5 g/l glucose, 100 units/ml penicillin, 100 g/ml streptomycin, and 0.45 mg/ml geneticin. For experiments cells were grown on 35-or 60-mm dishes in the absence of geneticin to ϳ75% confluence, and 10 mM sodium butyrate was added overnight to increase GHR expression (12).
tTA-HeLa cell lines stably transfected with the HA-tagged temperature-sensitive mutant of dynamin (dyn TS ) or wild-type dynamin (wtdyn) were kindly provided by Dr. S. Schmid (The Scripps Research Institute, La Jolla, CA). Dyn TS carries a mutation of the glycine at position 273 to aspartic acid. Cells were cultured at 37°C in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum, 100 units/ml penicillin, 100 g/ml streptomycin, 0.4 mg/ml geneticin, 2 g/ml tetracycline, and 200 ng/ml puromycin. For transfection experiments subconfluent cultures of tTA-HeLa cells were washed with phosphate-buffered saline (PBS), detached with trypsin-EDTA, plated on 60-mm dishes, and incubated at the permissive temperature of 32°C in the absence of tetracycline for 24 h. Cells were 30 -40% confluent when transfected with 1 g cDNA/dish, using LipofectAMINE according to the manufacturer's protocol. After 24 h, cells were washed with medium free of tetracycline. Cells were used for experiments 48 h after transfection and 72 h after removal of tetracycline.
GH Binding and Internalization-125 I-Human GH was prepared using chloramine T (12). For internalization studies, cells were grown in 12-well cluster plates, washed with MEM␣ supplemented with 20 mM Hepes, pH 7.4, and 0.1% bovine serum albumin (BSA), and incubated in a water bath. 125 I-GH (8 nM) was bound on ice for 2 h, and the cells were washed free of unbound GH and incubated for 0 -30 min at 30°C. Membrane-associated GH was removed by acid wash (0.15 M NaCl, 50 mM glycine, 0.1% BSA, pH 2.5) on ice. Internalized GH was determined by measuring the radioactivity after solubilization of the acid-treated cells in 1 M NaOH using an LKB gamma counter.
Metabolic Labeling-Cells were grown in 60-mm dishes and incubated in methionine-and cysteine-free MEM. Then 3.7 MBq/ml [ 35 S]methionine (Tran 35 S Label, , 40 TBq/mmol; ICN, Costa Mesa, CA) was added, and the incubation was continued at 30°C in a CO 2 incubator. The radioactivity was replaced with medium containing 100 M unlabeled methionine, 0.1% BSA, and 8 nM GH and chased for 0 -60 min. Cells were lysed, and samples were immunoprecipitated (see below). Radioactivity was determined using a Storm imaging system (Molecular Dynamics, Sunnyvale, CA) and quantified with Molecular Dynamics Image Quant software, version 4.2a.
Cell Lysis, Immunoprecipitation, and Western Blotting-Immunoprecipitations were performed as described previously (12). For GHR immunoprecipitations, cells were lysed on ice in 0.3 ml of lysis mix containing 1% Triton X-100, 1 mM EDTA, 50 mM NaF, 1 mM Na 3 VO 4 , 10 g/ml aprotinin, 10 g/ml leupeptin, 2 M carbobenzoxy-L-leucyl-Lleucyl-L-leucinal, and 100 mM phenylmethylsulfonyl fluoride in PBS. In ubiquitin immunoprecipitation experiments the cells were lysed in 0.3 ml boiling lysis buffer containing 1% SDS, 1 mM EDTA, 50 mM NaF, and 1 mM Na 3 VO 4 . The lysate was heated for 5 min at 100°C, after which the DNA was sheared using a 25-G needle. Immunoprecipitation of the supernatant was carried out in 1% Triton X-100, 0.5% SDS, 0.25% sodium deoxycholate, and 0.5% BSA in PBS plus various inhibitors. For anti-GH immunoprecipitations of GH-GHR complexes, the cells were lysed on ice in a lysis mix containing 1% Triton X-100, 150 mM NaCl, 10% glycerol, 50 mM Tris-HCl, pH 8.0, 10 mM N-ethylmaleimide, and various inhibitors. The immunoprecipitation was carried out in the same buffer. The lysates were incubated with the indicated antibodies for 2 h on ice, and immune complexes were isolated using protein A-agarose beads (Repligen Co., Cambridge, MA). The immunoprecipitates were washed twice with the same buffer and twice with 10-fold diluted PBS. Immune complexes were subjected to SDS-polyacrylamide gel electrophoresis and immunoblotting as described (14). For detection the enhanced chemiluminescence system (Roche Molecular Biochemicals) was used. To reprobe blots, the membranes were incubated for 1 h at room temperature in 0.15 M NaCl, 50 mM glycine, pH 2.5, buffer. The efficiency of the stripping procedure was checked and was found to remove Ͼ95% of the signal.
Potassium Depletion-Cells were subjected to potassium depletion as described previously (14,18). All incubation steps were performed at 30°C. Cells were washed twice with isotonic, potassium-free buffer A, (0.14 M NaCl, 20 mM Hepes, 1 mM CaCl 2 , 1 mM MgCl 2 , 1 g/l glucose, and 0.1% BSA, pH 7.4) and subjected to a hypotonic shock for 5 min in buffer A diluted 1:1 with H 2 O. After incubation in buffer A for 30 min, GH was added for an additional 30 min. Parallel cultures, which had also been hypotonically shocked, were incubated in buffer A supplemented with 10 mM KCl.
Light Microscopy-Fluorescently labeled GH (Cy3-GH) was prepared as described before (16). Cells grown on coverslips were incubated in MEM␣ supplemented with 20 mM Hepes, pH 7.4, and 0.1% BSA. Cy3-GH (0, 8 g/ml) was added, and the incubation was continued. Cells were washed with PBS to remove unbound label and fixed for 2 h in 3% paraformaldehyde in PBS. After fixation, the cells were embedded in Mowiol, and confocal laser scanning microscopy was performed using a Leica TCS 4D system.
Immunogold Electron Microscopy-Cells were incubated in MEM␣ plus 0.1% BSA and 8 nM biotinylated GH (21) in a CO 2 incubator. Before fixation and processing for immunoelectron microscopy, cells were washed three times with MEM␣ and 0.1% BSA. The cells were fixed in 2% paraformaldehyde and 0.2% glutaraldehyde in 0.1 M phosphate buffer, pH 7.4, for 30 min on ice followed by 3 h at room temperature. Further processing for ultrathin cryosectioning and labeling according to the protein A-gold method was done as described previously (22). To pick up ultrathin cryosections, a 1:1 mixture of 2.3 M sucrose and 1.8% methylcellulose was used (23). A rabbit polyclonal antibody against biotin was directly visualized by protein A-gold. A mouse monoclonal antibody against clathrin was first incubated with rabbit anti-mouse IgG, to provide binding sides for protein A, which binds poorly to mouse antibodies.
The effect of M␤CD on clathrin-coated pit morphology was quantified by determining the number of clathrin-coated structures at the plasma membrane. M␤CD-treated and control cells were labeled for clathrin, and the plasma membrane of 50 cell profiles with visible nucleus was screened for clathrin-coated structures, which were subdivided into four categories: 1) flat or slightly invaginated coated pits, 2) invaginated coated pits with a wide opening, 3) invaginated coated pits with a constricted neck, and 4) vesicles in close vicinity to the plasma membrane, which were not connected to the plasma membrane in the plane of the section. The frequency of each category was expressed as a percentage of the total number of clathrin-coated structures at the plasma membrane.

M␤CD Inhibits the Endocytosis of the GHR-
The GHR enters the cell via clathrin-coated pits (14,24,25). Previously we used methods that deplete the cytosol of free clathrin triskelions and interfere with the early stages of coated pit formation (18,26). These methods inhibited GHR internalization and abolished GHR ubiquitination (14), suggesting that GHR ubiquitination occurs during or after coated pit formation. Acute cholesterol depletion of the plasma membrane using M␤CD inhibits clathrin-coated pit budding. This method reduces the internalization of the transferrin receptor by Ͼ85% (27). Cholesterol depletion results in the formation of shallow coated pits, indicating that cholesterol is essential for clathrin coated vesicle formation (28). To monitor GH uptake, GHR-expressing ts20 cells were incubated for 30 min in the presence of Cy3labeled GH, which resulted in the presence of Cy3-GH in endosomal and lysosomal compartments (Fig. 1A, Control). As expected, the Cy3-GH was protected against acid treatment, confirming that the label is in intracellular structures (Fig. 1A, Acid Wash). When cells were preincubated in the presence of M␤CD, virtually no Cy3-GH entered the cells (Fig. 1A, M␤CD), and the majority of label could be removed on acid treatment (Fig. 1A, Acid Wash). Neither uptake nor binding was observed when excess unlabeled ligand was added together with Cy3-GH or when untransfected cells were used (data not shown).
To confirm and quantify the effect of M␤CD on endocytosis, uptake of 125 I-GH was measured in a time course experiment. As seen in Fig. 1B, M␤CD inhibited the internalization of GH efficiently. There was no effect of M␤CD on the total binding of 125 I-GH to the cells (data not shown). Two control experiments were performed to ascertain that M␤CD treatment did not affect other relevant cellular processes. The effect of M␤CD on GHR biosynthesis was measured using pulse-chase labeling with [ 35 S]methionine ( Fig. 2A). The receptor was synthesized as a 110-kDa glycoprotein precursor (double band; Fig. 2A, p) and on "complex glycosylation" in the Golgi complex converted to a 130-kDa mature species ( Fig. 2A, m). Quantification of the radioactivity showed that the GHR signal in the M␤CD cells is ϳ85-90% of the control cells, indicating a slight inhibition of protein synthesis. Conversion of precursor to mature receptor was detectable after 30 min of chase both in control and in M␤CD-treated cells, indicating that transport to the Golgi compartment was not affected by the cholesterol depletion. To examine the effect of M␤CD on GHR phosphorylation, a second control experiment was performed. Allevato and colleagues (29) showed that a mutated GHR, deficient in internalization, was capable of stimulating transcription of the serine protease inhibitor 2.1 promoter, and we showed that the GHR cytosolic tail is tyrosine-phosphorylated, while internalization is inhibited (14,20). From these studies it was concluded that GHR phosphorylation is independent of GHR endocytosis. M␤CD-pretreated cells were incubated for various periods with GH. As seen in Fig. 2B, phosphorylation became detectable after 5 min of incubation, reaching a maximum after 15 min for both control and M␤CD-treated cells. The blot was reprobed with anti-GHR to show that comparable amounts of receptor were loaded in each lane. From these results we conclude that M␤CD treatment has no effect on GH-induced phosphorylation of the GHR, indicating that activation of the tyrosine kinase (Janus kinase 2) and receptor dimerization can take place in the presence of M␤CD.

FIG. 1. Effect of M␤CD on GH internalization.
A, GHR-expressing ts20 cells were incubated for 30 min at 30°C in the absence (Control) or presence of 10 mM M␤CD and incubated with Cy3-GH for 30 min. Cells were fixed before (30 min, 30°C) or after acid wash (Acid Wash). Cy3-GH was visualized by confocal microscopy. B, GHR-expressing ts20 cells were incubated for 30 min at 30°C in the absence (control) or presence of 10 mM M␤CD and put on ice for 2 h with 125 I-GH. Unbound label was removed, and the cells were incubated at 30°C in the absence or presence of M␤CD as indicated. Background label was determined in the presence of excess unlabeled GH and subtracted. The amounts of 125 I-GH internalized are plotted as a percentage of the cell-associated radioactivity at the start of the incubation. Each point represents the mean value of two experiments performed in duplicate Ϯ S.D. q, control; E, M␤CD.

FIG. 2. Effect of M␤CD on the biosynthesis and the tyrosinephosphorylation of the GHR.
A, GHR-expressing ts20 cells were incubated for 30 min at 30°C in methionine-free medium in the absence (Control) or presence of 10 mM M␤CD. [ 35 S]Methionine was added, and the incubation was continued for 10 min. Cells were chased in MEM␣ supplemented with 0.1% BSA, 100 M methionine, and 8 nM GH in the absence or presence of M␤CD for the periods indicated. Upper panel, GHR was immunoprecipitated using anti-T; p, precursor GHR (110 kDa.); m, mature GHR (130 kDa.). Lower panel, radioactivity was quantified and expressed as a percentage of the radioactivity incorporated in the precursor GHR. q, precursor GHR; E, mature GHR. B, GHR-expressing ts20 cells were incubated at 30°C in the absence (control) or presence of 10 mM M␤CD. All dishes were incubated for 60 min in total; GH was present during the last 5, 15, or 30 min. Upper panel, cells were lysed, and the GHR was immunoprecipitated with anti-T and immunoblotted using anti-pY. Lower panel, the same blot was reprobed using anti-GHR (mAb5). p, precursor GHR (110 kDa.); m, mature GHR (130 kDa.).

M␤CD Does Not Affect Accessibility of Clathrin-coated Pits
for GHR-Previously, it was shown that M␤CD does not interfere with the association of clathrin to the plasma membrane but has an inhibitory effect on the invagination and fission of clathrin-coated pits (28). For the present study it was important to determine whether the GHR enters these clathrincoated areas at the plasma membrane of M␤CD-treated cells. Immunogold labeling of clathrin in GHR-expressing ts20 cells revealed that the majority of plasma membrane-associated clathrin was present on deeply invaginated pits and coated vesicles in close vicinity to the plasma membrane ( Fig. 3A and Table I). GH was regularly found in the deeply invaginated clathrin-coated pits and vesicles (Fig. 3, C and D), as well as in later compartments of the endocytic pathway (data not shown). In M␤CD-treated cells, the total number of clathrin-coated structures at the plasma membrane was increased twice compared with that in control cells. In contrast to control cells, Ͼ85% of the clathrin-containing structures were flattened coated pits (Fig. 3B). Deeply invaginated coated pits and coated vesicles were rarely seen (Table I). GH accumulated at the plasma membrane, where it regularly but not exclusively occurred in the flattened clathrin-coated pits (Fig. 3E). These findings suggest that the GH-GHR complex in M␤CD-treated cells indeed enters the clathrin-coated areas of the plasma membrane.
Effect of M␤CD on GHR Ubiquitination-To address the question of whether the GHR is ubiquitinated after M␤CD treatment, cells were incubated with or without GH, and ubiquitinated proteins were immunoprecipitated and analyzed by Western blotting, as indicated in Fig. 4. Ubiquitinated GHR appeared as high molecular weight species in the top part of the gel (Fig. 4A). Control cells (lanes 3 and 4) showed increased GHR ubiquitination on ligand binding. Both in unstimulated and stimulated cells the level of GHR ubiquitination increased when M␤CD was present (Fig. 4A, compare lane 3 with lane 9  and lane 4 with lane 10). The use of untransfected cells resulted, as expected, in the absence of signal for ubiquitinated GHR (ts20; Fig. 4A, lanes 1 and 2). When endocytosis was inhibited by potassium depletion, GHR ubiquitination was almost completely abolished (lanes 5 and 6). Ubiquitination was restored to control values by adding 10 mM KCl (lanes 7 and 8). Reprobing the blot from Fig. 4A with anti-ubiquitin showed that the amount of immunoprecipitated, ubiquitinated protein in each lane was comparable (Fig. 4B). These results show that the GHR is ubiquitinated at the cell surface before constriction of the coated pit occurs and suggest that assembly of the clathrin coat is a requirement for GHR ubiquitination.
Effect of a Temperature-sensitive Dynamin Mutation on GHR Ubiquitination-To examine the effect of inhibition of clathrinmediated endocytosis on GHR ubiquitination by an independent method, we used a HeLa cell line expressing a temperature-  sensitive mutant of human dynamin, dyn TS . Dyn TS carries a point mutation corresponding to the Drosophila shibire ts1 allele (30). For this dynamin mutant it has been shown that transferrin internalization is inhibited at the nonpermissive temperature, that the phenotype is rapid and reversible, and that invaginated but not constricted coated pits accumulate on the cytoplasmic surface of the plasma membrane. HeLa cells were transiently transfected with wild-type GHR cDNA and incubated in the presence of GH; afterward the GH-GHR complex was immunoprecipitated with anti-GH. Using this approach, only mature GHR species from the cell surface were immunoprecipitated. At the permissive temperature, ubiquitinated protein was detected in both the wild-type and mutant dynamin cells (Fig. 5A, anti-Ubi). After shifting the cells to the nonpermissive temperature, an increase in the amount of ubiquitinated GHR was detected in dyn TS and wild-type dynamin cells (Fig. 5A, anti-Ubi). Ubiquitination of proteins is dynamic with rapid addition and removal of ubiquitin (31). The increase in ubiquitination in the wild-type dynamin cells might reflect the higher endocytotic activity of the cells caused by the elevated temperature or a temperature-dependent shift in balance between ubiquitinating and deubiquitinating enzymes. The increase in amount of ubiquitinated GHR in the dyn TS cells at the nonpermissive temperature was consistently severalfold higher compared with wild-type dynamin cells. The data indicate that GHR ubiquitination precedes GHR endocytosis, which is in agreement with the experiments in which M␤CD was used. Reprobing the blot with anti-GHR showed that the amounts of mature receptor expressed in dyn TS and wild-type dynamin cells were comparable (Fig. 5A, anti-GHR). The amount of ubiquitinated GHR is only a small fraction of the total amount of GHR bound to GH (see below). The expression of HA-tagged dynamin in the HeLa cells was controlled on a Western blot of cell lysate using anti-HA (Fig. 5A, anti-HA). Both dyn TS and wild-type dynamin cells expressed the HAtagged dynamin, albeit in different amounts, whereas cells  1-4), in the absence (ϪK ϩ ; lanes 5 and 6) or presence (ϩK ϩ ; lanes 7 and 8) of potassium after hypotonic shock, or with 10 mM M␤CD (lanes 9 and 10) for 30 min at 30°C. The incubation was continued for 30 min with (lanes 2, 4, 6, 8, and 10, ϩGH) or without GH. A, cells were lysed with boiling lysis buffer; equal amounts of cell lysates were immunoprecipitated with anti-ubiquitin and analyzed by immunoblotting with anti-GHR (mAb5). B, the same blot was reprobed with antiubiquitin (Ubi). Relative molecular weight standards (M r ϫ 10 Ϫ3 ) are shown at the left.

FIG. 5. Effect of overexpression of the dyn TS mutant on GHR ubiquitination and internalization.
A, Wild-type (wtDyn) and mutant (Dyn TS ) dynamin cells were transfected with GHR cDNA. Cells were incubated at the permissive or nonpermissive temperature for 30 min; afterward, GH was added to all dishes, and the incubation was continued for 30 min. Left panel, cells were lysed on ice; equal amount of cell lysates were immunoprecipitated with anti-GH; and the immunoprecipitates were analyzed by immunoblotting with anti-ubiquitin (anti-Ubi). Middle panel, the anti-ubiquitin blot was reprobed with anti-GHR (mAb5). Right panel, equal amounts of cell lysates were loaded on a gel and immunoblotted with anti-HA. Relative molecular weight standards (M r ϫ 10 Ϫ3 ) are shown at the left. B, wild-type (wtDyn) and mutant (Dyn TS ) dynamin cells were transfected with GHR cDNA. Cells were incubated at the permissive or nonpermissive temperature for 30 min; afterward, 125 I-GH was added to all dishes, and the incubation was continued for 30 min. Background signal was measured in the presence of excess unlabeled GH and subtracted. Internalization in the dyn TS mutant was compared with the internalization in the wtdyn; specifically internalized 125 I-GH was calculated as a percentage of total cell-associated label and set to 100% for the wtdyn cells. The amount of cell surface expression of GHR was comparable in the two cell lines; mock-transfected cells showed no detectable specific binding of 125 I-GH (data not shown). The values are the mean Ϯ S.D. of two experiments performed in duplicate. C, mutant (dyn TS ) dynamin cells were transfected with wild-type GHR (wt) or GHR(F327A) cDNA (F327A). Experimental conditions and immunoprecipitations were as described in A. Left panel, the Western blot was detected with antiubiquitin (anti-Ubi). Right panel, the same blot was reprobed with anti-GHR (mAb5). cultured in the presence of tetracycline showed no detectable HA signal (results not shown). To ascertain that indeed GH uptake was inhibited, we measured internalization of 125 I-GH. At the permissive temperature the percentage of GH uptake was comparable in the two cell lines, whereas at the nonpermissive temperature GH internalization was strongly inhibited in dyn TS cells compared with wild-type dynamin cells (Fig. 5B). The results demonstrate that overexpression of a dominant negative mutant of dynamin-1 inhibits the clathrin-mediated endocytosis of the GHR. Because we analyzed a complex of proteins immunoprecipitated with anti-GH to monitor ubiquitination of the GHR, the possibility exists that other ubiquitinated proteins coimmunoprecipitate in this complex. To show that the ubiquitination of this complex depends on the GHR, we transfected the internalization-deficient GHR(F327A) mutant in dyn TS cells. Previously, we have shown that this mutant is not ubiquitinated because of a defective ubiquitin-dependent endocytosis motif (15). After immunoprecipitation of the GH-GHR(F327A) complex, almost no ubiquitinated protein was isolated either at the permissive or the nonpermissive temperature (Fig. 5C, left panel). This result shows that the ubiquitination of the GH-GHR complex is dependent on an intact ubiquitin-dependent endocytosis motif and that most likely the receptor itself is ubiquitinated, perhaps in complex with other ubiquitinated proteins. Control experiments with mock-transfected cells and incubations without GH showed no signal on the ubiquitin blots (results not shown). As seen in the Fig. 5C, right panel, mature GHR and GHR(F327A) were detected in the complex, and virtually no GHR signal is present at the top of the lanes, indicating that only a small percentage of total GHR is ubiquitinated. Most likely, the GHR is ubiquitinated during a very short period, presumably the resident time in the coated pit.

DISCUSSION
In this study two independent methods were used to inhibit clathrin-mediated endocytosis at the level of clathrin-coated vesicle formation. Both methods inhibited the endocytosis of the GHR, resulting in an accumulation of ubiquitinated receptors at the cell surface. A morphological approach detected the GHR in deeply invaginated coated pits in control cells and in flattened clathrin-coated pits in M␤CD-treated cells. Disruption of clathrin-mediated endocytosis by hypertonic medium treatment, cytosol acidification, or potassium depletion resulted in the accumulation of nonubiquitinated receptors at the cell surface (Ref. 14 and Fig. 4A). Why is the GHR not ubiquitinated under these conditions? The ubiquitination state of a protein is the result of a dynamic process of ubiquitination and deubiquitination. Changing the intracellular milieu by depleting potassium or modifying the pH could alter the balance between those two processes. Do the used methods interfere with the ubiquitination machinery itself and cause a complete inhibition of cellular ubiquitination? Analysis of cell lysates from potassium-depleted or hypertonic medium-treated cells showed no reduction in total cellular ubiquitin conjugates but rather an increase in high molecular mass ubiquitinated proteins (Ref. 14 and Fig. 4B). The amount of free ubiquitin under these circumstances as measured with Western blotting was reduced (data not shown). However, cellular cytosol acidification showed increased free and less conjugated ubiquitin (data not shown). Cytosol acidification causes the same precipitation of small clathrin microcages as seen after hypertonicity and potassium depletion (32). Recently, it was shown that after hypertonic treatment or cytoplasmic acidification, free clathrin triskelions within the cytosol are depleted, and all of the clathrin becomes associated with membranes (26). Because the presence of free clathrin triskelions is required for the stabilization of AP-2 coated pit nucleation sites, depletion of clathrin inter-feres with coated pit formation. Because the methods used have a varying effect on the ratio of free versus conjugated ubiquitin, it is most likely that the inhibition of GHR ubiquitination is the result of the interference with the coated pit formation rather than with ubiquitin conjugation itself.
The observation that GHR ubiquitination coincides with the recruitment of the GHR to clathrin-coated membrane areas suggests that the ubiquitin conjugation system and the endocytosis machinery act together in the endocytosis of the GHR. The earlier observation that ubiquitination of the receptor itself is not important for endocytosis suggests that ancillary proteins might be ubiquitinated or that factors of the ubiquitin conjugation system itself might act as adaptors for the endocytosis machinery. Ubiquitin-protein ligases have been implicated in endocytosis. For the epithelial sodium channel, it was shown that the ubiquitin-protein ligase Nedd4 mediates the down-regulation of the Na ϩ channel activity by ubiquitinating the channel, which leads to its endocytosis and degradation (33). The yeast homologue of Nedd4, Npi1/Rsp5, participates via its C2 domain in the endocytosis of Gap1 permease. A truncated Npi1 protein lacking the C2 domain can still promote ubiquitination but not the endocytosis of Gap1 permease, which is consistent with direct participation of Npi1 in endocytosis of the permease (34). Whether an E2/E3 ubiquitin ligase directly serves as an endocytic adaptor for GHR, analogous to the role of arrestin for the ␤2-adrenergic receptor (8) or binding of the ubiquitin conjugation system, results in the interaction with an endocytic adaptor (e.g. AP-2), remains to be established. Recently, GH-dependent association of AP-2 with the chicken GHR was reported (35). Also, a possible role for the ubiquitin polypeptide itself, conjugated to a GHR-associated protein, cannot be excluded, as has been described for Ste2p (36) and Ste3p (37). Ligand-induced ubiquitination was shown for Eps15, a clathrin-coated pit associated protein that is ubiquitinated on epidermal growth factor (EGF) receptor activation (38). Eps15 is required for clathrin-mediated endocytosis, and perturbation of Eps15 function inhibits receptor-mediated endocytosis of transferrin (39). The biological significance of Eps15 monoubiquitination is not known. Recently, it was shown that polyubiquitination of the EGF receptor occurs at the plasma membrane on ligand-induced activation (40). Inhibition of endocytosis caused by overexpression of mutant dynamin resulted in a transient polyubiquitination of the EGF receptor. The mechanisms for GHR and EGF receptor ubiquitination are probably different. Ubiquitination of the EGF receptor is mediated by Cbl adaptor proteins, and both EGF receptor and Cbl must undergo phosphorylation on specific sites for productive ubiquitination (41), whereas GHR ubiquitination occurs in the absence of GHR tyrosine phosphorylation and is dependent on the ubiquitin-dependent endocytosis motif (20).
The amount of ubiquitinated GHR is very low compared with the total amount of cell surface GHR. The fact that only a small percentage of total GHR is ubiquitinated indicates a coated pit restricted function of the ubiquitin conjugation system. Whether the ubiquitinated GHR is (partially) degraded soon after endocytosis or perhaps rapidly deubiquitinated is not clear at present. A role for deubiquitinating enzymes in regulating endocytosis cannot be excluded, because deubiquitination has been recognized as an important regulatory step (31,42). Recently, genetic data were presented that support a model whereby the Drosophila fat facets deubiquitinating enzyme removes ubiquitin from the product of the liquid facets gene. The liquid facets locus encodes epsin, a protein involved in clathrin-mediated endocytosis (43). Thus the ubiquitin system may add another layer of complexity to the membranesorting machinery at the plasma membrane and regulate, to-gether with the classical endocytosis machinery, the time span of the GHR at the cell surface.