Interaction of the Growth Hormone Receptor with Cytokine-Induced Src Homology Domain 2 Protein in Rat Adipocytes

GH stimulates the phosphorylation of tyrosine residues in the GH receptor (GHR), Janus kinase 2 (JAK2), and other signaling proteins in a transient manner that subsides within 1 h. To assess the possible roles of cytokine-induced Src homology domain 2 (SH2) (CIS/SOCS) proteins in these transient responses, we studied the expression and disposition of CIS/SOCS proteins in rat adipocytes, a physiological target of GH action. A tyrosine-phosphorylated protein that appears to be the GHR was coprecipitated from extracts of GH-treated adipocytes with α-CIS. In contrast, no tyrosine-phosphorylated adipocyte proteins were recovered after immunoprecipitation with α-SOCS3, although coprecipitation of GHR with SOCS3 was readily detected in extracts of 3T3-F442A fibroblasts. Interaction of GHR with CIS peaked between 2 and 10 min after adipocytes were treated with GH, when tyrosine phosphorylation of the GHR was maximal. By 60 min after GH, tyrosine phosphorylation of the GHR declined to very low levels, and its interaction with CIS was reduced correspondingly. Proteasome inhibitors prevented the decline in tyrosine-phosphorylated GHR and prolonged interaction of GHR and CIS for at least 1 h. These findings demonstrate the interaction of CIS with the GHR in vivo and suggest that CIS may enhance degradation of the receptor by a proteasomal pathway.

ACTIVATION OF signal transducer and activator of transcription (STAT) proteins and tyrosine phosphorylation of both Janus kinase 2 (JAK2) and the GH receptor (GHR) are well recognized responses of adipocytes and other cells to GH (1). However, these events are short-lived (2, 3) and disappear within minutes despite continued presence of the hormone in the incubation medium. Several mechanisms, including recruitment and activation of phosphoprotein phosphatases (3), down-regulation of GHRs on the cell surface (4), and quenching of signals by CIS/SOCS proteins (5–7), might contribute to the transience of these early aspects of the GH signal. Inhibition of protein synthesis with cycloheximide postpones the decline of these responses (2, 8) and prolongs the early insulin-like metabolic responses of adipose tissue (9), suggesting that either a labile protein or a protein induced by GH may contribute to the termination of these initial effects. Recognition of the importance of STAT proteins in GH signaling and discovery of the cytokine-induced inhibitors of STAT signaling (CIS/SOCS) proteins suggested that this family of proteins might limit the duration of GH signaling. In support of this idea, it was soon found that GH stimulated the expression of several CIS/SOCS family proteins, including CIS, JAK2-binding protein (JAB; also called SOCS1), and SOCS3 in 3T3-F442A fibroblasts within 1 h (5). When overexpressed in Chinese hamster ovary (CHO) cells, two of these proteins, JAB and SOCS3, inhibited trans-activation of a reporter gene in response to GH, but SOCS2 appeared to enhance the expression of the reporter gene, whereas CIS itself had no effect. These findings raised the possibility that CIS/SOCS proteins may interact with a variety of signaling proteins, compete with one another, or both. When expressed in COS cells along with other proteins involved in the GH signaling pathway, CIS appeared to displace STAT5b from phosphotyrosine residues at the carboxyl end of the GHR and to accelerate the proteasomal-dependent degradation of signaling proteins (7). However, as both mechanisms may reflect the high level expression of the signaling proteins in COS cells, the relevance of these mechanisms in other cells remains to be demonstrated. Nevertheless, the dwarf stature of transgenic mice that overexpress CIS (10) and the giant stature of mice lacking SOCS2 (11) suggest that these proteins may be important negative factors that regulate GH signaling in many tissues.

To achieve their inhibitory effects, CIS/SOCS family proteins use common structural features in a variety of ways (12–14). The CIS/SOCS proteins all include an SH2 domain that recognizes and allows binding to specific phosphotyrosine groups on cytokine receptors or JAK2 (12). The SOCS proteins may thus competitively inhibit the interaction of downstream signaling molecules with the GHR. The SOCS proteins share a second defining motif, termed the SOCS box, which is also found in some proteins that lack SH2 domains. One such protein, elongin C, functions as a ubiquitin E3 ligase (15) and acts as an adapter protein that promotes transfer of ubiquitin from a ubiquitin E2 ligase to a target protein. Microsequencing of fragments of the purified GHR indicated the presence of ubiquitin (16), and ubiquitination appears to play an important role in GH signaling (17). The present studies were undertaken to explore the role of CIS/SOCS proteins in GH signaling in rat adipocytes.

Materials and Methods

Mouse CIS, JAB, SOCS2, SOCS3, and SOCS4 cDNA cloned into the expression plasmid pcDNA3 with an N-terminal-fused myc tag (18) and mouse polyclonal antibodies (Abs) for CIS and SOCS3 (19) were described previously. Rabbit polyclonal Ab 2941 was raised against the intracellular domain of the rat GHR fused to the maltose-binding protein (20) and was used for Western blotting analysis and immunoprecipitation. Rabbit polyclonal BB74 was raised against the GH-binding protein (21), which includes the extracellular domain of the GHR, and was provided by Dr. W. R. Baumbach (American Cyanamid, Princeton, NJ). Horseradish peroxidase (HRP)-coupled α-phosphotyrosine monoclonal antibodies 4G10 and PY99 were purchased from Upstate Biotechnology, Inc. (Lake Placid, NY), and Santa Cruz Biotechnology, Inc. (Santa Cruz, CA), respectively. HRP-linked antirabbit or antimouse Ig were purchased from Amersham Pharmacia Biotech (Piscataway, NJ) and used as secondary antibodies in Western blotting.

Rat adipocytes

Male rats of the CD strain were obtained from the Charles River Laboratories, Inc. (Kingston, NY), and studied when they attained body weights of 160–200 g. Isolated adipocytes were prepared from epididymal and perirenal fat according to the procedure of Rodbell (22) as modified in this laboratory (23). After digestion for 20 min with 1 mg/ml collagenase (lot 143710, type A, Roche Molecular Biochemicals, Indianapolis, IN) in KRPG [Krebs-Ringer phosphate buffer enriched with 5.5 mm glucose and 40 mg/ml BSA (Metrix fraction IV, Reheis Chemical Co., Phoenix, AZ)], the cells were washed four times in KRPG containing 10 mg/ml BSA, resuspended 1:3 (vol/vol) in the same buffer, and incubated at 37 C. GH-deprived adipocytes, which are sensitive to the insulin-like effects of GH, were obtained by incubating freshly isolated cells in GH-free KRPG buffer for 3 h (24).

Cell culture and transfection

Human endothelial kidney cells (HEK293)-GHR cells are HEK293 cells stably expressing the rat GHR as established in this laboratory from the cDNA for the rat GHR provided by Dr. Nils Billestrup (Hagedorn Research Institute, Gentofte, Denmark). The cells are maintained in DMEM with 4.5 g/liter glucose supplemented with 10% fetal bovine serum (Life Technologies, Inc., Gaithersburg, MD).

RNA extraction and Northern blot analysis

Total RNA was extracted from rat adipocytes using guanidinium thiocyanate-phenol (25). Final RNA concentrations were determined spectrophotometrically at 260 nm. For Northern blotting, total RNA samples were fractionated by electrophoresis through 1% agarose-formaldehyde gels and transferred to nylon membranes. Membranes were hybridized for 2.5 h at 65 C with random primed, ³²P-radiolabeled probes derived from full-length cDNAs encoding CIS, JAB, and SOCS3 in Rapid-Hyb buffer (Amersham Pharmacia Biotech). Membranes were washed sequentially in 2× SSC [0.15 m sodium chloride and 15 mm sodium citrate (pH 7.0)]/0.1% sodium dodecyl sulfate at room temperature and in 0.1 × SSC/0.1% sodium dodecyl sulfate at 65 C. Membranes were then exposed to x-ray film with an intensifying screen at −70 C.

Cell stimulation and protein extraction

Rat adipocytes were incubated 1:3 (vol/vol) in KRPG at 37 C for 2 min with or without 500 ng/ml hGH except when indicated otherwise. They were washed with ice-cold PBS [10 mm sodium phosphate (pH 7.0) and 0.15 m sodium chloride] that contained 0.4 mm sodium orthovanadate and were lysed in 1:2 (vol/vol) cell lysis buffer [50 mm Tris-HCl (pH 8.0), 150 mm NaCl, 1% Triton X-100, 10% glycerol, 100 mm NaF, 2 mm EDTA, 1 mm phenylmethylsulfonylfluoride, 1 mm sodium orthovanadate, 10 mm benzamidine, and 10 μg/ml aprotinin]. After centrifugation at 12,000 × g for 10 min at 4 C, the supernatants were subjected to immunoprecipitation as described below.

Immunoprecipitation and Western blotting

Cell lysates were incubated with the appropriate antibody and protein A-agarose beads overnight at 4 C. The immune complexes were then washed with TSA [10 mm Tris (pH 7.5), 0.15 m sodium chloride, and 0.02% sodium azide] plus 0.1% Triton X-100. Precipitated proteins were released from the protein A-agarose beads by boiling in Laemmli sample buffer (26) for 2 min, resolved by SDS-PAGE, and transferred to Polyvinylidene Difluoride Plus membranes (Osmonics, Inc., Westborough, MA). After blocking with 20% horse serum/TSA (for α-phosphotyrosine) or 5% milk in washing buffer [10 mm Tris (pH 7.4), 150 mm NaCl, and 0.2% Tween 20] for 1 h or overnight, the membranes were immunoblotted with appropriate primary antisera (Ab 2941, 1:5000; BB74, 1:5000; αCIS: 1:2000; αSOCS3, 1:2000) and HRP-coupled secondary antibodies (1:2000) or HRP-coupled antibodies (4G10, 1:5000; PY99, 1:1000; αmyc, 1:500). Detection of proteins was achieved by Enhanced Chemiluminescence Plus reagents (Amersham Pharmacia Biotech). For stripping and reprobing, the membranes were washed with TSA, followed by incubating in stripping buffer [60 mm Tris (pH 6.8), 2% sodium dodecyl sulfate, and 0.7% 2-mercaptoethanol] at 50 C for 30 min. The stripped membranes were then rinsed with washing buffer and reprobed with appropriate antibodies. Data were quantified by densitometric scanning of the films and were analyzed statistically by t test.

Immobilization of antibody to protein A-agarose

To avoid possible interference by the approximately 30-kDa light chain of the immunoprecipitating antibody in Western blot when detecting CIS or SOCS3 (29–40 kDa), the antisera used for immunoprecipitation were immobilized on protein A-agarose beads using the cross-linking reagent dimethylpimelimidate. Briefly, protein A-agarose beads were suspended in TSA, and CIS or SOCS3 antisera were added and allowed to bind at 4 C for 2 h with gentle rocking. The agarose beads were then washed with TSA and 0.1 m sodium borate buffer, pH 9.0, and resuspended in sodium borate buffer. Freshly prepared dimethylpimelimidate (0.2 m) was added, and the mixture was incubated at 23 C for 30 min on a rocker. After treatment with 0.2 m ethanolamine to cap any partially reactive cross-linking reagent, the beads were washed and resuspended in TSA.

Deglycosylation reaction

To evaluate glycosylation status, immune complexes were boiled for 2 min in 15 μl 0.4% sodium dodecyl sulfate and 1% 2-mercaptoethanol before digestion overnight with endoglycosidase F (Roche Molecular Biochemicals), which removes N-linked carbohydrates. Digestion was carried out in 40 μl of a solution containing 1% Nonidet P-40, 67 mm sodium phosphate (pH 7.4), 2 mm sodium orthovanadate, protease inhibitor mixture, and 2 μl N-glycosidase F (0.043 U/μl). The reaction was stopped by the addition of Laemmli sample buffer and boiling for 2 min.

Generation of glutathione-S-transferase (GST)-CIS fusion protein and GST-CIS assay

The EcoRI/XbaI fragment of pcDNA3/myc-CIS, which contains the entire coding region sequence of CIS cDNA, was digested and inserted into the pBluescript vector (Stratagene, La Jolla, CA) to acquire appropriate restriction sites for cloning. The EcoRI/NcoI fragment was then excised from pBluescript/CIS and ligated to pGEX4T-3 (Amersham Pharmacia Biotech) digested by the same enzymes. The DNA junctions in pGEX4T-3/CIS so generated were confirmed by sequencing analysis. Expression of this construct produces a GST-CIS fusion protein with GST at the N terminus of CIS. Fusion protein induction and affinity purification on glutathione-Sepharose 4B beads (GST beads) were performed as suggested by the manufacturer (Amersham Pharmacia Biotech). The fusion protein was bound to 20 μl GST beads (1:1 slurry in PBS), followed by extensive washing before mixing with cell lysates and incubation at 4 C overnight. To assess the bound GST-CIS fusion protein, after washing with cell lysis buffer four times, the fusion protein was dissociated and denatured in Laemmli sample buffer and resolved on SDS-PAGE. The fusion protein was examined by Coomassie Blue staining and destaining or Western blotting.

Immunofunctional assays

The immunofunctional assay described previously (20) was used to quantify GHR in extracts of rat adipocytes. Samples of extract representing approximately 150 μl packed cells were immunoprecipitated overnight with Ab 2941 in the presence of 1 nm [¹²⁵I]GH (35 μC/μg). The recovered proteins were washed with Tris-buffered saline plus 0.1% Triton X-100 to remove unbound ligand and counted. To prevent isotopic dilution by unlabeled GH, adipocytes to be used for the preparation of these extracts were incubated for 2 min with 25 nm [¹²⁵I]human (h) GH with a specific activity identical to that added during immunoprecipitation.

Results

To determine whether GH stimulates the expression of CIS/SOCS proteins in primary adipocytes as reported for cultured cells, we began our studies by determining the effect of GH on the expression of several CIS/SOCS family proteins. To evaluate gene expression, total RNA samples from cells incubated with GH for various times were analyzed on Northern blots. Transcripts reflecting the expression of CIS and SOCS3 were readily detected, and with prolonged exposure of the films, the expression of JAB was also detected, as shown in Fig. 1. Although RNA protection assays suggested that SOCS2 and SOCS3 are expressed at similar levels in adipose tissue (27), no SOCS2 was detected by Northern blot analysis of three independent RNA samples prepared from isolated adipocytes. The expression of CIS, SOCS3, and JAB was not altered significantly in adipocytes treated with GH for 30 min, but small increases in CIS and JAB were seen after 1 h. Because tissue levels of the corresponding proteins depend upon rates of translation and protein degradation as well as mRNA levels, the effect of GH on CIS/SOCS family protein levels was evaluated by Western blotting. Although CIS, JAB, and SOCS3 could not be detected by Western blot analysis of whole cell extracts (results not shown), two proteins with electrophoretic mobilities corresponding to 32 and 37 kDa, which have been shown to correspond to monoubiquitinated and nonubiquitinated CIS (28), were detected by first immunoprecipitating and then blotting with α-CIS (Fig. 2). No significant changes were seen in the abundance of either form of CIS after adipocytes were incubated with GH for 30 or 60 min or for shorter times (results not shown). However, the abundance of both forms of CIS was somewhat diminished after incubating cells in vitro for 3 h. SOCS3 appeared as a single broad band of 30 kDa, and no changes in its abundance were detected after incubation with GH for 30 or 60 min, or when cells were incubated without additions for 3 h (results not shown). We were unable to detect JAB in adipocyte extracts by Western blot analysis, even after enrichment by immunoprecipitation of the protein from extracts representing 1 ml packed cells, a sample 2–5 times larger than that needed to detect CIS or SOCS3 (results not shown).

Figure 1.

GH increases CIS and JAB gene expression in rat adipocytes. A, Total RNA (15 μg/lane) was isolated from freshly isolated adipocytes after incubation in vitro with 25 nm hGH for the indicated times. Samples were resolved by agarose gel electrophoresis, transferred to nylon membranes, and hybridized with individual ³²P-labeled cDNA probes. The exposure time to detect CIS and SOCS3 by autoradiography was 18 h, and for JAB it was 3 d. B, Quantitative evaluation of scanned images from the number of experiments shown in parentheses (mean ± sem). *, P < 0.05.

Figure 2.

Effects of GH on CIS protein levels in rat adipocytes. A, Freshly isolated cells were incubated with 25 nm hGH for the indicated times or without additions for 3 h. Proteins were immunoprecipitated from cell lysates with α-CIS immobilized on protein A-agarose beads, and recovered proteins were analyzed by SDS-PAGE followed by immunoblotting with α-CIS. B, Quantitative evaluation of scanned images from the number of experiments shown in parentheses (mean ± sem). *, P < 0.05.

As CIS/SOCS proteins have SH2 domains that can interact with phosphotyrosines in other proteins, we determined whether tyrosine phosphorylated proteins in extracts of GH-treated adipocytes coprecipitate with CIS or SOCS3. As shown in Fig. 3A, a 116-kDa tyrosine-phosphorylated protein was readily detected by Western blot analysis of proteins recovered with α-CIS. However, in contrast to findings made with GH-treated 3T3-F442 fibroblasts (29), no GH-dependent tyrosine-phosphorylated proteins were immunoprecipitated with α-SOCS3 from adipocyte extracts. To confirm that this negative result is not simply due to inadequacy of the SOCS3 antiserum, the experiment was repeated using 3T3-F442A fibroblasts that were maintained under serum-replete or serum-starved conditions. In either case, GH-dependent interaction of the GHR with SOCS3 was clearly evident, as judged by coimmunoprecipitation of tyrosine-phosphorylated GHR and SOCS3 when either α-GHR or α-SOCS3 was used as the precipitating antibody (Fig. 3C). In agreement with earlier findings (29), the abundance of tyrosine-phosphorylated SOCS3 in serum-replete fibroblasts exceeded that in serum-starved cells. Tyrosine phosphorylation of the GHR in serum-replete fibroblasts was substantial, even in the absence of GH (Fig. 3C, lane 1). Coprecipitation of SOCS3 with Ab 2941 in the absence of GH was also substantial and presumably reflects interaction of the SOCS3 SH2 domain with tyrosine-phosphorylated GHR. It thus appears likely that failure to detect interaction of SOCS3 and the GHR in primary adipocytes reflects the low abundance of SOCS3 rather than inadequacy of the antiserum to immunoprecipitate and detect SOCS3.

Figure 3.

A tyrosine-phosphorylated 116-kDa protein (p116) coimmunoprecipitates with CIS, but not SOCS3, following GH stimulation of adipocytes. Freshly isolated rat adipocytes were treated with hGH for the indicated times. The cells were then lysed, and the lysates were subjected to immunoprecipitation with α-CIS (A) or α-SOCS3 (B). The recovered proteins were analyzed by SDS-PAGE and detected on Western blots with α-phosphotyrosine 4G10. The results shown are representative of two experiments. C, A tyrosine-phosphorylated 116-kDa protein coprecipitates with SOCS3 after GH stimulation of 3T3-F442A fibroblasts. Cells were grown to confluence in 100-mm dishes and serum-starved overnight in medium containing 2% BSA where indicated. After incubation with or without 25 nm hGH for 5 min, the cells were lysed in 1 ml buffer containing 1% Triton X-100, and proteins recovered by immunoprecipitation with Ab 2941 or α-SOCS3 were analyzed by Western blot with α-phosphotyrosine PY99. The results shown are representative of two experiments.

The apparent size of the 116-kDa protein recovered with α-CIS is consistent with the possibility that it is the GHR, but attempts to detect it by stripping and reblotting with Ab 2941 were unsuccessful. However, the 37- and 32-kDa forms of CIS both coprecipitated with the GHR, as shown in Fig. 4. As CIS coprecipitated with the GHR only in extracts of adipocytes that were treated with GH, it appears that CIS does not interact with the GHR constitutively, but is recruited to the tyrosine-phosphorylated receptor.

Figure 4.

CIS coprecipitates with the GHR. Adipocyte lysates were treated with α-CIS or Ab 2941 immobilized on protein A-agarose. The recovered proteins were analyzed by immunoblotting with α-CIS. The results shown are representative of two experiments.

Sequential immunoprecipitations with α-CIS followed by Ab 2941 suggest that only a small fraction of the tyrosine-phosphorylated GHR coprecipitates with CIS (Fig. 5). Immunoprecipitation first with α-CIS under optimal conditions resulted in partial recovery of phosphorylated GHR and left a substantial fraction behind, as indicated by its recovery in a subsequent immunoprecipitation with Ab 2941. However, when the order of the immunoprecipitations was reversed, no tyrosine-phosphorylated proteins that could be recovered with α-CIS remained in the supernatant after immunoprecipitation with Ab 2941. One interpretation of this result is that p116 is a fraction of the tyrosine-phosphorylated GHR, but the possibility that p116 is another protein that is tyrosine phosphorylated in response to GH and coimmunoprecipitates with the GHR cannot be excluded by these results.

Figure 5.

Sequential immunoprecipitation of p116 and the GHR. Freshly isolated adipocytes were treated with 25 nm hGH for 2 min. Cell lysates were immunoprecipitated with either α-CIS or Ab 2941 at 4 C overnight and recovered on protein A beads. The supernatant was then immunoprecipitated with the other antibody. The recovered proteins were analyzed by Western blotting with α-phosphotyrosine 4G10. The results shown are representative of two experiments.

To further characterize p116, we determined whether it is a glycoprotein and compared its electrophoretic mobility with that of the GHR after digestion with endoglycosidase F. Tyrosine-phosphorylated p116 and GHR were recovered from extracts of GH-treated adipocytes by immunoprecipitation with α-CIS or Ab 2941, respectively, and half of each recovered protein was digested overnight with endoglycosidase F. As shown in Fig. 6, the products resulting from deglycosylation of both proteins had identical electrophoretic mobilities, suggesting that each contained carbohydrate moieties that contributed approximately 20 kDa to the apparent mass. This estimate of the carbohydrate component of the GHR is in agreement with earlier findings in this laboratory (30) and those reported by Kim et al. (31). Similar results were obtained with samples immunoprecipitated from extracts of GH-treated HEK293 cells that express the rat GHR at high levels (∼10⁵ receptors/cell; Fig. 6B).

Figure 6.

p116 is a glycoprotein. Freshly isolated rat adipocytes (A) or HEK293-GHR cells (B) were stimulated with 25 nm hGH for 2 min. p116 and GHR were precipitated from the lysates with α-CIS or Ab 2941, and half of each precipitate was digested with endoglycosidase F. The other half served as the control. The samples were then analyzed by SDS-PAGE, followed by immunoblotting with α-phosphotyrosine 4G10. The results shown are representative of three experiments.

To recover larger amounts of p116 than could be obtained from adipocytes by immunoprecipitation, CIS fused to GST was bound to immobilized glutathione and used in a pull-down procedure to purify p116. Using this procedure, a protein similar to that coprecipitated by α-CIS was recovered from both adipocyte extracts and HEK293-GHR cell extracts (Fig. 7). Reblotting with antibodies that recognize either the cytosolic domain (Ab 2941) or the extracellular domain (BB74) of the GHR identified the 116-kDa protein in HEK extracts as the GHR, although once again attempts to identify p116 recovered from adipocyte extracts with antibodies to GHR failed. These results support the supposition that the 116-kDa protein is the GHR and suggest that because of its lower abundance in adipocyte extracts, the tyrosine-phosphorylated GHR recovered from adipocyte extracts using the CIS fusion protein may still be below the threshold of detectability by Ab 2941. Detection of p116 by blotting with α-phosphotyrosine might afford substantially greater sensitivity than can be achieved with Ab 2941.

Figure 7.

Recovery of p116 and the GHR by affinity chromatography using a GST-CIS fusion protein. The fusion protein, prepared as described in Materials and Methods, was immobilized on glutathione beads and incubated with lysates from 2 × 10⁶ adipocytes or 7 × 10⁷ HEK293-GHR cells that had been incubated with or without 25 nm hGH for 2 min. Proteins eluted with glutathione were analyzed by SDS-PAGE, followed by immunoblotting with α-phosphotyrosine PY99 or one of two α-GHR antisera: BB 74, which recognizes the extracellular domain, or Ab 2941, which recognizes the intracellular domain of the receptor. The results shown are representative of two experiments.

It is possible that only a small fraction of the receptor is available to react with CIS either because much of the receptor is not tyrosine phosphorylated or because other proteins compete with CIS for binding to the phosphorylated receptor. To distinguish between these possibilities, we isolated the phosphorylated receptors from extracts of GH-treated adipocytes and compared their abundance with that of the total recoverable pool of receptors. Extracts of adipocytes that had been treated with GH for 2 min, a time when tyrosine phosphorylation of the GHR was maximal (see Fig. 9), were applied to an affinity column comprised of α-phosphotyrosine PY99 immobilized on agarose beads. The column was repeatedly washed to remove proteins lacking phosphotyrosine before tyrosine-phosphorylated proteins were eluted with buffer containing 2,4-dinitrophenylphosphate. The abundance of the GHR in protein samples applied to and eluted from the column was estimated using an immunofunctional assay (20) in which [¹²⁵I]GH coprecipitated with the GHR. Proteins in the flow-through and wash buffers were analyzed by Western blot with α-phosphotyrosine to verify that the affinity column retained all of the tyrosine-phosphorylated GHR in the sample. No tyrosine-phosphorylated GHR could be immunoprecipitated from the flow-through and wash buffers (Fig. 8), but most of that present in the initial extract was recovered in the eluate displaced by 2,4-dinitrophenylphosphate. Quantitative estimates of the total amount of GHR eluted from the column indicated that most of the receptor is not tyrosine phosphorylated and was recovered in the combined flow-through and wash buffers. In two experiments the nontyrosine-phosphorylated GHR was 71% or 77% of the total GHR in the initial extract, and the tyrosine-phosphorylated GHR eluted with 2,4-dinitrophenylphosphate was 22% or 18%, respectively. Thus, less than 25% of the GHR is tyrosine phosphorylated under experimental conditions found to produce maximal tyrosine phosphorylation. As less than half of the tyrosine-phosphorylated receptor appears to coprecipitate with CIS (Figs. 4 and 5), the amount of GHR recovered by immunoprecipitation with α-CIS may be less than 10% of that recovered by immunoprecipitation with Ab 2941, which could account for our failure to detect the GHR by Western blotting after immunoprecipitation with α-CIS.

Figure 8.

Affinity chromatography of GHR using immobilized α-phosphotyrosine. A cell extract prepared from approximately 7 × 10⁶ adipocytes (350 μl packed cells) that were incubated with 25 nm [¹²⁵I]GH for 2 min was applied to an affinity column (0.25 ml bed volume) containing 500 μg immobilized PY99. The column was rinsed with Tris-buffered saline, and then proteins containing phosphotyrosine were eluted with buffer containing 0.1 m 2,4-dinitrophenylphosphate. Samples of the cell lysate (A), the flow-through and wash buffer combined (B), and the proteins displaced by 2,4-dinitrophenylphosphate (C) were immunoprecipitated with Ab 2941 in the presence of 1 nm [¹²⁵I]GH, and the amount of GHR present was estimated from the coprecipitation of [¹²⁵I]GH. Samples of the recovered proteins representing 150 μl packed cells were also analyzed by Western blot with α-phosphotyrosine PY99. The results shown are representative of two experiments.

Figure 9.

The effect of GH on CIS and tyrosine phosphorylation of key GH signaling molecules in freshly isolated and GH-deprived adipocytes. CIS, GHR, JAK2, or STAT5 was immunoprecipitated from lysates of adipocytes that had been incubated with GH for the indicated times. The recovered proteins were analyzed by immunoblotting with α-phosphotyrosine 4G10. CIS protein levels were monitored by blotting the membrane containing immunoprecipitated CIS with α-CIS. The results shown are representative of two experiments.

Another estimate of how much GHR coprecipitates with α-CIS was obtained by using the immunofunctional assay to compare the amounts of GHR recovered with Ab 2941 and α-CIS. In four experiments, the amount of [¹²⁵I]GH that coprecipitated with α-CIS was 3.22 ± 0.62% of that recovered by immunoprecipitation with Ab 2941. For comparison, the amount of [¹²⁵I]GH recovered with α-JAK2, which also binds to the cytoplasmic domain of the GHR, was 1.42 ± 0.36% of that recovered with Ab 2941. Of course, these results may underestimate the extent to which the GHR binds to CIS or JAK2 because protein-protein interactions between GH, GHR, the immunoprecipitating antiserum, and CIS or JAK2 may not be strong enough to prevent some dissociation during the experiment. Nevertheless, these results are consistent with our estimate that only a small fraction of the GHR coprecipitates with CIS based on the recovery of tyrosine-phosphorylated GHR.

Overexpression of CIS in COS-1 cells interfered with GH signaling, as indicated by several techniques, including electrophoretic mobility shift assays in which the ability of STAT5b to bind to DNA was measured (6, 7). Although CIS appeared to interfere with GH signaling by more than one mechanism, these studies compliment those in which the GHR was expressed in CHO cells to demonstrate the importance of ubiquitination (17) and proteasomal degradation (4, 32) in GH signaling. All of these studies are consistent with the possibility that CIS may facilitate degradation of the GHR, which might be expected to hasten the termination of GH signaling. To evaluate the importance of CIS levels in adipocytes for GH signaling, the time course of GH-dependent phosphorylation of GHR, JAK2, and STAT5 was determined in freshly isolated adipocytes that express relatively high levels of CIS and in adipocytes whose expression of CIS was reduced by preincubation for 3 h in buffer lacking GH. Maximum phosphorylation of the GHR immunoprecipitated with either Ab 2941 or α-CIS was seen at 2 min after GH treatment in both adipocyte populations and was sharply curtailed by 10 min (Fig. 9). A similar, but less pronounced, decline in tyrosine phosphorylation of JAK2 was also evident at 10 min in both cell populations, whereas tyrosine phosphorylation of STAT5 continued unabated for at least 20 min and appeared to be somewhat more pronounced in the GH-deprived adipocytes. Consistent with the results shown in Fig. 2, the abundance of CIS, especially the ubiquitinated 37-kDa form, was greater in freshly isolated than in GH-deprived cells, but despite the apparently lower abundance of CIS in the GH-deprived cells, there was not a significant decrease in either the CIS/p116 association or its rate of disappearance. Incubation with GH for 20 min did not alter the abundance of CIS in either freshly isolated or GH-deprived adipocytes, so the precipitous decline in the CIS/p116 complex is not attributable to the loss of CIS.

To evaluate the role of proteosomes in the transient phosphorylation of the GHR, the cells were incubated with GH for various times in the presence or absence of the proteasomal inhibitor MG-132 (Fig. 10A), or the more specific inhibitor clasto-lactacystin-β-lactone (33). Tyrosine phosphorylation of the GHR and p116, immunoprecipitated with Ab 2941 or α-CIS, respectively, was prolonged for at least 1 h by both inhibitors, consistent with the possibility that the GHR pool that interacts with CIS may be destined for proteasomal degradation.

Figure 10.

Tyrosine phosphorylation of proteins recovered by immunoprecipitation with α-CIS or Ab 2941 is sustained in the presence of proteasome inhibitors. Freshly isolated adipocytes were incubated with 25 nm GH for the indicated times in the presence or absence of 20 μm clasto-lactacystin-β-lactone. Cell lysates were immunoprecipitated with Ab 2941 or α-CIS, and the recovered proteins were analyzed by Western blot with α-phosphotyrosine PY99. Samples represent proteins recovered from 0.125 ml packed cells (Ab 2941) or from 0.5 ml packed cells (α-CIS). The results shown are representative of three experiments conducted with either 20 μm clasto-lactacystin-β-lactone or 50 μm MG-132.

Discussion

In their survey of SOCS family gene expression in rat tissues, Tollet-Egnell et al. (27) reported that CIS mRNA is nearly 5 times as abundant as the mRNAs for SOCS2 and SOCS3 in adipose tissue. The absence of SOCS2 mRNA in extracts of freshly isolated adipocytes may indicate that its prominence in whole tissue extracts reflects its expression in nonadipocyte elements in the tissue. CIS mRNA virtually disappeared after hypophysectomy (27), whereas the abundance of SOCS3 was undiminished, suggesting that in vivo expression of SOCS3 in adipose tissue is not solely dependent on GH. These observations are consistent with the present findings that in GH-treated adipocytes, CIS, but not SOCS3, associates with a prominent tyrosine-phosphorylated protein that almost certainly is the GHR.

Although the abundance of the tyrosine-phosphorylated adipocyte protein that coprecipitates with α-CIS is below the detection threshold of our GHR antisera in Western blots, the presence of CIS in immunoprecipitates of the activated GHR provides strong evidence for their association. The tyrosine-phosphorylated protein that associates with CIS has the same electrophoretic mobility as the GHR both before and after removal of its N-linked carbohydrates. Additionally, α-CIS failed to immunoprecipitate any tyrosine-phosphorylated protein from adipocyte extracts that were first cleared of the GHR by treatment with GHR antiserum. Finally, the GHR was recognized by two different antisera in proteins pulled down by a CIS-GST fusion protein from extracts of HEK293 cells that express high levels of the GHR.

The ability of CIS to bind to the tyrosine-phosphorylated GHR in vitro was demonstrated using recombinant fragments of the cytosolic domain of the GHR (7) and was also demonstrated in COS cells that were cotransfected with both CIS and the GHR (6, 7). The present findings indicate that these two proteins can interact in a GH-dependent manner within a physiological target for GH, the primary adipocyte, in which these proteins are present at physiological concentrations and stoichiometric proportions. Furthermore, this interaction occurs equally well in freshly isolated adipocytes and in fat cells whose levels of CIS are somewhat reduced after incubation without GH for 3 h. Hence, although transcription of the CIS gene apparently depends on GH (27), turnover of CIS protein is slow enough that concentrations needed for recruitment to phosphorylated tyrosine residues in the GHR are sufficient throughout periods extending between bursts of GH secretion (34).

Adipocytes that are deprived of GH for 3 h or longer acquire the ability to respond to GH in an insulin-like manner (24), whereas even brief exposure of the cells to GH renders them refractory to this effect when reexposed to GH within the next few hours. However, blockade of RNA synthesis with actinomycin D or of protein synthesis with cycloheximide prevents the onset of refractoriness and prolongs sensitivity to the insulin-like effects of GH (9). These observations raised the possibility that CIS/SOCS family proteins might account for the refractory phenomenon because they are acutely induced by GH and have negative effects on signaling by cytokine family receptors. However, although interaction between CIS and the GHR might interfere with the signal that evokes the insulin-like metabolic responses and might abbreviate the duration of tyrosine phosphorylation of the GHR and JAK2, the interaction of CIS with the GHR is equally evident in freshly isolated cells, which are refractory to insulin-like actions of GH, and in adipocytes that are responsive to these actions after incubation for 3 h in GH-free medium. These observations make it unlikely that the CIS/GHR interaction accounts for the loss of sensitivity to the insulin-like action of GH. Ji et al. (35) also concluded that although down-regulation of the GHR might account for the transient tyrosine phosphorylation of the GHR and activation of the JAK/STAT pathway, downstream, postreceptor effects account for refractoriness of other signaling pathways.

The expression of genes that encode SOCS3 and JAB and the GH-dependent increase in their mRNAs were confirmed in rat adipocytes. However, these proteins appear to have less effect on signaling by the GHR in primary adipocytes than has been described for other cells (5–7, 29, 36), as tyrosine-phosphorylated proteins such as the GHR or JAK2 were not detected in extracts of GH-treated adipocytes after immunoprecipitation with antisera that recognize SOCS3 or JAB. The apparent low abundance of JAB in adipocytes may account for this result. The failure to detect interaction of the GHR with SOCS3, which, like CIS, also binds to the tyrosine-phosphorylated recombinant GHR (7) contrasts sharply with the readily detectable interaction of SOCS3 with the GHR in 3T3-F442A fibroblasts demonstrated here and by others in these (29) and other cells (6, 36). It is also noteworthy that the expression of SOCS3 protein was rapidly up-regulated by GH in serum-starved 3T3-F442A adipocytes (29), but changed little or not at all in primary adipocytes. It is possible that the prevailing levels of proteins in rat adipocytes may allow some other protein to compete effectively with SOCS3 for binding to the GHR or that the expression level of SOCS3 is simply too low to associate with detectable amounts of phosphorylated GHR.

Transient tyrosine phosphorylation of the GHR and/or JAK2 similar to the results presented here has been widely reported (3, 8, 31, 37), but differences in the rates at which these proteins are dephosphorylated in the continued presence of GH suggest that these transient responses may be governed differently in various cell types. The prolonged response seen in cells that expressed a mutant GHR lacking tyrosine 595 suggested that recruitment and activation of phosphatase SHP-2 may be important for the dephosphorylation of GH signaling proteins (3), but the interference with GH signaling seen when a mutant SHP-2 that lacked catalytic activity was expressed suggested that SHP-2 has a positive, rather than a negative, role in GH signaling (31). Prolonged tyrosine phosphorylation of JAK2 was also seen in mice lacking the phosphatase SHP1, suggesting that this enzyme might contribute to the transient GH signal by dephosphorylating JAK2 (37), but this enzyme does not appear to interact with the GHR (31, 37).

Studies by Verdier et al. (38) demonstrated that receptor degradation and diversion, rather than phosphatase activation, account for the transient tyrosine phosphorylation of the erythropoietin receptor, whose signaling pathway resembles that of the GHR. The ability of proteasomal inhibitors to prolong the otherwise transient tyrosine-phosphorylated GHR-CIS complex suggests that turnover of the GHR in rat adipocytes is similar to that in cultured hepatocytes (7) and in CHO cells that express rabbit GHR (4), where proteasomes degrade the cytosolic portion of the receptor soon after it is internalized. By initiating the recruitment of CIS and accelerating subsequent degradation of the GHR, ligand binding may effect down-regulation of receptors on the cell surface and reduce the stimulation of JAK2 activity by the GHR. The finding that half of the detectable GHR in rat hepatoma cells disappears within 10 min of treatment with 25 nm GH (35) is consistent with rapid degradation of the GHR in response to ligand binding, but suggests that the fraction of the GHR pool that is tyrosine phosphorylated and then down-regulated may be larger in these cells than in adipocytes.

The observation that the SOCS box confers a ubiquitin ligase activity on elongin C (15) suggests that interaction of the GHR with CIS might account for the finding that the GHR is ubiquitinated (16). It is unclear whether diversion of newly synthesized receptors to proteasomes, as has been shown to occur with the erythropoietin receptor (38) might also occur in adipocytes. Receptor degradation and dephosphorylation may both contribute to the transient kinetics of this signaling pathway.

Although the fraction of the GHR that binds to CIS may be underestimated by the procedures employed here, it appears that only a small subset of the cellular pool of the GHR in adipocytes participates in signaling, as more than 70% of the GHR is not even tyrosine phosphorylated in the presence of a saturating concentration of GH and at a time when tyrosine phosphorylation appears to be maximal. As there are eight intracellular tyrosines in the rat GHR, some receptors may be phosphorylated on membrane-proximal tyrosines rather than the C-terminal residues that interact with CIS (6), and it is expected that only a fraction of the tyrosine-phosphorylated receptors bind to CIS. Perhaps only a fraction of the tyrosine-phosphorylated GHR may be selected for proteasomal degradation by binding to CIS, but we cannot exclude the possibility that eventually the entire tyrosine-phosphorylated GHR pool is degraded via a CIS-mediated pathway. The consistent finding that proteasome inhibitors preserve but do not result in the time-dependent accumulation of more tyrosine-phosphorylated GHR suggests that other factors, such as phosphatase activity, incomplete inhibition of proteasomes, or the availability of alternative degradative pathways, may contribute to the observed results.

Acknowledgments

This work was supported by NIH Grant R01-DK-19392.

The findings represent the views of the authors and not necessarily those of the NIH. A preliminary report of these findings was presented at the 81st Annual Meeting of The Endocrine Society, San Diego, CA, 1999. These studies were included in a thesis presented by L.D. to the Graduate School of Biomedical Sciences at the University of Massachusetts Medical School in partial fulfillment of the requirements for the Ph.D.

Abbreviations

 
• Ab

  Antibody;

 
• CHO

  Chinese hamster ovary cells;

 
• CIS

  cytokine-induced Src homology domain 2 protein (also called SOCS, supressors of cytokine signaling, or SSI, STAT-induced STAT inhibitor);

 
• GHR

  GH receptor;

 
• GST

  glutathione-S-transferase;

 
• HEK

  human endothelial kidney cells;

 
• hGH

  human GH;

 
• HRP

  horseradish peroxidase;

 
• JAB

  Janus kinase 2-binding protein, also called SOCS1;

 
• JAK

  Janus kinase;

 
• KRPG

  Krebs-Ringer phosphate buffer enriched with 5.5 mm glucose and 40 mg/ml BSA;

 
• SH2

  Src homology domain 2;

 
• SHP-2

  SH2 containing phosphatase 2;

 
• STAT

  signal transducer and activator of transcription.