Growth hormone (GH) and a GH antagonist promote GH receptor dimerization and internalization.

It has previously been shown that a human growth hormone (hGH) analog, hGH-G120R, acts as a GH antagonist (Chen, W. Y., Wight, D. C. , Wagner, T. E., and Kopchick, J. J. (1990) Proc. Natl. Acad. Sci. U. S. A. 87, 5061-5065; Chen, W. Y., White, M. E., Wagner, T. E., and Kopchick, J. J. (1991) Endocrinology 129, 1402-1408; Chen, W. Y., Chen, N-Y., Yun, J., Wang, X. Z., Wagner, T. E., and Kopchick, J. J. (1994) J. Biol. Chem. 269, 15892-15897). In this study, we report the ability of hGH and hGH-G120R to be internalized by GH receptor expressing cells. Additionally, results of chemical cross-linking experiments revealed that both native hGH and hGH-G120R form complexes similar in size to that expected for hGH when bound to recombinant hGH-binding protein (bp). The molecular mass of the complex was determined to be approximately 280 kDa which is consistent with multiple receptors interacting with the ligand. The predominant radiolabeled band detected was a complex of approximately 140 kDa which probably represents one GH molecule bound to one GH receptor. The cross-linked complexes were not detected in the presence of excess unlabeled hGH or hGH-G120R and were not observed in cells which do not express detectable levels of GH receptors. Also, GH induced tyrosine phosphorylation of a complex of proteins of approximately 95 kDa in these cells whereas hGH-G120R did not. Thus, we have separated the hGH or hGH-G120R/GHR binding and internalization capabilities from the ability to stimulate tyrosine phosphorylation of intracellular proteins.

Recently, the crystal structure of hGH with the extracellular domain of the hGH-binding protein (hGHbp) was solved. It was found to exist as a hGH⅐(hGHbp) 2 complex (4,29). Additionally, GH dependent dimerization of this Escherichia coli expressed hGHbp was found to occur sequentially (29). In this model, hGH was shown to bind one hGHbp molecule through site 1 on hGH and then a second hGHbp molecule binds through site two of hGH subsequently establishing a hGH⅐(hGHbp) 2 complex. It was suggested that formation of the 1-ligand⅐2-receptor dimeric complex may be important in the GH-induced signal transduction system (29,30). This type of GH/GHR signal transduction mechanism is analogous to that used by many tyrosine kinase receptors, such as epidermal growth factor and platelet-derived growth factor, in which binding of one hormone molecule to its receptor is thought to induce formation of a dimer, through a 2-hormone⅐2-receptor complex (31). Binding of these growth factors to their cognate receptors activate tyrosine kinases and receptor autophosphorylation. Therefore, receptor dimerization may be a critical step in mediating biological activities for a number of growth factors.
It has clearly been demonstrated that the third ␣-helix of GH is critical for its biological activity (1)(2)(3). In a structure-function study of this helical region, bGH-Gly 119 and hGH-Gly 120 were identified as amino acids critical for growth promotion (2,3). Transgenic mice expressing these GH antagonist genes possessed a dwarf phenotype (1,32,33). In fact, substitution of several amino acids (except alanine) of this Gly residue resulted in GH antagonists which were found to be active both in vitro and in vivo (2,3,12,33).
Similarly, a hybrid receptor containing the extracellular binding domain of the hGHR linked to the transmembrane and intracellular domains of the murine granulocyte colony-stimulating factor receptor was generated and expressed in a myeloid leukemia cell line, FDC-P1 (34). Treatment of these cells with hGH resulted in cellular proliferation. However, treatment of these cells with hGH-G120R, which contains a functional site 1 but a sterically blocked site 2, failed to promote cellular proliferation. Therefore, it was suggested that hGH-G120R acted as a GH antagonist, presumably by its inability to dimerize GHRs (34).
It has recently been reported that hGH down-regulates GHRs in IM-9 lymphocytes whereas hGH-G120R does not (30). This result may imply that hGH and hGH-G120R are "recognized" differently by the cell and suggests that hGH-G120R is not internalized following binding to GHRs.
In order to test the ability of hGH-G120R to dimerize and internalize, hGH-G120R was purified and iodinated. The iodinated forms of hGH and hGH-G120R were analyzed by binding and cross-linking to cells expressing the GHRs. Additionally, the ability of the GH antagonist to be internalized was determined.

MATERIALS AND METHODS
Iodination of Hormones-Human GH and a GH antagonist, hGH-G120R, were purified as described previously (35) and were labeled with Na 125 I by the lactoperoxidase method to a specific activity of 80 -105 Ci/g (36). Briefly, 1.0 mCi of Na 125 I was added to 1.0 mg of hGH or hGH-G120R. Lactoperoxidase (10 g dissolved in 10 l of 0.4 mol/liter acetate buffer, pH 5.6) and H 2 O 2 (5 l of 1.76 mmol/liter) were then added. After 30 min, the reaction was terminated by the addition of 100 l of transfer buffer (0.47 mol/liter sucrose, 0.06 mol/liter KI, sodium azide 0/02%, pH 7.6). Radiolabeled hGH was then separated by Sephadex G-100 chromatography.
Radioimmunoassay-A radioimmunoassay hGH kit (Hybritech, San Diego, CA) was used in order to determine the concentrations of hGH and hGH-G120R. However, only hGH was detected by the radioimmunoassay kit (3). hGH-G120R was not detectable by radioimmunoassay but was recognized by immunoblotting with a polyclonal hGH antibody (kindly provided by Eli Lilly, IN). Therefore, the epitope for the monoclonal antibody from the Hybritech radioimmunoassay kit may be near Gly 120 or in the third helical region. A second hGH radioimmunoassay kit (Nichols Institute Diagnostics) was able to recognize both hGH as well as hGH-G120R in order to determine GH concentrations (3).
GH/GHR Internalization-An acid extraction procedure (37) was used to differentiate surface bound 125 I-GH and internalized 125 I-GH. Briefly, mouse L cells (MLCs), pGHR-W10 (MLCs which express the porcine GHR), and IM-9 cells were maintained in DMEM or RPMI 1640 plus 10% fetal bovine serum. For the internalization studies the cells were grown to confluence in six-well 35-mm culture plates and depleted of GH in DMEM for 16 h as described previously (38,39). One ml of ice-cold phosphate-buffered saline, 1% bovine serum albumin solution containing 100,000 cpm of 125 I-hGH or 125 I-hGH-G120R was added to each culture dish which were pre-chilled on ice to prevent internalization. We have previously established that the binding affinity of hGH and hGH-G120R for the pGHR are approximately 1.0 nM. 2 The cells were allowed to reach equilibrium at 4°C for 24 h, after which the unbound ligand was removed by rinsing (three times) with ice-cold phosphate-buffered saline/1% bovine serum albumin. The cells were then incubated at 37°C for various time intervals (0, 5, 10, 15, 30, 45, 60, 90, and 120 min) and exposed to 0.5 ml of ice-cold solution of 50 mM glycine, 150 mM NaCl, pH 3.0, for 5 min. This solution was collected and represented the surface bound ligand. The cell monolayer was washed with 0.5 ml of 50 mM glycine, 150 mM NaCl, pH 3.0, and added to the previous collection. The cell monolayer was then solubilized with 0.5 ml of 0.1 N NaOH, 1% SDS and represented the internalized ligand. The radioactivity of the surface bound and internalized ligand were measured using a ␥-counter (Beckman 5500).
GH/GHR Cross-linking-125 I-hGH and 125 I-hGH-G120R were chemically cross-linked to MLCs, pGHR-W10 and pGHR-TR1 cells. Briefly, cells were plated in 100-mm culture dishes and grown to confluence. Cells were depleted of GH overnight in DMEM minus serum. Cells were then incubated in phosphate-buffered saline, 1% bovine serum albumin containing 300,000 cpm of either 125 I-hGH or 125 I-hGH-G120R for 2 h at 25°C with gentle agitation. Additionally, unlabeled hGH and hGH-G120R (500 ng/ml) were each used to specifically compete with the 125 I-hGH and 125 I-hGH-G120R. The bound radiolabeled ligand was chemically cross-linked to the pGHR with 2.5 mM bis(sulfosuccinimidyl)suberate, BS 3 (Pierce), for 90 min at 4°C. The cells were then scraped and homogenized in 0.25 M sucrose in 10 mM Tris-Cl, pH 7.4, 1 mM EDTA, containing a mixture of the following protease inhibitors: phenylmethylsulfonyl fluoride (Sigma), tosyl lysine chloromethyl ketone (Sigma), and tosylamide phenylethylchloromethyl ketone (Sigma) at a final concentration of 2 mM each. The homogenate was centrifuged at 14,000 ϫ g for 15 min. The membrane fraction was suspended in loading buffer (625 mM Tris, 1% SDS, 15% glycerol, 1% ␤-mercaptoethanol) and subjected to 5-15% SDS-PAGE (40). After electrophoresis, the gel was fixed in 10% methanol and 7.5% glacial acetic acid for 30 min, soaked in Enlightening (DuPont) for 30 min, dried, and exposed to XAR-II film (Kodak).
Similarly, a pGHR truncation analog consisting of amino acids 1-291, pGHR-TR1 (41), was chemically cross-linked to 125 I-hGH or 125 I-hGH-G120R in order to obtain a more precise determination of the M r of the GH⅐GHR complex as well as the dimeric complex.
GH-induced Protein Tyrosine Phosphorylation Assays-The cell lines described above were plated in 6-well tissue culture plates and propagated to confluence. For the protein tyrosine phosphorylation assays, which we have termed the pp95 assay, GH in the medium was removed by incubating the cells in serum-free medium (DMEM) overnight. Subsequently, the pp95 assay was performed as described previously (21,39,41). Briefly, cells were treated with or without hGH or hGH-G120R (500 ng/ml) at 37°C for 10 min. The cells were lysed in SDS-PAGE lysis buffer, separated by 7% gel electrophoresis, and transferred to a Hybond-ECL membrane (Amersham). The membrane was blocked with phosphate-buffered saline, 4% bovine serum albumin solution for at least 6 h and incubated with a phosphotyrosine antibody, PY20 (ICN, Costa Mesa, CA). The membrane was developed by the conventional ECL procedure (Amersham). The signal was visualized by exposure to x-ray film.

RESULTS
Internalization of 125 I-hGH and 125 I-hGH-G120R-In order to determine whether hGH-G120R was able to be internalized, experiments were conducted using cell lines which express GHR, i.e. pGHR-W10, and human IM-9 lymphocytes. Previous results have shown that pGHR-W10 cells were able to internalize approximately 85% of the specific bound 125 I-pGH (38,39). Results of the internalization assay of hGH and the hGH antagonist are shown in Fig. 1 (A and B). Both the 125 I-hGH and 125 I-hGH-G120R were internalized by pGHR-W10 cells and IM-9 lymphocytes with radioactivity increasing from 5 to 65% within 30 min at 37°C. Approximately 85% of the bound radiolabeled ligand was internalized by 1 h. MLCs did not significantly internalize hGH or hGH-G120R (data not shown).
Cross-linking of 125 I-hGH and 125 I-hGH-G120R-In the presence of 125 I-hGH or 125 I-hGH-G120R and a chemical crosslinker, BS 3 , radiolabeled bands with molecular masses of approximately 70, 140, and 280 kDa were observed in cells expressing the pGHR (Fig. 2, lanes 5 and 11). These radiolabeled complexes were not seen in MLCs which do not express pGHRs (Fig. 2, lanes 2-4 and 8 -10). Additionally, the observed complexes were able to be specifically competed with unlabeled 2 X. Wang and J. J. Kopchick, unpublished results.

FIG. 1. Internalization time course of 125 I-hGH and 125 I-hGH-G120R into pGHR-W10 cells (Panel A) and IM-9 lymphocytes (Panel B).
Cells were grown to confluence in six-well culture plates, depleted of serum, and incubated with 100,000 cpm of 125 I-hGH or 125 I-hGH-G120R for 24 h at 4°C. Subsequently, the cells were incubated at 37°C for 0, 5, 10, 15, 30, and 120 min during which time GH was internalized. Each point on the graph represents the mean value of three experiments performed in duplicate. See "Materials and Methods" for details.
A radiolabeled band with a molecular mass of approximately 75 kDa was observed for pGHR-TR1 when cross-linked to either 125 I-hGH or 125 I-hGH-G120R (Fig. 3, lanes 5 and 11). Upon subtraction of the molecular mass of hGH, 22 kDa, from the complex, pGHR-TR1 was found to possess a mass of approximately 53 kDa. Additionally, a larger radiolabeled complex of 150 kDa was observed with pGHR-TR1 upon crosslinking to 125 I-hGH or 125 I-hGH-G120R. Both the 75-and 150-kDa complexes were able to be specifically competed with unlabeled hGH or hGH-G120R (Fig. 3, lanes 6 and 12). Also, a radiolabeled complex of approximately 140 kDa was detected in pGHR-W10 cells (Fig. 3, lanes 3 and 9) that were specifically competed with excess hGH or hGH-G120R (Fig. 3, lanes 4 and  10). No complexes were seen in MLC (Fig. 3, lanes 1, 2, 7, and  8).
pp95 Induction Assay-The ability of hGH or hGH-G120R to induce tyrosine phosphorylated proteins of approximately 95 kDa in pGHR-W10 cells was examined. hGH and pGH were able to induce pp95 in pGHR-W10 cells (Fig. 4, lanes 2 and 4,  respectively). No pp95 induction was observed when pGHR-W10 cells were incubated with hGH-G120R (Fig. 4, lane 3). DISCUSSION It has been established that E. coli-derived hGHbp is able to form a dimer when incubated with GH (4,29). This dimeric complex consists of one hGH molecule bound to two hGHbp with hGH containing two binding sites for the hGHbp. It was proposed that GH site 1 binds with one GHbp molecule and then a second GHbp interacts with GH site 2 (29). De Vos et al. (4) have demonstrated that the hGH site 2 is found on the exposed sides of helices 1 (amino acids Phe 1 , Ile 4 , and Arg 8 ) and 3 (Asp 116 ) (4). Additionally, hGH-G120R, which retains a functional site 1 but a sterically blocked site 2, has been shown to inhibit proliferation of FDC-P1 cells transfected with a hybrid receptor that contained the extracellular domain of the hGHR linked to the transmembrane and intracellular domain of the murine granulocyte macrophage colony-stimulating factor receptor (34). It was hypothesized that once hGH-G120R is bound to its receptor, it could not dimerize and induce a GH signal. Therefore, these results suggest that GHR dimerization is critical for GH-induced signal transduction (19,34).
GH treatment of NIH 3T3-F442A fibroblasts resulted in tyrosine phosphorylation of GHRs as a result of association with an associated tyrosine kinase, JAK2, a non-receptor associated tyrosine kinase (13,14,18,20). It has been identified as a GH-activated tyrosine kinase responsible for self-phosphorylation and tyrosine phosphorylation of GHRs and possibly other intracellular proteins (20,42).
A 95-96-kDa protein (pp95) of unknown identity has been shown to be induced in MLCs stably transfected with the pGHR (pGHR-W10 cells) and 3T3-F442A cells upon treatment with physiological concentrations of GH (21, 38, 41, 43). Addi-  5 and 11) and in the presence of excess hGH or hGH-G120R (lanes 6 and 12). Additionally, pGHR-W10 cells were treated with 125 I-hGH or 125 I-hGH-G120R (lanes 3 and 9) and in the presence of excess hGH or hGH-G120R (lanes 4 and 10). MLCs were also incubated with 125 I-hGH or 125 I-hGH-G120R l (lanes 1 and 7) and in the presence of excess hGH or hGH-G120R (lanes 2 and 8). Lower and upper arrows on the left indicate the migration of the putative GH:GHR monomer and dimer, respectively.
FIG. 4. pp95 induction assay on pGHR-W10 cells that express the pGHR. Cells were plated in six-well culture plates. GH in the medium was removed by incubating the cells in serum-free medium overnight. Subsequently, cells were treated without or with hGH, hGH-G120R, or pGH at 37°C for 10 min and processed as described under "Materials and Methods." Lanes 1-4 represent no treatment, hGH, hGH-G120R, and pGH treatment, respectively. The arrow on the left indicates the position of pp95. tionally, the GH antagonists, bGH-M8 and hGH-G120R, did not induce tyrosine phosphorylation of pp95 (3,21). These results suggest that GH-induced tyrosine phosphorylation of pp95 plays an integral role in GH signal transduction.
GH treatment of a human lymphocyte cell line (IM-9), which contain endogenous hGHRs, results in tyrosine phosphorylation of two proteins with molecular masses of approximately 93 and 120 kDa (19). Additionally, treatment of IM-9 cells with hGH analogs (nM concentrations) directed toward site 1 (K172A, F176A) and site 2 (G120R) were unable to stimulate tyrosine phosphorylation of the 93-kDa protein but stimulated low level tyrosine phosphorylation of the 120-kDa protein (19). Incubation of IM-9 lymphocytes with 0.5 nM recombinant hGH and increasing amounts of hGH-G120R antagonized the ability of hGH to stimulate tyrosine phosphorylation. It was concluded that hGH-G120R antagonized tyrosine phosphorylation due to its inability to dimerize because of a defective site 2. High concentrations (M) of hGH inhibited tyrosine phosphorylation, presumably by inhibiting dimer formation and favoring the monomeric form (e.g. GH⅐GHR complex) (19,34). It should be noted that hGH-G120R treatment of IM-9 lymphocytes was able to stimulate tyrosine phosphorylation of the 120-kDa protein (19). This protein may possibly be JAK2 (20) which would imply that a GH antagonist can bind and activate JAK2 while not activating subsequent signal transduction events.
It has been established that 125 I-hGH is internalized by IM-9 cells at physiological temperatures (44,45). The fate of the internalized hormone is not well understood; however, it has been shown that 25% of the hormone is released into the extracellular environment (46). The fate of the GHR is either lysosomal degradation or exocytosis into the medium as a soluble GHbp. Ilondo et al. (30) proposed a model of GH:GHR internalization followed by exocytosis of the extracellular domain of the full-length GHR, which is the equivalent of the GHbp. Therefore, internalization may be a key pathway in generation of GHbp.
Our studies were directed at determining the ability of a GH antagonist, hGH-G120R, to promote GHR dimerization and to be internalized by use of cells which express GHRs. We hypothesized that we would see receptor dimerization and internalization with hGH but would not detect receptor dimerization and internalization with hGH-G120R. Surprisingly, both 125 Ilabeled hGH and hGH-G120R were able to be internalized following binding to GHR. Our results indicated that 75% of the bound hormone was internalized within 40 min. Therefore, it appears that there is no significant difference in the internalization process between hGH and hGH-G120R (refer to Fig. 1). It may be that 1) GHR internalization is constitutive independent of GH binding or 2) once GH is bound to its receptor, the GH⅐GHR complex is internalized. Regardless of the internalization pathway, we found that hGH-G120R was internalized in a manner similar to that of hGH.
Additionally, we report the ability of hGH and the hGH antagonist, hGH-G120R, to dimerize in what appears to be a (GH) 2 ⅐(GHR) 2 complex, unlike the GH⅐(GHR) 2 complex previously described (34). Three major bands were observed upon cross-linking analysis of 125 I-hGH or 125 I-hGH-G120R with the pGHR with molecular masses corresponding to 70, 140, and 280 kDa (refer to Fig. 2). These radiolabeled complexes were specifically competed with excess unlabeled hGH or hGH-G120R. The predominant radiolabeled band was the 140 kDa protein, which may correspond to the GH⅐GHR monomeric complex. It has been determined that the molecular mass of the pGHR was approximately 118 kDa (38,39). Cross-linking of this protein with either 125 I-hGH or 125 I-hGH-G120R (molecular mass of 22 kDa) would corresponded to the monomeric complex (GH⅐GHR) of 138 kDa. The less abundant, higher molecular mass form of the cross-linked complex is approximately 280 kDa which is consistent with a dimeric complex of two GH molecules bound to two GHR molecules (GH) 2 ⅐(GHR) 2 . Evidence leading to this conclusion is based on cross-linking of 125 I-hGH to a truncated form of the pGHR, pGHR-TR1, which is composed of the pGHR extracellular domain, transmembrane domain, and 3 amino acids of the intracellular domain. Upon cross-linking of GH to these cells, radiolabeled complexes of 75 and 150 kDa were observed (refer to Fig. 3). We believe that the 75-kDa complex is the monomeric complex, one GH molecule and one truncated GHR molecule. Additionally, the 150-kDa complex is consistent with a complex of two GH molecules and two truncated GHR molecules (2GH:2GHR). If the complex were in the ratio of 1GH:2GHR, based on the hypothesis of Fuh et al. (34), the expected molecular mass would be 130 kDa. Similarly, if the full-length pGHR cross-linked to hGH were in the ratio of one ligand and two receptors, the expected molecular mass would be 260 kDa and not the 280 kDa complex observed in Fig. 2. An alternative explanation for the unexpected molecular masses of the cross-linked complexes may be that the presence of 1% ␤-mercaptoethanol in the SDS-PAGE buffer is affecting migration. Previous results have demonstrated that cross-linking of 125 I-hGH to rat adipocytes GHRs resulted in a complex of 127-135 kDa when in the presence of 0.8 -1.2 mM dithiothreitol (47). However, when dithiothreitol concentrations were lowered, (0 -0.4 mM), the GH⅐GHR complex migrated at approximately 116 -125 kDa (47). Therefore, the reduction of GHR disulfide bonds may possibly be responsible for the larger cross-linked complexes seen in our results.
Additionally, we detected a 70-kDa band, which may represent the GH⅐GHbp complex. The GHbp may be the extracellular domain of the full-length pGHR generated by proteolytic cleavage. The 246-amino acid extracellular domain would produce a protein of approximately 25 kDa. We and others have shown that the GHR is heavily glycosylated on asparagine residues (39,48). N-Linked glycosylation comprised approximately 24 kDa of the total molecular mass of the pGHR (39). Therefore, cross-linking of GH or G120R to the extracellular domain (i.e. GHbp) would result in a complex of approximately 70 kDa which can be seen in Fig. 2.
Together, our results indicate that the GH antagonist, hGH-G120R, is able to bind, dimerize, and internalize GHRs in a similar manner as hGH. However, the dimeric complex formed with hGH-G120R is not functional as is the hGH:GHR dimer. Perhaps the hGH-G120R⅐GHR dimeric complex is not biologically equivalent to the native GH:GHR dimer and may be unable to associate with yet to be discovered intracellular elements in order to elicit GH's signal transduction system. Another possible explanation for the separation of GH binding and internalization from subsequent intracellular signaling is the hypothesis that the intracellular "trafficking" pattern is different for hGH⅐receptor versus hGH-G120R⅐receptor complexes. Ultimately, the data suggests that GHR internalization and dimer formation are not positively correlated with GH induced intracellular signaling.