Growth Hormone-promoted Tyrosyl Phosphorylation of a 12 1-kDa Growth Hormone Receptor-associated Protein*

Previous work in multiple cell types has shown that endogenous GH receptors, as well as the cloned liver GH receptor, associate with a tyrosine kinase. How-ever, in SDS-PAGE gels of highly purified, kinase-active GH receptor preparations from 35S-labeled 3T3-F442A cells, only one broad band was detected corresponding to the molecular weight of the GH receptor rather than two bands which might be expected to result from a kinase-receptor heterocomplex. In the present study, a transfected Chinese hamster ovary (CHO) cell line (CH04) that expresses an 84-kDa GH receptor rather than a 121-kDa GH receptor was used to examine whether the GH receptor might form a complex with a protein (e.g. tyrosine kinase) that comigrates on SDS-polyacrylamide gel electrophoresis gels with the endogenous GH receptor (Mr 121,000) in 3T3-F442A cells. GH-GH receptor complexes were immunoprecipitated with anti-GH antibody from GH-treated CH04 cells and incubated with [T-~'P]ATP. 32P was incorporated into a 121-kDa protein as well as the 84-kDa GH receptor. Phosphorylation of both the 84-kDa GH receptor and the 121-kDa protein

Previous work in multiple cell types has shown that endogenous GH receptors, as well as the cloned liver GH receptor, associate with a tyrosine kinase. However, in SDS-PAGE gels of highly purified, kinaseactive GH receptor preparations from 35S-labeled 3T3-F442A cells, only one broad band was detected corresponding to the molecular weight of the GH receptor rather than two bands which might be expected to result from a kinase-receptor heterocomplex. In the present study, a transfected Chinese hamster ovary (CHO) cell line (CH04) that expresses an 84-kDa GH receptor rather than a 121-kDa GH receptor was used to examine whether the GH receptor might form a complex with a protein (e.g. tyrosine kinase) that comigrates on SDS-polyacrylamide gel electrophoresis gels with the endogenous GH receptor (Mr 121,000) in 3T3-F442A cells. GH-GH receptor complexes were immunoprecipitated with anti-GH antibody from GHtreated CH04 cells and incubated with [T-~'P]ATP. 32P was incorporated into a 121-kDa protein as well as the 84-kDa GH receptor. Phosphorylation of both the 84-kDa GH receptor and the 121-kDa protein was on tyrosyl residues as determined by Western blotting with anti-phosphotyrosine antibody. The 121-kDa protein does not appear to bind GH. It was also not detected in the immunoprecipitate when cells had not been incubated with GH or when untransfected CHO cells were used. These findings suggest that in CH04 cells, the 121-kDa protein is precipitated by the GH antibody because of its ability to form a complex with the GH receptor (p84). Western blot analysis of whole cell lysates using anti-phosphotyrosine antibody revealed that GH promotes the tyrosyl phosphorylation of a 121-kDa protein and several other proteins (p97, p42, p39) in a dose-and time-dependent fashion. Taken together, these findings are consistent with either p121 being the tyrosine kinase that complexes with the GH receptor and is activated in response to GH binding or with p121 forming a ternary complex with both the GH receptor and a tyrosine kinase and serving as a substrate of the GH receptor-associated tyrosine kinase. Although GH' has been known for many years to be a primary determinant of body growth and a modulator of overall lipid, protein, and carbohydrate metabolism (1-41, little is known about the cellular mechanisms responsible for these diverse responses. The GH receptor cloned from liver shows no sequence homology to receptors with known signal transduction pathways (5). However, results from previous studies from our laboratory indicated that in multiple cell types, endogenous GH receptor as well as the cloned liver GH receptor copurifies with tyrosine kinase activity and is phosphorylated on tyrosyl residues (6)(7)(8)(9). In cell lines transfected with the cDNA for the liver GH receptor, phosphorylation of the receptor did not correlate with the amount of GH receptor expressed. Instead, phosphorylation was found to vary substantially with the cell type (9). Furthermore, in the one cell type tested (3T3-F442A cells), tyrosine kinase activity associated with the GH receptor was increased in response to GH binding (10). Taken together, these findings support the hypothesis that the GH receptor forms a complex with a tyrosine kinase and raise the possibility that activation of a GH receptor-associated tyrosine kinase is an early or perhaps initiating step in GH signal transduction.
If the GH receptor forms a complex with a tyrosine kinase, then one would expect to see two proteins (i.e. the GH receptor and the tyrosine kinase) in SDS-PAGE gels of kinase-active GH receptor preparations. However, only one rather broad band was consistently visible on SDS-PAGE gels of highly purified (estimated 30,000-40,000-fold) GH receptor isolated from "S-labeled 3T3-F442A cells (7,IO). Proteins in this band, which migrate with a size appropriate for the endogenous GH receptor in these cells ( M , 121,000), incorporated 32P into tyrosyl residues when the purified GH receptor preparation was incubated with [y3'P]ATP. Since the GH receptor and tyrosine kinase appear to be separate proteins, these findings suggest that the GH receptor-associated kinase: 1) has the same apparent molecular weight as the GH receptor and thus cannot be distinguished from the GH receptor by one-dimensional SDS-PAGE; 2) has a low rate of turnover or contains few or no methionine or cysteine residues and has therefore incorporated little 35S during the overnight incubation; or 3) is present in the highly purified GH-receptor preparations a t such low levels as to be undetectable, indicating a kinase:receptor stoichiometry substantially less than 1:l. To determine if the GH receptor forms a complex with a nonreceptor protein (e.g. kinase) of M , -121,000, we have taken advantage of a CHO cell line (CH04) which was stably transfected with a cDNA containing the entire coding region ' The abbreviations used are: GH, growth hormone; hGH, human growth hormone; DTT, dithiothreitol; DSS, disuccinimidyl suberate; SDS, sodium dodecyl sulfate; PAGE, polyacrylamide gel electrophoresis; IL, interleukin; CHO, Chinese hamster ovary; GM-CSF, granulocyte-macrophage colony-stimulating factor.

3573
of the cloned rat liver GH receptor but which expresses an 84-kDa GH receptor (p84) (11). This receptor is smaller than both endogenous and transfected rodent GH receptors (Mr 113,000-130,000) expressed in other cell lines (8,9). T h e molecular basis for the smaller size of t h e 84-kDa GH receptor in CH04 cells is not known since other CHO clones transfected simultaneously with the same GH receptor cDNA express the larger 121-kDa GH receptor and the size of t h e GH receptor mRNA expressed in t h e CH04 cells is the same as that expressed by the CHO cells expressing the larger 121-kDa receptor (Ref. 9, data not shown). Despite its smaller size, this GH receptor is a functional receptor, since it binds GH and elicits effects of GH on protein synthesis, lipid synthesis, lipolysis, a n d MAP kinase activity (11,12). ' We show that when GH. GH receptor complexes are immunoprecipitated with anti-GH antibody from GH-treated C H 0 4 cells, the 84-kDa GH receptor copurifies with a 121-kDa protein (~121). In an i n uitro kinase assay, both t h e GH receptor and t h e 121-kDa protein are phosphorylated on tyrosyl residues. Tyrosyl phosphorylation of a 121-kDa protein in intact cells was also shown to be GH-dependent. We speculate that the 121-kDa GH receptor-associated protein is t h e GH receptor-associated tyrosine kinase, although we cannot rule out the possibility that p121 is a substrate of the kinase that forms a ternary complex with the GH receptor and a GH receptor-associated tyrosine kinase. . Recombinant protein A-agarose was from Repligen, Triton X-100 (Surfact-Amps X-100) and DSS from Pierce Chemical Co., and SDS-PAGE and protein assay reagents from Bio-Rad. Aprotinin and leupeptin were purchased from Boehringer Mannheim. Molecular weight standards and chicken egg ovalbumin were purchased from Sigma, the enhanced chemiluminescence detection system and antimouse or anti-rabbit IgG conjugated to horseradish peroxidase were from Amersham, and polyvinylidene difluoride (PVDF) paper (Immobilon) was from Millipore.
Formation of Cross-linked lZ5I-hGH-Receptor Complexes-Human GH was labeled with by Dr. J. Smart (University of Michigan) to an estimated specific activity of 250-300 pCi/pg using a modification of the chloramine-T procedure (13). IZ5I-hGH was cross-linked to GH receptors as described previously (6). Briefly, cells were deprived of serum overnight as described above. lZ51-hGH (5 X lo6 CPM/lOO-mrn dish, -6 ng/ml) in Krebs-Ringer phosphate buffer, 1% bovine serum albumin was added to the cells in the presence or absence of 4 pg/ml unlabeled hGH and incubated for 1 h at 25 "C. After extensive washing, DSS (0.4 mM final concentration) was added for 15 min at 8 "C to cells in Krebs-Ringer phosphate buffer. Cells were solubilized using HVTDP, and cellular proteins were analyzed by SDS-PAGE.

Copurification and in Vitro Phosphorylation of a 121-kDa Protein (~121) and a Cloned Liver GH Receptor (p84)"
To determine if the inability to detect additional proteins (e.g. tyrosine kinase) in highly purified, kinase-active GH receptor preparations isolated from 3T3-F442A cells resulted from their comigration on SDS-PAGE with the 121-kDa GH receptor, we examined whether any protein(s) with molecular weight approximately 120,000 was detectable in purified GH receptor preparations from transfected CH04 cells. These cells express a GH receptor of a smaller size ( M r 84,000) t h a n endogenous GH receptors in 3T3-F442A cells ( M , 121,000) as illustrated in Fig. 1, which shows the results of experiments in which "'I-hGH was cross-linked to GH receptors in intact cells. C H 0 4 cells, as well as CHOA cells, which express a 121- is believed to represent partially proteolyzed "'1-hGH-receptor complexes, since it was present in all cell types in this particular experiment. This includes 3T3-F442A cells in which 12sI-labeled proteins of this size only appear under conditions in which proteolysis is likely to occur (e.g. experiments in which the sample processing time is extended due to a large number of samples) (compare to Fig. 1 of Ref. 6). kDa GH receptor, untransfected CHO cells, and 3T3-F442A cells (used as a positive control) were incubated with or without 100 ng/ml hGH for 1 h at 25 "C, and GH. GH receptor complexes were purified by immunoprecipitation utilizing anti-GH antibody. This method was estimated previously to purify the GH receptor from 3T3-F442A cells -8,000-fold (7).
The immunomatrices were incubated in vitro with [?-"' PI ATP in the presence of Mn" for 10 min a t 30 "C. Fig. 2 (lunes  1,4, and 7) illustrates that when CHOA and 3T3-F442A cells, which both express a 121-kDa GH receptor, are incubated with GH, '"P-labeled proteins of M, -121,000 are detected, as reported previously (7, 9). Consistent with it being a GH receptor, "*P-labeled 121-kDa protein was not present when cells had not been incubated with GH (Fig. 2, lune 8 ) nor when untransfected CHO cells, which lack endogenous GH receptors (Fig. 2, lanes 2 and 3 ) , were used. Additionally, Western blot analysis using anti-GH receptor antibody has identified proteins of this size in GH receptor preparations from 3T3-F442A cells (10).
Based upon these findings in 3T3-F442A and CHOA cells, the 84-kDa GH receptor (p84) was expected to be the only protein phosphorylated in the in vitro kinase assay when CH04 cells were used as starting material. However, when CH04 cells were incubated with GH and the GH receptor was immunoprecipitated on immobilized anti-GH antibody and incubated with [-y-"P]ATP, two phosphoproteins (p121 and p84) were detected (Fig. 2, lane 5). Neither phosphoprotein was detected when cells had not been incubated with GH (Fig.   2, lune 6 ) , consistent with both proteins being precipitated specifically by the anti-GH antibody. The identity of the lower band (p84) as the expressed cloned liver GH receptor is based on the observation that cross-linked '*'I-hGH-receptor complexes in CH04 cells migrate with M, 101,000 (Fig. 1,   lane 3 ) which suggests that the molecular weight of the GH receptor is -80,000. Furthermore, anti-GH receptor antibody recognizes an 84-kDa protein in Western blots (data not shown). The other band, pp121, is unlikely to be a GH receptor. Based on the cross-linking studies in Fig. 1 ( l a n e 3 ) and Western blots using anti-GH receptor antibody, pp121 is too big to be the expressed liver GH receptor in these cells. It is also unlikely to be an endogenous GH receptor, because CHO cells do not express endogenous GH receptors. No GH binding is detectable in untransfected CHO cells (Ref. 11 and data not shown), nor are any 134-kDa cross-linked ""I-hGHreceptor complexes observed when untransfected CHO cells are incubated with '*'I-hGH and cross-linking reagent (Fig.   1, lane I). Although a small amount of ' *' I migrating with M, -130,000 was observed for CH04 cells in some cross-linking experiments (Fig. 1, lune 3 ) , the finding that p121 is specifically immunoprecipitated by GH antibody and prominently phosphorylated in the presence of [y"P]ATP (Fig. 2, lune 5 ) but at best only inefficiently cross-linked to '*'I-hGH (Fig. 1,   lune 3 ) is most consistent with p121 being a non-GH binding protein that complexes with the GH receptor.

Both p121 and the GH Receptor ($4) Isolated from CH04 Cells Are Phosphorylated on Tyrosyl Residues in an in Vitro
Kinase Assay-Western blot analysis using anti-phosphotyrosine antibody was used to determine whether phosphorylation of p121 and p84 during the in vitro kinase assay is on tyrosyl residues and whether these proteins are phosphorylated on tyrosyl residues in the intact cell. CH04 cells were incubated with 100 ng/ml GH and GH.GH receptor complexes were immunoprecipitated using anti-GH antibody. Half of the sample was analyzed directly by Western blot analysis using an anti-phosphotyrosine antibody to evaluate whether p84 and p121 are phosphorylated in vivo on tyrosyl residues. The other half of the sample was incubated with unlabeled ATP (4 PM) for 10 min a t 30 "C as described for the in vitro kinase assay and then analyzed by Western blot assay using anti-phosphotyrosine antibody to determine whether the phosphorylation of p84 and p121 observed in the in vitro kinase assay is on tyrosyl residues. GH-receptor complexes from 3T3-F442A and CHOA cells were analyzed in a similar fashion. Fig. 3 (lunes 2,5, and 8 ) shows that p121 from CH04, CHOA, and 3T3-F442A cells incubated with GH  (+GH, lanes 2, 3, 5, 6, 8, and 9) or without  (-GH, lanes I , 4, and 7) 100 ng/ml GH were immunoprecipitated with anti-GH antibody as described for Fig. 2 except that the samples were not normalized for protein. Some samples (-ATP, lanes I , 2,4, 5, 7, and 8) were eluted with SDS-PAGE sample buffer and subjected to SDS-PAGE. Other samples (+ATP, lanes 3,6, and 9 ) were incubated with unlabeled ATP (4 PM) a t 30 "C for 10 min before elution with SDS-PAGE sample buffer. After electrophoresis, proteins were transferred to PVDF membranes for 1 h a t 1 amp. The membrane was then probed with aPY-4G10, and proteins were visualized by the enhanced chemiluminescence detection system. Molecular weights ( X of the protein standards are indicated on the right. The arrows indicate the migration of p121 and p84. is phosphorylated on tyrosyl residues in uiuo. Fig. 3 (lanes 3,  6, and 9) also reveals that in uitro incubation of GH receptor preparation with ATP results in increased tyrosyl phosphorylation of pp121 in all three cell types and the emergence of phosphorylated tyrosyl residues in p84 in the CH04 cells treated with GH, indicating that both p121 and p84 are phosphorylated on tyrosyl residues in the in uitro kinase assay. Tyrosyl phosphorylation resulting from the in uitro kinase reaction decreased migration of a portion of p121, consistent with the in uitro kinase reaction phosphorylating additional tyrosyl residues in proteins in which at least 1 tyrosyl residue had already been phosphorylated in uiuo. Interestingly, in uiuo tyrosyl phosphorylation of p84 from CH04 cells was not detectable in Fig. 3, lane 5. However, in some experiments in which GH-receptor preparations were analyzed by Western blotting using aPY-4G10, a faint band of M , 84,000 was detectable in overexposed autoradiograms, indicating that the 84-kDa GH receptor from GH-treated cells is most likely also phosphorylated on tyrosyl residues (data not shown). To determine whether the difficulty in detecting the presence of phosphorylated tyrosyl residues in p84 may reside at least in part from poor binding of p84 to PVDF membranes used in the Western blot procedure, p121 and p84 were labeled in uitro with 32P as described for Fig. 2, subjected to SDS-PAGE, and either analyzed directly by autoradiography or transferred to PVDF paper and washed as for the Western blot assay before autoradiography. A comparison of the relative intensities of 32P-labeled p121 and p84 in both autoradiograms indicated that "P-labeled 84-kDa GH receptor binds poorly to PVDF paper when compared to 32P-labeled 121-kDa protein (data not shown). The tyrosylphosphorylated 84-kDa GH receptor also bound poorly to nitrocellulose (data not shown). Poor binding of the GH receptor to PVDF may explain in part why the amount of tyrosyl-phosphorylated p121 present in the immunoprecipitate from CH04 cells appears to be as great as or greater than that in CHOA cells in Fig. 3, even though p121 in the CHOA cells consists of two proteins (GH receptor and p121 GHRassociated protein), whereas p121 in the CH04 cells consists only of p121 GHR-associated protein. However, it should also be noted that: ( a ) despite the same number of confluent plates being used for both CHOA and CH04 cells, the samples are not directly comparable, because the samples were not nor-malized for cell or GH receptor number; and ( b ) the CH04 cells express more GH receptors than the CHOA cells (see Fig. 1) so that if the cell number were the same in the two samples and if the stoichiometry of GH receptor:pl21 GHRassociated protein approaches 1:1, then one would expect to see more tyrosyl-phosphorylated p121 GHR-associated protein in the CH04 cells than in the CHOA cells.

A 121-kDa Protein Is Tyrosyl Phosphorylated in Response
to GH-To examine whether p121 is tyrosyl-phosphorylated in response to GH, CH04 cells were incubated a t 25 "C with various concentrations of GH (0, 1, 10, 100, and 1000 ng/ml) for 15 min (Fig. 4) or with 100 ng/ml GH for various lengths of time (0, 5, 15, and 60 min) (Fig. 5). Cells were lysed with  [1][2][3][4][5] and transfected CH04 cells (lanes [6][7][8][9][10] expressing the 84-kDa GH receptor were incubated with the indicated concentration of hGH for 15 min at 25 "C and lysed with boiling SDS lysis buffer. Proteins were boiled for 5 min and separated by SDS-PAGE. Proteins were then transferred to PVDF membranes, probed with aPY-4G10, and visualized by enhanced chemiluminescence. Molecular weights ( X lo-') of the protein standards are indicated. The arrows indicate the migration of p121, p97, p42, and p39.  (lanes 1-4), CH04 cells expressing an 84-kDa GH receptor (lanes 5-8), and CHOA cells expressing a 121-kDa GH receptor (lanes [9][10][11][12] were incubated with 100 ng/ml GH at 25 "C for the indicated time. Cells were lysed, and proteins were subjected to Western blot analysis using aPY-4G10 ( A ) or aPY-Shafer ( B ) . Molecular weights (X loy3) of the protein standards are indicated. The arrows in A indicate the migration of p121, p97, p42, and p39. The arrow in B indicates the migration of p121. In lane 10, the absence of the constitutively phosphorylated protein migrating above p121 (M, -140,000) and the simultaneous appearance of several bands migrating below p97 suggests that this sample underwent partial proteolysis during its preparation.
boiling SDS lysis buffer, and aliquots of the whole cell lysates were subjected to SDS-PAGE followed by Western blot analysis using anti-phosphotyrosine antibody (aPY-4G10 (Figs. 4 and 5A) or aPY-Shafer (Fig. 5B)) to visualize tyrosyl-phosphorylated proteins. Figs. 4 and 5 show that GH stimulates tyrosyl phosphorylation of a 121-kDa protein in CH04 cells in a dose-dependent (Fig. 4) and time-dependent (Fig. 5) fashion. This 121-kDa protein comigrates with the 121-kDa protein present in anti-GH receptor immunoprecipitates,3 consistent with pp121 in the Western blots and pp121 in the anti-GH antibody immunoprecipitates being the same protein. Tyrosyl phosphorylation of p121 is detectable after 15 min with 10 ng/ml GH and is maximal with 100 ng/ml GH (Fig. 4, lunes 6-10). When cells are incubated with 100 ng/ml hGH, the stimulatory effect is apparent at the earliest time tested (5 min) (Fig. 5, lane 6). Increased tyrosyl phosphorylation of p121 in response to GH was also observed in CHOA cells expressing the larger sized (121 kDa) GH receptor (Fig.  5B, lunes 9-12) although in these cells the relative contributions to that increase of the 121-kDa GH receptor and the nonreceptor 121-kDa protein cannot be determined. GH did not stimulate the tyrosyl phosphorylation of a 121-kDa protein in untransfected CHO cells (Fig. 4, lanes 1-5 and Fig.  5B, lanes 1-4) as predicted from the absence of GH receptors in these cells.
When CH04 and CHOA cells were incubated with GH, three proteins (p97, p42, and p39) in addition to p121 showed increased tyrosyl phosphorylation in response to GH (Figs. 4 and 5A). The 42-and 39-kDa proteins are thought to be ERKS 1 and 2 based.upon their comigration with proteins identified by anti-MAP kinase antibodies in CH04 and 3T3-F442A cells (12,15,16). GH-dependent tyrosyl phosphorylation of an 84-kDa protein is not detectable, presumably due at least in part to the poor ability of the GH receptor to bind to PVDF paper (see above).

DISCUSSION
Identification of a 121-kDa Protein That Forms a Complex with GH Receptor-In the present study, we demonstrate that anti-GH antibody specifically precipitates two proteins (p84 and p121) from CH04 cells incubated with GH. Both proteins are phosphorylated on tyrosyl residues when incubated in uitro with [Y-~~PIATP. We also demonstrate that the 121-kDa protein, as well as several other proteins, are tyrosylphosphorylated in response to GH. The 84-kDa protein is believed to be the GH receptor expressed in CH04 cells based on the findings that: 1) cross-linked lZ5I-hGH-receptor complexes in these cells migrate with a M, (101,000), consistent with a GH receptor of M, -80,000; and 2 ) p84 comigrates with a protein that binds in Western blot assays to an antirat GH receptor antibody. The findings that neither GH binding nor 134-kDa cross-linked 1251-hGH-receptor complexes are detectable in untransfected CHO cells and that the anti-GH receptor antibody does not detect a 121-kDa protein in Western blots of CH04 cells suggest that p121 in the CH04 cells is a nonreceptor protein. Interestingly, a faint band of -130,000 labeled by lZ51 was sometimes detected in cross-linking studies of CH04 cells. We feel it unlikely that this faint band represents endogenous 121-kDa GH receptor cross-linked to lZ5I-hGH since it was not present in untransfected CHO cells. One could hypothesize that GH binding to the cloned liver GH receptor expressed in CH04 cells somehow induces the expression of the endogenous 12l-kDa GH receptor. However, in mouse L cells in which the endogenous G. S. Campbell and C. Carter-Su, unpublished observation.
GH receptor mRNA can be distinguished from expressed cloned liver GH receptor mRNA by size, expression of the cloned mouse liver GH receptor decreased, rather than increased, levels of endogenous GH receptor mRNA (data not shown). Similarly, expression of a truncated rat liver GH receptor in RIN5-AH cells decreased rather than increased the number of endogenous GH receptors as assessed by crosslinking lZ51-hGH to GH receptors in intact cells (17). A more likely explanation for the presence of a faint band migrating with M, 134,000 in CH04 cells is that some lz5SI-hGH became directly cross-linked to p121 due to its proximate location in the GH receptor/pl21 complex. Precedence for this is provided by the observation that in cross-linking studies, a small amount of 1251-insulin can be cross-linked to the non-insulin binding subunit of the insulin receptor (18). In light of this evidence favoring p121 being a non-GH binding protein, the finding that p121 is immunoprecipitated by anti-GH antibody from lysates of GH-treated cells is most consistent with it forming a complex with the GH receptor (p84).
Both p121 GHR-associated Protein and the GH Receptor Are Phosphorylated on Tyrosyl Residues-The present studies provide strong evidence that p121 GHR-associated protein is tyrosyl-phosphorylated in response to GH and is phosphorylated in the in uitro kinase assay by the GH receptor-associated tyrosine kinase. It also provides strong evidence that, at least in CH04 cells, the GH receptor is phosphorylated by the GH receptor-associated tyrosine kinase in the in vitro kinase assay. Evidence also supports the tyrosyl phosphorylation of the GH receptor from 3T3-F442A cells by the GH receptorassociated tyrosine kinase in the in vitro kinase assay. In experiments not shown: GH-receptor complexes were immunoprecipitated using anti-GH antibody, incubated with [y-32P]ATP, boiled for 2 min in 1% SDS, 10 mM EDTA to dissociate protein complexes, and then incubated with anti-GH receptor antibody. The ability of immobilized anti-GH receptor antibody to precipitate a substantial portion of the 32P-labeled p121 provides evidence that at least a portion of the 32P was incorporated into the GH receptor rather than into p121 GHR-associated protein. In contrast to these results using the in uitro kinase assay, the Western blot results shown in Figs. 3-5 would seem to suggest that, when GH is added to the intact cell, very few if any GH receptors become phosphorylated on tyrosyl residues since very little 84-kDa protein is detected by the a-PY antibody in the Western blot assays. However, the Western blot results are misleading because poor binding of the GH receptor to PVDF membranes results in an artefactually low signal. The finding that even with poor binding of the GH receptor to PVDF membranes, a faint signal corresponding to the 84-kDa GH receptor (discussed under "Results") or a more substantial signal corresponding to an 80-kDa truncated GH receptor missing approximately half its cytoplasmic domain5 is detectable in some Western blots using a-PY antibody indicates the presence of tyrosyl residues in these GH receptors. Experiments using crosslinked lZ5I-hGH.GH receptor complexes also indicate the presence of phosphorylated tyrosyl residues in the GH receptor in 3T3-F442A cells. In previous studies showing specific binding (3-20%) of 1251-hGH. GH receptor complexes to immobilized anti-phosphotyrosine antibodies (6,9), solubilization conditions were sufficiently mild (0.1 or 1% Triton X- tyrosyl-phosphorylated p121 GHR-associated protein. To counteract this argument, we solubilized cross-linked (with disuccinimidyl suberate) lZ5I-hGH-receptor complexes from intact cells with 0.1% Triton X-100 and then boiled the cell lysate in a buffer containing 0.1% Triton X-100, 0.5% SDS, and 10% P-mercaptoethanol to dissociate protein complexes. The sample was diluted 1 2 6 with buffer containing 0.1% Triton X-100 and then incubated with immobilized aPY-Shafer antibody. Even under these harsh solubilization conditions, some (-1%) '251-hGH-receptor complexes were recognized by the aPY-Shafer antibody. Taken together, these results provide evidence that at least a portion of the GH receptors are tyrosyl-phosphorylated in GH-treated cells. However, they do not provide insight into the relative proportions of phosphorylated tyrosyl residues present in the GH receptor and in p121 GHR-associated protein. This delineation will require development of techniques that separate the two proteins without affecting their relative recoveries or phosphorylation states. Determination of whether tyrosyl phosphorylation of the GH receptor is important for eliciting responses to GH awaits an examination of whether receptors lacking specific tyrosyl residues are able to function normally.

p121 GHR-associated Protein Is the GH Receptor-associated Tyrosine Kinase or a Substrate of the GH Receptor-associated Tyrosine Kinase That Forms a Ternary Complex with the Kinase and the GH Receptor-While
the current study revealed the presence of a 121-kDa GHR-associated protein which can be tyrosyl-phosphorylated in vitro and in vivo, whether p121 is the GH receptor-associated kinase is not yet clear. Consistent with p121 being the GH receptor-associated tyrosine kinase, several lines of evidence indicate that the tyrosine kinase responsible for GH receptor phosphorylation in CH04 and 3T3-F442A cells is present in a complex with GH receptor and p121 GHR-associated protein. First, in CH04 cells the GH receptor-pl21 complex purified by immunoadsorption to anti-GH antibody catalyzes the in vitro phosphorylation of tyrosyl residues in both proteins in the complex. Second, dilution by 100-fold of GH-receptor preparations from GH-treated 3T3-F442A cells does not decrease the rate of GH receptor/pl21 phosphorylation in an in vitro kinase assay (10). Finally, the amount of tyrosine kinase activity present in anti-GH receptor immunoprecipitates from 3T3-F442A cells, assessed by the amount of 32P transferred from [y-"PIATP to a synthetic substrate or to GH receptor/ p121, is greater when cells are incubated with GH (10). Since previous studies indicate that the GH receptor does not possess intrinsic tyrosine kinase activity (9), it is appealing to hypothesize that p121 is the GH receptor-associated tyrosine kinase which phosphorylates the GH receptor. A tyrosine kinase of M, 121,000 would explain why two proteins, one corresponding to the GH receptor and the other to the tyrosine kinase responsible for GH receptor phosphorylation, could not be distinguished in SDS-PAGE gels of highly purified, kinase-active, GH receptor preparations from 35S-labeled 3T3-F442A cells; cells in which the GH receptor itself migrates with a M, 121,000. The following findings using 3T3-F442A cells (19) are consistent with p121 being a tyrosine kinase with a central role in GH signaling: 1) in time course studies, tyrosyl phosphorylation of 121-kDa protein is detected in response to GH as early as or earlier than tyrosyl phosphorylation of other proteins; 2 ) increased tyrosyl phosphorylation of the 121-kDa protein is detected at lower concentrations of GH (0.5 ng/ml) than required for detection of other phosphoproteins; 3) p121 is one of the few proteins phosphorylated on tyrosyl residues in response to GH that is not phosphorylated on tyrosyl residues in response to other growth factors whose receptors have intrinsic tyrosine kinase activity (e.g. receptors for epidermal growth factor, plateletderived growth factor, insulin-like growth factor-1); and 4 ) all three of the tyrosine kinase inhibitors identified as inhibitors of the GH receptor-associated tyrosine kinase using the in vitro kinase assay prevent GH-promoted tyrosyl phosphorylation of p121 when added to cells. In contrast, other inhibitors of tyrosine kinases that are ineffective against the GH receptor-associated kinase do not block tyrosyl phosphorylation of p121 in response to GH.
The cloned liver GH receptor is a member of the cytokine/ hematopoietic receptor superfamily, with the prolactin receptor being the member most closely related to the GH receptor. This receptor family, which includes receptors for ILs 2-7, erythropoietin, granulocyte colony stimulating factor, GM-CSF, and ciliary neurotrophic factor (20-30) is defined by homologies in the extracellular domain. However, receptors in this family may also share some common characteristics in their intracellular signaling. In fact, binding of ligand to a number of these receptors stimulates tyrosine phosphorylation of cellular proteins (e.g. receptors for GH, prolactin, ILs 2-4, and 7, GM-CSF, erythropoietin (16,19,(31)(32)(33)(34)(35)). In a manner similar to that shown for the GH receptor (7), the receptors for IL-2 and erythropoietin have been shown to associate with a tyrosine kinase (36,37). The tyrosine kinase has been tentatively identified (as Lck, a 56-kDa member of the Src family) for only one member of this receptor family, IL-2 receptor (38,39). Whether a member of the Src family of kinases associates with the GH receptor is not known. However, no phosphoproteins migrating with a size appropriate for members of this family have been observed in kinaseactive GH receptor preparations (6)(7)(8)(9).
Proteins similar in size (Mr 130,000-140,000) to the 121-kDa protein identified in this study have been identified in association with the receptors for other members of the cytokine/hematopoietic receptor family, including the receptors for IL-3, IL-5, IL-6, GM-CSF, and erythropoietin (37,(40)(41)(42). These proteins which associate with the receptor as a complex are thought to play a role in ligand binding and/or signal transduction. While the 130-kDa proteins that associate with the receptors for IL-3, IL-5, IL-6, and GM-CSF have been cloned and are assumed not to be tyrosine kinases based upon sequence analysis (40-42), it is not known whether they are phosphorylated on tyrosyl residues in response to ligand binding. In contrast, erythropoietin receptor-associated p130 is tyrosyl-phosphorylated in response to erythropoietin (37) and is a likely candidate for the erythropoietin receptor-associated tyrosine kinase. It is therefore intriguing to speculate that p121 identified in this study as a GH receptor-associated protein is a member of a family of proteins that associate with members of the cytokine/hematopoietic receptor family and are important for signal transduction, either as tyrosine kinases or as signal transducing proteins that are activated upon phosphorylation by the appropriate ligand-activated tyrosine kinase. While we favor the interpretation that p121 is the GH receptor-associated tyrosine kinase, our data cannot rule out the alternative hypothesis that p121 forms a ternary complex with the GH receptor and the tyrosine kinase, and serves as a substrate of the GH receptor-associated tyrosine kinase. This is a less appealing hypothesis, since, if p121 is not the kinase, the inability to detect more than one band when highly purified kinase-active GH receptor is purified from 3sS-labeled 3T3-F442A cells and subjected to SDS-PAGE and autoradiography (7, 10) would imply that the GH receptor-associated tyrosine kinase has an extremely low rate of turnover; has few or no cysteine or methionine residues; or that the ratio of kinase to GH receptor/pl21 complex in the highly purified kinase-active GH preparations is so low that the kinase is not detectable.
GH Signaling through GH Receptor-associated Tyrosine Kinase-In addition to p121, we show here at least three other cellular proteins (p97, p42, and p39) that become tyrosylphosphorylated in response to GH in CHOA and CH04 cells expressing the cloned liver GH receptor. Two of these proteins (p42 and p39) comigrate with proteins identified in 3T3-F442A cells as ERKl and ERK2 (16,15). This confirms the finding of Moller and colleagues (12) that GH stimulates tyrosyl phosphorylation of MAP kinase in CH04 cells expressing the cloned liver GH receptor. The identity of the 97-kDa protein is unknown, although a protein of similar size was also shown to be tyrosyl-phosphorylated in 3T3-F442A cells upon stimulation by GH (16,19). The finding that binding of GH to the cloned liver GH receptor triggers tyrosyl phosphorylation of a set of proteins which are also phosphorylated in 3T3-F442A cells upon binding of GH to the endogenous receptor suggests that the cloned liver GH receptor and the endogenous GH receptor in 3T3-F442A cells trigger similar signal transduction pathways involving tyrosine kinases and indicate that it may be possible to identify the GH receptor-associated tyrosine kinase and to dissect at least some of these pathways using CHO cells and the cloned liver GH receptor.