Growth hormone-dependent phosphorylation of tyrosine 333 and/or 338 of the growth hormone receptor.

Many signaling pathways initiated by ligands that activate receptor tyrosine kinases have been shown to involve the binding of SH2 domain-containing proteins to specific phosphorylated tyrosines in the receptor. Although the receptor for growth hormone (GH) does not contain intrinsic tyrosine kinase activity, GH has recently been shown to promote the association of its receptor with JAK2 tyrosine kinase, to activate JAK2, and to promote the tyrosyl phosphorylation of both GH receptor (GHR) and JAK2. In this work, we examined whether tyrosines 333 and/or 338 in GHR are phosphorylated by JAK2 in response to GH. Tyrosines 333 and 338 in rat full-length (GHR1-638) and truncated (GHR1-454) receptor were replaced with phenylalanines and the mutated GHRs expressed in Chinese hamster ovary cells. These substitutions caused a loss of GH-dependent tyrosyl phosphorylation of truncated receptor and a reduction of GH-dependent phosphorylation of the full-length receptor. Consistent with Tyr333 and/or Tyr338 serving as substrates of JAK2, these substitutions resulted in a loss of tyrosyl phosphorylation of truncated receptor in an in vitro kinase assay using substantially purified GH•GHR•JAK2 complexes. The Tyr to Phe substitutions did not substantially alter GH-dependent JAK2 association with GHR or tyrosyl phosphorylation of JAK2. These results suggest that Tyr333 and/or Tyr338 in GHR are phosphorylated in response to GH and may therefore serve as binding sites for SH2 domain-containing proteins in GH signal transduction pathways.

Many signaling pathways initiated by ligands that activate receptor tyrosine kinases have been shown to involve the binding of SH2 domain-containing proteins to specific phosphorylated tyrosines in the receptor. Although the receptor for growth hormone (GH) does not contain intrinsic tyrosine kinase activity, GH has recently been shown to promote the association of its receptor with JAK2 tyrosine kinase, to activate JAK2, and to promote the tyrosyl phosphorylation of both GH receptor (GHR) and JAK2. In this work, we examined whether tyrosines 333 and/or 338 in GHR are phosphorylated by JAK2 in response to GH. Tyrosines 333 and 338 in rat full-length (GHR 1-638 ) and truncated (GHR 1-454) receptor were replaced with phenylalanines and the mutated GHRs expressed in Chinese hamster ovary cells. These substitutions caused a loss of GH-dependent tyrosyl phosphorylation of truncated receptor and a reduction of GH-dependent phosphorylation of the fulllength receptor. Consistent with Tyr 333 and/or Tyr 338 serving as substrates of JAK2, these substitutions resulted in a loss of tyrosyl phosphorylation of truncated receptor in an in vitro kinase assay using substantially purified GH⅐GHR⅐JAK2 complexes. The Tyr to Phe substitutions did not substantially alter GH-dependent JAK2 association with GHR or tyrosyl phosphorylation of JAK2. These results suggest that Tyr 333 and/or Tyr 338 in GHR are phosphorylated in response to GH and may therefore serve as binding sites for SH2 domain-containing proteins in GH signal transduction pathways.
Ligand binding to membrane receptors with intrinsic tyrosine kinase activity has been shown to result in the phosphorylation of multiple tyrosines in the receptors themselves (reviewed in Ref. 1). Once phosphorylated, these tyrosines have been shown to bind the Src homology 2 (SH2) domain (a region of ϳ100 amino acids) of both enzymes and noncatalytic proteins involved in signal transduction (2)(3)(4)(5). Thus, phosphorylated tyrosines in receptors are believed to serve a vital role linking extracellular stimuli and cellular responses. Although GHR 1 itself is not a tyrosine kinase, GH has recently been shown to promote the association of GHR with the tyrosine kinase JAK2, to activate JAK2, and to promote phosphorylation of tyrosyl residues in both GHR and JAK2 (6,7). Since JAK2 is activated (and thus phosphorylated) in response to multiple ligands that bind to members of the cytokine receptor superfamily (6, 8 -10), it has been presumed that specificity in ligand response resides at least in part in the tyrosines that are phosphorylated in individual receptors. We were therefore interested in determining which tyrosines in GHR are phosphorylated in response to GH and thereby could potentially serve as binding sites for SH2 domain-containing signaling molecules that mediate responses to GH.
Previous studies showed that rat GHR  (numbering system of Ref. 11), which lacks approximately half of the cytoplasmic domain of GHR, is phosphorylated on tyrosines in response to GH in intact cells (7). Furthermore, this truncated receptor is phosphorylated on tyrosines when GH⅐GHR⅐JAK2 complexes are substantially purified from GH-treated cells and incubated with [␥-32 P]ATP. These results indicate that one or more of the 4 intracellular tyrosines (Tyr 333 , Tyr 338 , Tyr 391 , Tyr 437 ) present in this truncated GHR is phosphorylated in response to GH, presumably by JAK2. Based upon predictions of secondary structure, hydrophilicity, chain flexibility, and surface probability (12)(13)(14)(15)(16)(17), Tyr 333 of GHR was judged to be the most likely of the 10 cytoplasmic tyrosines in the receptor to be phosphorylated. It is also the only tyrosine in the truncated receptor that is conserved among species (cow, human, rabbit, sheep, rat, mouse, chicken, pig) (18 -25), suggesting that Tyr 333 may play an important role in GH action. In this work, we use site-directed mutagenesis to examine whether Tyr 333 and/or its close neighbor Tyr 338 is phosphorylated in response to GH. Our findings provide strong evidence that one or both of these tyrosines is phosphorylated in response to GH, serves as a substrate for JAK2, and is not required for association of the receptor with JAK2 or GH activation of JAK2. The accompanying paper (42) provides evidence that Tyr 333 and/or Tyr 338 may be involved in at least some actions of GH.

EXPERIMENTAL PROCEDURES
Materials-Recombinant DNA-derived 22,000-dalton hGH was a gift of Lilly. Recombinant protein A-agarose was from Repligen, and the protein assay (BCA) was from Pierce. Triton X-100, aprotinin, and leupeptin were purchased from Boehringer Mannheim. Molecular weight standards (unstained) and ovalbumin were purchased from Sigma, prestained molecular weight standards were from Life Technol-ogies, Inc., nitrocellulose membranes were from Schleicher & Schuell, and [␥-32 P]ATP (6000 Ci/mmol) was from DuPont NEN. The enhanced chemiluminescence detection system and x-ray film were from Amersham Corp.
Antisera-Antibody to GH (␣GH) (NIDDK-anti-hGH-IC3, lot C11981) came from the National Institute of Diabetes and Digestive and Kidney Diseases/National Hormone and Pituitary Program. Antiphosphotyrosine antibody (␣PY) (4G10) was purchased from Upstate Biotechnology, Inc. Antibody to JAK2 (␣JAK2) was prepared either in our laboratory in conjunction with Pel-Freez Biologicals or the laboratory of J. Ihle (St. Jude Children's Research Hospital, Memphis, TN) in rabbits against a synthetic peptide corresponding to amino acids 758 -776 as described previously (9). Antibody to GHR (␣GHBP), kindly provided by W. Baumbach (American Cyanamid, Princeton, NJ), was produced in rabbits using recombinant rat GH-binding protein produced in Escherichia coli (26).
Mutagenesis, Transfection, and Cell Culture-CHO cells were cotransfected with plasmids pLM108 and pIBP-1 (27,28). Plasmid pIBP-1 contains a thymidine kinase promoter fused to the bacterial neomycin phosphotransferase gene conferring G418 resistance. Plasmid pLM108 contains the simian virus 40 enhancer and the Zn 2ϩ -inducible human metallothionein IIa promoter driving the expression of the cDNA coding for full-length rat liver GHR (GHR 1-638 ) (gift of G. Norstedt, Center for Biotechnology, Karolinska Institute, Huddinge, Sweden), the same cDNA with the termination codon replacing the lysine codon 455 (GHR 1-454 ) (gift of G. Norstedt); the same cDNA with phenylalanines replacing tyrosines at positions 333 and 338 (GHR 1-638 Y333F,Y338F) and the same cDNA with the termination codon replacing the lysine codon 455 and with phenylalanines replacing tyrosines at positions 333 and 338 (GHR 1-454 Y333F,Y338F) ( Fig. 1). The cDNAs encoding the truncated and mutated GHR were generated by polymerase chain reaction (28,29,7). CHO cells were cultured and screened for GHR expression as described previously (7,30,31). Two cell lines expressing different levels of GHR 1-638 Y333F,Y338F were used in these studies, designated clones 23 (level of expression similar to cells expressing GHR 1-638 ) and 3 (level of expression only 20% that of cells expressing GHR 1-638 ). Experiments were performed using clone 23 unless noted otherwise.
Immunoprecipitation and Western Blotting-Confluent CHO cells were incubated in the absence of serum overnight (32). Cells were incubated for the indicated times with hGH at 37°C in 95% air, 5% CO 2 , rinsed with three changes of ice-cold PBSV (10 mM sodium phosphate, pH 7.4, 137 mM NaCl, 1 mM Na 3 VO 4 ) and scraped on ice in lysis buffer (50 mM Tris, pH 7.5, 0.1% Triton X-100, 137 mM NaCl, 2 mM EGTA, 1 mM Na 3 VO 4 , 1 mM phenylmethylsulfonyl fluoride, 10 g/ml aprotinin, 10 g/ml leupeptin). Cell lysates were centrifuged at 12,000 ϫ g for 10 min, and the resulting supernatants were incubated on ice for 2 h with the indicated antibody. Immune complexes were collected on protein A-agarose during a 1-h incubation at 8°C, washed three times with wash buffer (50 mM Tris, pH 7.5, 0.1% Triton X-100, 137 mM NaCl, 2 mM EGTA), and boiled for 5 min in a mixture (80:20) of lysis buffer and SDS sample buffer (250 mM Tris, pH 6.8, 5% SDS, 10% ␤-mercaptoethanol, 40% glycerol). The immunoprecipitates and lysates were subjected to SDS-PAGE (with the amount of sample normalized to protein) followed by Western blot analysis with the indicated antibody using the enhanced chemiluminescence detection system (33). In some experiments, the blots were rinsed and reprobed with a second antibody.
Formation of Cross-linked 125 I-hGH Receptor Complexes and GH Binding Assay-Human GH was labeled with 125 I to an estimated specific activity of ϳ90 Ci/g using chloramine T by the University of Michigan Reproductive Sciences Training Grant Core Facility. As described previously (34), cells were incubated in serum-free medium overnight and then washed with Krebs-Ringer phosphate buffer (KRP) containing 1% bovine serum albumin. 125 I-hGH (12 ϫ 10 6 counts/min/ 100-mm dish, 8 -40 ng/ml) in KRP, 1% bovine serum albumin was added to the cells in the presence or absence of 1 g/ml unlabeled hGH and incubated for 1 h at 25°C. After extensive washing with KRP, 1% bovine serum albumin, disuccinimidyl suberate (0.4 mM final concentration) dissolved in dimethyl sulfoxide was added, and cells were incubated for 15 min at 8°C. Cells were solubilized using HVTDP buffer (25 mM HEPES, 0.1% Triton X-100, 2 mM Na 3 VO 4 , 0.5 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride, pH 7.4, 10 g/ml aprotinin, 10 g/ml leupeptin) and mixed (80:20) with SDS sample buffer (32) and analyzed by SDS-PAGE. Relative affinities of the GHR for 125 I-hGH were estimated as described previously (28). Because the affinities of the different GHR appeared not to differ (data not shown), receptor numbers were compared by incubating cell monolayers with 125 I-hGH (1-6 ng/ml) for 1 h at 25°C. Cells were washed with ice-cold KRP and lysed with 1 N NaOH. Radioactivity associated with the cells was determined by counting cell lysates in a ␥ counter. Results were normalized to protein content and corrected for nonspecific binding, which was determined by incubating cell monolayers with 125 I-hGH in the presence of 2 g/ml unlabeled hGH. The relative abilities of the different cell lines to bind 125 I-hGH were similar when binding was determined by incubating the cells with 125 I-hGH for 1 h at 25°C or overnight at 4°C. 2 In Vitro Kinase Assay-Cells were grown to confluence and deprived of serum overnight as described previously (35). Cells were then incubated in the absence or presence of 100 ng/ml (4.5 nM) GH at 25°C for 1 h (Fig. 3) or 100 ng/ml GH at 37°C for 15 min (Fig. 4) and lysed in HVTDP buffer. Cell lysates were centrifuged at 230,000 ϫ g for 1 h. Supernatants were then incubated with ␣GH (1:10,000) for 2 h at 8°C. Immune complexes were precipitated with immobilized protein A, and immunomatrices were extensively washed with NHT (50 mM HEPES, 150 mM NaCl, 0.1% Triton X-100, pH 7.6) plus 0.5 mM dithiothreitol. The immunomatrices were washed in buffer C (50 mM HEPES, 100 mM NaCl, 6.25 mM MnCl 2 , 0.5 mM dithiothreitol, 0.1% Triton X-100, pH 7.6) and resuspended in 200 l of buffer C containing 250 g/ml aprotinin and 250 g/ml leupeptin. In vitro phosphorylation was carried out by adding 20 l of 50 mM HEPES, 0.1% Triton X-100, pH 7.6, containing vehicle, unlabeled ATP (5 M), or unlabeled ATP (5 M) plus [␥-32 P]ATP (ϳ200 Ci) and incubating for 10 min at 30°C as described previously (35,36). The reaction was stopped by the addition of 12 ml of ice-cold NHT buffer containing 10 mM EDTA, pH 7.6, followed by extensive washing. Immunoprecipitated proteins were eluted from the immunomatrices by boiling in 200 l of 150 mM Tris, pH 6.8, 3% SDS, 3% ␤-mercaptoethanol, 30% glycerol, and 0.03 mg/ml bromphenol blue and analyzed by SDS-PAGE followed by either autoradiography or Western blotting using ␣PY as described above.

Ability of GH to Promote Tyrosyl Phosphorylation of GHR
Lacking Tyr 333 and Tyr 338 -Previous results indicated that GHR 1-454 is tyrosyl-phosphorylated in response to GH (7), indicating that 1 or more of the 4 tyrosines present in the cytoplasmic domain of this truncated receptor is a substrate of the GH-activated, GHR-associated JAK2 tyrosine kinase. To determine which of the 4 tyrosines present in this truncated receptor is phosphorylated in response to GH, we replaced Tyr 333 and Tyr 338 in both wild-type (GHR 1-638 ) and GHR 1-454 with phenylalanines ( Fig. 1), expressed the mutated receptors in CHO cells, compared the relative levels of 125 I-hGH binding in the different cell lines, and examined whether GH stimulates tyrosyl phosphorylation of these mutated receptors. 125 I-hGH binding to cells expressing GHR 1-638 Y333F,Y338F, GHR 1-454 , and GHR 1-454 Y333F,Y338F was 88 Ϯ 2% (n ϭ 2), 71 Ϯ 4% (n ϭ 12), 41 Ϯ 5% (n ϭ 10), respectively, that of cells expressing wild-type GHR 1-638 (Fig. 1). GH⅐GHR⅐JAK2 complexes were prepared from GH-treated cells using ␣GH. Because GHR is not phosphorylated in the absence of GH (6, 7), the amount of tyrosyl phosphorylated GHR observed in the ␣GH precipitates from GH-treated cells reflects the amount of GH-dependent phosphorylation. When GH⅐GHR⅐JAK2 complexes are precipitated from CHO cells expressing truncated receptor and Western-blotted with ␣PY, two tyrosyl phosphorylated proteins are detectable (Fig. 2, lane F) as reported previously (7). The lower, relatively diffuse band (M r ϳ80,000) has been identified as receptor and the upper, relatively narrow band as JAK2 because of their sizes and their ability to be recognized in Western blots by ␣GHR and ␣JAK2, respectively (7). In contrast, when GH⅐GHR complexes are precipitated from CHO cells expressing the truncated Y333F,Y338F receptor and blotted with ␣PY, a band corresponding to JAK2 is observed, but no band corresponding in size to the truncated receptor is observed (Fig. 2, lane H), even in experiments in which the signal for GHR 1-454 is ϳ20 times that in Fig. 2 (data not shown). Although binding was reduced (by 41 Ϯ 5%, see Fig. 1) in the cells expressing the mutated, truncated receptor compared with cells expressing its unmutated counterpart, it was not sufficiently reduced to account for the total absence of a band corresponding to GHR. Thus, this result is consistent with Tyr 333 and/or Tyr 338 being the tyrosine in the N-terminal half of the cytoplasmic domain of GHR that is phosphorylated in response to GH.
When GH⅐GHR⅐JAK2 complexes are precipitated using ␣GH from GH-treated CHO cells expressing wild-type receptor and Western-blotted with ␣PY, a broad band migrating with M r ϳ120,000 -130,000 is observed (Fig. 2, lane B), as reported previously (7). Western blotting with ␣GHR and ␣JAK2 indicates that this band contains both GHR and JAK2, with JAK2 migrating as a rather narrow band (M r ϳ130,000) (see Fig. 6) and GHR migrating as a diffuse band (M r ϳ120,000) with and just below JAK2 (6,7). In ␣PY blots of ␣GH precipitates from GH-treated CHO cells expressing Y333F,Y338F full-length receptor, a diffuse band migrating with a M r appropriate for both GHR and JAK2 is also observed (Fig. 2, lane D). The diffuseness of the band indicates that the mutated receptor is phosphorylated, suggesting that tyrosines other than 333 and/or 338 in GHR are phosphorylated in response to GH. However, the significantly reduced intensity of this band (by 76 Ϯ 2%, n ϭ 3) compared with that obtained with wild-type receptor suggests that phosphorylation of Tyr 333 and/or Tyr 338 contributes to the level of GHR phosphorylation observed in wild-type receptor.
Ability of Substantially Purified GH⅐GHR Complexes to Incorporate Phosphate in an in Vitro Kinase Assay-To examine whether Tyr 333 and/or Tyr 338 are likely to be phosphorylated by the GHR-associated JAK2 kinase, GH⅐GHR⅐JAK2 complexes were substantially purified from GH-treated CHO cells using ␣GH and incubated with [␥-32 P]ATP. Fig. 3a illustrates that 32 P is incorporated into proteins migrating with molecular weights appropriate for both GHR and JAK2 when GH⅐GHR⅐JAK2 complexes are prepared from CHO cells expressing GHR   (Fig. 3a, lane B), as reported previously (7). When a similar experiment is performed using the Y333F,Y338F mutated receptor, 32 P is incorporated almost exclusively into a band corresponding in size to JAK2 (Fig. 3a,  lane D). The amount of mutated receptor is assumed to be less (by 40%) than the amount of unmutated receptor based upon 125 I-hGH binding data (Fig. 1). To ensure that a 40% reduction in the number of GHR could not account for the inability to detect phosphorylated Tyr 3 Phe truncated receptor, lanes C and D were exposed to film for a longer period of time sufficient to almost triple the phosphorylation signal for GHR  . Even with the longer exposure, no band co-migrating with phosphorylated truncated receptor was observed (Fig. 3a, lane F). A faint band of unknown origin may be seen migrating slightly ahead of where truncated GHR would be (Fig. 3a, lane F).
To confirm that in the in vitro kinase assay, there is a difference between mutated and unmutated receptor in the amount of phosphate incorporated into tyrosyl, as opposed to seryl and threonyl, residues, GH⅐GHR⅐JAK2 complexes were prepared from CHO cells treated with 100 ng/ml (4.5 nM) GH for 15 min at 37°C and incubated in the absence and presence of unlabeled ATP at the same concentration of ATP (5 M) used in the [␥-32 P]ATP experiment. Kinase assay-dependent changes in the amount of tyrosyl phosphorylation of GHR were assessed by Western blotting with ␣PY. An ATP-dependent tyrosyl phosphorylation of a protein migrating with appropriate M r was observed when GHR was prepared from CHO cells expressing truncated receptor (Fig. 3b, compare lanes B and C), but not when it was prepared from cells expressing truncated receptor with the Tyr to Phe substitution (Fig. 3b, compare lanes E and F). As in Fig. 3a, the intensity of the JAK2 band from mutated versus unmutated cells was reduced approximately to the same extent as binding of 125 I-hGH (Fig. 1). Even when lanes E and F were exposed to film for a substantially longer period of time (Fig. 3b, lanes H and I) to compensate for the 40% decrease in GH binding in the cells expressing mutated receptor and making the JAK2 signal comparable for the mutated and unmutated receptors, no band corresponding to the mutated receptor was detectable. 3 Size of GHR Lacking Tyr 333 and Tyr 338 -The experiments described for Figs. 2 and 3 provide evidence that the amount of phosphate incorporated into GHR lacking Tyr 333 and Tyr 338 both in vivo and in the in vitro kinase assay is reduced compared with nonmutated GHR by more than can be accounted for by differences in the amount of GHR expressed in the corresponding cell lines. This is consistent with GH promoting the tyrosyl phosphorylation of Tyr 333 and/or Tyr 338 and with one or both of these tyrosines serving as a substrate of a GHR-associated kinase, presumably JAK2. However, the reduced level of phosphorylation observed for GHR lacking Tyr 333 and Tyr 338 could also potentially arise if in comparison to their nonmutated counterparts, the mutated receptors: 1) were more susceptible to proteolysis so that they lacked tyrosines other than 333 and 338 that are sites of phosphorylation or 2) had a substantially reduced ability to associate with or to activate JAK2.
To verify that mutation of Tyr 333 and Tyr 338 to Phe did not result in adventitious proteolysis of full-length and truncated receptors to the extent that potential alternative phosphorylation sites were deleted, CHO cells expressing the various GHRs were incubated with 125 I-hGH, followed by the cross-linking reagent disuccinimidyl suberate. Fig. 4 illustrates that when the molecular weight of hGH (22,000,Ref. 37) is taken into account, the various cross-linked 125 I-hGH⅐GHR complexes migrate as proteins of the appropriate size (M r ϳ134,000 for full-length, M r ϳ95,000 for truncated receptor). In Fig. 4, a significantly greater portion of the 125 I-hGH⅐GHR 1-638 Y333F, Y338F complexes compared with other 125 I-hGH⅐GHR complexes appeared to be degraded, migrating as if the receptor was truncated at amino acid ϳ415. However, this large difference was not reproducible. In the three cross-linking experiments performed, the amount of 125 I-hGH cross-linked to full-length mutated GHR was only 26 Ϯ 37% less than the amount of 125 I-hGH cross-linked to full-length GHR. This suggests that the reduced (by ϳ80%) phosphorylation observed for GHR 1-638 Y333F,Y338F⅐JAK2 complexes compared with GHR 1-638 ⅐JAK2 complexes cannot be attributed to a comparable reduction in the amount of intact receptor. Despite the presence of substantial amounts of degraded 125 I-hGH⅐GHR 1-638 complexes corresponding in size to GHR 1-415 in Fig. 5, one does not see in Fig. 2 a phosphorylated band corresponding in size to this truncated receptor. This provides ad- 3 In contrast to Fig. 2, in vivo tyrosyl phosphorylation of truncated receptor is not evident in Fig. 3b (i.e. no 80-kDa band is detected in lane  B). This apparent discrepancy results from the difference in GH concentrations used in the two experiments. In Fig. 2, cells were incubated with a maximally stimulatory concentration of GH (500 ng/ml, 22 nM) to maximize the amount of GHR phosphorylation. For Fig. 3b, cells were incubated with a submaximal concentration of GH (100 ng/ml, 4.5 nM) to enable additional phosphorylation of these proteins to occur in the in vitro kinase assay. The result was the same whether the incubation with GH was for 15 min for 37°C as in Fig. 3b or for 1 h at 25°C as for Fig. 3a (data not shown).  638 (lanes A, B, and G), GHR   (lanes C and D), GHR 1-454 Y333F,Y338F (lanes E and F), and GHR 1-638 Y333F,Y338F (lane H) were incubated for 1 h at 25°C with 125 I-hGH in the absence (TOT,  lanes B, D, F, G, and H) or presence (NS, lanes A, C, and E) of 1 g/ml unlabeled hGH. Disuccinimidyl suberate (0.4 mM) was then added and the incubation continued for 15 min at 8°C. Samples were analyzed by SDS-PAGE followed by autoradiography. The migration of molecular weight standards (ϫ 10 Ϫ3 ) is indicated between lanes F and G. The migration of GHR 1-454 , GHR 1-638 , and degraded GHR 1-638 Y333F, Y338F is indicated. ditional evidence that Tyr 391 (predicted to be the only tyrosine present in Tyr 3 Phe mutated GHR 1-415 ) is not phosphorylated in response to GH, supporting our overall conclusion that Tyr 333 and/or Tyr 338 , but not Tyr 391 (or Tyr 437 ) are phosphorylated in response to GH.
Ability of GHR Mutants to Mediate GH-promoted Association of JAK2 with GHR and JAK2 Activation-The results of Figs.  2 and 3 showing the presence of a tyrosyl-phosphorylated protein migrating with a molecular weight appropriate for JAK2 in ␣GH immunoprecipitates from GH-treated cells indicate that JAK2 co-precipitates with the mutated full-length and truncated GHR. This suggests that JAK2 is capable of associating with these mutant receptors. To provide more direct evidence of the ability of JAK2 to associate with these mutant receptors, and to compare the ability of the mutated and nonmutated receptors to bind JAK2, GH⅐GHR⅐JAK2 complexes were immunoprecipitated using ␣GH and Western-blotted with ␣JAK2 (Fig. 5). JAK2 was found to associate with both the nonmutated (lanes C and E) and mutated (lanes A and G) receptors. Although the amount of JAK2 associated with GHR 1-454 Y333F,Y338F compared with GHR 1-454 in Fig. 5 appeared to be substantially reduced, this finding was not reproducible. In three experiments, the amount of JAK2 associated with GHR 1-454 Y333F,Y338F was 115 Ϯ 31% the amount associated with GHR 1-454 . The amount associated with mutated, full-length GHR was 67 Ϯ 1% the amount associated with wild-type GHR. Thus, JAK2 association with GHR does not depend upon Tyr 333 and/or Tyr 338 . Furthermore, the substantial reduction in the amount of phosphorylated receptor observed when Tyr 333 and Tyr 338 are mutated to Phe cannot be attributed to a comparable decreased level of JAK2 associated with these GHR.
To determine whether Tyr 333 and/or Tyr 338 are required for activation of JAK2, we compared the abilities of the mutated and unmutated GHR to mediate GH-dependent tyrosyl phosphorylation of JAK2. Tyrosyl phosphorylation of JAK kinases is thought to be due to autophosphorylation and thus to reflect JAK activation (6,8), a hypothesis supported by the finding that tyrosyl phosphorylation of JAK2 correlates well with JAK2 activation by GH. 4 JAK2 was precipitated from CHO cells expressing the various GHRs using ␣JAK2 and tyrosyl phosphorylation of JAK2 was assessed by Western blotting with ␣PY. Fig. 6 illustrates that for CHO cells expressing the four different GHR, including two CHO cell lines expressing different levels of GHR 1-638 used in the accompanying paper (42) (clone 23 used in all other experiments (relatively high) and clone 3 (relatively low)), GH stimulates tyrosyl phosphorylation of JAK2, roughly in proportion to the amount of GH binding detected in the different cell lines. Phosphorylation of JAK2 was 129 Ϯ 29% (n ϭ 2) for mutated versus unmutated full-length receptor and 48 Ϯ 20% (n ϭ 2) for mutated versus unmutated truncated receptor. The corresponding ratios for 125 I-hGH binding were 88 Ϯ 2% and 60 Ϯ 6% (n ϭ 10), respectively. ␣JAK2 Western blots of the ␣JAK2 immunoprecipitates revealed similar amounts of JAK2 expressed in the different cell lines (data not shown), indicating that differences in JAK2 phosphorylation in response to GH cannot be attributed to differences in levels of JAK2 expressed in the different cell lines.
Ability of GHR Mutants to Stimulate Tyrosyl Phosphorylation of Cellular Proteins-To provide additional evidence of the ability of the various mutant GHR to mediate GH-dependent JAK2 activation, we examined the ability of the mutated GHR to stimulate tyrosyl phosphorylation of cellular proteins, a phenomenon thought to be dependent upon JAK2 activation (33). Consistent with previous results (7,35), when lysates from CHO cells that express wild-type GHR are lysed with boiling SDS lysis buffer and then analyzed by ␣PY Western blot, GH-dependent tyrosyl phosphorylation of four proteins (p121, p97, p42, and p39, designated by the arrows on the left of Fig.  7) is observed (Fig. 7, lanes A-C). p121 has been identified as JAK2 (Ref. 6; data not shown). The identity of p97, which sometimes appears as a doublet (lanes B and C) and may actually represent two proteins, is currently unknown. p42 and p39 have been tentatively identified as the mitogen-activated protein kinase isoforms designated extracellular signal-regulated kinases (ERKs) 1 and 2 (27,35,38,39). p121, p97, p42, and p39 were all phosphorylated in response to GH in CHO cells expressing GHR 1-638 Y333F,Y338F (Fig. 7, lanes D-F). In contrast, in CHO cells expressing truncated receptor, GH stimulates tyrosyl phosphorylation of p121, p42, and p39 but not p97 (Fig. 6, lanes G-I) as shown previously (7). The same three proteins are tyrosyl phosphorylated in response to GH in CHO cells when Tyr 333 and Tyr 338 in the truncated receptor are mutated. Phosphorylation was reduced, consistent with the reduced number of receptors in these cells. These results provide evidence that Tyr 333 and Tyr 338 are not required for the phosphorylation of p97, ERK 1, and ERK 2. They also provide 4 E. Adkins, G. Campbell, and C. Carter-Su, unpublished observation. additional evidence that GH-dependent JAK2 association and activation occurs with the full-length or half-truncated GHR in the absence of Tyr 333 and Tyr 338 . DISCUSSION The results presented in this work provide strong evidence that Tyr 333 and Tyr 338 are not required for JAK2 association with GHR or for JAK2 activation. This is consistent with the finding using human GHR expressed in FDC-P1 cells (published while the present paper was under review) that no tyrosines in GHR are required for JAK2 phosphorylation in response to GH (41). The results of this study also provide evidence that Tyr 333 and/or Tyr 338 are phosphorylated in response to GH, since receptors lacking these tyrosines are phosphorylated to a significantly reduced extent (full-length) or not at all (truncated receptor) compared with the same sized receptors retaining these tyrosines. Presumably, it is the GHRassociated JAK2 tyrosine kinase that phosphorylates these tyrosine(s), since the truncated receptor is phosphorylated when GH⅐GHR⅐JAK2 complexes are precipitated with ␣GH and incubated with [␥-32 P]ATP, whereas the truncated receptor lacking Tyr 333 and Tyr 338 is not. Although Tyr 333 seems the most likely candidate based upon sequence analysis, additional studies will be required to determine which of the two tyrosines (333 or 338) is phosphorylated. Whether or not these are the only tyrosines in GHR 1-454 that are phosphorylated is not known, since it is possible that phosphorylation of Tyr 333 and/or Tyr 338 is required for the subsequent phosphorylation of Tyr 391 and Tyr 437 . Although it seems unlikely given the conservative nature of the amino acid substitution, our results cannot rule out the alternative possibility that Tyr 333 and Tyr 338 are not themselves phosphorylated but rather, mutating Tyr 333 and Tyr 338 to Phe alters the ability of Tyr 391 and/or Tyr 437 to be phosphorylated.
Tyr 333 and/or Tyr 338 appear not to be the only tyrosines phosphorylated in response to GH, since GH appears to stimulate the tyrosyl phosphorylation of the full-length, Tyr 3 Phe mutated receptor. The 6 tyrosines between amino acids 454 -638 are the most likely candidates because tyrosines other than 333 and/or 338 present in GHR 1-454 appear not to be phosphorylated to any great extent. Multiple phosphorylated tyrosines in GHR would be consistent with multiple sites of phosphorylation in receptors with intrinsic tyrosine kinase activity (e.g. receptors for insulin, epidermal growth factor, platelet-derived growth factor) (reviewed in Ref. 1). Multiple sites of phosphorylation with differing affinities for various SH2 domains would provide a mechanism by which GH could initiate several signaling pathways simultaneously.
The finding that Tyr 333 and/or Tyr 338 are likely to be phosphorylated in response to GH raises the question of whether either or both serve as binding sites for specific SH2 domains. Neither tyrosine, with its surrounding amino acids, closely resembles a high affinity binding site (as currently defined) of the SH2 domains of Csk, SHC, Syk, Vav, Grb2, 3BP2, HCP, or fps/fes (5). Consistent with this, neither tyrosine appears to be required for GH-dependent SHC phosphorylation (40). Since SHC lies downstream of GHR and upstream of mitogen-activated protein kinases, the SHC data are consistent with the data presented here and in the accompanying paper (42) that Tyr 333 and Tyr 338 are not required for GH activation of the mitogen-activated protein kinases ERKs 1 and 2. Tyr 333 and Tyr 338 also appear not be required for GH stimulation of insulin receptor substrate-1 5 or for activation of signal transducer and activator of transcriptions (Stats) 1 and 3 , 6 although the data do not exclude the possibility that one or more of these SH2 domain-containing proteins binds to Tyr 333 and/or Tyr 338 . However, as described in the accompanying paper (42), Tyr 333 and/or Tyr 338 do seem to be required for other actions of GH (lipid synthesis, protein synthesis), cellular responses for which the signaling pathways are less well defined. Verification that Tyr 333 and/or Tyr 338 are tyrosyl-phosphorylated using phosphopeptide analysis and identification of the signaling molecules that bind to the corresponding phosphorylated tyrosine(s) is likely to provide important information about signaling pathways used not only by GH but by other ligands that signal via tyrosine kinases.