Mapping of a cytoplasmic domain of the human growth hormone receptor that regulates rates of inactivation of Jak2 and Stat proteins.

It has been previously demonstrated that growth hormone (GH)-stimulated tyrosine phosphorylation of Jak2 and Stat5a and Stat5b occurs in FDP-C1 cells expressing either the entire GH receptor or truncations of the cytoplasmic domain expressing only the membrane-proximal 80 amino acids. However, other receptor domains that might modulate rates of GH activation and inactivation of this cascade have not been examined. Here we have defined a region in the human GH receptor between amino acids 520 and 540 in the cytoplasmic domain that is required for attenuation of GH-activated Jak/Stat signaling. Immunoprecipitations with antibodies to Jak2 indicate that the protein tyrosine phosphatase SHP-1 is associated with this kinase in cells exposed to GH. To address the possibility that SHP-1 could function as a negative regulator of GH signaling, liver extracts from motheaten mice deficient in SHP-1 or unaffected littermates were analyzed for activation of Stats and Jak2. Extracts from motheaten mice displayed prolonged activation of the Stat proteins as measured by their ability to interact with DNA and prolonged tyrosine phosphorylation of Jak2. These results delineate a novel domain in the GH receptor that regulates the inactivation of the Jak/Stat pathway and appears to be modulated by SHP-1.

It has been previously demonstrated that growth hormone (GH)-stimulated tyrosine phosphorylation of Jak2 and Stat5a and Stat5b occurs in FDP-C1 cells expressing either the entire GH receptor or truncations of the cytoplasmic domain expressing only the membrane-proximal 80 amino acids. However, other receptor domains that might modulate rates of GH activation and inactivation of this cascade have not been examined. Here we have defined a region in the human GH receptor between amino acids 520 and 540 in the cytoplasmic domain that is required for attenuation of GH-activated Jak/Stat signaling. Immunoprecipitations with antibodies to Jak2 indicate that the protein tyrosine phosphatase SHP-1 is associated with this kinase in cells exposed to GH. To address the possibility that SHP-1 could function as a negative regulator of GH signaling, liver extracts from motheaten mice deficient in SHP-1 or unaffected littermates were analyzed for activation of Stats and Jak2. Extracts from motheaten mice displayed prolonged activation of the Stat proteins as measured by their ability to interact with DNA and prolonged tyrosine phosphorylation of Jak2. These results delineate a novel domain in the GH receptor that regulates the inactivation of the Jak/Stat pathway and appears to be modulated by SHP-1.
Growth hormone (GH) 1 exerts its pleiotropic actions on a variety of tissues including fat, bone, soft tissue, and liver. One of the earliest events that occurs after GH binds to its cell surface receptor is the tyrosine phosphorylation of several cellular proteins, including the SH2 domain-containing transcription factors termed signal transducers and activators of transcription or Stats (1)(2)(3)(4). Tyrosine phosphorylated Stat proteins bind enhancers that are present in genes whose transcription is rapidly induced by the treatment of cells with GH and other cytokines. One of these enhancers is the gamma response region (GRR) present in the promoter of the Fc␥RI receptor gene. This enhancer, which is required for IFN␥-activated transcription of the Fc␥RI receptor gene, has a sequence similar to those of enhancers that are required for the activation of cellular genes by a variety of other cytokines. GRR binding activity can be measured in many cells in response to growth hormone treatment, and it serves as an assay for the tyrosine phosphorylation of Stat proteins (1,5). Most of the cytokine receptors interact with members of the Jak family of tyrosine kinases, and Jak activation closely parallels and in many cases is required for Stat protein phosphorylation on tyrosine. Tyrosine phosphorylated Jak2 has been shown to associate with the GHR after the addition of ligand, which allows Stat1, Stat3, Stat5a, and Stat5b to also be phosphorylated (6).
The receptors for growth hormone and other members of this cytokine receptor superfamily have several conserved features including cysteine residues within their extracellular domains and two intracellular subdomains (termed box 1 and box 2) adjacent to the transmembrane region. To elucidate the domains in the GH receptor required for activation of Stat(s) and Jak2, cell lines containing deletions in the cytoplasmic domain of the human receptor have been analyzed for GH-stimulated tyrosine phosphorylation of Jak2 and GRR binding activity. These studies demonstrated the importance of box 1 and box 2 in GH activation of Jak2 kinase and the Stat transcription factors (5,(7)(8)(9)(10)(11). However, little if any information has been reported concerning the role of other domains within the cytoplasmic region of the receptor in the modulation of GH activation of the Jak/Stat pathway. It has been shown that the SH2 domain-containing PTP SHP-1 (PTP1C, SHPTP1, and HCP) plays a role in the dephosphorylation of Jak2 after erythropoietin (EPO) stimulation and functions to down-regulate the proliferative effects of both EPO and IL-3, activators of Jak2 (12)(13)(14). In the case of EPO activation of Jak2, SHP-1 is recruited through its SH2 domain to the receptor as a consequence of the tyrosine phosphorylation of the later (13). Several reports have also implicated a role for tyrosine phosphatases in IFN regulation of the Jak/Stat pathway, including the role of SHP-1 as a negative regulator of IFN signaling and PTP1D (SHP-2) as a positive activator of both interferon and prolactin stimulation of the Jak/Stat pathway (14 -16). These results suggested that it would be worthwhile to examine whether other components modulate GH stimulation of the Jak/Stat pathway.

Cells-
The FDC-P1 cell line was transfected with cDNAs of the human growth hormone receptor and cytoplasmic truncations thereof (17). Cell lines were grown in RPMI 1640 supplemented with 10% fetal calf serum, 50 M 2-mercaptoethanol, 50 g/ml gentamicin, 700 g/ml G418, and 5 nM human growth hormone (17). Cells were starved overnight in the absence of GH and then incubated for 1-2 h in fresh medium minus serum prior to being treated with 10 nM GH for the times indicated.
Whole Cell Extracts-Cells (5 ϫ 10 7 ) were collected by centrifugation, washed with phosphate-buffered saline, and resuspended in ice cold * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Electrophoretic Mobility Shift Assay-The EMSA was performed as described previously using whole cell extracts (see above) (1,18,19). The GRR (gamma response region) (5Ј-AGCATGTTTCAAGGATTT-GAGATGTATTTCCCAGAAAAG-3Ј) of the promoter of the Fc␥RI gene was end-labeled using T4 polynucleotide kinase and [ 32 P-␥]ATP and used in all EMSAs.
Liver Extracts-15-20-day-old me/me mice or their unaffected littermates were injected intraperitoneally with GH (10 g/10 g of body weight) and were sacrificed 15-20 min later. Livers were removed, and a portion was snap frozen in liquid nitrogen prior to preparation of whole cell extracts (20). The remaining tissue was placed in Dulbecco's modified Eagle's medium and incubated at 37°C. After 10, 15, or 30 min, aliquots of tissue were snap frozen, and extracts were prepared.

RESULTS
Experiments were initiated to examine whether regions of the cytoplasmic domain of the GH receptor other than the previously described box1-box2 Jak2-binding domain might be involved in the modulation of the GH-stimulated Jak/Stat signaling pathway. As an initial screen, lysates of FDC-P1 cells that express either full-length or carboxyl-terminal truncations of the GH receptor (see Fig. 1) were analyzed for GH-induced activation of Stat proteins by their ability to bind to the GRR of the high affinity Fc␥R1 receptor. Several of these lines have been previously characterized and were found to express approximately equal numbers of GH receptors (17). We determined that Stat5a and Stat5b are the only known Stats to become tyrosine phosphorylated in these cell lines as a result of incubation with GH (data not shown). To determine whether the removal of any region of the cytoplasmic domain of the GHR might affect the duration of the GH signal, cells were incubated with GH for 10 min at 37°C, diluted, pelleted, and then washed once with warm medium before being resuspended in fresh medium without GH. The cells were incubated at 37°C for varying times without GH, and extracts were prepared. In all cell lines tested, a robust induction of GRR binding activity (labeled GHSF in Fig. 2A) was detected after incubation with GH for 10 min (Fig. 2A, compare lanes 1 and 2).
In cells expressing the full-length receptor or a receptor that contained only the first 539 amino acids of the GHR (P540stop), most of GHSF complex disappeared after a 30-min incubation in the absence of GH and was nearly absent after 60 min ( Fig.  2A, compare lanes 2 with lanes 4 and 5). However, in cell lines that contained only the amino-terminal 461 or 350 residues (S462stop or D351stop, respectively), no loss of the GH-induced Stat complex was observed after 1 h in the absence of GH, and much of the activated Stat was still present after 4 h ( Fig. 2A,  lanes 5 and 6).
To further map the region in the cytoplasmic domain of the GHR responsible for this down-regulation of the Jak/Stat signaling cascade, a series of carboxyl-terminal truncations of the receptor were generated between amino acids 462 and 540, where the change in the rate of decay of the activated Stats occurred (Fig. 2B). Cell lines that expressed 520 amino acids or less of the GHR all showed delayed rates of attenuation of the GHSF, suggesting that the region between amino acids 520 and 539 mediates this function. Although this particular domain of the receptor contains no tyrosines, SHP-1 has two SH-2 do- mains that have been implicated in binding to a phosphorylated tyrosine in the EPO receptor (13). We decided to mutate the two tyrosines located at amino acids 469 and 516 to ensure that these residues were not involved in altering the half-life of activation of the GH signal. A PhosphorImager was used to determine the amount of GHSF Stat5-containing complex in cell lines where these tyrosines were replaced with phenylalanines. The results of these experiments are shown in Fig. 2C. Compared with the cell line that expresses 350 amino acids of the GHR, the cell lines that either expressed the full-length receptor or mutations of one or both tyrosines (Y469F, Y516F, or Y469F/Y516F) all displayed similar rates of Stat inactivation. It therefore appears that tyrosine phosphorylation of the receptor in the 462-540 region is not involved in the mechanism for down-regulating the Stat-containing GHSF complex.
The Jak2 tyrosine kinase is activated by tyrosine phosphorylation as a result of treatment of cells with GH, and activation of this kinase is linked to GH-stimulated tyrosine phosphorylation of the Stat proteins (1)(2)(3)(4). To determine whether the differential rates of inactivation of the Stat proteins paralleled different rates of inactivation of Jak2, cells were treated with GH for 10 min and washed in medium as described above. Cellular extracts were prepared, and Jak2 was immunoprecipitated and examined on blots by probing with antiphosphotyrosine antibodies (Fig. 3). In all of the cell lines examined, a 10-min incubation of cells with GH stimulated the tyrosine phosphorylation of Jak2 (Fig. 3A, compare lanes 1 and 2). After removing GH, cells that expressed the full-length or aminoterminal 539 amino acids of the receptor displayed rapid dephosphorylation of Jak2, which was complete within 30 min. However, in cells expressing either the proximal 520 or 350 amino acids of the receptor, a delayed dephosphorylation of the enzyme was observed. Reprobing the blots with Jak2 antiserum confirmed the presence of approximately equal amounts of Jak2 protein in each sample. These results correlated with the presence of the GH-induced Stat complex seen in Fig. 2 and indicated that a region in the receptor between 521 and 540 is required to inactivate GH stimulation of the Jak/Stat signaling cascade.
Recent evidence has implicated the protein tyrosine phosphatase SHP-1 as a negative regulator of IFN␣/␤, EPO, and IL-3 signaling by Jak1 or Jak2 (12)(13)(14). In the case of IFN␣/␤ activation of the Jak/Stat pathway, SHP-1 is constitutively associated with the ␣ subunit of the IFN␣ receptor and is displaced from the signaling complex after the addition of IFN␤ (14). To determine whether SHP-1 might be responsible for inactivation of the GH-stimulated Jak/Stat pathway, experiments were performed to determine whether SHP-1 was associated with the GHR. In co-immunoprecipitation experiments, SHP-1 was often constitutively associated with the full-length or truncated GHR and was lost after treatment of cells with GH; however, this result was not consistent (data not shown).
To examine this interaction by an alternative approach, immunoprecipitations were performed to determine whether SHP-1 associated with Jak2 because Jak2 is activated and becomes associated with the GHR as a result of treatment of cells with GH (6). Extracts made from GH-stimulated cells were immunoprecipitated with Jak2 antiserum, and the resulting immunoblots were probed with either antiphosphotyrosine (Fig. 4A) or SHP-1 antibodies (Fig. 4B). SHP-1 associated with Jak2 after incubation of cells with GH at a time when the kinase became tyrosine phosphorylated (Fig. 4, A and B, compare lanes 1 and 3), suggesting that SHP-1 might function to shut off signaling by dephosphorylating Jak2. Fig. 4C is a reprobe of Fig. 4A with Jak2 antiserum to demonstrate that approximately equal amounts of protein were present in each sample.
Although the association/dissociation of SHP-1 with Jak2 was demonstrated to be ligand-dependent, it was possible that the changes in down-regulation of the signaling cascade that were observed with the truncated receptors were not directly correlated with the actions of SHP-1. To examine this issue in greater detail, experiments were performed using motheaten mice (me/me). The me/me phenotype is a result of a mutation in the SHP-1 gene such that this PTP is absent in these mice (21,22). The lack of expression of SHP-1 causes multiple hematopoietic abnormalities, including hyperproliferation and inappropriate activation of macrophages resulting in widespread inflammation. Previous studies have shown that injection of rats with GH activates the Jak/Stat pathway in the liver (4). To examine the role of SHP-1 in GH signaling, livers were isolated from me/me mice and their unaffected littermates after the mice were injected with either GH or saline. Cellular extracts were prepared from a portion of the liver at the time the animals were sacrificed (Fig. 5A, lanes 1, 2, 5, and 6). The . Extracts were prepared as described by David et al. (14) and incubated with Jak2-specific antiserum. The immunoblots were either probed with antiphosphotyrosine antibody (A) or with antibody against SHP-1 (B). C, the blot shown in A was reprobed with Jak2 antiserum to show equal protein loading.
remaining tissue from animals injected with GH was incubated for varying times at 37°C, and portions of the liver were extracted for analysis of activated Stats by EMSA. GH-stimulated Stat activation was assayed by EMSA in equivalent protein loadings and was found to be approximately 1.5-fold greater in livers isolated from me/me mice compared with livers from unaffected littermates (Fig. 5A, lanes 2 versus 6). The decay in the GHSF complex was markedly delayed in livers from me/me mice after incubation at 37°C (Fig. 5A, lanes 3  versus 7). The results of several experiments are displayed in Fig. 5B, where the amount of GHSF was quantitated by a PhosphorImager. Tyrosine phosphorylation of Jak2 was also assayed in liver extracts from mice injected with GH (Fig. 6), and its disappearance was found to be delayed in me/me mice. Reprobing the blot for Jak2 protein showed that approximately equal amounts of protein were present in each lane (data not shown). DISCUSSION Previous studies suggested that PTPs have both positive and negative functions in cytokine and growth factor signal transduction (23,24). SHP-1 controls EPO and IL-3 signal transduction by regulating receptor-associated Jak2 tyrosine phosphorylation; SHP-1 also controls Jak1 tyrosyl-phosphorylation in response to IFN␣/␤ (12)(13)(14). We have shown here that SHP-1 is one negative regulator of the GH signaling pathway in liver, and it is likely that at least some of the regulatory actions of this PTP are modulated by a domain in the cytoplasmic region of the GHR that is distinct from that required for activation of the Jak/Stat pathway by GH. These results support the concept that inactivation of receptor-associated Janus PTKs may be a general mechanism by which SHP-1 regulates multiple cytokine receptor signaling pathways. SHP-1 association with Jak2 appears to be dependent on stimulation of cells with GH ( Fig.  4). At the moment it is unclear whether SHP-1 interacts with the GHR or associates with tyrosine phosphorylated Jak2 independent of Jak2 association with the GHR. Immunoprecipitations with an antibody that recognizes the GHR revealed an association of SHP-1 with the GHR in the absence of treatment of cells with GH (data not shown). However, this result has not been consistent, suggesting that the interaction is weak. Our inability to detect a strong interaction between SHP-1 and the GHR is also consistent with the fact that tyrosine phosphorylation of the GHR does not correlate with the changes in the rate of Jak2 dephosphorylation when the region between amino acids 521 and 540 is deleted from the receptor. The association of SHP-1 with the IFN␣/␤ receptor is constitutive. Concomitant with activation of Jak1 and Tyk2, SHP-1 transiently dissociates from the complex and then returns at later time points (14). In contrast to the effects of SHP-1 in the IFN signaling cascade, stimulation of cells with EPO permits this PTP to directly associate with the EPO receptor, and this association is dependent upon tyrosine phosphorylation of the receptor (13). It thus appears that GH modulation of SHP-1 activity combines mechanisms used in both the IFN␣/␤ system where tyrosine phosphorylation of the receptor is not required for recruitment of the PTP, but like with Epo, Jak2 dephosphorylation by SHP-1 is enhanced by GH treatment of cells (13,25).
The likely presence of SHP-1 in the GHR signaling complex does not appear to be necessary to prevent gratuitous Jak2 activation, because basal activity of the enzyme is not elevated in livers from me/me mice (Fig. 6). Therefore, SHP-1 alone cannot account for all the PTP activity required to prevent spontaneous Jak/Stat activation. Rather it appears that this enzyme functions to limit the extent and duration of Jak/Stat activation. It is notable that in most experiments there is also enhanced GH-stimulated tyrosine phosphorylation of Jak2 in FIG. 5. Livers from motheaten mice injected with GH display prolonged activation of the GHSF. A, mice (me/me, lanes 6 -8, or unaffected littermates, lanes 2-4) were injected intraperitoneally with GH (10 g/10 g of body weight) and were sacrificed 15-20 min later. Livers were isolated, and a portion was snap frozen in liquid nitrogen prior to the preparation of whole cell extracts (20). The remaining tissue was placed in Dulbecco's modified Eagle's medium and incubated at 37°C. After 10, 15, or 30 min, aliquots of tissue were snap frozen, and extracts were prepared. EMSAs were performed using equal amounts of protein and the 32 P-labeled GRR probe. The GH-stimulated factor (GHSF) is marked with an arrow. Extracts from livers of mice injected with a saline control are shown in lanes 1 and 5. B, the results of several experiments such as shown in A were quantitated for the formation of GHSF as in Fig. 2C. The incubation time of the liver is plotted on the x axis. The amount of GHSF seen in the livers of animals at the time they were sacrificed was given an arbitrary value of 100%. WT, wild type; Me, me/me. 6. Livers from motheaten mice injected with GH display prolonged tyrosine phosphorylation of Jak2. Lysates (500 g) from liver extracts were incubated with anti-Jak2 antiserum. Immune complexes were collected on protein G beads, run on SDS-polyacrylamide gel electrophoresis (8% gel), immunoblotted with monoclonal anti-phosphotyrosine antibody 4G10, and detected by ECL. Samples in lanes 1 and 4 are liver lysates from saline-injected mice, lanes 2 and 5 are lysates from mice sacrificed 20 min after GH injection, and lanes 3 and 6 are lysates prepared from liver samples that were incubated in medium for 15 min at 37°C. Me, me/me. lines that demonstrate prolonged tyrosine phosphorylation of the enzyme (see Fig. 3). Whether another PTP functions to control basal activation of GH-stimulated Jak/Stat activity is not clear. However, vanadate does activate the cascade in the absence of ligand in macrophages isolated form me/me mice, suggesting that SHP-1 probably is not the only negative regulatory PTP in GH signaling (14). It is also clear that in livers from GH-treated mice as well as in all of the cell lines expressing the GHR, Jak2 is eventually dephosphorylated. This observation suggests that another PTP contributes to shutting off the system or can substitute for SHP-1 when it is not functional.
Although our results suggest a general model for SHP-1 regulation of GH signaling, questions still remain. The component(s) of the GHR/Jak2 complex that directly mediate association with SHP-1 as well as the molecular determinants of association (SH2-or non-SH2-mediated) remain to be defined. The region between 521 and 540 in the GHR which potentiates down-regulation of Jak2 activation by GH contains no tyrosine residues, and the two adjacent tyrosine residues at amino acids 469 and 516 appear to have no effect on down-regulation of signaling. It is therefore unlikely that tyrosine phosphorylation of the GHR is mediating this effect. However, it is possible that SHP-1 can interact directly or indirectly with the GHR at more than one site because signaling does eventually diminish in the truncated forms of the receptor. In fact, SHP-1 has been seen to associate with the GHR in the absence of ligand in lines expressing truncated forms of the receptor (data not shown). Alternatively, the carboxyl terminus of SHP-1, which has been implicated in its association with the insulin receptor (26), could be responsible. Studies using purified recombinant proteins should resolve this issue. Understanding the mechanisms by which SHP-1 is able to regulate cytokine signaling complexes is clearly of importance as its pivotal role in the regulation of cellular growth and differentiation becomes more and more evident.