The Growth Hormone-binding Protein Is a Location-dependent Cytokine Receptor Transcriptional Enhancer*

In the rat, a growth hormone-binding protein (GHBP) exists that is derived from the growth hormone (GH) receptor gene by an alternative mRNA splicing mechanism such that the transmembrane and intracellular domains of the GH receptor are replaced by a hydrophilic carboxyl terminus. In isolation, the GHBP is inac-tive, although it does compete with the receptor for ligand binding in the extracellular space and therefore inhibits the cellular response to GH. The GHBP is also located intracellularly and is translocated to the nucleus upon ligand stimulation. We show here that endogenously produced GHBP, in contrast to exogenous GHBP, was able to enhance the STAT5-mediated transcriptional response to GH. Interestingly, when the GHBP was targeted constitutively to the nucleus by the addition of the nuclear localization sequence of the SV40 large T antigen, greater enhancement of STAT5-mediated transcription was obtained. The transcriptional enhancement did not require GH per se and was not specific to the GH receptor, since similar enhancement of STAT5-mediated transcription by nuclear localized GHBP was obtained with specific ligand stimulation of both prolactin and erythropoietin receptors. Thus, the GHBP six-well plates on 50- (cid:2) m-thick sapphire glass coverslips. Transient transfection was performed in serum-free DMEM with DOTAP according to the manufacturer’s instructions. Cells were kept for 12 h in serum-free medium and then stimulated with 50 n M hGH or transferred into DMEM containing 10% fetal bovine serum as indicated. The cells were fixed by using an ethane-freezing/methanol fixa- tion (33). To enhance the GFP signal, the fixed cells were exposed to an anti GFP-antibody. Western Blot Analysis— Medium from BRL cells transiently transfected with the different GHBP constructs was collected and concen- trated. Fractions were normalized for protein content and loaded onto a 7.5% polyacrylamide gel as described (32). Proteins were transferred to nitrocellulose membranes using a semidry apparatus in Laemmli elec-trophoresis buffer containing 15% methanol. Membranes were blocked for 1 h with 5% skim milk powder in TBS (20 m M Tris-HCl, 150 m M NaCl, pH 7.4). mAb 4.3 at 0.25 (cid:2) g/ml in TTBS (TBS plus 0.1% Tween 20) with 1% skim milk powder was used for GHBP detection. Membranes were further processed and developed using the ECL system as previously described (32).

In the rat, a growth hormone-binding protein (GHBP) exists that is derived from the growth hormone (GH) receptor gene by an alternative mRNA splicing mechanism such that the transmembrane and intracellular domains of the GH receptor are replaced by a hydrophilic carboxyl terminus. In isolation, the GHBP is inactive, although it does compete with the receptor for ligand binding in the extracellular space and therefore inhibits the cellular response to GH. The GHBP is also located intracellularly and is translocated to the nucleus upon ligand stimulation. We show here that endogenously produced GHBP, in contrast to exogenous GHBP, was able to enhance the STAT5-mediated transcriptional response to GH. Interestingly, when the GHBP was targeted constitutively to the nucleus by the addition of the nuclear localization sequence of the SV40 large T antigen, greater enhancement of STAT5mediated transcription was obtained. The transcriptional enhancement did not require GH per se and was not specific to the GH receptor, since similar enhancement of STAT5-mediated transcription by nuclear localized GHBP was obtained with specific ligand stimulation of both prolactin and erythropoietin receptors. Thus, the GHBP exerts divergent effects on STAT5-mediated transcription depending on its cellular location. The use of a soluble cytokine receptor as a location-dependent transcriptional enhancer and the ligand-independent involvement of the extracellular domain of a polypeptide ligand receptor in intracellular signal transduction provide additional novel mechanisms of transcriptional control.
Growth hormone (GH) 1 is the major regulator of postnatal body growth and initiates its biological actions, including the induction of a number of RNA species in mammalian tissues, by interaction with a specific membrane-bound receptor (1,2).
The GH receptor was the first cloned member of the now extensive cytokine receptor superfamily, which includes the receptors for prolactin (PRL), erythropoietin (EPO), granulocyte colony-stimulating factor, granulocyte-macrophage colony stimulating factor, ciliary neutrophic factor, thrombopoietin, leptin, cardiotrophin I, and the ␤-chain of interleukin (IL)-2 through IL-7, IL-9, and IL-11 to IL-13 (3). Most receptors of the cytokine receptor superfamily exist in a soluble and transmembrane form (4 -7). The function of the transmembrane forms is well documented and includes signal transduction predominantly but not exclusively through the JAK-STAT pathway, resulting in gene transcription (8,9). The role of the soluble cytokine receptors, with the notable exception of the soluble forms of the IL-6 and ciliary neurotropitir factor (CNTF) receptors (10,11), appears confined to ligand sequestration in the extracellular space with a consequent impairment of the cellular response to exogenous ligand (4).
A soluble rat growth hormone-binding protein (GHBP) exists that is derived from the GH receptor gene by an alternative mRNA splicing mechanism such that the transmembrane and intracellular domains of the GH receptor are replaced by a hydrophilic carboxyl-terminal sequence (12). An analogous GHBP exists in other species (13), such as humans, but is derived by proteolytic cleavage of the full-length membranebound receptor (14), presumably by the action of specific metalloproteases (15). The GHBP has been demonstrated to compete with the receptor for ligand binding in the extracellular space and has been shown to inhibit the cellular response to GH in vitro (16,17). In vivo, the GHBP has been demonstrated to increase the biological activity of GH by prolongation of the half-life of plasma GH (18). The GHBP is also located intracellularly (19 -21) and has also been prominently localized to the nucleus (20). Other components of the GH signal transduction pathway are also located in the nucleus or translocate to the nucleus upon GH stimulation (21)(22)(23)(24)(25). Thus, the GH receptor is subject to ligand-dependent nuclear translocation (21), and constitutively nuclear JAK2 is phosphorylated by exogenous GH stimulation (24). Internalization of the GH receptor has been reported not to be necessary to achieve transcriptional activation by GH (26), and therefore the function of the nuclear localization of components of the GH signal transduction pathway is unknown.
We demonstrate here that nuclear localized GHBP functions as a potent enhancer of STAT5-mediated transcription, not only for GH but also for other members of the cytokine receptor superfamily. Thus, the GHBP exerts opposing effects on STAT5-mediated transcription depending on its extra-/intracellular location. The use of a soluble cytokine receptor as a location-dependent transcriptional enhancer and the ligandindependent involvement of the extracellular domain of a * 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.

EXPERIMENTAL PROCEDURES
Materials-Human growth hormone (hGH) was a generous gift of Novo Nordisk (Singapore). oPRL and rGH were obtained from NIDDK (National Institutes of Health), and mEPO was purchased from Roche Diagnostics. All cell culture medium and supplements (for culture medium) were obtained from Sigma. The luciferase assay system was purchased from Promega (Madison, WI). The ECL kit was obtained from Amersham Biosciences. The GH, PRL, and EPO receptor cDNAs used here were as described previously (27). Transfection reagent DOTAP, poly(dI/dC) and the DNA 3Ј-end labeling kit were purchased from Roche Diagnostics. Monoclonal antibody against hemagglutinin was obtained from Clontech, monoclonal antiserum against phospho-STAT5A/B were from Upstate Biotechnology, Inc. (Lake Placid, NY), monoclonal antisera against green fluorescent protein (GFP) were from Molecular Probes, Inc. (Eugene, OR), and polyclonal antibody against STAT5B were from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). mAb 4.3 and recombinant rat GHBP were generous gifts of Dr. W. R. Baumbach (Monsanto Corp.). The production and characterization of the recombinant rat GHBP has been previously described by us (20).
Generation of Stable Cell Transfectants-BRL cells were stably transfected with the complete rat GH receptor cDNA inserted into an expression vector containing the human cytomegalovirus enhancer and promoter (pcDNA1). The characterization and use of these cells has previously been described in detail (28). These cells will be referred to as BRL-GHR 1-638 cells.
Cell Culture-BRL cells were grown in DMEM (supplemented with 10% heat-inactivated fetal calf serum (FCS), 100 units/ml penicillin, 100 g/ml streptomycin, and 2 mM L-glutamine) at 37°C in 5% CO 2 . hGH, rGH, oPRL, and recombinant rat GHBP were prepared as a stock solution of 1 mg/ml in distilled water. For treatment of cells, hGH, rGH, oPRL, mEPO, and recombinant rat GHBP were diluted in fresh DMEM serum-free medium and added to the cells after transient transfection. Cells were treated with 100 nM hGH unless otherwise specified. oPRL was used at 100 nM. mEPO was used at 10 units/ml.
Construction of Expression Plasmids-The cDNA expression plasmid encoding the wild type GHBP under the control of the metallothionein Ia promotor was as previously described (17). In the XS-GHBP construct, the rat GHBP was PCR-amplified without its NH-terminal signal sequence, and an ATG was introduced in the primer just upstream of where the mature GHBP protein is coded. The nuclear localization sequence (NLS)-GHBP was constructed in a similar way, but a nuclear localization signal from the SV40 large T antigen (PKKKRKV) (29) was added upstream of where the mature GHBP is coded. For GHBP-GFP, wild type GHBP was subcloned into a N-terminal enhanced green fluorescent protein vector from Clontech (pEGFP-N3) under the control of a cytomegalovirus IE promoter. For the construction of epitopetagged GHBP mutants, wild type GHBP and NLS-GHBP with or without a stop codon at the position of amino acid 115 were subcloned into a pCI-neo vector with a double hemagglutinin tag at the N terminus. The integrity of the reading frame for the GHBP modifications was confirmed by sequence analysis. The construction of GH receptor cDNA expression plasmids containing a deletion, a deletion of box 1 (⌬297-311), or the individual substitution of proline residues 300, 301, 303, and 305 in box 1 for alanine has been described previously (30).
Transient Transfection and Reporter Assay-BRL and BRL-GHR 1-638 cells were cultured to confluence in six-well plates. Transient transfection was performed in serum-free DMEM with DOTAP according to the manufacturer's instructions. 1 g of reporter plasmid (SPI-GLE1-CAT) and 1 g of pSV2-LUC were transfected per well. The control or empty vectors served to normalize the amount of DNA transfected. For receptor cDNA transfection into BRL cells, 1 g of each receptor cDNA was used. Cells were incubated with DOTAP/DNA for 12 h before the medium was changed to serum-free DMEM containing either the respective hormones or GHBP at the indicated concentrations. After a further 24 h, cells were washed in PBS and scraped into lysis buffer. The protein content of the samples was normalized, and CAT and luciferase assays were performed as previously described (31). Results were normalized to the level of luciferase to control for transfection efficiency and calculated as the -fold stimulation of unstimulated (nonhormone-treated) cells.
Confocal Laser-scanning Microscopy-BRL cells were grown on glass coverslips in six-well plates and transiently transfected as described above. Fixation was performed with PBS, pH 7.4, containing 4% paraformaldehyde for 10 min at room temperature. Cells were permeabilized with PBS, 0.1% Triton X-100 for 1 min and processed for immunofluorescence as described (32). The location of the expressed GHBP was determined using the mAb 4.3 directed against the hydrophilic carboxyl terminus of the GHBP (20). Noncross-reactive mAbs 50.8 and 7 at the same concentration were used as control (20). Labeled cells were visualized with a Carl Zeiss Axioplan microscope equipped with epifluorescence optics and a Bio-Rad MRC1024 confocal laser system.
For the GFP-GHBP translocation, BRL-GHR 1-638 cells were cultured to confluence in six-well plates on 50-m-thick sapphire glass coverslips. Transient transfection was performed in serum-free DMEM with DOTAP according to the manufacturer's instructions. Cells were kept for 12 h in serum-free medium and then stimulated with 50 nM hGH or transferred into DMEM containing 10% fetal bovine serum as indicated. The cells were fixed by using an ethane-freezing/methanol fixation (33). To enhance the GFP signal, the fixed cells were exposed to an anti GFP-antibody.
Western Blot Analysis-Medium from BRL cells transiently transfected with the different GHBP constructs was collected and concentrated. Fractions were normalized for protein content and loaded onto a 7.5% polyacrylamide gel as described (32). Proteins were transferred to nitrocellulose membranes using a semidry apparatus in Laemmli electrophoresis buffer containing 15% methanol. Membranes were blocked for 1 h with 5% skim milk powder in TBS (20 mM Tris-HCl, 150 mM NaCl, pH 7.4). mAb 4.3 at 0.25 g/ml in TTBS (TBS plus 0.1% Tween 20) with 1% skim milk powder was used for GHBP detection. Membranes were further processed and developed using the ECL system as previously described (32).
Gel Electrophoretic Mobility Shift Assay-The gel electrophoretic mobility shift assay was performed according to standard protocols. The binding reactions were performed by preincubating 10 g of nuclear extracts or cytosolic extracts of both hGH-treated and -untreated control vector and NLS-GHBP-transfected cells with 3 g of poly(dI-dC) acid in 15 ml of buffer containing 20% Ficoll, 60 mM HEPES, pH 7.9, 20 mM Tris, pH 7.9, 0.5 mM EDTA, and 5 mM dithiothreitol for 15 min on ice. For supershift analysis, the extracts were incubated with the antibodies against STAT5B (Santa Cruz Biotechnology) or control antibodies for another 10 min on ice. 32 P-3Ј-end-labeled double-stranded SPI-GLE1 probe was added (5Ј-TGTTCTGAGAAATA-3Ј), and the mixture was incubated for 5 min on ice followed by another 10 min at room temperature. The samples were electrophoresed on 4.5% nondenaturing polyacrylamide gels in 0.25ϫ TBE (22.5 mM Tris borate, pH 8.0, 0.5 mM EDTA) at 250 V at 4°C for 2 h. The gel was dried, and the radioactive pattern was then visualized by autoradiography.
Statistics and Presentation of Data-All experiments were repeated at least three times. Figures presented for Western blot analyses are representative of multiple experiments. All numerical data are expressed as mean Ϯ S.E., and the data were analyzed using Instat 3.0 from GraphPad Software Inc.

Effect of Exogenous GHBP on GH-stimulated STAT5-mediated Transcription-
We have used a BRL (Buffalo rat liver) cell co-transfection assay (28) to study the role of the GHBP in the signal transduction pathway of GH. The characterization and use of BRL cells stably transfected with GH receptor cDNA has previously been described in detail (28), and these cells will be referred to as BRL-GHR 1-638 cells. We first used the BRL-GHR 1-638 cells to demonstrate the effect of exogenously added recombinant rat GHBP on GH-stimulated STAT5-mediated transcription utilizing a reporter plasmid containing the STAT5 binding element of the serine protease inhibitor 2.1 gene promoter (SPI-GLE1-CAT) (31). Exogenously added recombinant rat GHBP decreased in a dose-dependent manner the GH stimulation of STAT5-mediated transcription (Fig. 1). With the GH concentration fixed at 1 nM, a dose-dependent inhibition of GH-stimulated STAT5-mediated transcription was observed over the range of 1-100 nM GHBP with an ED 50 of 10 nM. The inhibition of GH-stimulated STAT5-mediated transcription is in accord with previous demonstrations that exogenous GHBP inhibits GH-stimulated function (16).
Effect of Endogenous GHBP on GH-induced STAT5-mediated Transcription-The GHBP has also previously been local-ized intracellularly both attached to intracellular membranes (19,20) and soluble in the cytoplasm and nucleoplasm (20). To determine the effect of endogenously produced GHBP on the GH-stimulated STAT5-mediated transcriptional response, we transiently transfected BRL cells with both GH receptor cDNA and WT-GHBP cDNA and examined the STAT5-mediated transcriptional response to GH. The transfected WT-GHBP cDNA resulted in GHBP protein expression in a predominantly perinuclear location and also secretion of the GHBP to the extracellular space (see Fig. 2). In contrast to exogenously added GHBP, 1 g of transiently transfected WT-GHBP cDNA resulted in a significant increase in the STAT5-mediated transcriptional response to GH (Fig. 3A). This effect was observed at concentrations up to 100 nM GH, and the subsequent decrease in GH-stimulated STAT5-mediated transcription at GH concentrations higher than 100 nM is consistent with antagonism of the GH effect at high ligand concentrations (34). Transfection of increasing amounts of WT-GHBP cDNA resulted in less enhancement of the GH-stimulated STAT5 transcriptional response presumably due to increased secretion of GHBP to the medium with the consequent inhibition of GH function. Transient transfection of WT-GHBP cDNA exerted no significant effect on STAT5-mediated transcription in the absence of concomitant transfection of GH receptor cDNA (Fig. 3B). Transient transfection of WT-GHBP cDNA into BRL-GHR 1-638 cells produced similar results as transient transfection of both WT-GHBP and GH receptor cDNA (data not shown). These results indicated that extracellular and intracellular GHBP exerted opposing effects on GH-stimulated STAT5-mediated transcription.
Cytoplasmic GHBP Enhances GH-stimulated STAT5-mediated Transcription-To determine whether nonsecreted cytoplasmically localized GHBP would enhance GH-stimulated STAT5-mediated transcription, we removed the secretion sequence from the GHBP cDNA (XS-GHBP) (12). In the XS-GHBP construct, the rat GHBP was PCR-amplified without its signaling peptide, and an ATG was introduced in the primer just upstream of where the mature GHBP protein is coded. XS-GHBP is expressed throughout the cytoplasm of the cell as observed by confocal laser-scanning microscopy and is not secreted to the extracellular space (Fig. 2). Transient transfection of XS-GHBP cDNA also increased the STAT5-mediated tran-scriptional response to GH. Transfection of increasing amounts of XS-GHBP cDNA further enhanced GH-stimulated STAT5mediated transcription. As observed with WT-GHBP, transient transfection of XS-GHBP cDNA exerted no significant effect on STAT5-mediated transcription in the absence of concomitant transfection of GH receptor cDNA (Fig. 4). Thus, the intracellular GHBP enhances GH-stimulated STAT5-mediated transcription independent of its secretion to the extracellular space.
Ligand-dependent Nuclear Translocation of the GHBP-It has been previously reported that the GHBP is located intracellularly in both the cytoplasm and the nucleus in both an insoluble form attached to membranes and a soluble form in the cytoplasm or nucleoplasm. To determine whether the GHBP was subject to ligand-dependent nuclear translocation, we subcloned the WT-GHBP into a N-terminal enhanced fluorescent protein vector. The WT-GHBP was therefore expressed as a fusion protein to the N terminus of the enhanced GFP. The integrity of the reading frame was confirmed by sequence analysis, and protein expression was examined by Western blot analysis and confocal laser-scanning microscopy (Fig. 5A). The WT-GHBP-GFP was expressed as a protein with a molecular mass of 67 kDa in contrast to the native GFP, which was expressed as a protein of 27 kDa. Examination of the cellular distribution of the WT-GHBP-GFP demonstrated distinct perinuclear localization in contrast to the native GFP, which exhibited a diffuse cytoplasmic distribution. We transfected GHBP-GFP in BRL-GHR 1-638 cells to demonstrate the effect of exogenous GH stimulation on the cellular distribution of GHBP. Stimulation of cells with 50 nM hGH first resulted in a contraction of the perinuclear distribution of the WT-GHBP-GFP and subsequent nuclear translocation, which intensified to 30 min after stimulation with GH. In contrast, the native GFP was not translocated to the nucleus upon cellular stimulation with GH. Cellular stimulation with fetal bovine serum also resulted in nuclear translocation of the WT-GHBP-GFP (Fig. 5) but not the native GFP. Thus, the GHBP is subject to ligand-dependent nuclear translocation indicative of an intracellular/intranuclear function. We therefore focused our attention on the nuclear GHBP. Effect of Nuclear Localized GHBP on GH-stimulated STAT5mediated Transcription-To examine the function of the nuclear localized GHBP, we introduced the NLS of the SV40 large T antigen at the N terminus of the GHBP (NLS-GHBP). The NLS-GHBP was constructed by replacement of the secretion sequence of the WT-GHBP with the nuclear localization signal from the SV40 large T antigen (PKKKRKV) (29). The integrity of the reading frame for the GHBP modification was confirmed by sequence analysis, and translation and location of the protein product was determined by confocal laser-scanning microscopy (Fig. 2). The NLS-GHBP was predominantly localized to the nucleus and was not secreted from the cell and therefore could be utilized to study the functional effect of nuclear localized GHBP on GH-stimulated STAT5-mediated transcription. Transient transfection of BRL cells with both GH receptor cDNA and NLS-GHBP cDNA resulted in a dramatic enhancement of GH-stimulated STAT5-mediated transcription (Fig.  6A). Transient transfection of BRL-GHR 1-638 cells with NLS-GHBP cDNA also resulted in dramatic enhancement of GHstimulated STAT5-mediated transcription (data not shown).
The transcriptional enhancing effect of NLS-GHBP on GHstimulated STAT5-mediated transcription was increased with the transfection of increasing amounts of NLS-GHBP cDNA (Fig. 6A). No STAT5-mediated transcriptional response to GH was obtained upon transfection of NLS-GHBP cDNA without concomitant transfection of GH receptor cDNA (Fig. 6A). The ability of NLS-GHBP to enhance GH-stimulated STAT5-mediated transcription was observed over a wide concentration range of both the homologous rat GH (Fig. 6B) and human GH (Fig. 6C). The ability of the nuclear localized GHBP to enhance GH-stimulated transcription was not observed when a STAT1/ 3-responsive reporter plasmid (c-fos-SIE-CAT) (35) was utilized instead of the STAT5 reporter (SPI-GLE1-CAT) (data not shown). Thus, the transcriptional enhancing effect of the nuclear localized GHBP has some specificity for GH-stimulated STAT5-mediated transcription. Similarly transient transfection of NLS-GHBP cDNA does not alter the activity of reporter plasmids that constitutively express either CAT (pCMV-CAT) or luciferase (pSV2-LUC).
Regions in the Intracellular Domain of the GH Receptor Required for NLS-GHBP to Enhance GH-stimulated STAT5mediated Transcription-As described above, no STAT5mediated transcriptional response to GH was obtained upon transfection of NLS-GHBP cDNA without concomitant transfection of GH receptor cDNA. We therefore next examined the regions in the intracellular domain of the GH receptor required for NLS-GHBP to enhance GH-stimulated STAT5-mediated transcription. NLS-GHBP cDNA was co-transfected into BRL cells with the cDNA encoding various receptor mutations or deletions as indicated in Fig. 7. No GH stimulation of STAT5mediated transcription was observed with GH receptor mutations that lacked the proline-rich box 1 region of the GH receptor (GHR 1-294 , GHR 1-638 ⌬297-311, and GHR 1-638 P300A, P301A,P303A,P305A) required for association and activation of JAK2 (30,36)  scription. Thus, the ability of the nuclear localized GHBP to enhance GH-stimulated STAT5-mediated transcription requires the activation of JAK2 and residues of the intracellular domain of the GH receptor located between amino acids 541 and 638 that are required for binding and full activation of STAT5 (38,39).

NLS-GHBP Increases the Rate of GH-stimulated STAT5mediated Transcription-
To determine the potential molecular mechanisms by which nuclear localized GHBP enhanced GHstimulated STAT5-mediated transcription, we first examined the effect of the NLS-GHBP on the rate of GH-stimulated STAT5-mediated transcription. As is observed in Fig. 8, the rate of GH-stimulated STAT5-mediated transcription was dramatically increased in the presence of NLS-GHBP. The differential rate of GH-stimulated STAT5-mediated transcription was limited to the first 6 h after cellular stimulation with GH. Thus, nuclear localized GHBP enhances the rate of GH-stimulated STAT5-mediated transcription.
Effect of NLS-GHBP on Tyrosine Phosphorylation and DNA Binding Activity of STAT5-We subsequently examined whether the GHBP-enhanced rate of GH-stimulated STAT5mediated transcription was due to alteration in the phosphorylation state of STAT5. Tyrosine phosphorylation of STAT molecules is requisite for their dimerization, nuclear translocation, and DNA binding (40). We examined both the nuclear translocation of STAT5 and the appearance of tyrosine-phosphorylated STAT5 in the nucleus. Cellular stimulation with GH resulted in the nuclear translocation of STAT5 concordant with the concomitant appearance of tyrosine-phosphorylated STAT5 in the nucleus. The presence of NLS-GHBP did not alter the ability of GH to stimulate either the tyrosine phosphorylation or nuclear translocation of STAT5 (Fig. 9A). Furthermore, the presence of the NLS-GHBP did not alter the rate of removal of tyrosine-phosphorylated STAT5 from the nucleus as observed by the equal reduced amount of tyrosine-phosphorylated STAT5 180 min after stimulation with GH (Fig. 9A) nor after a prolonged period to 8 h (data not shown). We next examined the effect of nuclear localized GHBP on the ability of GH to stimulate binding of STAT5 to its DNA response element. Electrophoretic mobility shift assay with use of the GASlike element from the SPI 2.1 gene promoter used in reporter assays demonstrated two distinct binding species in response to cellular stimulation with GH. These two SPI-GLE1 binding species have previously been identified as STAT5 (slower migrating) and STAT1 (faster migrating) (41) in BRL-GHR 1-638 cells stimulated with GH. By supershift analysis, we have also demonstrated that the slower migrating DNA binding species does indeed contain STAT5 (Fig. 9B). GH-stimulated STAT5 DNA binding activity was evident 5 min after GH stimulation and declined thereafter to 3 h when DNA binding of STAT5 was minimal (Fig. 9B). NLS-GHBP did not alter the ability of GH activated STAT5 to bind to its DNA response element (Fig. 9B). The GHBP was not present in the STAT5-containing DNA binding complex as indicated by failure of supershift of the complex with monoclonal antibody to the GHBP (data not shown). Thus, NLS-GHBP does not enhance GH-stimulated STAT5-mediated transcription by alteration of STAT5 tyrosine phosphorylation and subsequent DNA binding.
A Truncated Version of NLS-GHBP Also Enhances GH-stimulated STAT5-mediated Transcription-To determine whether GH was required to bind the GHBP to observe the NLS-GHBP enhancement of GH-stimulated STAT5-mediated transcription, we simply truncated the NLS-GHBP at amino acid number 115. We introduced a stop codon at the amino acid position 115 in the NLS-GHBP construct, thereby destroying the ability of the GHBP mutant to bind GH (42). Truncation of the GHBP would also result in loss of the 17-amino acid hydrophilic tail, which contains the epitope for the monoclonal antibody (mAb 4.3) used so far to examine the expression of the transfected GHBP constructs. We therefore introduced an epitope tag (hemagglutinin) at the amino terminus of either the wild type GHBP, the nuclear localized GHBP, or the truncated nuclear localized GHBP (NLS-GHBP  ). The integrity of the reading frame was determined by sequence analysis. Transient transfection and Western blot analyses of these constructs demonstrated expression of the WT-GHBP, NLS-GHBP, and NLS-GHBP 1-115 at the appropriate molecular weights (Fig. 10A). Subsequent analysis of STAT5-mediated transcription demonstrated that NLS-GHBP 1-115 was able to enhance GH-stimulated STAT5-mediated transcription to a similar extent as the full-length NLS-GHBP (Fig. 10B). Thus, the nuclear localized GHBP is able to function independently of GHBP-bound ligand to enhance GH-stimulated STAT5-mediated transcription.

NLS-GHBP Enhances STAT5-mediated Transcription Stimulated by Other Members of the Cytokine Receptor Superfam-
ily-Since the GHBP could enhance STAT5-mediated transcription without bound ligand, we examined whether NLS-GHBP could function as a transcriptional enhancer for other cytokine receptor superfamily members that also utilize STAT5 for transcriptional activation (43). We therefore transiently FIG. 5. Expression of GHBP-GFP in BRL-GHR 1-638 cells and its cellular distribution after stimulation with exogenous GH and fetal bovine serum. Cells were transiently transfected with GHBP-GFP and serum-starved for 12 h before they were treated as indicated with either 50 nM hGH or 10% fetal bovine serum. A, Western blot analysis of cell extract from BRL-GHR 1-638 cells transiently transfected with GFP-vector or GHBP-GFP. B, localization of the green fluorescent protein or GHBP-GFP expressed in BRL-GHR 1-638 cells in minutes after stimulation with GH or fetal bovine serum. Cells were fixed using an ethane-freezing/methanol fixation, and the GFP signal was further enhanced by using an anti-GFP antibody. transfected either the PRL or EPO receptors into BRL cells (27) and determined the STAT5-mediated transcriptional response in the presence of NLS-GHBP (Fig. 11). Human GH is also a ligand for the PRL receptor (44), and therefore an activation of the PRL receptor and a STAT5-mediated transcriptional response to hGH via the PRL receptor can be expected (27). STAT5-mediated transcription, stimulated specifically through the PRL receptor either with hGH or with ovine PRL, was also enhanced in the presence of NLS-GHBP to a similar extent as the enhancement of STAT5-mediated transcription by NLS-GHBP through the GH receptor. Enhancement of STAT5-mediated transcription by NLS-GHBP was also observed upon activation of the EPO receptor with erythropoietin, whereas GH stimulation of cells transfected with the EPO receptor failed to activate STAT5 with or without the presence of the NLS-GHBP (Fig. 11). An expression plasmid encoding for nuclear localized hGH (NLS-hGH) also did not result in enhancement of EPO-stimulated STAT5-mediated transcription, indicating that the NLS-GHBP-enhanced STAT5-mediated transcription is not due to any effect of the NLS of the SV40 large T antigen (data not shown). Thus, the nuclear localized GHBP functions as an enhancer of STAT5-mediated transcription for members of the cytokine receptor superfamily. DISCUSSION We have described here a new functional and ligand-independent role for the soluble extracellular domain of the growth hormone receptor otherwise known as the GHBP. Exogenously applied GHBP behaves as expected (16) and inhibits the cellular response to GH in vitro. In contrast, endogenously produced GHBP functions as an enhancer of cytokine receptor-stimulated STAT5-mediated transcription. Such enhancement of cytokine receptor-stimulated STAT5-mediated transcription is mediated predominantly in the nucleus and does not require The intracellular (19 -21) and nuclear localization of the GHBP (19,21) have been reported previously. The localization of the GHBP to the nucleus has been observed both in vivo (19,21) and in vitro in experimental systems (21). The localization of the GHBP to the nucleus was heterogenous (19), suggesting that the nuclear localization of the GHBP was dynamic rather than constitutive. We have now demonstrated here that a GHBP-GFP fusion protein translocates to the nucleus upon cellular stimulation with GH or serum. Thus, the GHBP, in addition to secretion to the extracellular space, is also specifically localized to the nucleus. This localization is not constitutive but requires exposure of the cells to a stimulus and is therefore presumably an active process. We had also previously reported that both GH (23) and the GH receptor (21) are subject to a rapid and transient nuclear translocation. At least one function of the nuclear translocation of the growth hormone and its receptor appears to be the phosphorylation of nuclear localized JAK2 (24,25) (45). Interestingly, however, the nuclear translocation of both GH and the GH receptor are independent of JAK2, 2 suggesting that nuclear translocation of the ligand-receptor/binding protein complexes are distinct from classical JAK-STAT pathways. Whether the GHBP requires ligand for nuclear translocation remains to be determined, as does the mechanism of the nuclear translocation. The characterization of the GHBP-GFP reported here should greatly facilitate delineation of the mechanism of secretion of the GHBP and also translocation of the GHBP to the nucleus.
We have demonstrated here that nuclear localized GHBP functions as an enhancer of STAT5-mediated transcription not only for GH but also for other members of the cytokine receptor superfamily, which utilize STAT5 for transcriptional responses. STAT5 has been demonstrated to be utilized by a FIG. 9. Effect of transient transfection of NLS-GHBP cDNA on the tyrosine phosphorylation and nuclear translocation of STAT5 (A) and on STAT5 DNA binding activity after GH stimulation (B). BRL-GHR 1-638 cells were cultured to confluence and transiently transfected with NLS-GHBP or the control vector and treated with 50 nM GH for the indicated times. A, nuclear extracts were prepared and analyzed for phosphorylated STAT5 and total STAT5 by Western blotting. B, nuclear extracts were prepared, and DNA binding activity to the SPI-GLE1 probe was analyzed by electrophoretic mobility shift assay. variety of hormones and cytokines including interleukin-2, -3, -5, -7, and -15, erythropoietin, granulocyte-macrophage colony stimulating factor, thrombopoietin, and epidermal growth factor (43,46,47). Thus, there must exist a complex interaction between different cytokines at the cellular level for regulation of STAT5-mediated transcription. Ligands that do not bind to the GHBP, such as EPO, would not be subject to the extracellular binding and subsequent inhibition by GHBP as would GH. It is plausible, however, that EPO or other factors may modulate the production and/or secretion of the GHBP such that the response of the cell to GH and subsequent STAT5mediated transcription is altered. The EPO receptor and the GH receptor and their respective ligands, in addition to the endocrine distribution of the ligands, are widely co-expressed in various tissues such as placenta, mammary gland, the central nervous system, and smooth muscle (48 -52). Thus, the final hormonal response of the cell would depend on a complex interplay of the ratio of extracellular to intracellular (nuclear) GHBP and the identity of the stimulating ligand. Thus, physiological factors promoting the nuclear localization of the GHBP would enhance the otherwise limited transcriptional response of the cell to various ligands. Other physiological factors, which could up-regulate GHBP production and secretion, would presumably increase the STAT5-mediated transcriptional responses to other ligands such as EPO. Many analogous soluble isoforms of various cytokine and growth factor receptors have also been reported (4 -7). Presumably, the complexity of the cellular STAT5-mediated transcriptional response would be further increased if other soluble cytokine receptors/binding proteins (such as the PRL-binding protein (53)) function as transcriptional enhancers similar to GHBP. This regulatory strategy may also be one mechanism by which the cell can filter multiple redundant signals initiated by cytokine molecules sharing the same signal transduction pathway (9,47,54). Such regulatory mechanisms would presumably play an important role during physiological states such as puberty, pregnancy, and lactation.
The molecular mechanism by which the NLS-GHBP enhances cytokine receptor-stimulated STAT5-mediated transcription remains to be determined. We have observed that the nuclear localized GHBP does not alter GH-stimulated tyrosine phosphorylation, nuclear translocation, or DNA binding of STAT5. We are also unable to detect an association between the GHBP and STAT5. 3 Tyrosine phosphorylation, nuclear translocation, and even DNA binding of STAT5 is not sufficient for STAT5 to induce transcriptional activity (55), 3 suggesting that additional factors are involved in the activation of STAT5. Indeed, multiple co-activators and repressors that interact with STAT5 have been identified (56 -58). One possibility for the observed GHBP-enhanced STAT5-mediated transcription could be that GHBP binds to a repressor of STAT5-mediated transcription, thereby preventing an inhibitory association between the repressor and STAT5, releasing STAT5 to function as an activator of transcription. A family of STAT transcriptional repressors has been identified (59). The protein inhibitor of activated STAT (PIAS) family consists of five members including PIAS1, PIAS3, PIASx␣, PIASx␤, and PIASy (59). Members of the PIAS family bind to activated STAT molecules and prevent STAT binding to DNA and subsequent transcription (59,60). Indeed, nuclear localized PRL has just been demonstrated to bind to PIAS3 independent of the PRL receptor, thereby allowing STAT5 to bind preferentially to its DNAresponsive element instead of the PIAS protein (61). Alternatively, the GHBP may participate in the formation of the transcriptional complex required for STAT5-mediated transcription. Several enhancers of STAT5-mediated transcription have been identified, including the p300-CBP complex (62), the Nmi (N-Myc-interacting protein) (63), and the glucocorticoid receptor (64). PRL has been demonstrated to stimulate the association between STAT5 and the histone acteyltransferase p300-CBP, enhancing STAT-mediated transcription by linkage with the transcriptional complex (62). Nmi has been demonstrated to enhance STAT5-mediated transcription by increased formation of STAT5-p300-CBP complexes (63). STAT5 cooperates with the glucocorticoid receptor for transcriptional activation without a need for the DNA or ligand binding domains of the nuclear receptor (65), supporting the idea that even minimal promoter sites are enough to attract complex transcriptosomes without DNA binding of all components. Another transcription factor interacting with STAT5, YY1, requires additional DNA recognition sites on its promoter region to cooperate with STAT5 in transcriptional regulation (56). The SPI-GLE1 sequence used here is alone sufficient for formation of the GH-induced STAT5 DNA binding complexes and does not contain other consensus sequences used by other transcription factors (28). Since we observed no interaction of the GHBP with STAT5, it is therefore not surprising that the GHBP was not contained in the DNA binding complex. Further studies to delineate proteins, which interact with the GHBP, should allow identification of the mechanism by which the GHBP enhances cytokine receptor-stimulated STAT5-mediated transcription.
In conclusion, we demonstrate here that nuclear localized GHBP functions as a potent enhancer of STAT5-mediated transcription, not only for GH but also for other members of the cytokine receptor superfamily. Thus, the GHBP exerts opposing effects on STAT5-mediated transcription depending on its extra-/intracellular location. The use of a soluble cytokine receptor as a location-dependent transcriptional enhancer and 3 R. Graichen and P. E. Lobie, unpublished observations. FIG. 11. Effect of transient transfection of NLS-GHBP cDNA on the STAT5-mediated transcriptional response to hGH and rGH, oPRL, and mEPO in BRL cells transiently transfected with the GH receptor, PRL receptor, or EPO receptor cDNA, respectively. BRL cells were cultured to confluence and transiently transfected with GH receptor, PRL receptor, or EPO receptor cDNAs, SPI-GLE1-CAT, and 5 g of NLS-GHBP cDNA. Cells were treated for 24 h with 50 nM hGH or rGH, 100 nM oPRL, or 10 units/ml mEPO, respectively. Results are presented as the mean Ϯ S.E. of triplicate determinations of the -fold stimulation above non-hormone-stimulated cells. the ligand-independent involvement of the extracellular domain of a polypeptide ligand receptor in intracellular signal transduction provide additional novel mechanisms of transcriptional control. What remains to be determined is the mechanism by which the nuclear localized GHBP functions as a transcriptional enhancer. Identification of proteins interacting with the GHBP should be useful in this regard, and such experiments are in progress.