STAT5b Down-regulates Peroxisome Proliferator-activated Receptor α Transcription by Inhibition of Ligand-independent Activation Function Region-1trans-Activation Domain*

Growth hormone-activated STAT5b inhibits by up to 80% the transcriptional activity of peroxisome proliferator-activated receptor (PPAR) α, a nuclear receptor activated by diverse environmental chemicals and hypolipidemic drugs classified as peroxisome proliferators. This inhibitory cross-talk between STAT5b and PPAR is now reported for PPAR forms γ and δ and for thyroid hormone receptor, indicating a more general potential for inhibitory cross-talk between JAK/STAT and nuclear receptor signaling pathways. Further investigations revealed that SOCS-3, a growth hormone-inducible negative regulator of cytokine signaling to STAT5b, abolished the STAT5b inhibitory response. A constitutively active STAT5b mutant failed to inhibit PPARα activity, indicating that STAT5b does not induce synthesis of a more proximal PPARα inhibitor. STAT5b inhibition was not reversed by overexpression of the heterodimerization partner of PPAR (retinoid X receptor) or the nuclear receptor coactivators P300 and SRC-1, suggesting that STAT5b does not inhibit PPARα by competing for these limiting cellular cofactors. STAT5b did not inhibit a chimeric receptor comprised of yeast GAL4 DNA-binding domain linked to the ligand binding/AF-2 trans-activation domain of PPARα, indicating that the COOH-terminal AF-2 domain of PPAR is not the target of STAT5b inhibition. Rather, STAT5b inhibited transcription driven by the NH2-terminal ligand-independent AF-1 trans-activation domain of PPARα in a GAL4-linked chimera by ∼80%. The conservation of this AF-1trans-activation function in many nuclear receptors suggests that AF-1 may serve as an important target for inhibitory cross-talk between STAT transcription factors and nuclear receptors in a variety of signaling pathways.

80% the transcriptional activity of peroxisome proliferator-activated receptor (PPAR) ␣, a nuclear receptor activated by diverse environmental chemicals and hypolipidemic drugs classified as peroxisome proliferators. This inhibitory cross-talk between STAT5b and PPAR is now reported for PPAR forms ␥ and ␦ and for thyroid hormone receptor, indicating a more general potential for inhibitory cross-talk between JAK/STAT and nuclear receptor signaling pathways. Further investigations revealed that SOCS-3, a growth hormone-inducible negative regulator of cytokine signaling to STAT5b, abolished the STAT5b inhibitory response. A constitutively active STAT5b mutant failed to inhibit PPAR␣ activity, indicating that STAT5b does not induce synthesis of a more proximal PPAR␣ inhibitor. STAT5b inhibition was not reversed by overexpression of the heterodimerization partner of PPAR (retinoid X receptor) or the nuclear receptor coactivators P300 and SRC-1, suggesting that STAT5b does not inhibit PPAR␣ by competing for these limiting cellular cofactors. STAT5b did not inhibit a chimeric receptor comprised of yeast GAL4 DNA-binding domain linked to the ligand binding/AF-2 trans-activation domain of PPAR␣, indicating that the COOHterminal AF-2 domain of PPAR is not the target of STAT5b inhibition. Rather, STAT5b inhibited transcription driven by the NH 2 -terminal ligand-independent AF-1 trans-activation domain of PPAR␣ in a GAL4linked chimera by ϳ80%. The conservation of this AF-1 trans-activation function in many nuclear receptors suggests that AF-1 may serve as an important target for inhibitory cross-talk between STAT transcription factors and nuclear receptors in a variety of signaling pathways.

STATs 1 and PPARs are distinct families of transcription
factors that mediate cellular responses to diverse endogenous and environmental stimuli. STATs are activated by multiple cytokines and peptide hormones via receptor-associated JAK tyrosine kinases by a process that involves tyrosine phosphorylation and dimerization of the SH2 domain-containing STAT proteins, followed by STAT translocation to the nucleus, binding to the specific DNA enhancer elements, and activation of target gene expression (1). Two closely related STAT forms, STAT5a and STAT5b, share striking amino acid sequence homology (ϳ90% identity) and can be activated by the same set of hormones and cytokines, which includes GH, prolactin, interleukins 2, 3, and 5, and erythropoietin (2). STAT5a and STAT5b both activate target gene expression via STAT5 response elements, albeit with some differences in binding specificity (3,4). Mouse gene knock out studies demonstrate, however, that STAT5a and STAT5b have distinct and largely nonredundant tissue-specific functions respectively related to mammary gland development and liver metabolic function (5)(6)(7).
PPARs are transcription factors that belong to the nuclear receptor superfamily, and, like other nuclear receptors, they contain conserved domains that function in DNA binding, ligand binding, transcriptional activation, and dimerization (reviewed in Refs. 8 and 9). PPARs form high affinity DNAbinding complexes with the heterodimerization partner RXR, which facilitates their binding to specific PPAR response elements upstream of target genes. Mammalian PPARs include three subtypes (␣, ␥, and ␦), which are characterized by unique functions, ligand specificities, and tissue distributions (10). A wide variety of structurally distinct PPAR ligands have been identified. These include certain environmental chemicals, fibrate hypolipidemic drugs, fatty acids, and eicosanoids for PPAR␣; anti-diabetic thiazolidinediones and 15-deoxy-⌬ 12,14 prostaglandin J 2 for PPAR␥; and the fatty acid-like compound L631033 for PPAR␦ (11,12). This diverse spectrum of PPAR ligands suggests that these receptors mediate a wide variety of biological functions. Although the function of PPAR␦ is unknown, PPAR␣ and PPAR␥ regulate diverse biological processes, including lipid metabolism, cell differentiation, carcinogenesis, and apoptosis (13)(14)(15).
Extensive investigations have advanced our understanding of JAK/STAT and PPAR signaling pathways. However, the extent to which these two pathways interact remains unclear. The observation that hypophysectomy enhances hepatic peroxisome proliferation in female rats (16) suggests that the pituitary secretes an endocrine factor that inhibits this PPAR␣-dependent (17) liver response. GH and prolactin are two major pituitary hormones that directly activate JAK-STAT signaling pathways (18). GH treatment suppresses expression in rat liver in vivo of certain peroxisome proliferator-inducible liver P450 genes (19), and GH inhibits peroxisomal ␤-oxidation induced by the PPAR␣ activator clofibrate in rat primary hepatocyte culture (20,21). We previously demonstrated that activation of STAT5b signaling by either GH or prolactin inhibits, by up to 80%, PPAR␣-dependent reporter gene transcription stimulated by both endogenous and foreign chemical peroxisome proliferators (22). This inhibition required both GH receptor (GHR) and STAT5b and was not seen with STAT1, a second GH-activated STAT. STAT5b-PPAR␣ complexes could not be detected by anti-STAT5b supershift analysis of PPAR␣-DNA complexes, suggesting that the mechanism that underlies this inhibitory cross-talk between STAT5b and PPAR␣ is distinct from the cross-talk between STAT5 and glucocorticoid receptor, where protein-protein complexes form (23,24).
In the present study, we investigate the mechanism that underlies the inhibitory cross-talk between JAK/STAT and PPAR signaling pathways. We report that GH-activated STAT5b inhibits the transcriptional activity of PPAR␥, PPAR␦, and T3R in addition to PPAR␣ and that the GH-dependent inhibition of PPAR␣-stimulated gene transcription is blocked by co-expression of SOCS3, a GH-inducible inhibitor (25) that belongs to a recently described family of cytokine signaling inhibitors (26 -28). We evaluate the requirement for STAT5binduced gene transcription, the role of transcriptional coactivators, and the role of the DNA-binding and ligand-binding and trans-activation domains of PPAR␣ for STAT5b inhibition. Our findings suggest that STAT5b inhibition of PPAR transcriptional activity proceeds via a novel mechanism that involves the NH 2 -terminal AF-1 trans-activation domain of the nuclear receptor, which also serves as the target for insulin stimulation of PPAR␣ activity (29), albeit by a distinct mechanism.
Cell Culture and Transfection-COS-1 cells were grown in Dulbecco's modified Eagles medium with 10% fetal calf serum. Transfection of COS-1 cells grown in 48-well tissue culture plates was carried out using FuGENE 6 transfection reagent (Roche Molecular Biochemicals; catalog number 1814443). FuGENE 6 (0.5 l) was used for each well in 48-well plates. 24 h after addition of the DNA-FuGENE 6 mixture to the cells, the medium was changed to nonserum Dulbecco's modified Eagles medium. PPAR activators (e.g. Wy-14,643 at 1-20 M) were then added to the culture medium in combination with GH at concentrations specified in the figure legends. Cell were lysed 24 h later, and firefly luciferase and Renilla luciferase activity were measured using a dual luciferase assay kit (Promega, Madison, WI; catalog number E1910). EMSA and Western Blot Analysis-EMSA analysis using probes for STAT5 and PPAR DNA binding activity was performed as described previously (22). For Western blotting, whole cell lysates from transfected COS-1 cells were subjected to 10% SDS/polyacrylamide gel electrophoresis and then transferred to nitrocellulose membranes. Western blotting was performed using STAT5b-specific antibody (Santa Cruz Biotechnology, sc-835), as described elsewhere (22).

STAT5b Inhibits PPAR␥, PPAR␦, and T3R Transcriptional
Activity-GH-activated STAT5b inhibits PPAR␣-dependent gene transcription by ϳ80% when evaluated in a reporter gene trans-activation assay (22). To determine whether this inhibition is specific to PPAR␣, we examined the other two mammalian PPAR forms, PPAR␥ and PPAR␦, as well as T3R, all of which belong to family NR1 of the nuclear receptor superfamily (40). GH signaling and PPAR or T3R transcriptional activation pathways were reconstituted in COS-1 cells by co-transfection of key components: GHR, JAK2 kinase, and STAT5b for GH signal transduction; PPAR␥ or PPAR␦ and the reporter plasmid pHD(X3)Luc for PPAR signaling; and T3R and the reporter plasmid TRE-CAT for thyroid hormone signaling. The transfected COS-1 cells were stimulated for 24 h with the nuclear receptor form-specific ligands troglitazone (1 M), L631033 (5 M) or T3 (74 nM) to activate PPAR␥, PPAR␦, and T3R, respectively, either in the presence or the absence of GH. Fig. 1 shows that activation of PPAR␥, PPAR␦, and T3R transcriptional activity by the respective ligand of each receptor is substantially decreased in the presence of GH. The extent of STAT5b inhibition of these three nuclear receptors is similar to that seen previously for PPAR␣ (80% inhibition for PPAR␣ (22) compared with 89, 60, and 73% inhibition for PPAR␥, PPAR␦, and T3R, respectively). This finding suggests that these nuclear receptors share a common target for STAT5b inhibition.
STAT5b Does Not Inhibit Ligand-dependent AF-2 Transcriptional Activation Domain of PPAR␣-PPAR and other nuclear receptors contain a central DNA-binding domain and a COOHterminal ligand-binding domain that displays ligand-dependent trans-activation activity, termed AF-2. The hydrophobicity of this region is conserved in all nuclear receptors, and mutational analysis suggests that AF-2 contains an interaction surface for coactivators, such as SRC-1 (41). To determine whether STAT5b exerts its inhibitory effect on nuclear receptor activity by inhibiting the COOH-terminal AF-2 region, we evaluated the effect of STAT5b on the chimeric receptor GAL4-PPAR␣ (167-468), where the COOH-terminal region of mouse PPAR␣ (amino acids 167-468) is fused to the DNA-binding domain of the yeast transcription factor GAL4 (33). Activation of this chimeric receptor by peroxisome proliferators was monitored in transfected COS-1 cells using the reporter gene (UAS) 5 -tk-CAT, which contains five copies of the DNA response element of GAL4 upstream of the thymidine kinase promoter driving a CAT reporter gene. Stimulation of the cells with the PPAR␣ activator and potent peroxisome proliferator Wy-14,643 strongly activated the GAL4-PPAR␣ chimera; however, GHactivated STAT5b failed to inhibit receptor activity (Fig. 2). Thus, the presence of the COOH-terminal AF-2 trans-activa-tion/ligand-binding domain of PPAR␣ is not sufficient for STAT5b to inhibit receptor activity, suggesting that STAT5b does not directly target this region of PPAR␣.
STAT5b Inhibits the NH 2 -terminal Ligand-independent AF-1 trans-Activation Domain of PPAR␣ via a MAP Kinaseindependent Mechanism-A ligand-independent trans-activation domain, designated AF-1, has recently been identified within the NH 2 -terminal 92 amino acids (A/B region) of PPAR␣ (29). We therefore investigated whether GH-activated STAT5b can inhibit the AF-1 function of PPAR␣, which was assayed using a fusion construct linking the GAL4 DNA-binding domain to the first 92 amino acids of hPPAR␣, designated GAL4-PPAR␣ (1-92). Fig. 3A shows that the AF-1 region of hPPAR␣ exhibits strong ligand-independent transcriptional activity (vehicle control versus GAL4 DNA-binding domain alone; compare the first two bars), in agreement with recent studies by Meier and co-workers (29). Moreover, GH activation of STAT5b inhibited this trans-activation activity in a dose-dependent manner. The extent of inhibition, ϳ80%, is consistent with the extent of STAT5b inhibition of full-length PPAR␣ that we reported previously (22).
Serine residues 12 and 21 within the AF-1 region of PPAR␣ have been identified as targets of MAP kinase, and phosphorylation of both of these serines is required for insulin stimulation of PPAR␣ activity (29). To examine whether these serines participate in STAT5b inhibition of the AF-1 activity of PPAR␣, we tested the corresponding serine to alanine mutants of the GAL4-PPAR␣ (1-92) chimera. Fig. 3 (B and C) show that mutation of these serine residues does not prevent STAT5b inhibition, indicating that the GH-dependent inhibition of the AF-1 activity of PPAR␣ does not involve changes in the phosphorylation of these residues.
Coactivators P300 and SRC-1 Do Not Reverse STAT5b Inhibition of PPAR␣ Activity-The observation that GH-activated STAT5b inhibits activation of all three PPAR isoforms as well as T3R suggests that STAT5b may compete with a common regulatory factor of the PPARs, perhaps a nuclear receptor coactivator (42). This possibility is further supported by our identification of PPAR␣'s AF-1 region, which in the case of PPAR␥ interacts with coactivators (43,44), as a target of STAT5b. P300 and SRC-1 are widely expressed nuclear receptor coactivators that interact directly with PPAR␣ (45). P300 can interact with both the AF-1 and AF-2 regions of PPAR (44) and can also coactivate STAT5 (46). To investigate whether these coactivators become limiting for PPAR␣ transcriptional activity in cells that simultaneously signal via STAT5b, we examined the effects of P300 and SRC overexpression on the STAT5b-dependent inhibition of PPAR␣ activity. Fig. 4A shows that cotransfection of P300 or SRC-1 substantially enhances both basal and Wy-14,643-inducible PPAR␣ activity. This confirms that P300 and SRC-1 can both coactivate PPAR␣ and indicates that their endogenous expression levels in transfected COS-1 cells are indeed limiting. Fig. 4B shows, however, that overexpression of these coactivators does not block the GH-induced inhibition of PPAR␣ activity. Therefore, STAT5b does not inhibit PPAR␣ by competing for the limiting endogenous P300 or SRC-1. Of note, STAT5b expression alone, in the absence of GH treatment, reversed the stimulation of PPAR␣ activity by co-expressed P300 or SRC-1; whereas P300 and SRC-1 alone enhanced Wy-14,643-stimulated PPAR␣ activity by 5-and 8-fold, respectively (Fig. 4A), these enhancing effects were abolished in cells cotransfected with GHR, JAK2, and STAT5b but not treated with GH (cf. data points on y axis of Fig. 4B). This result is consistent with the report that P300 directly interacts with the trans-activation domain of STAT5 (46) and may be explained by STAT5b competing with PPAR␣ for the transfected P300 and perhaps also SRC-1.
Effects of STAT5b on PPAR␣ DNA Binding Activity-We next investigated whether PPAR␣ DNA binding activity is affected by STAT5b activation. Fig. 5A shows that Wy-14,643 stimulation of PPAR␣ leads to a 3.3-fold increase in PPAR␣ DNA binding activity, assayed by EMSA using a PPRE probe derived from the rabbit CYP4A6 gene (lane 2 versus lane 1). In the absence of GHR, JAK2, and STAT5b, GH had no effect on Wy-14,643-induced PPAR␣ DNA-binding (Fig. 5B, lanes 3 and  4 versus lane 2). In contrast, under conditions where GHR, JAK2, and STAT5b are cotransfected with PPAR␣, GH treatment led to a small (ϳ25%) decrease in the abundance of the PPAR␣-DNA binding complex (Fig. 5A, lanes 3 and 4 versus  lane 2). Because the DNA binding activity of PPAR␣ is obligatorily dependent on the presence of its heterodimerization part- ner RXR␣ (47), we investigated whether GH activation of STAT5b might decrease PPAR␣ DNA binding activity by interfering with PPAR/RXR heterodimerization. To test this possibility, the cells were co-transfected with RXR to augment the limited amount of endogenous RXR in COS-1 cells. RXR overexpression significantly increased basal PPAR␣ DNA binding activity, which, as expected (47), is no longer further increased by Wy-14,643 treatment (Fig. 5, A, lanes 6 -8, and B, lanes 5-7). Nevertheless, GH still decreased PPAR DNA binding activity by ϳ30% (Fig. 5A, lane 8 versus lane 7), despite the co-expression of RXR. This suppression varied from 24 to 50% in different experiments and was not seen in the absence of GHR, JAK2 and STAT5b (Fig. 5B, lanes 7 versus lane 6). Overexpression of RXR␣ significantly increased both basal and Wy-14,643-induced luciferase reporter gene activity; however, it did not reverse STAT5b inhibition of PPAR␣ activity (Fig. 5C). We conclude that STAT5b does not inhibit PPAR␣ by decreasing cellular RXR levels or the availability of RXR for PPAR␣ heterodimerization.
Constitutively Active STAT5 Mutants Do Not Inhibit PPAR␣-STAT5a and STAT5b site-specific mutants with amino acid substitutions H299R and S711F are constitutively phosphorylated on tyrosine and undergo nuclear translocation and transcriptional activation in the absence of hormone or cytokine stimulation (38). These constitutively active STAT5 mutants, designated STAT5a1*6 and STAT5b1*6, were used to investigate whether STAT5b transcriptional activity is sufficient for PPAR␣ inhibition. STAT5a1*6 and STAT5b1*6 were confirmed to be constitutively active when expressed in COS-1 cells, as shown by their strong activation of a luciferase reporter gene containing four copies of a STAT5 response element from the ntcp gene (39) in the absence of GHR expression and GH stimulation (37-fold activation by STAT5b1*6 and 7-fold by STAT5a1*6) (Fig. 6A). Cotransfection of GHR with STAT5b1*6 followed by GH stimulation did not result in further activation of the reporter plasmid (data not shown), suggesting that STAT5b1*6 is already fully active in COS-1 cells without GH stimulation. We next examined whether these constitutively active STAT5 mutants inhibit PPAR␣ activity. Fig. 6B shows that neither STAT5a1*6 nor STAT5b1*6 blocked PPAR␣ activity stimulated by Wy-14,643. STAT5b1*6 also failed to inhibit PPAR␣ activity in GHR-and JAK2-transfected cells stimulated with GH (data not shown). Because these constitutively active STAT5 mutants failed to affect the inhibition of PPAR␣ activity observed with wild-type STAT5b, we conclude that STAT5b does not inhibit PPAR␣ transcriptional activity by stimulating transcription of a more proximal inhibitory factor.

SOCS-3 Blocks GH-stimulated STAT5b
Inhibition of PPAR␣-SOCS/CIS proteins constitute a novel family of cytokine-inducible suppressors that negatively regulate the JAK/ STAT signaling pathways (26 -28). One of these proteins, SOCS-3, may, in part, inhibit GH signaling by binding to the tyrosine phosphorylated cytoplasmic domain of GHR (48). To investigate whether the STAT5b-dependent inhibition of PPAR␣ is subjected to SOCS protein negative regulation, we first examined whether SOCS protein expression blocks GHstimulated STAT5b activation. COS-1 cells were transiently transfected with GHR, STAT5b, and SOCS expression plasmids. 24 h later, the cells were stimulated with GH for 30 min, and cell lysates were assayed for STAT5b DNA binding activity by EMSA. Fig. 7A shows that SOCS-3 fully blocks GH activation of STAT5b EMSA activity (lanes 9 and 10 versus lanes 3 luciferase activity 24 h later. Activities are expressed as firefly luciferase normalized by the reporter activity of a Renilla luciferase internal standard.  1-4) and without (lanes 6 -8) RXR␣ coexpression. COS-1 cells were transfected with expression plasmids as indicated. 24 h after transfection, cells were treated with Wy-14,643 (10 M) and GH (100 or 500 ng/ml) for 24 h. Whole cell lysates (10 g) were assayed for EMSA activity using a CYP4A6 PPRE probe. Lanes 5 and 9, untransfected Cos-1 cell control. The experiment shown in lanes 1-4 of A and B used three times more PPAR␣ expression plasmid/well than in the standard methods (i.e. 15 versus 5 ng PPAR␣/well of a 48-well plate). This was required to provide sufficient PPAR␣ DNA binding activity to visualize the EMSA complexes, which are weak in the absence of cotransfected RXR␣ (cf. left-hand panels versus right-hand panels of A and B). C, overexpression of RXR␣ does not reverse STAT5b inhibition of PPAR␣ transcriptional activity. COS-1 cells were transfected with pHD(X3)Luc reporter plasmid and expression plasmids encoding GHR, JAK2, STAT5b, and PPAR in the presence or absence of RXR␣ expression plasmid. 24 h after transfection, cells were treated with Wy-14,643 (20 M) and GH (25 ng/ml). Cell lysates were prepared and assayed for and 4). Partial or no inhibition was observed with three other SOCS/CIS proteins (CIS, SOCS-2, and SOCS-6; Fig. 7A). Western blot analysis (Fig. 7B) confirmed that STAT5b protein expression levels were unaffected by SOCS transfection, although in the case of SOCS-3 a slight increase in STAT5b protein mobility, consistent with the inhibition of STAT tyrosine phosphorylation, could be discerned (lane 6). To investigate the impact of SOCS-3 on PPAR␣ signaling, GHR/JAK2/ STAT5b and PPAR␣ signaling were reconstituted in COS-1 cells in the absence or presence of SOCS-3. As seen in Fig. 7C, the inhibition of PPAR␣ transcriptional activity by STAT5b was fully blocked by SOCS-3. These findings confirm that STAT5b must be activated to its DNA-binding form (Fig. 7A) to exert its inhibitory effect on PPAR␣ and suggest that STAT5b inhibition of PPAR␣ activity is subjected to negative regulation by SOCS-3. DISCUSSION GH-activated STAT5b inhibits the transcriptional activation of PPAR␣ by structurally diverse peroxisome proliferators (22), suggesting a mechanism whereby hormones and cytokines that activate STAT5b (2) may inhibit the pleiotropic responses of liver and perhaps other tissues to the numerous environmental chemicals, hypolipidemic agents, and other structurally di-verse chemicals that are classified as peroxisome proliferators (13). The present study reports that GH also exerts inhibitory effect on PPAR␥, PPAR␦, and T3R transcriptional activity. STAT5b transcriptional activity alone (in the form of the constitutively active STAT5b1*6) was found to be insufficient for inhibition of PPAR␣, excluding the possibility that STAT5b acts by stimulating transcription of a more proximal PPAR␣ inhibitor. Chimeric GAL4-PPAR␣ chimeras were utilized to FIG. 6. Effect of constitutively active STAT5 mutants on Wy-14,643 activation of PPAR␣. A, constitutively active STAT5a1*6 and STAT5b1*6 induce transcription from the ntcp promoter in the absence of GH stimulation. COS-1 cells were transfected with 4X-pT109-Luc reporter plasmid (39) and expression plasmids encoding STAT5a1*6 or STAT5b1*6. pCI empty vector was used as a DNA control. 24 h after transfection, cells were lysed and assayed for firefly luciferase activity normalized to a Rellina luciferase control. B, constitutively active STAT5 mutants do not inhibit PPAR␣ activation by Wy-14,643. COS-1 cells were transfected with pHD(X3)Luc reporter plasmid and expression plasmids encoding PPAR␣, STAT5a1*6 or STAT5b1*6. Transfected cells were treated with vehicle or Wy-14,643 (10 M) for 24 h. Transcriptional activity is expressed as relative firefly luciferase activity normalized to Renilla luciferase control. Values shown are the means Ϯ S.E. (n ϭ 3).

FIG. 7. SOCS-3 abolishes the GH inhibition of PPAR␣ by blocking GH activation of STAT5b.
A, expression of SOCS-3 blocks GH activation of STAT5b DNA binding activity in transfected COS-1 cells. COS-1 cells were transfected with expression plasmids encoding GHR, STAT5b and each of the indicated SOCS/CIS proteins. 24 h later, cells were treated with GH (200 ng/ml) for 30 min followed by preparation of whole cell extracts. EMSA assays for STAT5 DNA binding activity were carried out using a 32 P-labeled ␤-casein promoter DNA probe, which contains a STAT5b binding site. The specific STAT5b-DNA complex is marked STAT5b EMSA, and the nonspecific protein binding to the ␤-casein probe is labeled n.s. B, STAT5b protein expression level is not affected by overexpression of SOCS-3 protein. The same cell lysates assayed in A were analyzed for STAT5b protein expression by anti-STAT5 Western blotting (Santa Cruz antibody sc-835). C, SOCS-3 blocks STAT5b-dependent GH inhibition of PPAR␣ activity. COS-1 cells were cotransfected with expression plasmids encoding GHR, JAK2, STAT5b, PPAR␣, SOCS-3, and the reporter plasmid pHD(X3)Luc as described above. Transfected cells were treated with Wy-14,643 (Wy, 20 M), either alone or in the presence of GH (5, 25, or 100 ng/ml) for 24 h. Cell extracts were then prepared and assayed for firefly luciferase activities normalized to a Renilla luciferase transfection control. establish that the NH 2 -terminal AF-1 trans-activation domain of PPAR␣, which is active in a ligand-independent fashion (29), is the target of STAT5b inhibition, whereas the COOH-terminal, ligand-dependent AF-2 trans-activation domain is not. EMSA analysis suggested that PPAR␣ DNA binding activity may be partially inhibited in cells containing GH-activated STAT5b, although it is difficult to exclude the possibility that experimental variation in protein yield and transfection efficiency lead to variation in the amount of PPAR␣ protein available for DNA complex formation. GH-activated STAT5b may inhibit these processes by competing for a common factor required for nuclear receptor DNA binding and/or trans-activation, such as a nuclear receptor coactivator (42), although our data indicate that the coactivators P300 and SRC-1 are not the targets of STAT5b inhibition.
Ligand-dependent transcriptional activation by nuclear receptors is mediated by a large, complex COOH-terminal domain that integrates several critical functions: ligand binding, dimerization, trans-activation, and interaction with transcriptional coactivators (49). The three mammalian PPAR isoforms share high homology in the DNA-binding domain and weaker homology in the ligand-binding domain. The AF-2 domain of the receptor, a highly conserved hydrophobic region within the distal COOH-terminal portion of the ligand binding domain, is required for transcriptional activation. Mutations in AF-2 abolish trans-activation function, whereas DNA binding activity is preserved (41). AF-2 is likely to serve as an adapter surface for interaction with coactivators such as SRC-1 (41). Indeed, amino acids within the COOH terminus of PPAR␣ as well as residues within the hinge region of the receptor are required for liganddependent interaction with P300 (45). However, our data show that, unlike full-length PPAR␣, the isolated PPAR␣ COOHterminal region, when fused to a GAL4 DNA-binding domain, is not subjected to inhibition by STAT5b. The AF-2 domain of PPAR␣ is thus not a direct target of STAT5b inhibition, making it unlikely that STAT5b competes for coactivators that interact selectively with the AF-2 domain. AF-2-independent inhibition has also been reported for T3R and the transcription factors Jun and Fos at an AP-1 site, and mutation of the AF-2 domain of that receptor does not affect the inhibition (50).
Interdomain communication has been described as a novel mechanism to regulate PPAR␥ activity (43). PPAR␥ ligand binding activity can be modulated by interactions between the NH 2 -terminal A/B domain (AF-1 domain) of this nuclear receptor and its COOH-terminal ligand-binding domain. Modification of the A/B domain by MAP kinase phosphorylation reduces PPAR␥ ligand-binding affinity and inhibits the transcriptional activity of that receptor. A strong ligand-independent transactivating domain (AF-1-like domain) was recently localized to human PPAR␣ amino acids 1-92 (29). In contrast to the inhibition that characterizes MAP kinase phosphorylation of PPAR␥, insulin-induced, MAP kinase-dependent phosphorylation of two serine residues within the AF-1 region stimulates PPAR␣ activity, leading to the suggestion that the strong AF-1 region of PPAR␣ is partially silenced by corepressor proteins (29). In the present study, we show that the AF-1 function of PPAR␣ is inhibited by GH-activated STAT5b to an extent that is similar to the STAT5b inhibition of full-length PPAR␣. This inhibition of the AF-1 domain of PPAR␣ is not dependent on the serine residues at positions 12 and 21, and thus the corepressor proteins that are proposed to dissociate upon MAP kinase phosphorylation of PPAR␣ (29) are probably not the target of STAT5b inhibition. These findings suggest that the inhibition of PPAR␣ transcriptional activity by GH-activated STAT5b occurs by a mechanism that is distinct from the MAP kinase-dependent regulation of PPAR␣ activity by insulin, even though both hormones target the AF-1 region of PPAR␣. This latter conclusion is supported by our earlier observation that the MAP kinase inhibitor PD98059 does not block the inhibition of PPAR␣ activity by GH-activated STAT5b (22).
STAT5b may directly inhibit the AF-1 activity of PPAR␣, or STAT5b may inhibit PPAR␣ transcription by modulating the binding of coactivators or corepressors that interact with the AF-1 region in a manner that is independent of MAP kinase phosphorylation. PPAR␣ ligand binding and DNA binding activity could conceivably also be affected by such an interaction. Our finding that STAT5b may partially inhibit PPAR␣ DNA binding activity (Fig. 5) may be explained by two possibilities. First, STAT5b interaction with PPAR␣ AF-1 region may modulate the interdomain communication of the receptor, as suggested for PPAR␥ (43), such that PPAR␣ ligand binding and DNA binding activity are inhibited. Alternatively, STAT5b may alter the binding of PPAR␣ to its consensus PPRE, in addition to its inhibition of the NH 2 -terminal AF-1 function of the receptor. The PPRE recognized by PPAR is composed of a direct repeat of two AGGTCA half-site separated by a single nucleotide (DR-1 sequence motif). PPAR obligatorily binds to DNA as a heterodimer with RXR (47), which is also a heterodimerization partner for many other nuclear receptors. RXR can thus define a point for convergence and cross-talk between PPAR and other intracellular signaling pathways. In fact, competition for RXR is the basis for the cross-talk between PPAR and T3R (51). In our study, however, overexpression of RXR␣ did not block the STAT5b inhibition of PPAR␣ activity.
Cells respond to GH through a signal transduction cascade that is initiated at the cell surface GHR. GHR dimerizes and is activated upon GH binding, leading to activation of the receptor-associated JAK2 kinase, tyrosine phosphorylation, and activation of STAT proteins, which translocate to the nucleus and ultimately stimulate target gene expression. Although protein tyrosine phosphatases play a role in terminating the GH response, the recently identified family of SOCS/CIS proteins (26 -28) sheds new light on the mechanism of negative regulatory control. GH induces a rapid and transient expression of several SOCS/CIS genes, including SOCS-3 (25), which appears to establish a classic negative feedback loop that downregulates GH signaling. In our studies, SOCS-3 expression fully blocked the STAT5b-dependent inhibition of PPAR␣ activity in GH-treated cells by preventing the activation of STAT5b. This suggests that the cross-talk between PPAR␣ and STAT5b is subjected to negative feedback regulation by SOCS-3. Mechanistically, SOCS-3 appears to inhibit JAK2 kinase signaling to STAT5b by binding via its SH2 domain to tyrosine phosphorylated residues on GHR (25,48). This may, in turn, prevent the recruitment of STAT5b or facilitate ubiquitination of the GHR-JAK2 kinase complex, as suggested in the case of the erythropoietin receptor and the SOCS/CIS family member CIS-1 (52).
Functional interactions between STAT5a and glucocorticoid receptor, another nuclear receptor family member, have been described (23,24). STAT5a and glucocorticoid receptor form a molecular complex that inhibits transcription from a glucocorticoid response element. Although the coactivator P300 enhances STAT5a-dependent, prolactin-stimulated transcription through a direct interaction with STAT5a, P300 cannot reverse the STAT5a-dependent inhibition of the glucocorticoid response (46), a finding that is analogous to that reported here for STAT5b and PPAR␣. Direct protein-protein interactions between STAT5a and glucocorticoid receptor have been proposed as the mechanism that underlies the STAT5a inhibitory effect of that nuclear receptor. However, in contrast to the STAT5aglucocorticoid receptor interaction, no STAT5b-PPAR␣ com-plex was detected on a PPRE element by supershift assay using STAT5b-specific antibody (22). This suggests that an indirect mechanism, one that does not involve direct STAT5b-PPAR␣ interactions, may account for STAT5b inhibition of PPAR␣ activity. One possibility is that STAT5b induces transcription of a secondary gene that yields a negative regulatory protein that inhibits PPAR␣ activation. This model is unlikely, however, in view of our finding (Fig. 6) that a constitutively active STAT5b mutant (STAT5b1*6) (38) failed to inhibit PPAR␣ activity, even though this mutant STAT strongly induced reporter gene expression driven by STAT5 response elements, even in the absence of GH stimulation. However, given that a single amino acid exchange between STAT5a and STAT5b can interconvert their distinguishable DNA binding activities on STAT5 form-specific promoter sequences (3), the possibility remains that, in combination, the two amino acid changes that confer constitutive activity to STAT5b (38) alter the DNA binding specificity of STAT, such that induction of the hypothetical inhibitory factor is lost.
An alternative model for STAT5b inhibition is that STAT5b competes for an essential coactivator or other interacting protein of PPAR␣, which may interact with the AF-1 domain of PPAR␣, leading to loss of PPAR␣ ligand-independent transcriptional activity. This model is supported by our finding that STAT5b can inhibit trans-activation by all three PPAR isoforms and by T3R, which suggests that activated STAT5b interferes with the action of a common regulatory factor required for full activity of each of these nuclear receptors. (Of note, however, STAT5b did not inhibit an endogenous PPAR-like activity in the same COS-1 cell model (22).) The coactivators P300 and SRC-1 were both considered as candidate factors, because both proteins can interact with and coactivate PPAR␣ (Fig. 4) (45,53). Interaction with the nuclear receptor A/B domain has been described for the coactivator SRC-1 (54); however, overexpression of SRC-1 did not block STAT5b inhibition of PPAR␣. Moreover, P300 has been demonstrated to interact with STAT1 (55), STAT2 (56), and STAT5a (46). P300 interacts with transcriptional machinery and RNA polymerase II, and it also interacts with the nuclear factor P/CAF, which exhibits histone acetylase activity. Histone acetylation induces conformational changes in nucleosome structure and facilitates the entry of core transcription factors. Thus, P300 serves as an integrator of multiple signaling pathways. Indeed, P300 has been implicated in the antagonism between STAT and Ap-1 signaling (57). P300 plays an essential coactivator role for an expanding number of regulated transcription factors, including CREB, Jun, and Fos, in addition to STATs (46,56). However, the present study shows that P300 cannot reverse STAT5b inhibition of PPAR␣ activity. Although it is thus unlikely that STAT5b inhibits PPAR␣ by competing for P300 or SRC-1, STAT5b and PPAR␣ may compete for other nuclear receptor coactivators and/or other interacting factors. Further study will be needed to fully elucidate the mechanism for this inhibitory cross-talk between STAT transcription factors and nuclear receptor and to better understand its consequences in terms of the impact of hormonal factors on PPAR signaling initiated by diverse pharmacological agents and environmental chemicals.