This Article Abstract Full Text (PDF) All Versions of this Article: 21/1/106    most recent Author Manuscript (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager NURSA Molecule Pages Link Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Buser, A. C. Articles by Edwards, D. P. Search for Related Content PubMed PubMed Citation Articles by Buser, A. C. Articles by Edwards, D. P.

Progesterone Receptor Repression of Prolactin/Signal Transducer and Activator of Transcription 5-Mediated Transcription of the ß-Casein Gene in Mammary Epithelial Cells

Department of Pathology (A.C.B., E.K.G.-H., S.A.L., H.W., S.M.A., D.P.E.), Program in Molecular Biology (S.M.A., D.P.E.) and Department of Microbiology (J.S.), University of Colorado Health Sciences Center, Aurora, Colorado 80045; Baylor College of Medicine (S.L.W., J.M.R., D.P.E.), Department of Molecular and Cellular Biology, Houston, Texas 77030; and Division of Medical Biochemistry (W.D.), Biocenter, Innsbruck Medical University, Innsbruck A-6020, Austria

Address all correspondence and requests for reprints to: Dean P. Edwards, Ph.D., Baylor College of Medicine, Department of Molecular/Cellular Biology, One Baylor Plaza, Room 145E, Mail Box BCM-130, Houston, Texas 77030. E-mail: deane{at}bcm.tmc.edu.

	ABSTRACT

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS REFERENCES

 
Prolactin (PRL) and glucocorticoids act synergistically to stimulate transcription of the ß-casein milk protein gene. Signal transducer and activator of transcription 5 (Stat5) mediates PRL-dependent trans-activation, and glucocorticoid potentiation occurs through cross talk between glucocorticoid receptor (GR) and Stat5 at the ß-casein promoter. In the mouse, progesterone withdrawal leads to terminal differentiation and secretory activation of the mammary gland at parturition, indicating progesterone’s role in repressing milk protein gene expression during pregnancy. To investigate the mechanism of the inhibitory action of progesterone, experiments were performed with cell culture systems reconstituted to express progesterone receptor (PR), the PRL receptor/Stat5 signaling pathway, and GR, enabling evaluation of PR, GR, and Stat5 interactions at the ß-casein promoter. With COS-1, normal murine mammary gland, HC-11, and primary mammary epithelial cells, progestin-PR directly repressed the PRL receptor/Stat5a signaling pathway’s mediation of PRL-induced ß-casein transcription. Progestin-PR also inhibited glucocorticoid-GR enhancement of PRL induced trans-activation of ß-casein. Inhibition depended on a functional PR DNA binding domain and specific PR-DNA interactions at the ß-casein promoter. Chromatin immunoprecipitation assays in HC-11 cells revealed recruitment of PR and Stat5a to the ß-casein promoter by progestin or PRL, respectively. Recruitment was disrupted by cotreatment with progestin and PRL, suggesting a mutual interference between activated PR and Stat5a. Without PRL, progestin-PR also recruited Stat5a to the ß-casein promoter, suggesting that recruitment of an unactivated form of Stat5a may contribute to inhibition of ß-casein by progesterone. These results define a negative cross talk between PR and Stat5a/GR that may contribute to the physiological role of progesterone to repress lactogenic hormone induction of the ß-casein gene in the mammary gland during pregnancy.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS REFERENCES

 
DEVELOPMENT OF THE mammary gland during pregnancy is under the control of progesterone and prolactin (PRL). Progesterone directly stimulates epithelial cell proliferation leading to extensive ductal side branching and alveolar morphogenesis. PRL is also required for alveolar development; however, it has both systemic effects, maintaining the corpus luteum required for progesterone production, and direct effects on development of alveoli (1, 2, 3, 4, 5). During pregnancy, progesterone has the additional function of inhibiting secretory activation and terminal differentiation until after parturition by inhibiting milk protein synthesis and tight junction closure; the latter is important for preventing reflux of accumulated milk into the interstitial space (5, 6, 7, 8, 9, 10). In the mouse, a precipitous decline in circulating progesterone at parturition leads to increased milk protein production and tight junction closure (11, 12, 13). Progesterone receptor (PR) expression in the mouse mammary gland declines progressively during pregnancy, suggesting that lowered levels of PR at parturition also contribute to secretory activation (14). However, a fall in circulating progesterone does not occur in all species. In humans, progesterone levels remain high during the onset of labor, and studies have suggested that PR function is inhibited at parturition through multiple molecular mechanisms (15, 16, 17). Thus release from inhibition by progesterone appears to be a common mechanism required for terminal differentiation of the mammary gland and, depending on the species, is due either to inhibition of PR function or to withdrawal of progesterone.

The primary hormone responsible for induction of milk protein gene expression is PRL with a strong potentiation effect by glucocorticoids (18, 19, 20). The most extensively studied milk protein gene is ß-casein, and cell culture systems have defined the mechanisms of hormonal regulation. Induction of ß-casein transcription by PRL is mediated by Stat5 (signal transducer and activator of transcription), a member of a transcription factor family that transmits signals from cytokine and growth factor receptors to nuclear target genes (21, 22, 23). PRL binding activates PRL receptor (PRLR) through dimerization and tyrosine phosphorylation by Janus kinase 2 (JAK2) (24, 25). Stat5 is recruited by activated PRLR and phosphorylated by JAK2 on a conserved tyrosine (Y694 or Y699, depending on species), resulting in its dimerization, translocation to the nucleus, and binding to a specific 9-bp palindromic consensus sequence (TTCnnnGAA). The ß-casein promoter contains Stat5 binding sites, and PRL activation has been shown to be dependent upon Stat5 binding. Stat5 contains a carboxyl-terminal transcriptional activation domain that binds coactivators such as p300/cAMP reponse element-binding protein-binding protein) (26) and is phosphorylated on serine residues (27). Additionally, the corepressor, silencing mediator of retinoid and thyroid receptor, interacts with Stat5, and its dissociation has been suggested to be required for cytokine-induced activation of Stat5 (28). Of the two closely related Stat5 proteins, Stat5a has been demonstrated to be more important for PRL-dependent mammary gland development and lactation than Stat5b, which appears to be required for GH signaling (29, 30).

Mammary gland-specific expression of milk protein genes does not occur by the action of tissue-specific transcription factors, but rather through combinatorial interactions with composite response elements present in the promoter and upstream enhancer regions. Stat5 and other transcription factors that interact with the ß-casein gene are not mammary gland specific; rather, they are expressed in a broad range of tissues (18, 19). The ß-casein promoter contains closely spaced binding sites for Stat5, CCAAT/enhancer-binding protein-ß (C/EBPß), and YY-1 and has multiple half-glucocorticoid response elements (GREs) that are potential binding sites for glucocorticoid receptor (GR) (Fig. 1A). Cooperative interactions between these elements and their functional importance for tissue-specific gene activation by lactogenic hormones have been shown by site-directed mutagenesis studies and gene deletions in mice (31, 32). Although PRL alone can induce ß-casein expression through Stat5, GR and C/EBPß are not capable of inducing transcription in the absence of a Stat5. However, these two transcription factors strongly synergize with Stat5 to generate a much more robust induction than exhibited by PRL alone (31). Glucocorticoid potentiation has been reported to occur through multiple mechanisms. At high concentrations of GR, potentiation was shown to require interaction of GR with DNA-bound Stat5 and the presence of the N-terminal activation function domain 1, suggesting that GR functions as a Stat5 coactivator (33, 34). At low receptor concentrations, potentiation required integrity of the glucocorticoid response element (GRE) half-sites, suggesting that GR-DNA binding is also required (35, 36). Potentiation has also been reported to involve GR, promoting sustained tyrosine phosphorylation of Stat5a and stabilizing Stat5-DNA binding (37). In addition to positive trans-acting factors, the transcription factor YY1 (yin and yang), which has the capacity to act as a repressor or an activator, has been shown to repress ß-casein expression in the absence of lactogenic hormone and to be dissociated from its binding site in response to PRL, indicating that induction of ß-casein transcription also involves a relief from inhibition (31, 38, 39).

View larger version (29K): [in this window] [in a new window]   Fig. 1. Progestin-Dependent PR-Mediated Inhibition of PRL Induction of ß-Casein Gene Activation in a Reconstituted Cell System A, Schematic representation of the ß-casein promoter with binding sites for different transcription factors. Arrows depict putative GRE half-sites, diamonds are C/EBPß sites, hexagons are YY1 sites, triangles are Stat5 sites, and ovals are OCT-1 sites. B and C, COS-1 cells were cotransfected with a ß-casein (–2300/+490)-LUC reporter gene and expression plasmids for rat PRLRL (100 ng), Stat5a (100 ng), and varying amounts (100–500 ng) of human PR-B (B) and human PR-A (C). At 24 h after transfection, cells were treated with ovine PRL (1 µg/ml) alone or in combination with R5020 (100 nM) for 24 h, and lysates were prepared and assayed for reporter gene activity. Luciferase activity corrected for variation in transfection efficiency with an internal CMV-ß-galactosidase reporter was set to 100% for PRL treatment alone in the absence of PR, and all other data were calculated as luciferase activity relative to the 100% control. Values are averages ± SEM from four independent determinations. D, Levels of ectopically expressed human PR-A and PR-B proteins in COS-1 cells relative to endogenous PR in T47D breast cancer cells as detected by immunoblot with the anti-PR mAb, 1294. OCT, Octamer-binding transcription factor.

The suppressive effect of progesterone on milk protein gene expression was initially shown by experiments in which ovariectomy of pregnant animals resulted in transient lactogenesis and accumulation of ß-casein mRNA and protein. This phenomenon was prevented by the administration of progesterone at the time of ovariectomy, but not by other hormones, including estrogen, PRL, or glucocorticoids (6, 7, 8, 9). Experiments with organ cultures from mammary gland explants of pregnant mice similarly demonstrated that progesterone inhibited ß-casein mRNA accumulation stimulated by PRL (47, 48, 49). The mechanism by which progesterone inhibits lactogenic hormone induction of ß-casein gene expression is not well defined. Because PR is expressed in only a fraction of mammary epithelial cells (MECs) in late pregnancy, this suggests that some of the inhibitory effect may occur indirectly through progesterone regulation of paracrine factors that are capable of acting on PR-negative cells. For example, progesterone decreases the sensitivity of mammary epithelium to PRL during pregnancy by suppressing PRL up-regulation of the PRLR (50, 51). Progesterone has also been reported to induce a DNA binding activity, termed pregnancy-specific nuclear mammary factor, that has been proposed to act as a transrepressor of the ß-casein gene promoter (52, 53).

To explore whether this suppressive function of progesterone involves direct cross talk of PR with GR and PRLR/Stat5 signaling pathways, we have examined interactions of PR, GR, and Stat5a at the ß-casein promoter in different cell culture systems including transiently transfected COS-1, normal murine mammary gland (NmuMG), and HC-11 cell lines, and primary mouse MECs. Under all experimental conditions, PR in a progestin-dependent manner directly interfered with the PRLR/Stat5 signaling pathway and its ability to induce transcriptional activation of the ß-casein gene. PR also inhibited the potentiation effect of glucocorticoids-GR. Inhibitory activity was partially dependent on a functional PR DBD and binding of PR to GRE-half-sites in the ß-casein promoter. Using a chromatin immunoprecipitation (ChIP) assay with HC-11 cells revealed that PR and Stat5 are recruited independently to the ß-casein promoter in response to their cognate activating hormone, whereas assembly of PR and Stat5 was disrupted by cotreatment of cells with PRL and progestin. These results suggest that a mutual interference between PR and Stat5a at the ß-casein promoter contributes to the repressive effect of progesterone on lactogenic hormone induction of ß-casein. An additional sequestering mechanism was suggested by the finding that progestin in the absence of PRL, recruited unactivated Stat5a to the ß-casein promoter. These data contribute to defining the combinatorial transcription factor interactions that mediate the repressive effect of progesterone on transcriptional activation of the ß-casein gene.

	RESULTS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS REFERENCES

 
Progesterone Inhibition of PRL Induction of ß-Casein Gene Transcription Involves Progesterone Receptor (PR) Interference with PRLR/Stat5 Signaling
To examine whether progesterone and PR have a direct influence on PRL induction of ß-casein gene transcription, a reconstitution cell cotransfection system was used. PRLR, Stat5a, and human PR were ectopically expressed in COS-1 cells that are deficient in these components of progesterone and PRL signaling, along with a ß-casein-LUC reporter gene. The LUC reporter gene was linked to the –2300 to +490 promoter region of the rat ß-casein gene that contains proximal Stat5 response elements (S5REs), multiple half-glucocorticoid response elements (GREs), and sites for various other transcription factors (Fig. 1A). PRL treatment resulted in a 6- to 7-fold induction of ß-casein-LUC expression, and this was inhibited by the synthetic progestin R5020 in a manner dependent on expression of either PR-B or PR-A (Fig. 1, B and C). Furthermore, inhibition was dependent on the amount of PR transfected, with PR-A mediating a complete inhibition and PR-B mediating a reduction to approximately 35% that induced by PRL at the highest amount of transfected PR. This difference in PR isoforms was not due to a different level of protein expression, because PR-A and PR-B were expressed in comparable amounts as determined by immunoblotting (Fig. 1D). PRL induction was dependent on expression of Stat5a, indicating that progesterone inhibits the PRLR/JAK2/Stat5 signaling pathway (Fig. 2). ß-Casein-LUC expression was not stimulated by progestin in the absence of PRL, whereas R5020, under the same cell conditions, stimulated induction of a PRE₂-tk-luciferase reporter containing a classical palindromic PRE (Fig. 2). These results, taken together, indicate that progesterone is capable of inhibiting PRL induction of ß-casein gene transcription due to PR interfering with the PRL/Stat5 signaling pathway in a cell-autonomous fashion.

View larger version (23K): [in this window] [in a new window]   Fig. 2. PR Inhibits Stat5a-Dependent PRL Induction of ß-Casein Gene Activation COS-1 cells were cotransfected as in Fig. 1, with the ß-casein (–2300/+490)-LUC reporter and expression plasmids for rat PRLRL, Stat5a, and human PR-B. Cells were treated with vehicle (ethanol), PRL (1 µg/ml), or R5020 (100 nM), and luciferase activity was determined and expressed as relative activity as in Fig. 1, B and C. Under the same transfection conditions, cells were also cotransfected with PRE2-tk-LUC and human PR-B. Values represent averages ± SEM from three independent experiments.

View larger version (19K): [in this window] [in a new window]   Fig. 3. PR Inhibits Glucocorticoid Potentiation of PRL-Induced Activation of ß-Casein Gene Activation COS-1 cells were cotransfected as in Fig. 1, B and C, with ß-casein (–2300/+490)-Luc reporter, rat PRLRL, Stat5a, and human PR-A (500 ng) or GR (100 ng). At 24 h after transfection cells were treated another 24 h with ovine PRL (1 µg/ml), PRL (1 µg/ml) + Dex (100 nM), Dex (100 nM), PRL (1 µg/ml) + R5020 (100 nM) or PRL (1 µg/ml) + Dex (100 nM) + R5020 (100 nM). Luciferase was assayed and expressed as relative activity as in Fig. 1, B and C. Values are averages ± SEM from triplicate independent determinations.

View larger version (27K): [in this window] [in a new window]   Fig. 4. Inhibition of PRL-Induced ß-Casein-LUC Gene Activation Involves the DBD of PR COS-1 cells were cotransfected as in Fig. 1 with ß-casein (–2300/+490)-LUC reporter and expression plasmids for PRLRL, Stat5a, PR-A(500 ng), PR-B (500 ng), PR-Bmut587 (cys to ala) (500 ng), PR-Amut587 (cys to ala) (500 ng), or an empty vector. Cells were treated with vehicle, PRL (1 µg/ml), or PRL (1 µg/ml) + R5020 (100 nM), and luciferase was measured and expressed as relative activity to the 100% PRL-induced level in the absence of PR. The values are averages ± SEM from three independent determinations.

View larger version (58K): [in this window] [in a new window]   Fig. 5. PR Binds to GRE Half-Sites in the Proximal ß-Casein Gene Promoter A, EMSA of PR binding to a double-stranded oligonucleotide probe corresponding to –110 to –70 bp of the rat ß-casein promoter containing the S5RE and 3'- and 5'-flanking GRE half-sites. Varying concentrations (10–200 nM) of purified baculovirus-expressed PR-A (bound to R5020) were reacted with a limiting concentration of 32P-labeled DNA probe (0.3 ng) in the absence and presence of HMGB-2 (500 ng). Free and bound DNA complexes were separated by electrophoresis on native 5% polyacrylamide gels and detected by autoradiography. Supershift of the complex with PR-specific antibody (1294) is shown in the far right lane. B, Schematic of double-stranded oligonucleotide 32P-labeled ß-casein promoter probes containing point mutations in either the Stat5 response element or 3'- and 5'-GRE half-sites. Mutations are indicated by bold letters. C, EMSAs of purified Stat5 and PR binding to the 32P-labeled DNA probes shown in panel B above. Only regions of gels with upshifted DNA complexes are shown; the free DNA has been cut off. Left panels, Varying concentrations (25–100 nM) of PR were reacted with a limiting concentration (0.3 ng) of 32P-labeled DNA probes. Supershift of PR-DNA complexes with PR-specific antibody (1294) is in the far right lane. Right panels, Varying concentrations (500 ng to 1.25µg) of Stat5a were reacted with a limiting concentration of 32P-labeled DNA probe. D, EMSA of PR-B and Stat5a binding individually and together as a ternary complex with a DNA probe containing the ß-casein promoter sequence (–110 to –70). The 32P-labeled DNA probe (0.3 ng) was reacted with PR-B (75 nM) alone without and with a PR-specific antibody (1294), with Stat5a (200 ng) alone without and with a Stat5a-specific antibody (mouse monoclonal) or a mixture of PR-B (75 nM) and Stat5a (200 ng), without and with antibodies to PR (1294), Stat5a, or a control unrelated antibody. Ab, Antibody.

To map the region of the promoter required for the inhibitory effect of progesterone, a LUC reporter was used that contains a truncated ß-casein promoter from –344 to +1 bp. In the COS-1 cell reconstitution assays, the proximal promoter between –344 to +1 was sufficient for full PRL induction and for mediating inhibition by progesterone (Table 1). Mutations in the 3'-GRE half-site had minimal influence on R5020 inhibition of PRL induction, whereas mutations in the 5'-GRE half-site significantly diminished the inhibitory effect of R5020 mediated by PR-B, but had minimal effect on PR-A-mediated inhibition (Table 1). These functional results correlate with in vitro DNA results, supporting the conclusion that PR binding to the 5'-GRE half-site is important for the inhibitory effect of progesterone on PRL induction of ß-casein gene transcription, whereas the PR-A isoform requires an additional DNA-independent mechanism.

View larger version (31K): [in this window] [in a new window]   Fig. 6. Both Mouse and Human PR Mediate Inhibition of PRL Induction of ß-Casein Gene Transcription in Mammary Epithelial-Derived Cells NMuMG cells were cotransfected with rat PRLR (100 ng), Stat5a (5 ng), ß-casein luciferase (200 ng), and varying amounts (as indicated in figure) of mouse or human PR-A and PR-B expression vectors. At 24 h post transfection, cells were treated for another 24 h with ovine PRL (1 µg/ml), or PRL (1 µg/ml) + R5020 (10 nM), and luciferase activity was measured. The data were calculated as percent inhibition of PRL-induced luciferase activity by R5020, and the values represent averages ± SEM from three independent determinations. ms, Mouse; h, human.

View larger version (45K): [in this window] [in a new window]   Fig. 7. Progestin Inhibits PRL Induction of the Endogenous ß-Casein Gene in Primary Mouse MECs A, Western blot of msPR expressed in primary MECs. MECs were grown on collagen for 2 d before resuspension and were transduced for 1 h at 37 C with recombinant adenoviruses encoding LacZ, mouse PR-A, or mouse PR-B. Cells were then replated on a thin layer of matrigel and treated with 100 nM R5020 or EtOH (– control) for 48 h. MECs were harvested and PR protein was assayed by immunoblotting using a polyclonal antibody to mouse PR. *, Nonspecific band. B, Primary mouse MECs were transduced as in panel A above and treated for 48 h with either EtOH (– control), ovine PRL (3 µg/ml), or ovine PRL (3 µg/ml) + R5020 (100 nM). Cells were harvested and ß-casein protein expression was assayed by immunoblotting using a polyclonal antibody to mouse ß-casein. C, Quantitation of ß-casein protein expression. Immunoblots of ß-casein protein expression were scanned with a Storm phosphorimaging system and data were analyzed using ImageQuant 5.0 software. Data were calculated as the percent intensity of ß-casein protein bands in the presence of PRL and R5020 to that in the presence of PRL alone,which was set to 100%. Ratios were calculated separately for each treatment group: LacZ, mouse PR-A, and mouse PR-B. The values are averages ± SEM from three independent determinations. ms, Mouse; HC, hydrocortisone.

View larger version (36K): [in this window] [in a new window]   Fig. 8. Recruitment of PR and Stat5a to the ß-Casein Promoter in HC-11 Cells Detected by ChIP Assay Was Disrupted by the Presence of Both Hormones, Progestin and PRL A, HC-11 cells stably transfected with the ß-casein (–2300/+490)-LUC reporter construct (HC-11.7E) were transduced for 24 h with an adenovirus expression vector for human PR-B (MOI of 50 pfu/cell). Cells were then incubated for another 24 h with EtOH (– control), ovine PRL (3 µg/ml), or ovine PRL (3 µg/ml) + R5020 (100 nM), and luciferase was measured and data expressed as fold ß-casein induction. Values are average determinations ± SEM from triplicate indepdendent determinations. B, HC-11.7E cells were transduced with 50 pfu/cell of human PR-B encoding adenovirus and then treated for 1 h with either EtOH (– control), ovine PRL (3 µg/ml), R5020 (100 nM), or ovine PRL (3 µg/ml) + R5020 (100 nM). Cells were then harvested for ChIP assay using PCR primers flanking the S5RE and the two GRE half-sites within the proximal promoter of ß-casein. Upper left, Schematic of the primers for ChIP assay (upper right), representative agarose gel of PCR products of input DNA, and from immunoprecipitation by control antibody, PR mAB (1294), or Stat5 polyclonal antibody (L-20). Lower panel, Quantification of ChIP assays from multiple experiments using Syngene Genetools software. Data represent average density of immunoprecipitations as a percentage of total input DNA adjusted after subtracting signals obtained with control antibody. The values are averages ± SEM from three determinations. C, ChIP assay described in panel B was repeated with a second set of PR (B-30) and Stat5 (N-20) antibodies. The values are averages ± SEM from three determinations. D, ChIP assay, as described in panel B above with 1294 PR antibody and L-20 Stat5a antibody, was performed with endogenous ß-casein gene in parental nontransfected HC-11 cells. The agarose gel is representative of three independent experiments. Ab, Antibody; IP, immunoprecipitation; GFP, green fluorescent protein; R, R5020.

Progestin-PR Dependent Recruitment of Stat5a to the ß-Casein Promoter in the Absence of PRL
To further investigate the mechanism by which progestin promotes recruitment of Stat5a to the ß-casein promoter in the absence of PRL, we asked whether Stat5a was recruited by PR to a PRE-controlled MMTV target gene that lacks Stat5 response elements. As expected, ChIP assay with T47D breast cancer cells containing a stably integrated MMTV gene (61) detected recruitment of PR to the promoter in response to treatment of cells for 1 h with R5020. No recruitment of Stat5a was detected in response to either PRL or R5020, suggesting that progestin-PR recruitment of Stat5 is promoter context dependent, requiring a composite PR/Stat5 binding site (Fig. 9).

View larger version (28K): [in this window] [in a new window]   Fig. 9. PR Does Not Recruit Stat5a to Classical PRE of the MMTV Gene T47D breast cancer cells containing a stably integrated MMTV gene fused to chloramphenicol acetyltransferase (CAT) were treated for 1 h with either EtOH (– control), ovine PRL (3 µg/ml), R5020 (100 nM), or ovine PRL (3 µg/ml) + R5020 (100 nM). Cells were then harvested for ChIP assay using PCR primers to the GRE/PRE sites located between –190 and –80 in the LTR of the MMTV promoter. Upper left, Schematic of the primers used for ChIP assay of the hormone response element (HRE) in the MMTV promoter. Upper right, Representative agarose gel of PCR products from input DNA and from immunoprecipitation by control antibody, PR mAB (1294), and Stat5 polyclonal antibody (L-20). Lower panel, Quantification of ChIP assays using Syngene Genetools software. Data represent average density of immunoprecipitations as a percentage of total input DNA after subtraction of signal obtained with a control antibody. The values are averages ± SEM from three determinations. Ab, Antibody; IP, immunoprecipitation; R, R5020.

View larger version (36K): [in this window] [in a new window]   Fig. 10. Progesterone Does Not Affect PRL-Induced Tyrosine Phosphorylation of Stat5a A, Human breast carcinoma T47D cells were treated with vehicle (ETOH), PRL (1 µg/ml), PRL (1 µg/ml) + R5020 (100 nM), or R5020 (100 nM) for 30 min. The cells were then lysed, and equal amounts of protein were immunoprecipitated with anti-Stat5, L-20 antibody (+), or a control unrelated antibody (–). The immunoprecipitates were then probed with antiphosphotyrosine Stat5 antibody (Tyr694/699). B, Monkey kidney COS-1 cells were transiently transfected with Stat5a alone or in combination with either human PR-B or human PR-A and then treated with vehicle, PRL (3 µg/ml), R5020 (100 nM), or PRL (3 µg/ml) + R5020 (100 nM) for 1 h. The cells were then lysed and equal amounts of protein were probed by Western blot. Upper panel, Tyrosine-phosphorylated Stat5a (Tyr694/699); lower panel, total Stat5a (L-20). C, Mouse mammary epithelial HC-11.7E cells were transduced with either adenovirus encoding GFP as a control vector, or human PR-B at an MOI of 50 pfu/cell, or no adenovirus (Mock), and were then treated with vehicle, R5020 (100 nM), PRL (3 µg/ml), or PRL (3 µg/ml) + R5020 (100 nM) for 1 h. The cells were then lysed, and equal amounts of protein were analyzed by immunoblot for PR (1294) (upper panel) or were immunoprecipitated with antibody to total Stat5a and the immunoprecipitate was analyzed by immunoblot for tyrosine-phosphorylated Stat5a (Tyr694/699) (midpanel) or total Stat5a (L-20) (lower panel). GFP, Green fluorescent protein; IB, immunoblotting; IP, immunoprecipitation.

	DISCUSSION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS REFERENCES

Because PR and GR can recognize the same or similar DNA sequence response elements, we explored whether PR can bind to GRE half-sites in the ß-casein promoter and whether this interaction was important for mediating the inhibitory effects of progestins. By EMSA, PR bound to a ß-casein proximal promoter DNA probe in a manner dependent on the integrity of the 5'-GRE half-site, indicating that half-GREs are indeed potential binding sites for PR (Fig. 5). By ChIP assay, PR in a progestin-dependent manner was also recruited to the proximal ß-casein promoter in HC-11 cells, confirming PR interaction in nuclei in the context of chromatin (Fig. 8, B–D). The importance of PR interaction with GRE half-sites in mediating functional response to progestin was shown by the fact that a DNA binding-deficient PR-B with a point mutation in the DBD exhibited a complete loss of repression activity and a partial loss in the context of PR-A (Fig. 4). Point mutations in GRE half-sites in the ß-casein reporter gene also resulted in a substantial loss of the repression activity of PR-B, whereas the activity of PR-A was minimally affected, indicating the involvement of additional DNA-independent mechanisms for the A isoform of PR (Fig. 4 and Table 1). It should be noted that the more efficient and complete inhibition of PRL transactivation obtained with human PR-A in COS-1 cells was not observed with mouse PR-A or with human PR-A in MECs, suggesting this may be a cell type-specific or conditional property of human PR-A.

Using in vitro DNA binding assays with a DNA probe, we found that PR activated by R5020 and Stat5a activated by phosphorylation with JAK-2 can bind simultaneously to the proximal ß-casein promoter to form a ternary PR/Stat5/DNA complex (Fig. 5D). By ChIP assay PR or Stat5a was recruited to the ß-casein promoter in response to treatment of cells with either progestin or PRL, respectively; however, neither PR nor Stat5 was recruited after 1 h treatment of cells with both hormones (Fig. 8, B–D). Because the same results were obtained by ChIP assay with antibodies to different regions of PR and Stat5a, this suggests that adding PRL and progestin together prevents recruitment of both PR and Stat5a to the promoter, rather than inducing a change in the conformation of these proteins that obstructs antibody recognition. There are several possible explanations for the apparent difference between in vitro DNA binding and ChIP results. First, the orientation of the adjacent Stat5a and PR binding sites (GRE half-sites) may simply be different in the DNA probe in vitro vs. DNA in the context of nucleosomes in nuclei. Second, treatment of cells with PRL and progesterone may stimulate posttranslational modifications (presumably phosphorylation) that affect protein-protein interactions in a manner that disrupts assembly of PR and/or Stat5a with the promoter. Third, PR and/or Stat5a in vivo may interact with other protein factors that disrupt their interaction with the ß-casein promoter. Because dissociation of YY1 is required for PRL induction of ß-casein, this raises the possibility that PR could stabilize YY1 interaction in a manner that reduces Stat5 binding, and Stat5a binding, in turn, may be required to stabilize PR interaction (32, 38, 39). Lastly, ChIP experiments were performed only at a single time point of hormone addition (1 h), raising the possibility that a sequential mechanism may operate to expel Stat5a from the ß-casein promoter. PR may first contact the ß-casein promoter and cycle on and off with time of hormone treatment resulting in expulsion of Stat5a. A sequential mechanism would be consistent with the requirement of the PR DBD and PR-DNA for the repressive effect of progesterone on PRL induction of ß-casein gene transcription. These results, taken together, are consistent with the conclusion that progestin inhibition of PRL induction of ß-casein gene transcription involves a mutual interference of PR and Stat5 binding that is dependent on the presence of both PRL and progesterone.

Progesterone could potentially repress PRL/R/Stat5a signaling by inhibiting more proximal steps of Stat5a activation including tyrosine phosphorylation by Jak2, or subsequent dimerization and nuclear translocation of Stat5a. However, as determined by immunoblotting with an affinity-purified antibody to tyrosine 694/699 of Stat5a, progestin did not inhibit PRL-induced phosphorylation of Stat5a under any conditions examined including that of endogenous Stat5a in T47D breast cancer cells and HC-11 cells, and transfected Stat5a in COS-1 cells (Fig. 10). Paradoxically, we observed a progestin-PR-dependent recruitment of Stat5a to the ß-casein promoter in the absence of PRL as detected by ChIP assay (Fig. 8, B–D), indicating the formation of a ternary PR/Stat5a/DNA complex under these conditions. This result is consistent with a previous report of Richer et al. (62), showing progestin-induced nuclear translocation of Stat5 in T47D breast cancer cells. In cotransfected COS-1 cell experiments we similarly observed by immunofluorescence a progestin-PR-dependent nuclear translocation of Stat5a in the absence of PRL (Buser, A. C., and D. P. Edwards, unpublished data). Stat5a recruited to the ß-casein promoter by progestin appears to be unactivated because progestin did not stimulate tyrosine phosphorylation of total cellular Stat5 (Fig. 10). However, it is possible that progestin stimulates tyrosine phosphorylation of a small fraction of total cellular Stat5a that is recruited to the ß-casein promoter, but is below the level of detection by phosphoimmunoblotting. ChIP assay with antibodies to total and phosphorylated Stat5a will be required to distinguish between these possibilities.

How progestin promotes nuclear translocation and recruitment of Stat5a to the ß-casein promoter is not known. A physical interaction between PR and Stat5a has been reported previously by coimmunoprecipitation assay of Hela cell nuclear extracts cotransfected with Stat5a and Flag-tagged PR-B (62). However, the interaction did not appear to be hormone dependent (62). We did not observe an efficient interaction of PR and Stat5a by coimmunoprecipitation assay of T47D cells either in the absence of hormone or after short-term (30 min) treatment of cells with R5020, PRL or both hormones together. Yet, under the same conditions, an efficient coimmunoprecipitation of GR and Stat5a was detected in cells treated with glucocorticoid and PRL (Gass-Handel, E. K., and D. P. Edwards, unpublished data) consistent with previous reports of physical interaction of GR and Stat5 (33). Thus, whether PR and Stat5 make protein-protein interactions in the absence of DNA is not clear. Progestin-PR-dependent recruitment of Stat5a by ChIP assay was not observed with the PRE-containing MMTV promoter (Fig. 9), suggesting that PR recruitment of Stat5a does not rely simply on protein-protein interaction but is promoter context dependent. Thus a combination of protein-protein and protein-DNA interactions by PR and Stat5a at the ß-casein promoter may be required. The functional consequence of progestin-PR-dependent recruitment of Stat5a to the ß-casein promoter is not known. Stat5a does not appear to function as a PR coactivator, because progestin did not induce transcription of the ß-casein gene under conditions where PR recruits Stat5a. If, in fact, Stat5a recruited by progestin-PR in the absence of PRL has bypassed the PRLR/Jak2 signaling pathway and is not tyrosine phosphorylated, it could contribute to the inhibitory effect of progesterone by acting as a dominant negative for any active Stat5a. Further studies will be required to determine the role of progestin-PR-dependent recruitment of Stat5a to the ß-casein gene in the absence of PRL.

Several studies have reported interactions between Stats and different classes of steroid receptors indicating that cross talk between these two distinct families of transcription factors activated by different signaling pathways may be of general functional relevance. GR interaction with Stat5a results in a strong enhancement of Stat5a-mediated gene transcription and plays a role in potentiating PRL induction of milk protein gene expression in the mammary gland. The potentiation effect of glucocorticoids requires both GR-Stat5 protein interaction and the integrity of GRE half-sites in the ß-casein promoter, suggesting an involvement of GR-DNA binding (20, 33, 34, 35, 36). GR and Stat5a can both recruit the histone acetyl transferase (HAT) p300 coactivator to the ß-casein promoter in a manner that correlates with functional synergy between PRL and glucocorticoids and acetylation of core histones H3 and H4 (26, 32), suggesting that GR and Stat5a, as a unit, may be required for efficient recruitment of coactivators and subsequent chromatin remodeling and transcription. Because PR and GR are closely related and can bind to the same DNA sequences, this raises the interesting question of how these two receptors can have opposite effects on PRL induction of ß-casein, and how PR can inhibit the potentiation effect of glucocorticoids. It is possible that PR binds with higher affinity to the GRE half-sites and effectively competes with GR. Doppler et al. (36), by use of PR/GR chimeras, showed that the strong enhancement activity toward PRL-induced trans-activation of ß-casein that is lacking in PR resides in the N-terminal domain of GR. Also, the DBD of GR was shown to be dispensable for glucocorticoid potentiation activity (34), whereas the DBD of PR is important for the inhibitory activity of progesterone (Fig. 4). Clearly, further studies are required to determine the mechanism responsible for the opposing actions of GR and PR.

Cross talk between ER and Stat5 has also been reported and, depending on the cellular context or experimental conditions, estrogen and ER have been observed to either repress or potentiate Stat5-mediated gene transcription. In HC-11 cells, ER or ERß potentiated PRL-induced Stat5a-mediated transcription of a ß-casein reporter gene in a manner dependent on ER ligand, on an intact ER DBD, and on interaction of ER with Stat5 bound to DNA (63). ER potentiation of PRLR signaling mediated by Stat5a was also observed in MCF-7 breast cancer cells (64), but most other studies observed an estrogen-ER-dependent repression of Stat5-mediated transcriptional activity. In transfected human embryonic kidney 293 cells, ER and ERß repressed PRL-induced activation of a ß-casein reporter gene mediated by Stat5a/b (64, 65). Repression required protein interaction between ER DBD and Stat5 and appeared to be mediated by a decrease in PRL-induced tyrosine phosphorylation of Stat5a and reduced nuclear translocation and binding of Stat5a to DNA (64, 65). In HepG2 liver cells, an estrogen-ER-dependent repression of PRL activation of the Na+/taurocholate cotransporter polypeptide (ntcp) gene was demonstrated by a mechanism that also involved interference with the PRLR/Stat5a signaling pathway (66). In addition to cytokine receptors, growth factor receptors can mediate effects on gene expression by signaling through Stat5. In breast cancer cells, overexpressing EGF receptor, an estrogen-ER-dependent repression of EGF-induced activation of Stat5-mediated gene transcription was reported that appeared to account for the ability of estrogen to inhibit EGF-stimulated DNA synthesis (67). The physiological role of ER-Stat5 cross talk is not known other than to speculate that both ER and Stat5 are important in the mammary gland and that negative or positive functional cooperativity may have a role in specific developmental stages or functions of the mammary gland.

Much less is known about interactions of other steroid receptors with Stat5. Two survey experiments analyzed functional interactions of different classes of steroid receptors with Stat5 in COS-7 or CV-1 cell cotransfection assays (36, 68). Androgen receptor had no effect on Stat5a-mediated transcriptional activity, a minimal positive effect of mineralocorticoid receptor on Stat5a activity was observed, and conflicting results have been reported with PR. PR enhanced or had a minimal effect on PRL induction of Stat5-mediated transcription of ß-casein, as compared with the more substantial enhancement observed by GR. Another study of the influence of progesterone on PRL-Stat5 signaling in T47DY breast cancer cells stably transfected with PR-B reported that pretreatment of cells with R5020 for 48 h was required for PRL-induced tyrosine phosphorylation of Stat5 and PRL induction of a ß-casein-LUC reporter gene (62). Why progestin-PR-dependent repression of PRL/Stat5 signaling at the ß-casein promoter was not observed in these other studies is not known. As with studies of ER-Stat5 cross talk, the nature of the effect on Stat5 activity may be dependent on cell type-specific factors or experimental conditions. Much higher concentrations of expression plasmids were used in COS-1 and CV-1 cotransfection experiments as compared with the present study, suggesting differences may be the result of overexpression of steroid receptors and Stat5. In T47DY experiments, cells were pretreated for 48 h with R5020 before PRL addition, indicating that progesterone sensitized cells to PRL by persistent changes in gene expression, as opposed to mediating a direct cross talk between PR and Stat5 signaling pathways (62). A requirement of progesterone pretreatment for PRL stimulation of tyrosine phosphorylation of Stat5 and induction of ß-casein is not consistent with other results in the literature, including studies herein with parental T47D breast cancer cells, suggesting this may be a unique property of T47DY cells stably transfected with PR-B.

Immortalized cell lines derived from normal MECs, as well as primary MEC cultures, lose expression of PR and ER; thus, it was not possible in this study to evaluate endogenous PR in mediating effects of progesterone on the PRLR/Stat5a signaling pathway. However, an advantage of ectopically expressed PR is a versatility to show that inhibitory effects of progestin were dependent on expression of PR, to compare different forms of PR, and to analyze the dependence or influence of other transcription factors. A potential limitation of the experimental system is that overexpression may force PR interactions with other factors that may not occur with lower endogenous concentrations of PR. However, under all experimental conditions, a similar progestin-PR-dependent repression of PRL induction of ß-casein gene expression was observed including lipid-mediated transient transfection of COS-1 and NMuMG cells that can result in high protein expression in a fraction of the cell population, and expression in HC-11 and primary MECs from adenoviral vectors under conditions that result in a more homogenous lower level of expression. Similar results were also obtained with reporter genes and endogenous ß-casein, arguing against forced interactions. Moreover, the inhibitory activity of PR observed in these experimental systems is consistent with the role of progesterone in the mammary gland during pregnancy to suppress secretory activation and differentiation until after parturition.

To our knowledge, these are the first studies to extensively analyze cross talk between PR and the PRLR/Stat5 signaling pathway at the ß-casein promoter. These results define a direct interaction between PR and the PRLR/Stat5 signaling pathway in the nucleus that may contribute to the role of progesterone to repress lactogenic hormone induction of milk protein gene expression in the mammary gland during pregnancy. In the mammary gland PR is expressed heterogeneously in epithelial cells (3), and it is likely that a paracrine mechanism contributes to the inhibitory action of progesterone on PR-negative cells. Because of the functional importance of this action of progesterone, it is reasonable to speculate that inhibition by progesterone is multifactorial, i.e. mediated by indirect paracrine and direct nuclear actions of PR. Further studies will be necessary to determine the contribution of direct PR cross talk at the ß-casein promoter relative to paracrine effects of progesterone on PR-negative cells and to better define the paracrine factors and mechanisms.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS REFERENCES

 
Materials
R5020 (promegestone; 17,21dimethyl-19-norpregna-4,9-diene-3,20-one) was obtained from DuPont-New England Nuclear Research Products (Boston, MA), Dex was from Sigma (St. Louis, MO), and ovine PRL was provided by National Institute of Diabetes and Digestive and Kidney Diseases (Bethesda, MD). 1294/H9 is a mouse IgG1 mAb produced against purified human PR that recognizes both the A and B isoforms of PR (69), and B-30 is a mAb that recognizes the N-terminal segment unique to human PR-B (70). N-terminal Stat5 (N-20) (sc836) and C-terminal Stat5 (L-20) (sc1081) polyclonal rabbit anti-Stat5 antibodies were purchased from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). Rabbit polyclonal antiphosphotyrosine Stat5 antibody (Tyr694/699) (07–586) was purchased from Upstate Biotechnology (Charlottesville, VA). Polyclonal rabbit anti-ß-casein antibody used in primary MEC experiments was obtained from Margaret Neville [University of Colorado at Denver and Health Sciences Center, Aurora, CO].

Plasmids
The mammalian cell expression plasmids, pCMV-human PR-A and human PR-B, containing the cDNA for human PR-B under the control of the cytomegalovirus (CMV) enhancer/promoter, have been described previously (55, 56, 61). The rat Stat5a cDNA was subcloned into pRcCMV expression vector (Invitrogen, Carlsbad, CA). The rat PRLR-long form expression vector pECE-PRLRL and ß-casein (–2400/+490)-LUC reporter gene were described previously (32). Truncated ß-casein (–344/+1), and mutant ß-casein promoters driving chloramphenicol acetyltransferase gene were described previously (35). Baculovirus vectors for expression of murine Stat5a and murine Jak2 were obtained from James Ihle (St. Jude’s, Memphis, TN) and Andrew Kraft (University of Colorado Health Science Center, Aurora, CO) respectively. Mouse PR-B cDNA was provided by Shyamala Gopalan (Berkeley, CA) and was cloned into the CMV-based mammalian cell expression plasmid pcDNAI (Invitrogen, Carlsbad, CA) by insertion into the BspHI/EcoRI site located in the multiple cloning site. The cDNA for mouse PR-A was prepared by partial digest of mouse PR-B cDNA, ligated into the pBlueBacHis2B transfer plasmid and then inserted into the BamHI/EcoRV site in the multiple cloning cassettes of pcDNAI. pSV2neo was obtained from CLONTECH (Mountain View, CA).

Transient Transfections
Monkey kidney COS-1 cells were plated into six-well dishes (Falcon, Oxnard, CA) at a density of 1.65 x 10⁵ cells per well in DMEM (Life Technologies, Gaithersburg, MD) supplemented with 10% fetal bovine serum (FBS) (Hyclone Laboratories, Logan, UT). Cells were grown at 37 C for approximately 24 h and allowed to reach 50–60% confluency. Cells were then transfected using Lipofectamine according to the manufacturer’s instructions (Life Technologies, Inc., Gaithersburg, MD). DNA (1–2 µg total) and 2.25 µl of Lipofectamine were used per 35-mm well. After transfection, cells were maintained in serum-free DMEM containing 5 µg/ml insulin and 10 µg/ml apo-transferrin (Sigma, St. Louis, MO) throughout the experiment. After 24 h, cells were treated with the synthetic progestin R5020 and/or ovine PRL (1 µg/ml) in serum-free DMEM supplemented with insulin and apo-transferrin, as indicated in figure legends, and incubated for 24 h at 37 C.

Normal murine mammary gland (NMuMG) cells obtained from Gary Firestone (University of California, Berkeley, CA) were plated into six-well dishes (Falcon, Oxnard, CA) at a density of 1.5 x 10⁵ cells per well in DMEM supplemented with 10% FBS. Cells were grown at 37 C for approximately 24 h and reached 60–70% confluency. Cells were then transfected using the Lipofectamine protocol above. After transfection, cells were maintained in DMEM containing 5% FBS treated with dextran-coated charcoal (DCC) to remove steroids. After 24 h, cells were treated with vehicle, R5020 and/or PRL (1 µg/ml) in DMEM supplemented with 5% FBS-DCC, 5 µg/ml insulin, and 1 µg/ml hydrocortisone.

Reporter Gene Assays
Cell monolayers in six-well dishes were rinsed twice with wash buffer [40 mM Tris-HCl (pH 7.4), 150 mM NaCl, and 1 mM EDTA] and cells were lysed directly in the well with 300 µl of lysis buffer consisting of wash buffer plus 0.5% Triton X-100. The lysates were removed and centrifuged at 12,000 rpm for 10 min at 4 C to pellet cell debris. Luciferase activity in lysates was measured using a Monolight 2001 luminometer (Analytical Luminescence Laboratories, San Diego, CA). Lysates (25 µl) were added to 0.35 ml luciferase assay buffer [100 mM K2HPO4 (pH 7.8), 15 mM MgSO₄, 5 mM ATP, and 1 mM dithiothreitol (DTT)]. Luciferase-mediated light output was assessed for 1 mM luciferin in the reaction chamber. To compensate for variation in transfection efficiency, cells were cotransfected with a CMV-ß-gal reporter plasmid. Results were calculated as the ratio of LUC to ß-gal activity. Normalization of LUC activity to 1.0 in the absence of hormone or to 100% in the presence of PRL, or PRL plus glucocorticoids, enabled calculation of average values between multiple experiments. Values for relative luciferase activity were calculated as averages from at least three independent determinations ± SEM.

Protein Purification of PR, HMGB-2, and Stat5a
Polyhistidine-tagged PR and HMGB-2 were expressed from baculovirus vectors in Sf9 insect cells and prepared as previously described (55, 56). Briefly, cells were lysed in the following buffer: 20 mM Tris-HCl, pH 8.0; 350 mM NaCl; 10 mM imidazole; 5% glycerol; and a cocktail of protease inhibitors. Cell lysates were centrifuged at 100,000 x g for 30 min, and the supernatant was used as soluble whole-cell extract. Whole-cell extracts were passed over nickel affinity resins (NTA; QIAGEN, Valencia, CA) in a column at a flow rate of 1–2 ml/min. The resins were washed with lysis buffer until the OD₂₈₀ returned to the buffer baseline. Bound polyhistidine-tagged proteins were eluted under nondenaturing conditions by competition with 100 mM imidazole. For PR, DTT (1 mM), zinc chloride (1 µM), and MgCl₂ (1 mM) were added immediately to stabilize receptor DNA and steroid binding activity. For HMGB-2, 1 mm DTT was added to the eluate. Samples were dialyzed in elution buffer minus imidazole and were stored at –80 C in aliquots.

Murine Stat5a and murine Jak2 were coexpressed from baculovirus vectors in Sf9 insect cells to yield expression of tyrosine-phosphorylated Stat5a. Sf9 cells grown in spinner flask suspension culture were coinfected with baculoviruses encoding Stat5a and Jak2 with a multiplicity of infection of 2 and 10 PFU/cell, respectively. At 48 h post infection, cells were harvested and lysed for 15 min on ice in the following buffer: 20 mM HEPES (pH 7.9); 50 mM sodium fluoride; 1 mM EDTA; 1 mM sodium vanadate; 2 mM DTT; 0.4 M NaCl; and a cocktail of protease inhibitors. Lysates were centrifuged at 40,000 rpm for 30 min and dialyzed in lysis buffer containing 50 mM NaCl, and supernatant was loaded on a heparin Sepharose (Pharmacia, Uppsala, Sweden) column and eluted with a NaCl (50 mM to 1 M) gradient in lysis buffer. Fractions were analyzed by silver stain and immunoblotting using anti-Stat5 antibody L-20 (Santa Cruz Biotechnology). Fractions containing Stat5a protein were assayed for functional Stat5 protein using a phosphotyrosine-specific antibody to Stat5a/b (Tyr694/699; Upstate Biotechnology).

EMSA
A 28-bp oligonucleotide containing a PRE/glucocorticoid response element derived form the MMTV-long-term repeat and a 35-bp oligonucleotide containing the proximal region of the rat ß-casein gene promoter (–110 to –70) were used for EMSAs. The sequences of the coding strand: wild type, 5'-TAATCATGTGGACTTCTTGGAATTAAGGGACTTTT-3'; GASmut, 5'-TAATCATGTGGACTTCTTTTAATTAAGGGACTTTT-3'; GREmut, 5'-TAATCAAGCTTACTTCTTGGAATTAACAGACTTTT-3'; GRd, 5'-TAATCAAGCTTACTTCTTGGAATTAAGGGACTTTT-3'; GRe, 5'-TAATCATGTGGACTTCTTGGAATTAACAGACTTTT-3'. For PR/Stat5 cobinding experiments the sequence of the coding strand for ß-casein probe was 5'-GATCCATGTGGACTTCTTGGAATTAAGGGACTTTT-3'. Oligonucleotides were designed with 4-bp overhangs and were labeled by Klenow polymerase filling of the overhangs with ³²P-labeled deoxynucleoside triphosphates (NEN Life Science Products, Boston, MA). Recombinant purified PR and Stat5a were incubated for 1 h at 4 C with ³²P-labeled DNA (0.3 ng) in a total reaction volume of 25 µl. Also included was 100 ng of poly (dA-dT) nonspecific competitor DNA. The DNA binding buffer contained 10 mM Tris-base, pH 7.4, 50 mM NaCl, 5 mM dithiothreitol, 2 mM MgCl₂, 10% glycerol, and 50 ng/ml of carrier ovalbumin. Samples (25µl) were electrophoresed on 5% polyacrylamide gels prepared at 40:1 (wt/wt) acrylamide-bis-acrylamide ratio using 20 mM Tris-acetate, 0.5 mM EDTA as the electrode buffer. To maintain constant temperature during electrophoresis, 4 C water was recirculated through the gel apparatus. Gels were dried and subjected to autoradiography. Quantification of protein-DNA complexes was carried out by direct scanning of dried gels for radioactivity using a series 400 PhosphorImager (Molecular Dynamics, Sunnyvale, CA) and quantitated by Molecular Dynamics ImageQuant Program.

Immunoblot Analysis
Cell lysates containing equal amounts of total protein were solubolized in 1% sodium dodecyl sulfate (SDS) sample buffer and electrophoresed on 8–12% polyacrylamide SDS gels. Separated proteins were transferred to Immobilon-P PDVF membranes (Millipore Corp., Bedford, MA) and incubated with appropriate antibodies in TBS-T (50 mM Tris-HCl, pH 7.4; 150 mM NaCl; and 0.1% Tween 20). Antibody interaction was detected by enhanced chemiluminescence (Amersham Pharmacia Biotech, Arlington Heights, IL).

Adenovirus
Mouse PR-B and mouse PR-A cDNA were cloned into the pShuttle-CMV adenovirus shuttle vector at the HindIII/EcoRV sites of the multiple cloning cassette as previously described (71). Competent E. coli cells carrying pA