A common pro-opiomelanocortin-binding element mediates leukemia inhibitory factor and corticotropin-releasing hormone transcriptional synergy.

Using murine AtT20 pituitary cells transfected with a rat pro-opiomelanocortin (POMC) promoter (-706/+64) linked to the luciferase reporter, we showed leukemia inhibitory factor (LIF) to strongly potentiate corticotropin-releasing hormone (CRH) induction of POMC gene expression. We therefore tested mechanisms for molecular interactions between LIF and CRH. Although LIF and CRH synergized to induce an 8-fold induction of POMC transcription, CRH alone (but not LIF) induced cAMP response element-binding protein phosphorylation (5-fold) or an increase of c-fos mRNA levels (>100-fold), suggesting that these pathways are not implicated in LIF transcriptional synergistic effects. Using a DNase I footprint assay, POMC promoter regions protected by AtT20 cell nuclear extracts were identified (-121/-109, and -143/-134, and -173/-160). The protected -173/-160 element fused to a heterologous promoter conferred LIF-CRH synergy (6.5-fold induction of POMC) and formed a specific complex with AtT20 cell nuclear extracts. This complex was supershifted by an anti-phosphoserine antibody, and a serine/threonine kinase inhibitor also altered both this complex and LIF-CRH transcriptional synergy on the POMC promoter-luciferase reporter construct, indicating that these events depend on post-translational serine phosphorylations. LIF-CRH synergy on POMC transcription is therefore mediated at least in part by -173/-160 sequences conferring confluent transcriptional activity of both peptides.

activates the c-fos proto-oncogene, implying cAMP and protein kinase A or Ca 2ϩ and calmodulin kinase signaling (5). Furthermore, a c-fos-binding element in POMC exon 1 (indistinguishable from an AP-1-binding site) has recently been identified to be responsible for part of the CRH transactivating effect on POMC transcription (4). Another CRH-inducible element centered at position Ϫ166 on the POMC promoter has also been described and could be responsible for c-fos-independent induction of POMC transcription by CRH (6).
The development and function of differentiated pituitary neuroendocrine cells are regulated both by hypothalamic trophic hormones and by intrapituitary cytokines and growth factors. Several cytokines stimulate the hypothalamic-pituitaryadrenal axis in vivo and pituicyte ACTH production in vitro (7)(8)(9)(10). LIF is a pleiotropic cytokine, whose gene expression has recently been reported in human fetal and adult and murine pituitary cells (11,12). The biological functions of LIF are mediated through its binding to a high-affinity cell-surface LIF-receptor complex, consisting of a low-affinity LIF-binding subunit (LIF-R) and a gp130 subunit (13). Ligand activation of LIF-R is followed by gp130 and LIF-R heterodimerization and subsequent activation of tyrosine kinases (14). The Janus-activated kinase family of transmembrane receptor-associated tyrosine kinases is phosphorylated by LIF (15), leading to subsequent activation and translocation into the nucleus of STAT1␣ (signal transducer and activator of transcription) and STAT3 (16 -19).
LIF also synergizes with CRH to induce secretion of ACTH (20-fold) and to stimulate POMC gene expression (8-fold) in AtT20 cells (16). However, the mechanism for this potent synergy is difficult to explain as no specific cytokine response elements have thus far been identified in the POMC gene. In this study, we therefore explored mechanisms of interaction between the LIF and CRH signaling cascades to determine points of confluence that might explain the synergy exerted by both peptides on POMC transcription. We report here that LIF does not influence the cAMP/c-fos-dependent activation of POMC transcription by CRH. However, a LIF-and CRH-responsive element has been identified in the POMC gene (Ϫ173/ Ϫ160) and is defined to be responsible for mediating LIF-CRH synergy on POMC transcription.

MATERIALS AND METHODS
Cell Culture-AtT20 cells, obtained from the American Type Culture Collection (Rockville, MD), were grown in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum, 2 mM L-glutamine, 100 units/ml streptomycin, 100 units/ml penicillin, and 0.25 g/ml amphotericin B (Life Technologies, Inc.).
Western Analysis-AtT20 cells were grown in 10-ml dishes to 80% confluency and serum-deprived 24 h before a 30-min treatment with LIF (R&D Systems, Minneapolis, MN), CRH (American Peptide Co., Sunnyvale, CA), LIF and CRH, forskolin, or phorbol 12-myristate 13-acetate (Sigma). Cells were then lysed by scraping into 2 ϫ SDS-polyacrylamide * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. gel electrophoresis loading buffer. Samples were separated by electrophoresis on 10% SDS-polyacrylamide gel, and proteins were transferred to polyvinylidene difluoride membrane (Millipore Corp., Milford, MA). The membrane was blocked in 5% nonfat milk. Anti-phospho-CREB detection was carried out with polyclonal anti-phosphorylated CREB antibody (Upstate Biotechnology, Inc., Lake Placid, NY) in blocking buffer for 14 h at 4°C. Anti-rabbit Ig conjugated to horseradish peroxidase (Amersham Corp.) was incubated with the membrane in blocking buffer, and detection was accomplished using ECL reagent (Amersham Corp.) as suggested by the manufacturer. To reprobe the membrane, the antibody was stripped using 100 mM ␤-mercaptoethanol, 2% SDS, and 62.5 mM Tris-HCl (pH 6.7) at 50°C for 30 min.
Northern Analysis-Poly(A) RNA was prepared from AtT20 cells treated with LIF, CRH, or both using trizol reagent (Life Technologies, Inc.). 10 g of RNA/lane was size-fractionated under denaturing conditions using formaldehyde-agarose (1.2%) gel, transferred to nitrocellulose, and hybridized as described (20). A cDNA fragment of c-fos was used as a template for a 32 P-labeled probe generated using random primers.
Nuclear Extracts-AtT20 cells were grown to 80% confluency and serum-deprived 24 h before a 1-h treatment with or without LIF, CRH, or both. Cells were then harvested in cold phosphate-buffered saline. Nuclear extracts were prepared as described (21). Final concentrations were typically 3 g/l as determined by Bio-Rad protein assay.
DNase I Footprinting-A double-stranded DNA fragment was prepared by polymerase chain reaction. Briefly, the antisense primer was end-labeled using T 4 polynucleotide kinase and [␥-32 P]dATP (6000 mCi/ mmol; DuPont NEN), purified through a Bio-Spin 30 chromatography column (Bio-Rad), and used in the polymerase chain reaction. The Ϫ351/Ϫ7 region of the rat POMC promoter was also amplified and purified on a 1.3% agarose gel. DNA was recovered using a USBioclean kit.
Nuclear extracts (0 -50 g) were incubated for 20 min at room temperature with 1-2 ng of 32 P-labeled DNA (25,000 cpm) in binding buffer (22 mM HEPES (pH 7.9), 6 mM KCl, 10 mM dithiothreitol, 5 mM spermidine, 8% glycerol, and 2% Ficoll) with 2 g of poly(dI-dC). Samples were briefly treated (1 min at room temperature) with DNase I (Pharmacia Biotech Inc.), titrated to determine a partial digestion pattern. Digestion was terminated by addition of 3 volumes of 768 mM sodium acetate, 128 mM EDTA, 0.56% SDS, and 256 g/ml yeast RNA and by treatment (5 min at 55°C) with 2 g of proteinase K. DNA was also extracted with phenol/CHCl 3 and ethanol-precipitated. Reaction products were then analyzed on a standard 8 M urea, 8% polyacrylamide gel adjacent to a sequencing ladder (G ϩ A) generated from the same end-labeled DNA fragment to precisely identify borders of DNase protection.
10 pmol of double-stranded oligonucleotide were end-labeled with [␥-32 P]dATP using T 4 polynucleotide kinase, purified through a Bio-Spin 30 chromatography column, resolved on a 10% acrylamide gel, and eluted from the gel in 10 mM Tris-HCl (pH 8) and 1 mM EDTA. Approximately 30 pg of labeled DNA (20,000 -30,000 cpm) were added to the nuclear extracts.
For the binding assays, 7.5 g of nuclear extracts were preincubated for 15 min at room temperature in 20 l of binding buffer with 1 g of poly(dI-dC). 32 P-Labeled probe was added, and the binding reaction was left at room temperature for 15 min. In competition experiments, a 100-fold molar excess of unlabeled competitor oligonucleotides was added to the preincubation reaction. Antibody experiments were performed using 1 l of anti-phosphotyrosine antibody (Santa Cruz Bioreagents, Santa Cruz, CA) or 1 l of anti-phosphoserine antibody (Sigma), which was added to the preincubation reaction and incubated for 1 h at 4°C. The protein-DNA complexes were resolved on a 4% polyacrylamide gel in 0.5 ϫ Tris borate/EDTA. Gels were dried and autoradiographed with intensifying screens at Ϫ70°C with Amersham Hyperfilm-MP films.
Mutagenesis and Plasmid Construction-Deletion of the Ϫ167/Ϫ117 region of the POMC promoter was performed with an ExSite polymerase chain reaction-based mutagenesis kit (Stratagene Cloning Systems, La Jolla, CA) according to manufacturer's instructions. The Ϫ173/Ϫ160 element-luciferase plasmid was made by synthesizing two complementary DNA sequences (5Ј-tcgaTCTGCTGTGCGCGCAGC-3Ј), annealing, and ligating them into pGL3 (Promega) linearized at the KpnI/SacI site.
All constructs were sequenced for verification.
Transfection and Luciferase Assay-AtT20 cells were plated in 2-ml dishes (100,000 cells/well) and allowed to adhere for 24 h. Cells were transfected as described (16), and 24 h after transfection, they were treated in triplicate for 6 h. After cell lysis, the luciferase activity was measured in a Berthold Lumat LB 9501 luminometer (Wallac, Gaithersburg, MD). Cotransfection of the ␤-galactosidase reporter showed transfection efficiency to vary Ͻ15% within a given experiment.

RESULTS
Synergy of LIF-CRH on POMC Transcription-Using the rat POMC promoter (Ϫ706/ϩ64) fused to the luciferase reporter gene, we previously showed that LIF induced CRH action on rat POMC transcription (16). To determine the specificity of LIF action, the transcriptional effects of other cytokines, notably TNF␣ and IL-1␤, which are also induced by acute inflammatory insults (22,23), were examined. AtT20 cells transiently transfected with the POMC promoter-luciferase construct were treated for 6 h with LIF (1 nM), CRH (10 nM), TNF␣ (1 ng/ml), IL-1␤ (100 pg/ml), or combinations of CRH and one of the other cytokines. Neither TNF␣ nor IL-1␤ alone exerted transcriptional effects, nor did they alter CRH induction of POMC ( Fig.  1). Thus, there does not appear to be redundancy in the response of POMC to these inflammatory cytokines, i.e. the LIF synergistic effect with CRH on POMC transcription appears selective. Presumably, the specificity of the LIF response is determined by LIF receptor expression (11). Urocortin, a novel neuropeptide related to both urotensin and CRH, activates CRH receptors and, in vivo, is a more powerful stimulant of the hypothalamic-pituitary-adrenal axis than CRH (24). We therefore examined urocortin transcriptional activity and demonstrated that it could also activate POMC transcription as well as potentiate LIF effects on POMC (Fig. 1). These results provide evidence for the role of CRH receptor subtype I (25) in modulating LIF action. Therefore, to determine points of confluence between the two signaling cascades, we explored the hypothesis by which LIF could enhance CRH signaling pathways leading to POMC transcription and ACTH secretion.
LIF and CRH Effects on the cAMP Pathway-CRH induction of POMC transcription was reported to be mediated through a cAMP pathway (3, 26 -29), although LIF itself did not alter
CREB Phosphorylation-As CRH stimulates accumulation of cAMP and protein kinase A activation in target cells (16,26), we tested whether LIF altered CREB phosphorylation. Using an antibody that exclusively recognizes the phosphorylated form of CREB (33), CRH (10 nM), but not LIF (1 nM), induced a striking increase (Ͼ5-fold) in the proportion of intracellular phosphorylated CREB within 30 min (Fig. 2): a doublet that migrated with the 43-kDa marker, the expected size for CREB, was detected by an anti-phospho-CREB antibody. These two inducible bands may be CREB␣ and CREB⌬, which differ in sequence by 14 amino acids (33). These changes were not accompanied by increases in the amount of total intracellular CREB, as measured by an antibody recognizing all forms of CREB. Interestingly, although LIF and CRH powerfully synergize to stimulate POMC expression, LIF does not further enhance CREB phosphorylation over that observed with CRH treatment. We concluded that CREB phosphorylation is therefore not a molecular event responsible for LIF synergistic effects on CRH-induced POMC transcription.
c-fos Expression-c-fos is a member of a family of related acute-phase transcription factors that form heterodimers to bind DNA and activate transcription (34). c-fos mRNA accumulation occurs during treatment of AtT20 cells with CRH, and overexpression of c-fos results in enhanced POMC transcription (5). In addition, LIF may, in some systems, enhance c-fos transcription (35). Therefore, changes in c-fos mRNA were analyzed after cells were treated for 15 min with LIF and CRH (1 and 10 nM, respectively). Striking inductions (Ͼ100-fold) of c-fos mRNA after CRH treatment were observed, in keeping with previous reports (5). However, in these cells, LIF alone failed to induce c-fos expression and furthermore failed to enhance the CRH effect on c-fos mRNA levels (Fig. 3). Therefore, although c-fos may play a role in mediating CRH action, it does not appear to be important for LIF action or, more significantly, for mediating the strong synergy between LIF and CRH.
In addition, c-fos has been shown to stimulate POMC transcription through an AP-1 site located in exon 1 of the POMC promoter. Using this site as a probe in the gel-shift assay or a construct containing the POMC promoter mutated at this site and fused to the luciferase reporter, we showed this site not to be critical in the context of LIF and CRH synergy on POMC transcription (data not shown).
LIF Action on CRH-induced POMC Transcription, Implying a c-fos-independent Pathway-CRH may also induce POMC transcription and ACTH secretion by a c-fos-independent pathway (4). A region in the POMC promoter (Ϫ171/Ϫ160), different from the exon 1 AP-1 site, was shown to confer strong CRH stimulation on POMC transcription (6). We therefore explored whether this element could be the point of convergence of the CRH and LIF pathways.
Using a DNase I footprint assay, we examined POMC promoter regions whose DNase sensitivity was protected by bound nuclear proteins present in AtT20 whole cell extracts previously treated for 1 h with CRH (10 nM), LIF (1 nM), and LIF ϩ CRH (1 and 10 nM, respectively) (Fig. 4). We focused on Ϫ351/Ϫ7 footprinting sequences as this region contains the important regulatory element for the synergy of LIF-CRH on POMC transcription (16).
50 g of AtT20 cell nuclear extracts protected three elements (Ϫ109/Ϫ121, Ϫ134/Ϫ143, and Ϫ173/Ϫ160) in this region, both in the basal state and after stimulation of the cells by LIF, CRH, or both agents together. These regions have previously been designated as potential overlapping AP-2-and NF-Bbinding sites (Ϫ134/Ϫ143 element) (6) and as a mouse metallothionein metal regulatory element (Ϫ173/Ϫ160 element) (6). However, we also found that no additional regions were further protected after treatment of the cells with both LIF and CRH compared with LIF or CRH alone.
To further analyze whether these regions could be important for the synergy of LIF-CRH on POMC transcription, we per- formed functional assays by transfecting AtT20 cells treated with LIF (1 nM) and CRH (10 nM) for 6 h with the rat POMC promoter lacking the Ϫ167/Ϫ117 region and fused to a luciferase reporter gene (Fig. 5). This deletion reduced basal promoter activity and also severely reduced the LIF-CRH synergy (61.5%) on reporter activity. We therefore fused the Ϫ173/Ϫ160 element to the same luciferase reporter gene with a heterologous SV40 promoter (Fig. 5). We focused on this element first because of its identification during footprint analysis and also because of previous work demonstrating that this region was bound by a CRH-inducible factor and was the major CRHresponsive element in the POMC promoter (6). As expected, CRH exerted its positive regulation of transcription, but surprisingly, LIF and CRH together strongly enhanced this induction (6.5-fold), thus conferring LIF responsiveness to the heterologous construct.
To confirm these results, we analyzed binding protein states in the Ϫ173/Ϫ160 region by using AtT20 whole cell extracts (after cell treatment with LIF, CRH, or both) and a radioactive probe comprising this element using a gel-shift assay. Nuclear extracts from control, CRH-treated (10 nM, 60 min), LIF-treated (1 nM, 60 min), or (LIF ϩ CRH)-treated (1 and 10 nM, respectively; 60 min) AtT20 cells were prepared as described under "Materials and Methods." These extracts were also shown to bind efficiently to oligonucleotide A, containing the region (Ϫ173/ Ϫ160) important for the LIF ϩ CRH induction of POMC transcription (Fig. 6A). In fact, three complexes were shifted, and therefore, to determine the exact binding localization of each on the A probe, two oligonucleotides were designed: B and C. After addition of B and C sequences, the entire A nucleotide sequence was covered without overlap, while A-C were the same size. From these experiments, we concluded that only the slowest migrating band (designated Y) was specific for the Ϫ167/Ϫ158 region as it appeared only with the A and C probes (Fig. 6, A and C, respectively). The "intermediate" complex (designated I) seemed to bind the Ϫ175/Ϫ168 region (overlapping region between the A and B probes) (Fig. 6, A and B), but was not further considered because of its diffuse nature. The faster migrating band (designated L) was not specific for any region as it appeared with all three probes (A-C) (Fig. 6, A-C).
To analyze the binding specificity of the Y complex on the A probe, competition experiments were performed (Fig. 7). The Y complex was shown to be efficiently competed by unlabeled oligonucleotide A (100-fold molar excess) (Fig. 7, lanes 5 and 7), whereas an AP-1 oligonucleotide was inefficient in competing (lanes 9 -11). The binding specificity of the Y complex on the Ϫ167/Ϫ158 motif was further confirmed by a less efficient competition with a unlabeled oligonucleotide similar to oligo-nucleotide A except that it contained a mutation at GCGC (Ϫ167/Ϫ164) (Fig. 7, lanes 12-14). This Ϫ167/Ϫ164 region must therefore be at least a component of the binding site of the Y complex. In addition, competition experiments with unlabeled oligonucleotides B and C further strengthened the binding specificity of the Y complex on the Ϫ167/Ϫ158 region as oligonucleotide B did not compete this binding, whereas oligonucleotide C competed effectively (Fig. 6A, lanes 6 and 7).
In addition, Figs. 6A and 7 show that greater amounts of the shifted Y complex were detected in assays that contained nuclear extracts from AtT20 cells treated with CRH for 60 min as compared with untreated cell extracts. No quantitative enhancement of the shifted Y complex was observed with extracts from AtT20 cells treated with LIF ϩ CRH for 60 min as compared with those treated with CRH alone for 60 min. We also concluded that LIF, at the concentrations used and for the duration of the treatments, did not further enhance CRHinduced complex shifts or the amount of Y complex. LIF and CRH may also act through the same binding element (Ϫ173/ Ϫ160) by activating similar transcription factor(s) by modify-FIG. 5. Identification of a POMC promoter element mediating the synergy of LIF-CRH on POMC transcription. The luciferase assay was carried out using the rat POMC promoter-luciferase construct lacking the Ϫ167/Ϫ117 region (open bars) or the Ϫ173/Ϫ160 element of the POMC promoter fused to the SV40-luciferase reporter gene (shaded bars). AtT20 cells were transfected with these constructs and treated with 1 nM LIF, 10 nM CRH, or both. Luciferase activity was measured as described under "Materials and Methods." Closed bars, wild-type promoter.
ing their conformation or phosphorylation state. Indeed, no significant correlation was observed between the quantitative binding state of these proteins on the Ϫ173/Ϫ160 element after cell treatment with CRH, LIF, or both or in the corresponding stimulation of POMC transcription in experiments where the Ϫ173/Ϫ160 element was fused to the heterologous SV40-luciferase reporter gene.
The time-dependent activation of the Y complex by both cytokines suggested a post-translational regulation. To explore this hypothesis, we analyzed POMC transcription in the luciferase assay as well as Y complex formation in the gel-shift assay with or without pretreating AtT20 cells for 30 min with 10 g/ml cycloheximide. This inhibitor did not alter these LIFand CRH-induced events (data not shown), demonstrating that ongoing protein synthesis is not required. To investigate whether phosphorylated events are critical, supershift experiments were performed using anti-phosphoprotein antibodies and showed the Y complex to contain protein(s) phosphorylated on serine residues, but not on tyrosine. Indeed, the Y complex was supershifted by an anti-phosphoserine antibody (Fig. 8A,  lane 12) and not by an anti-phosphotyrosine antibody (lane 13) or by preimmune serum (lane 11). Several complexes were in fact supershifted by the anti-phosphoserine antibody; however, only the two upper bands (designated ␣ and ␤) are specific since they did not appear in a control reaction containing the A probe with the anti-phosphoserine antibody but without AtT20 cell nuclear extracts (data not shown). The appearance of at least two supershifted complexes suggests that different serinephosphorylated proteins are components of the Y complex with DNA. Furthermore, the presence of these complexes even in the untreated reaction reveals that the signaling pathway lead-ing to these serine phosphorylations is constitutively activated. To provide further functional confirmation for these observations, AtT20 cells were pretreated for 1 h with a 0.5 M concentration of a serine/threonine inhibitor, staurosporine, before adding LIF or CRH, and Y complex formation was tested or luciferase activity was measured (Fig. 8, B and C, respectively). LIF-, CRH-, or (LIF ϩ CRH)-induced POMC transcription was abolished by staurosporine, and Y complex formation was severely reduced after staurosporine treatment, whereas this agent only modestly altered these parameters in control cells. Furthermore, treatment of AtT20 cells with 100 g/ml genistein, a specific tyrosine kinase inhibitor, did not alter LIF and CRH induction (data not shown). These results indicate that LIF-and CRH-triggered signaling pathways involve the specific action of serine/threonine protein kinases.

DISCUSSION
Our earlier work had shown the pleiotropic cytokine LIF to demonstrate a very robust interaction with CRH in potentiating its effects on POMC transcription (16). However, these two peptides exerted opposing effects on a number of indices of cell proliferation, including S phase accumulation (36). These either synergistic effects on POMC transcription or antagonistic effects on the cell cycle indicated complex interactions between both peptides.
It is also relevant that the CRH type II receptor ligand urocortin (24) has similar synergistic effects on the rat POMC promoter, presumably acting through the type I receptor (25). However, the synergistic effect appears to be highly specific to LIF as two unrelated inflammatory cytokines, IL-1␤ and TNF␣, had no effect on POMC alone and did not alter the stimulation caused by CRH.
These studies were performed to dissect interactions between the LIF and CRH signaling cascades and to identify potential intracellular loci where the observed alterations in the target gene (i.e. POMC) activity could occur. Cross-talk between signaling cascades is being increasingly recognized as important for fine-tuning cellular responses to extracellular signals (37). The regulation of POMC expression by both CRH and LIF appears to be a novel model for studying cross-talk between two different signaling pathways. Previously, inhibition of cytokine signaling by the cAMP pathway has been described (38), making the potentiated responses we observed of even greater interest. The results shown here indicate a complex pattern of signaling interaction between LIF and CRH in AtT20 cells.
CRH is the major positive regulator of the POMC transcriptional response (2). Although its primary effects are an increase in cAMP levels and activation of protein kinase A (16,26), leading to phosphorylation of CREB, the induction of phospho-CREB by CRH has not previously been reported. However, our results show that CRH signaling through phospho-CREB is not altered by LIF alone or in combination with CRH, and consequently, this signal is apparently not a mechanism for both peptides to exert their synergy on POMC transcription.
A putative downstream target of both LIF and CRH is the proto-oncogene c-fos. c-fos mediates a component of the CRH activation of POMC transcription through a c-fos-responsive sequence in exon 1 of the POMC promoter and was therefore a rational end point to examine. We were surprised that LIF had no effect on c-fos mRNA in pituicytes, whereas CRH clearly was a powerful stimulant at the mRNA level. LIF did not enhance c-fos mRNA and, in fact, inhibited the marked increase observed with CRH. Thus, LIF and CRH exhibit antagonistic effects on c-fos mRNA, similar to the discordant effect we have previously observed for cell cycle progression in AtT20 cells (36).
Our previous studies had indicated that the cytokine-responsive region on the POMC gene was situated in the 5Ј-promoter sequence between positions Ϫ323 and Ϫ166 (16). To further dissect this region, a footprint analysis was performed and showed three elements whose DNase sensitivity was protected by AtT20 cell nuclear extracts. Also, further deletions between positions Ϫ167 and Ϫ117 led to severe abrogation of LIF-CRH synergy on POMC transcription. In addition, fusion of the Ϫ173/Ϫ160 element to a heterologous SV40-luciferase reporter gene led to enhancement of POMC transcription by both peptides to the same extent as the wild-type promoter. This implied that both LIF and CRH act mostly through closely apposed or similar DNA elements present within the Ϫ173/Ϫ160 region of the POMC promoter.
To confirm these results, dynamic binding assays, using a radiolabeled probe covering the Ϫ173/Ϫ160 sequences and AtT20 cell nuclear extracts, were performed and showed no significant differences between binding states induced by CRH alone or in combination with LIF. We conclude that both peptides may stimulate two signaling pathways, converging distally and leading to enhancement of similar transcription factor(s) binding to the Ϫ173/Ϫ160 element.
Involvement of a similar promoter element mediating two different signaling pathways to enhance transcription of a common target gene has been reported (39,40). For example, the acute-phase response element fused to a heterologous promoter conferred responsiveness to both interferon-␥ and IL-6 in HepG2 cells (39). Furthermore, an 18-base pair core region of the rat ␣ 2 -macroglobulin gene was sufficient to confer both 5, and 7). 32 P-Labeled oligonucleotide A was used as the probe. C, rat POMC reporter-luciferase-transfected AtT20 cells untreated or treated for 6 h with 1 nM LIF, 10 nM CRH, or a combination of both with (shaded bars) or without (closed bars) a 1-h pretreatment with 0.5 M staurosporine. Luciferase activity was measured.

FIG. 8. Phosphorylation events on serine/threonine residues are required for LIF-and CRH-enhanced POMC transcription.
A, gel-shift analyses with nuclear extracts (7.5 g) from untreated (control (C)); lanes 1, 2, and 9), LIF-treated (1 nM, 60 min; lanes 3 and 4), CRH-treated (10 nM, 60 min; lanes 5 and 6), and (LIF ϩ CRH)treated (1 and 10 nM, respectively; 60 min; lanes 7, 8, and 10 -13) AtT20 cells. 32 P-Labeled oligonucleotide A was used as the probe. An antiphosphoserine antibody (Ab ␣-PSerine) was added to the binding reaction in lanes 2, 4, 6, 8, and 12 and shifted the Y complex, while an anti-phosphotyrosine antibody (Ab ␣-PTyrosine; lane 13) or nonimmune serum (lane 11) added under the same conditions did not. B, gel-shift analyses with nuclear extracts (7.5 g) from untreated (lanes 1 and 2), LIF-treated (1 nM, 60 min; lanes 3 and 4), CRH-treated (10 nM, 60 min; lanes 5 and 6), and (LIF ϩ CRH)-treated (1 and 10 nM, respectively; 60 min; lanes 7 and 8) AtT20 cells. Before cytokine treatment, cells were preincubated for 60 min with 0.5 M staurosporine (Sigma) (lanes 1, 3, induction by IL-6 and synergy of IL-6 plus glucocorticoids to a minimal promoter (40). However, in these latter cases, two separate transduction pathways leading to the activation of distinct transcription factors converged on a common element on the promoter to give a synergistic effect. We questioned therefore whether LIF and CRH pathways converge upstream of the transcription factor level, leading to enhancement of the same binding protein, or downstream, directly on the Ϫ173/ Ϫ160 element by activating different transcription factors. The notion of different transcription factors whose homo-or heterodimerization or coordinated effects are responsible for transcriptional synergistic induction between cytokines is now well recognized (41,42). However, LIF-, CRH-, or (LIF ϩ CRH)induced activation of POMC transcription is insensitive to protein synthesis inhibition by cycloheximide, occurring therefore at the post-translational level, and could be prevented by the protein kinase inhibitor staurosporine, known to inhibit most serine/threonine protein kinases. Furthermore, a supershift binding assay with an anti-phosphoserine antibody showed the Y complex to contain several serine-phosphorylated proteins. However, since these proteins are phosphorylated and bind DNA constitutively in the basal unstimulated state, we propose that LIF and CRH synergy occurs through an enhancement of the phosphorylation level of these proteins and/or through other conformational modifications. We therefore conclude that LIF and CRH may activate either a common binding protein or distinct transcription factors, whose activation through serine/ threonine phosphorylation events may influence their common transcriptional activity. Requirement of serine phosphorylation events for formation of the activated homodimer STAT3-STAT3, which binds DNA, has similarly been reported (43). However, STAT3 and STAT1␣ are not implicated in LIF-and CRH-triggered synergy on POMC transcription since neither STAT3 nor STAT1␣ antibodies were able to supershift the Y complex in binding assays (data not shown).
In summary, we have demonstrated that LIF and CRH effectively synergize to induce POMC transcription in AtT20 cells and that LIF and CRH effects are not dissociated and are centered on the Ϫ173/Ϫ160 element. Interestingly, this element was previously shown to bind a novel CRH-inducible transcription factor, a metallothionein protein, whose binding may transactivate POMC gene expression (6). This protein could also be a molecular target for the convergence of LIF and CRH pathways. Furthermore, LIF and CRH synergy appears to occur post-translationally and to involve serine/threoninemediated phosphorylation events. We have thus identified and characterized a novel interaction between a G protein-coupled receptor and a cytokine receptor that results in synergy at the level of POMC gene transcription.