Cyclic AMP-induced Forkhead Transcription Factor, FKHR, Cooperates with CCAAT/Enhancer-binding Protein β in Differentiating Human Endometrial Stromal Cells*

Decidual transformation of human endometrial stromal (ES) cells requires sustained activation of the protein kinase A (PKA) pathway. In a search for novel transcriptional mediators of this process, we used differential display PCR analysis of undifferentiated primary ES cells and cells stimulated with 8-bromo-cAMP (8-Br-cAMP). We now report on the role of forkhead homologue in rhabdomyosarcoma (FKHR), a recently described member of the forkhead/winged-helix transcription factor family, as a mediator of endometrial differentiation. Sustained 8-Br-cAMP stimulation resulted in the induction and nuclear accumulation of FKHR in differentiating ES cells. Immunohistochemical studies revealed that endometrial stromal expression of FKHR in vivo is confined to decidualizing cells during the late secretory phase of the cycle and coincides with the expression of CCAAT/enhancer-binding protein β (C/EBPβ). Reporter gene studies showed that FKHR potently enhances PKA-dependent activation of the tissue-specific decidual prolactin (dPRL) promoter, a major differentiation marker in human ES cells. Transcriptional augmentation by FKHR was effected through functional cooperation with C/EBPβ and binding to a composite FKHR-C/EBPβ response unit in the proximal promoter region. Furthermore, FKHR and C/EBPβ were shown to interact directly in a glutathione S-transferase pull-down assay. These results provide the first evidence of regulated expression of FKHR and demonstrate that FKHR has an integral role in PKA-dependent endometrial differentiation through its ability to bind and functionally cooperate with C/EBPβ.

During the menstrual cycle, ovarian estradiol and progesterone stimulate the ordered growth and differentiation of endo-metrial tissue compartments. In the human, this includes synchronous growth and coiling of the spiral arteries, secretory transformation of glandular epithelium, migration of bone marrow-derived cells, and decidualization of the stroma, which is thought to be essential for blastocyst implantation and subsequent formation of a hemochorial placenta. Decidualization of human endometrial stromal (ES) 1 cells represents a process of morphological differentiation accompanied by distinct biochemical phenotypic changes. Decidual transformation is first apparent in stromal cells surrounding the spiral arteries approximately 10 days after the postovulatory rise in ovarian progesterone levels, indicating that the expression of deciduaspecific genes is unlikely to be under direct transcriptional control of activated steroid hormone receptors.
There is compelling evidence to suggest that initiation of the decidual process requires elevated intracellular cAMP levels and sustained activation of the protein kinase A (PKA) pathway (1)(2)(3)(4)(5). Expression of PRL, under control of the tissuespecific decidual PRL (dPRL) promoter (3,6), by ES cells coincides with decidual differentiation and is widely used as a biochemical marker of this process (2)(3)(4)(5)7). Previous studies have shown that CCAAT/enhancer-binding protein ␤ (C/ EBP␤), a member of the C/EBP subfamily of basic region/ leucine zipper transcription factors, is induced during ES cell differentiation (4). Furthermore, C/EBP␤ participates in the formation of a nucleoprotein complex that binds the proximal dPRL promoter region upon PKA activation (4). Although other members of the C/EBP family are expressed in cultured human ES cells, including C/EBP␣ and ␦, C/EBP␤ is the major and cAMP-inducible form (4).
To establish the identity of additional factors relevant to the decidual process, we used differential display PCR and isolated FKHR (forkhead homologue in rhabdomyosarcoma) as a cAMPinducible gene in differentiating human ES cells. Forkhead or "winged helix" transcription factors have been shown to play important roles in cell differentiation, embryogenesis, and oncogenesis (8,9). FKHR was first identified as a transcription factor involved in a translocation with PAX3 in alveolar rhab-domyosarcoma (10 -12). It also has an important role in apoptosis, glucose homeostasis, cell cycle regulation, and as a nuclear receptor cofactor (13)(14)(15)(16)(17)(18). We now report that FKHR is induced in decidualizing endometrium and participates in PKA signal transduction through its ability to interact and transcriptionally cooperate with C/EBP␤.

EXPERIMENTAL PROCEDURES
Primary ES Cell Culture-ES cells were isolated from normal proliferative endometrial tissues obtained from cycling women by endometrial biopsy at the time of diagnostic laparoscopy and hysteroscopy. The Hammersmith and Queen Charlotte's Hospital Research and Ethics Committee approved the study, and patient consent was obtained before biopsy. Samples were collected in Earle's buffered saline containing 100 units/ml penicillin and 100 g/ml streptomycin. After enzymatic digestion, the stromal cells were separated from epithelial cells and passed into culture as described previously (2,7). Proliferating ES cells were cultured in maintenance medium of Dulbecco's modified Eagle's medium/F-12 containing 10% dextran-coated charcoal-treated FBS (DCC-FBS) and 1% antibiotic-antimycotic solution. Confluent monolayers were treated in Dulbecco's modified Eagle's medium/F-12 containing 2% DCC-FBS with 0.5 mM 8-Br-cAMP to induce a differentiated phenotype. All experiments were carried out before the fourth cell passage.
RNA Isolation and Analysis-Total RNA was extracted, using RNAzol B (Biogenesis), from primary cultures of ES cells after 12-h treatment with or without 0.5 mM 8-Br-cAMP. The experiment was carried out in duplicate using different biopsy samples. The differential display technique was performed on total RNA (100 ng) using RNAimage (Gen-Hunter Corp., Nashville, TN), according to the manufacturer's instructions. The radioactive RT-PCR products were size-fractionated on denaturing 6% polyacrylamide gels and visualized by autoradiography. The resulting autoradiographs were examined to locate cDNA bands that exhibited differential intensity in treated and untreated cells. Satisfactory cDNA bands were cut from the dried polyacrylamide gels, re-amplified by PCR, cloned into the plasmid vector pGEM-T Easy (Promega), and sequenced. Clone identities were determined by performing BLAST searches against the GenBank TM data base.
A single tube duplex reverse transcriptase (RT)-PCR strategy was used to examine FKHR mRNA expression in differentiating ES cells. Briefly, 1 g of total RNA, obtained from untreated cultures and cells stimulated with 8-Br-cAMP for 24 h, was reverse-transcribed and amplified in a single reaction using the Access RT-PCR System (Promega) according to the supplier's instructions. Simultaneous amplification of FKHR and ␤-actin was performed by adding 50 pmol each of the following oligonucleotides to each reaction: FKHR-sense (5Ј-AAGAGC-GTGCCCTACTTCAA-3Ј), FKHR-antisense (5Ј-AACTGTGATCCAGG-GCTGTC-3Ј), actin-sense (5Ј-GGAGCAATGATCTTGATCTTC-3Ј), actin-antisense (5Ј-CCTTCCTGGGCATGGAGTCCT-3Ј). The ␤-actin cDNA, representing a non-regulated gene, served as an internal control. The reaction was allowed to continue for 30 cycles, which were within the exponential phase of the amplification reaction as determined by cycle profiling. Southern blots of the PCR products were successively hybridized with a 32 P-labeled FKHR PCR product, amplified from the expression vector pCMV5-FKHR using the oligonucleotides FKHRsense and FKHR-antisense, and a 32 P-labeled ␤-actin PCR product, amplified from RNA extracted from ES cells using the actin sense and antisense oligos.
SDS-PAGE, Western Blotting, and Immunodetection-A modified method of Rittenhouse and Marcus (19) was used for protein analysis. Protein concentrations were determined by Bradford assay (Bio-Rad Laboratories). Equal amounts of nuclear and cytosolic proteins (20 g) were separated on a 10% SDS-polyacrylamide gel before electrotransfer at 80 V onto a polyvinylidene difluoride membrane (Hybond P, Amersham Biosciences, Inc.). Even loading and transfer efficiency were confirmed by Ponceau S staining. Nonspecific binding sites were blocked by overnight incubation with 5% dried skimmed milk in Tris-buffered saline (TBS, 130 mM NaCl, 20 mM Tris, pH 7.6). For FKHR immunodetection, blots were exposed to a primary rabbit polyclonal anti-FKHR antibody (N-18, Santa Cruz Biotechnology), diluted 1:1000 in TBS with 5% dried nonfat milk, for 1 h at room temperature, and then incubated with secondary peroxidase-conjugated rabbit anti-goat IgG (Sigma Chemical Co.), also for 1 h at room temperature. The primary antibody for C/EBP␤ immunodetection was a rabbit polyclonal anti-C/EBP␤ antibody (C-19, Santa Cruz). Protein bands were visualized by enhanced chemiluminescence (ECL Western blotting detection, Amersham Biosciences, Inc.).
Recombinant Protein and Gel Shift Assay-The cDNA coding for amino acids 160 -266 of FKHR, encompassing the DNA binding domain (DBD) and 10 additional residues C-terminal to the DBD, was amplified, and NdeI and XhoI sites were created at the 5Ј-and 3Ј-ends, respectively, by PCR using the full-length FKHR cDNA as template and sense (5Ј-AGGTTGCCCCACATATGGCGTTGCGGCGGG-3Ј) and antisense (5Ј-GGCAGCTCGGCTCGAGGCTCTTAGCAA-3Ј) primers. The amplified fragment was cloned into pCR2.1 (Invitrogen) and sequenced. The NdeI-XhoI fragment was subcloned in-frame with a C-terminal 6x(His)-tag in pET-21b (Novagen). BL21-DE3 cells (Stratagene) were transformed with the expression vector, and protein expression was induced at mid-logarithmic growth by addition of 1 mM isopropyl-1-thio-␤-D-galactopyranoside. Washed cells were lysed with 8 M urea, pH 4.5, and recombinant protein was purified by nickel-agarose chromatography and renatured folding. Recombinant protein was quantified by Bradford dye-binding assay (Bio-Rad) and stored at Ϫ70°C in 10% glycerol, 0.1 M sodium phosphate buffer, pH 6.5.
Transfection Studies-Transient transfections of ES cells plated at a density of 2.5 ϫ 10 5 cells/well in 24-well plates were performed by calcium phosphate precipitation in medium supplemented with 2% DCC-FBS. Promoter-reporter constructs and expression constructs were transfected at concentrations of 0.5 g/well and 125 ng/well, respectively. The empty expression vectors pcDNA or pALTER were included as filler constructs when required. Cell extracts were harvested, and luciferase activity was measured with the luciferase reagent kit (Promega) and expressed as relative light units. Transfections were performed in triplicate and repeated at least three times. Representative experiments are shown (means Ϯ S.D.).
GST Pull-down Assays-GST pull-down assays were performed as described previously (7). 35 S-Labeled proteins were prepared by the in vitro transcription-translation method, using the TnT T7 Coupled Reticulocyte Lysate System following the supplier's protocol (Promega). The presence of [ 35 S]methionine (Ͼ1000 Ci/mmol, Amersham Biosciences, Inc.) in the incubation mixture was used to produce labeled FKHR protein from the plasmid pcDNA/FKHR.

RESULTS AND DISCUSSION
Cyclic AMP Induces FKHR Expression in Differentiating Human ES Cells-Differential display PCR analysis of mRNA obtained from untreated primary ES cell cultures and cells treated with the cell-permeable cAMP analogue 8-Br-cAMP for 12 h yielded 19 apparently differentially expressed cDNAs. One clone was found to be 99% homologous to the reported sequence for the FKHR cDNA between nucleotides 1834 and 2078 rela-tive to its start codon (20). The regulated expression of FKHR mRNA during cAMP-induced ES cell differentiation was confirmed by simultaneously amplified FKHR and ␤-actin mRNAs (Fig. 1A). Induction of FKHR protein upon 8-Br-cAMP treatment was apparent by Western blotting after 24 to 48 h of stimulation, with some variation between cultures, and its expression was sustained even after 8 days of treatment (Fig.  1B). To our knowledge, this is the first example of regulated expression of FKHR in mammalian cells.
FKHR and related members of the FOXO subgroup of the forkhead/winged helix family, including FKHRL1 and AFX (21,22), have been previously identified as targets of protein kinase B (PKB/Akt), a serine/threonine kinase located downstream of phosphatidylinositol 3-kinase (13,(23)(24)(25)(26). FKHR has three putative PKB/Akt phosphorylation sites (Thr-24, Ser-256, Ser-319), which are also conserved in DAF16, the nematode Caenorhabditis elegans homologue. Upon PKB/Akt phosphorylation, DAF16 and its human counterparts are retained in the cytoplasm, and their exclusion from the nucleus is associated with reduced transcriptional activity (13,14,23,24). Hence, we determined the subcellular localization of FKHR in untreated and 8-Br-cAMPstimulated human ES cell cultures. Immunofluorescence microscopy studies demonstrated that, upon 8-Br-cAMP treatment, FKHR accumulated predominantly in the nucleus (Fig. 1C). The absence of discernible cytoplasmic translocation suggests that FKHR is transcriptionally active in differentiating ES cells.
Cycle-dependent Expression of FKHR and C/EBP␤ in Human Endometrium-Previous studies have shown that C/EBP␤ is induced during cAMP-dependent differentiation of ES cells in culture in a similar manner to FKHR (4). To delineate a potential role for these transcription factors in vivo, we investigated if FKHR and C/EBP␤ are expressed in human endometrium in a cycle-dependent manner. Endometrial biopsies obtained at different phases of the menstrual cycle were immunohistochemically stained for either FKHR or C/EBP␤. Fig. 2 demonstrates weak immunoreactivity for FKHR but not C/EBP␤ in the glandular compartment during the proliferative phase of the cycle. The glandular expression of both factors increased in the early secretory phase and was most intense toward the end of the cycle. In contrast, stromal expression of FKHR and C/EBP␤ was confined to the late secretory phase of the cycle and most apparent in the decidualizing perivascular stroma. The distinct spatio-temporal expression of FKHR and C/EBP␤ in differentiating human endometrial stroma suggests a role for these transcription factors in the regulation of the expression of decidua-specific genes in vivo.
There are two isoforms of C/EBP␤, the full-length liverenriched activating protein (LAP) and the truncated liver-enriched inhibitory protein (LIP). The latter lacks the N-terminal transactivation domains of LAP and acts as a potent repressor of C/EBP-dependent transcription (27). Additional Western blot analysis studies showed the presence of LAP (1⍀33 and 36 kDa), but not LIP (1⍀16 kDa) in normal non-pregnant human endometrium (data not shown). This allowed us to conclude that the immunoreactive C/EBP␤ in vivo represents the activating isoform LAP but not the transcriptional repressor LIP.
FKHR Enhances dPRL Promoter Activity in Response to cAMP-The coordinated expression of FKHR in the endometrial stroma during the late secretory phase of the cycle suggested a putative role in decidualization. Expression of PRL, a cardinal phenotypical marker of decidualization, is detectable in culture after ϳ48 h of 8-Br-cAMP treatment (2). The pattern of induction and nuclear retention of FKHR upon 8-Br-cAMP treatment in vitro suggested the dPRL promoter is a potential target for FKHR action. To test this hypothesis, primary cultures were transiently transfected with a luciferase reporter gene construct under control of either 3 kb of the dPRL promoter region (dPRL-3000/luc3) or the minimal cAMP-responsive promoter region (dPRL-332/luc3). Cotransfection of a FKHR expression vector minimally stimulated the basal activity of these promoter-reporter constructs (Fig. 3A). However, FKHR markedly enhanced induction of dPRL-3000/luc3 activity upon cAMP stimulation or in response to coexpressed catalytic subunit, C␣, of the PKA holoenzyme (Fig. 3A, left panel). FKHR also enhanced cAMP-or C␣-dependent activation of the dPRL-332/luc3 construct, and, qualitatively, this response was indistinguishable to that observed with the dPRL-3000/luc3 construct (Fig. 3A, right panel). These observations indicated that the minimal cAMP-responsive promoter region could be a target for FKHR. Additional transfection studies with a series of truncated promoter-reporter constructs identified the region between positions Ϫ332 to Ϫ270 as critical for FKHR-mediated enhancement of dPRL promoter activity (Fig. 3B).
Dependent upon the cellular context, cAMP or its effector PKA have been suggested to either stimulate or inhibit the PKB/Akt signaling pathway (28 -30). This raised the possibility that cAMP could enhance the trans-activation potential of FKHR in differentiating ES cells by reducing PKB/Akt activity and thereby facilitating nuclear targeting of FKHR. However, overexpression of FKHR-(T/S/S)-A, a constitutively active mutant in which the three PKB/Akt phosphorylation sites (Thr-24, Ser-256, and Ser-319) are changed to alanines ((T/S/S)-A), only elicited a 3-fold increase in basal dPRL promoter activity (Fig.  3C). Furthermore, this FKHR mutant was still capable of enhancing promoter activity in response to 8-Br-cAMP treatment, indicating that phosphorylation of FKHR by PKB/Akt is not required for this effect. Overexpression of a DNA bindingdeficient FKHR mutant (FKHR-Helix3.2M), in which critical residues within helix 3 of the DNA binding domain are mutated, failed to augment luciferase activity (Fig. 3C). This observation provides further evidence for a transcriptional role FIG. 3. FKHR targets the dPRL promoter. A, FKHR enhances cAMP-dependent dPRL promoter activity in differentiating ES cells. Primary cultures were transfected with either the dPRL-3000/luc3 (left panel) or the dPRL-332/luc3 (right panel) reporter construct. In some cultures, the expression vector pC␣, encoding for the catalytic C␣ subunit of PKA, was cotransfected as indicated. Transfected cultures remained untreated or were stimulated with 0.5 mM 8-Br-cAMP for 40 h. Cellular extracts were used to measure luciferase activity, and the results show the mean Ϯ S.D. of triplicate measurements. B, FKHR enhances transcription through the cAMP-responsive Ϫ332/Ϫ270 dPRL promoter region. Cells were transiently transfected with one of the following luciferase reporter constructs under the control of various lengths of the proximal dPRL promoter: dPRL-332/luc3, dPRL-311/luc3, dPRL-301/luc3, dPRL-270/luc3, or dPRL-32/luc3. After transfection the cells were treated and harvested as described above. C, effect of FKHR mutants on cAMP-dependent dPRL promoter activity. ES cells were transfected with the reporter construct dPRL-3000/luc and an expression vector encoding for the wild-type FKHR (FKHR-wt), a constitutively active FKHR mutant (FKHR-T/S/S-A) in which the three PKB/ Akt phosphorylation sites (Thr-24, Ser-256, and Ser-319) are mutated to alanines, or a DNA binding-deficient FKHR mutant (FKHR-Helix3.2M). Transfected cultures were treated and harvested as described in A.
for FKHR in regulating dPRL gene expression and indicates a requirement for direct interaction with the dPRL promoter.
To determine whether the Ϫ332/Ϫ270 region of the dPRL promoter contains specific binding sites for FKHR, we performed gel shift assays with 32 P-labeled oligonucleotide probes (Table I) and a bacterially expressed recombinant protein, which contains the FKHR DNA binding domain (DBD). As shown in Fig. 4A, binding of an oligonucleotide probe containing residues Ϫ332 to Ϫ270 of the dPRL promoter to the FKHR DBD was detectable with as little as 1 ng of recombinant protein, and the formation of this nucleoprotein complex increased with the addition of more protein in a dose-dependent fashion. Studies with an excess of the unlabeled dPRL(332/270) oligonucleotide competitor confirmed that this binding is competitive (data not shown).
It has been previously reported that the Ϫ332/Ϫ270 dPRL promoter region contains two response elements (C/EBP D and B; Table I) that can form nucleoprotein complexes containing C/EBP␤ in combination with nuclear proteins prepared from differentiated ES cells (4). Additional gel shift studies with truncated oligonucleotide probes revealed that residues between Ϫ301 and Ϫ270 are sufficient to interact with the FKHR DBD (Fig. 4B). Interestingly, a mutation that disrupts both D and B sites (probe dPRL(317/277).DBMut) also disrupts the ability of the recombinant FKHR DBD to interact, indicating that critical FKHR binding sites are located in this region of the dPRL promoter (Fig. 4B). Based on a consensus sequence for FKHR binding (GGTAAACAA), derived from site-selected amplification studies (31), we identified two potential FKHR binding sites in this region of the dPRL promoter: FKHR-1 (GCTAAACAT) and FKHR-2 (TAGCAACAT) ( Table I). These sites are overlapping, contained within the C/EBP B site, and located on opposite strands of the dPRL promoter. As shown in Fig. 4B, mutation of both FKHR-1 and -2 sites (probe dPRL(301/281).FHMut) completely disrupts the ability of this promoter region to interact with the FKHR DBD. Mutation of the FKHR-1 site alone (probe dPRL (301/281).FHMut1) impairs binding to the FKHR DBD whereas selective mutation of the FKHR-2 site (probe dPRL(301/281).FHMut2) has no discernible effect. These results indicate that FKHR DBD interacts preferentially with the FKHR-1 site. This was confirmed by dose-response studies, shown in Fig. 4C, comparing the relative binding affinities of recombinant FKHR DBD for oligonucleotides that contain either the individual FKHR binding sites in the Ϫ332/Ϫ270 dPRL promoter region (probes dPRL(Ϫ310/Ϫ281).Mut1 and dPRL(Ϫ310/Ϫ281).Mut2) or another naturally occurring FKHR binding site, i.e. the insulin response sequence (IRS) of the insulin-like growth factor binding protein-1 (IGFBP-1) promoter (probe ⌬IRS.1) (17). Phosphorimaging analysis of gel shift assays showed that FKHR DBD binding to dPRL(Ϫ310/Ϫ281).Mut2 was ϳ3-fold stronger than its binding to dPRL(Ϫ310/Ϫ281).Mut1 and comparable to its interaction with ⌬IRS.1 (Fig. 4D). Together, these observations indicate that the FKHR-1 site is sufficient to account for FKHR DBD binding to the Ϫ332/Ϫ270 dPRL promoter region and that the binding affinity of FKHR for this site is comparable to another response element known to mediate the effects of FKHR on promoter function.
Functional Cooperation between FKHR and C/EBP␤-The observation that both C/EBP␤ and recombinant FKHR interact with overlapping and adjacent sequences within the Ϫ332/ Ϫ270 dPRL promoter region, and contribute to cAMP-stimu-TABLE I Oligonucleotides used for gel shift studies Double-stranded oligonucleotide probes containing residues Ϫ332/Ϫ270 of the dPRL promoter or smaller portions, with or without specific mutations, were prepared for gel shift studies. A, locations of C/EBP binding sites D and B and of two FKHR binding sites (FKHR-1, FKHR-2) in the dPRL promoter region Ϫ332/Ϫ270 are indicated by arrows. The binding sites in the dPRL promoter element, which were inactivated by mutations (boldface lowercase letters) are underlined. B, the ⌬IRS.1 probe contains the insulin response sequence (IRS; underlined) of the IGFBP-1 promoter. The relationship between the FKHR binding site identified in this promoter, the consensus binding sequence for FKHR and other FOXO proteins (FOXO), and the FKHR binding sites FKHR-1 and FKHR-2 of the dPRL promoter is shown. Core residues (AAAC) important for effective interaction with FKHR (14) are in boldface. lated promoter activity, suggested possible functional cooperation between these distinct transcription factors. To test this hypothesis, primary human ES cells were transiently transfected with a luciferase construct under control of the Ϫ332/ Ϫ270 region fused to the minimal dPRL promoter (dPRL(Ϫ332/ Ϫ270wt)/Ϫ32/luc3) (Fig. 5). Directed mutation of the distal C/EBP binding site (dPRL(Ϫ332/Ϫ270Dmut)/Ϫ32/luc3), the proximal composite FKHR/C/EBP␤ binding site (dPRL(Ϫ332/ Ϫ270FHmut)/Ϫ32/luc3), or both sites (dPRL(Ϫ332/Ϫ270DBmut)/Ϫ32/luc3) ( Table II) allowed assessment of their relative contributions. The minimal promoter construct dPRL-32/luc3 served as control reporter vector.
Overexpression of FKHR alone did not induce dPRL(Ϫ332/ Ϫ270wt)/Ϫ32/luc3 activity (Fig. 5A). In contrast, expression of full-length C/EBP␤ elicited a 26-fold induction in promoter activity (Fig. 5A). Furthermore, coexpression of FKHR and C/EBP␤ yielded a 52-fold increase in dPRL(Ϫ332/Ϫ270wt)/ Ϫ32/luc3 activity, demonstrating that FKHR enhances C/EBP␤ trans-activation of the Ϫ332/Ϫ270 promoter region. We previously reported that overexpression of C/EBP␤ modestly activates the control reporter construct dPRL-32/luc3 (5fold) as well as the promoterless construct pGL3-Basic (7). Activation of dPRL(Ϫ332/Ϫ270DBmut)/Ϫ32/luc3, which does not have functional C/EBP or FKHR binding sites, in the presence of C/EBP␤ was identical to that observed with the dPRL-32/luc3 construct. Furthermore, coexpression of FKHR and C/EBP␤ had no additional effect upon reporter activity. Selective ablation of the distal C/EBP␤ binding site (dPRL(Ϫ332/ Ϫ270Dmut)/Ϫ32/luc3) also markedly blunted C/EBP␤-dependent trans-activation of the Ϫ332/Ϫ270 region (7-fold). However, transcriptional cooperation between C/EBP␤ and FKHR was still apparent, because coexpression of both factors elicited a 16-fold induction in dPRL(Ϫ332/Ϫ270Dmut)/Ϫ32/luc3 activity (Fig. 5A). In contrast, targeted deletion of the FKHR binding sites (dPRL(Ϫ332/Ϫ270FHmut)/Ϫ32/luc3) not only impaired C/EBP␤ trans-activation of the Ϫ332/Ϫ270 promoter region but also abolished its cooperation with FKHR. Together, these results indicate that the proximal composite FKHR/CEBP␤ binding site is essential for C/EBP␤ trans-activation of the Ϫ332/ Ϫ270 promoter region and for transcriptional synergy with FKHR. In contrast, the distal C/EBP binding site is also nec- FIG. 4. Identification of FKHR binding sites in the dPRL promoter. A, dose-dependent binding. A double-stranded 32 P-labeled probe spanning residues Ϫ332 to Ϫ270 of the dPRL promoter was incubated with various amounts of recombinant protein (0 -30 ng) containing the DBD of FKHR (amino acids 160 -266), then loaded for non-denaturing polyacrylamide gel electrophoresis. Free and bound probe were identified by autoradiography of dried gels. B, localization of FKHR binding sites within the dPRL promoter. Gel shift studies were performed using probes containing smaller portions of the Ϫ332/Ϫ270 dPRL promoter region with or without targeted mutations as described in Table I. C, relative activity of the FKHR binding sites in the dPRL promoter. 32 P-Labeled oligonucleotide probes (20,000 cpm/lane) containing FKHR-1 (dPRL(310/281).FHMut2) or FKHR-2 (dPRL(310/ 281).FHMut1) sites from the dPRL promoter, or the insulin response sequence from the IGFBP-1 promoter were incubated with increasing amounts of recombinant protein containing the FKHR DBD prior to gel electrophoresis. Free and bound probes were identified by autoradiography. D, quantification of the binding activities for probes containing mutations of FKHR-1 (dPRL(310/281).FHMut1) or FKHR-2 (dPRL(310/ 281).FHMut2) sites from the dPRL promoter or IRSA (⌬IRS.1) from the IGFBP-1 promoter. Free and bound radioactivity, in the presence of various concentrations of recombinant protein, was quantified by phosphorimaging, and binding activity was defined as the percentage of bound probe.
The ability of FKHR and C/EBP␤ to regulate transcriptional activation of the Ϫ332/Ϫ270 promoter region through interaction with a composite response element raised the possibility of physical association between these distinct transcription factors. This was confirmed by in vitro protein binding studies demonstrating specific interactions between FKHR and the glutathione S-transferase (GST)-tagged full-length C/EBP␤ (GST-LAP). FKHR also interacted with the GST-fused truncated C/EBP␤ isoform (GST-LIP) (Fig. 5B), indicating that the N-terminal trans-activation domains of C/EBP␤ are not required for physical association with FKHR. This is in agreement with other studies demonstrating that the C terminal domain of C/EBP␤ mediates binding to other nuclear factors, including the phosphoprotein Nopp140, members of the nuclear factor-B family, Ets-1, and various members of the nuclear receptor superfamily (7,(32)(33)(34)(35).
Taken together, these in vivo and in vitro results indicate that FKHR is an important effector of the decidual response in the late secretory phase of the menstrual cycle. Interestingly, our results indicate that FKHR interacts with the PKA signal transduction pathway in at least two ways to contribute to the coordinated expression of decidua-specific genes in differentiating human ES cells. First, differential display studies revealed that FKHR mRNA is induced in ES cells after stimulation with cAMP, and subsequent studies confirmed that levels of FKHR protein also are increased during ES cell differentiation in vitro and in vivo. Previous studies have shown that the expression of FKHR and other members of the FOXO subfamily of forkhead/winged-helix transcription factors is tissue-specific (21,23,31). To our knowledge, the observation that FKHR expression is regulated through a cAMP-dependent pathway provides the first report indicating that FOXO proteins can be regulated in response to activation of a discrete signaling pathway. Studies are in progress to examine specific mechanisms mediating this effect of cAMP on FKHR expression and to determine whether FKHR gene expression is responsive to activation of the PKA pathway in different cells.
We also found that expression of FKHR enhances dPRL promoter activity in the presence of activated PKA. This result indicates that, once expressed, FKHR interacts with and enhances the function of other cellular factors mediating effects of cAMP on promoter activity. We find that FKHR functions cooperatively with C/EBP␤ to stimulate dPRL promoter function through a previously identified C/EBP response element in the proximal promoter region. Direct interaction between FKHR and C/EBP␤ may contribute to their ability to function cooperatively but does not exclude other potential mechanisms, including the recruitment of shared coactivators such as p300/ CBP (36,37). Previous in vitro binding studies indicate that C/EBP and forkhead proteins interact with overlapping elements in the phosphoenolpyruvate carboxyl kinase promoter (38,39). We and others have reported that a nucleoprotein complex containing C/EBP␤ interacts with a known FKHR binding site (IRSA) in the IGBP-1 promoter (40), and recent studies indicate that FOXO forkhead proteins may contribute to the formation of this complex. 2 These observations indicate that the functional interaction between FOXO forkhead family members and C/EBP transcription factors may be important for transcriptional activation of diverse genes.