Human chorionic gonadotropin-alpha gene is transcriptionally activated by epidermal growth factor through cAMP response element in trophoblast cells.

The purpose of this study was to analyze the mechanism of transcriptional activation of human chorionic gonadotropin-alpha (hCGalpha) gene by epidermal growth factor (EGF) in trophoblast cells. We stably transfected hCGalpha promoter-chloramphenicol acetyltransferase constructs into Rcho-1 trophoblast cells and monitored the promoter activities. -290-base pair hCGalpha promoter containing a tandem repeat of cAMP response element (CRE) was activated by EGF in a dose- and time-dependent manner. Deletion analysis of hCGalpha promoter suggested an involvement of CRE in EGF-induced hCGalpha transcriptional activation. Moreover, the hCGalpha promoter, of which both CREs were mutated, did not respond to EGF. These results indicate that EGF activates the hCGalpha gene transcription through CRE. Although EGF did not alter the amount of CRE-binding protein (CREB), EGF induced CREB phosphorylation. We next examined the mechanism of CREB phosphorylation by EGF. Protein kinase C inhibitors (H7, staurosporin, and chelerythrine) inhibited EGF-induced CREB phosphorylation, whereas either mitogen-activated protein kinase kinase-1 inhibitor (PD98059) or protein kinase A inhibitor (H8) showed no effect. Furthermore, H7 and staurosporin but not H8 inhibited hCGalpha promoter activation by EGF. In conclusion, EGF promotes hCGalpha gene transcription via the CRE region probably by phosphorylating CREB mainly through the protein kinase C pathway in trophoblast cells.

Epidermal growth factor (EGF), 1 which consists of 53 amino acids, stimulates proliferation and differentiation in many kinds of cells and tissues (1). An important role of EGF has been suggested in mouse (2)(3)(4)(5) and human (6 -9) pregnancies. Human placentas are extremely rich in EGF receptors (6 -8), suggesting that EGF is important in placental functions (9). EGF has been shown to promote functional differentiation of human trophoblast cells (10); EGF treatment of choriocarci-noma cells (11)(12)(13) and normal human trophoblasts (9,14,15) results in an increase in human chorionic gonadotropin (hCG) secretion. Human chorionic gonadotropin plays a critical role at the early stage of pregnancy and is progressively produced as human trophoblasts differentiate into syncytiotrophoblasts (10).
EGF has been shown to increase hCG␣ and hCG␤ mRNA level and their stability (16); however, it remains unclear whether or not EGF increases the transcriptional activity of hCG genes. hCG consists of hCG␣ and hCG␤ subunits (17). hCG␣ gene is present as a single copy gene on chromosome 6q21.1-23 (18), whereas hCG␤ consists of six closely spaced genes (19). The structure of hCG␣ promoter is simpler and is well studied (19). Therefore, in this study we analyzed the effect of EGF on the hCG␣ gene.
In this study, we used Rcho-1 cells to examine the molecular mechanism of EGF effect on the hCG␣ gene transcription. The Rcho-1 cell line was established from a transplantable rat choriocarcinoma (20) and can be manipulated to proliferate or differentiate along the trophoblast giant cell pathway (21). Several genes have been shown to be transcriptionally activated during Rcho-1 cell differentiation (22,23). Using Rcho-1 cells stably transfected with hCG␣ promoter-CAT construct, we studied the EGF effect on the transcriptional activity of hCG␣ gene and analyzed the mechanism of EGF action.

EXPERIMENTAL PROCEDURES
Cell Culture-The Rcho-1 cell line was routinely maintained in subconfluent conditions with NCTC-135 medium (Sigma) supplemented with 20% fetal bovine serum (20% FBS/NCTC) as reported previously (21). Differentiation was induced by growing to confluence in FBSsupplemented culture medium and then replacing the serum supplementation with 1% horse serum (1% HS/NCTC).
Plasmid Construction-Ϫ290h␣CAT was kindly provided by Dr. John Nilson (24). Deletion mutants of Ϫ290h␣CAT were generated by the polymerase chain reaction. Polymerase chain reaction products were inserted into pSV0CAT (Promega). CRE-tk-CAT and CRE-CREtk-CAT were generated by inserting one or two copies of CRE in the native orientation directly up-stream of the tk-CAT. Ϫ290h␣ (mCRE)-CAT, both CREs of which were mutated, was generated using a polymerase chain reaction-based site-directed mutagenesis kit (Stratagene). All constructs were sequenced using T7 DNA polymerase sequencing kits (Amersham).
Stable Transfections and CAT Assays-Rcho-1 cells were transfected using a liposome-mediated delivery system (Life Technologies, Inc.) as described previously (23). Stable transfectants were established by co-transfecting 9 g of promoter-reporter construct plasmid DNA with 1 g of pSV2Neo DNA. Selection of stable transfectants was performed by growth in the presence of G418 (250 g/ml).
Stably transfected Rcho-1 cells were incubated in 20% FBS/NCTC medium for 2 days, and then culture was shifted to 1% HS/NCTC. EGF from mouse submaxillary glands (Toyobo) was used for the EGF stimulation experiments. EGF treatment was started when the medium was replaced with 1% HS/NCTC. Protein concentrations of the whole cell extracts were determined by the Bio-Rad protein assay system. CAT reactions were carried out with 50 g of protein for 3 h at 37°C. The acetylated and nonacetylated forms of [ 14 C]chloramphenicol were separated by a thin layer chromatography, autoradiographed, and quantitated by an image analyzer system (BAS2000, FUJIX). All experiments were repeated at least three times with consistent results.
Electrophoretic Mobility Shift Assays-Nuclear extracts were prepared from Rcho-1 cells according to a previously described procedure (25). The extraction buffer contained an additional phosphatase inhibitor (NaF) at 10 mM. Protein concentrations of extracts were determined using the Bio-Rad protein assay system. Nuclear extracts (5 g/lane) were incubated for 10 min at room temperature with 2 g poly(dI-dC)poly(dI-dC) in a reaction mixture containing 10 mM HEPES, pH 7.8, 50 mM KCl, 1 mM EDTA, 5 mM MgCl 2 , 1 mM dithiothreitol, and 10% glycerol. A 32 P-labeled oligonucleotide probe (1 ϫ 10 4 cpm) was added, and the reaction mixture was incubated for 30 min at room temperature. To demonstrate binding specificity, unlabeled CRE or mCRE was used. DNA-protein complexes were resolved on 5% polyacrylamide gels in 0.5ϫ TBE and visualized by autoradiography.
Western Blot Analysis-Nuclear extracts (150 g) were resolved by 10% SDS-polyacrylamide gel electrophoresis (PAGE) and transferred to nitrocellulose membranes. After a blocking reaction (5% nonfat dry milk in Trisbuffered saline, pH 7.4, 0.05% Tween-20) for 1 h, membranes were incubated in a blocking buffer with antisera against rat CREB (1:3,000 dilution) or against rat phosphorylated CREB (1:2, 500 dilution) for overnight at 4°C. After incubation with horseradish peroxidase-linked rabbit IgG (Life Technologies, Inc., 1:3,000), the membranes were developed by using Enhanced Chemiluminescence System (Amersham) according to the manufacturer's instructions. For reprobing, the membranes were submerged in a stripping buffer (100 mM 2-mercaptoethanol, 2% SDS, 62.5 mM Tris-HCl, pH 6.7) and incubated at 50°C for 30 min with occasional agitation. After washing twice in Tris-buffered saline for 10 min at room temperature, the membranes were blocked for 1 h and incubated with a rat CREB antiserum (1:3,000). CREB and phosphorylated CREB antisera were raised in rabbits against a synthetic peptide (amino acids 1-205) and a synthetic phosphopeptide (amino acids 123-136), respectively (Update Biotechnology, Inc.).
Southwestern Blot Analysis-Nuclear extracts were resolved and transferred in the same way as above. Membranes were initially incubated in TNE-50 (10 mM Tris, pH 7.5, 50 mM NaCl, 1 mM EDTA, and 1 mM dithiothreitol) containing 5% nonfat dry milk for 2 h at room temperature, washed briefly in TNE-50 without milk, and then incubated in TNE-50 containing a 1 ϫ 10 6 cpm/ml CRE as a probe and 10 g/ml poly(dI-dC)-poly(dI-dC). After the incubation, the blots were washed three times for 5 min each with TNE-50, air dried, and then exposed to a Kodak XAR film.
Statistics-Statistical analysis was performed by unpaired t test. All experiments were performed in triplicate or quadruplicate and repeated at least three times with similar results.

EGF Promotes Differentiation-dependent Increase in hCG␣-
CAT Activity in Rcho-1 Cells-Rcho-1 cells morphologically and functionally differentiate in the differentiation medium (1% HS) throughout the culture period as reported previously (21).
Ϫ290h␣CAT activity showed a differentiation-dependent increase in stably transfected Rcho-1 cells (Fig. 1B, Control). To study the effect of EGF on the differentiation-dependent increase in hCG␣ promoter activity, Rcho-1 cells stably transfected with Ϫ290h␣CAT were treated with various concentrations of EGF (0 -10 nM) for days 2-8 of culture. Cells were harvested on day 8, and CAT activities were determined. EGF enhanced the hCG␣ promoter activity in a dose-dependent manner with a maximal effect at 10 nM (2.7-fold enhancement) (Fig. 1A). EGF at 30 nM or more did not show a further promotion (data not shown).
We next performed the time course study on the promotive effect of EGF on the differentiation-dependent increase in hCG␣ promoter activity for days 4 -10 of culture. EGF at 10 nM promoted the differentiation-dependent increase in hCG␣ promoter activity on days 8 and 10 of culture. Maximal EGF effect was observed on day 8 (Fig. 1B).
Clonal Rcho-1 cell culture under a differentiation condition contains a mixture of proliferative small cells and terminally differentiated giant cells. We investigated whether or not EGF affected the cell number and the proportion of the small and giant cells population throughout day 10 of culture. Neither the cell number nor the percentage of the small and giant cells changed by 10 nM EGF treatment up to day 10 (data not shown).
CRE on the hCG␣ Gene Promoter Is the Region Responsible for EGF Effect--The Ϫ290-base pair region of the hCG␣ promoter contains several consensus sequences: trophoblast-specific element, GATA element, and CRE (16). To determine the region responsible for EGF effect, various deletion mutants were generated and stably transfected into Rcho-1 cells. Fig. 2 FIG. 1. EGF promotes hCG␣ gene transcription. A, dose-dependent effect of EGF on hCG␣ promoter activity. Rcho-1 cells stably transfected with Ϫ290h␣CAT were plated at 3.6 ϫ 10 5 cells/plate (60 mm) and maintained in 20% FBS/NCTC. On day 2, culture was shifted to 1% HS/NCTC without or with various concentrations of EGF (0 -10 nM), and media containing or not containing EGF were changed every 2 days. On day 8, cells were extracted and CAT activities were determined. Relative CAT activity is shown as a ratio in CAT activities of treated/untreated cells. Each value is the mean Ϯ S.E. of the mean of triplicate measurements. **, p Ͻ 0.005 versus control. B, time-dependent effect of EGF on hCG␣ promoter activity. Rcho-1 cells stably transfected with Ϫ290h␣CAT were plated at 3.6 ϫ 10 5 cells/plate (60 mm) and maintained in 20% FBS/NCTC. On day 2, culture was shifted to 1% HS/NCTC without or with EGF (10 nM), and media were changed every 2 days. Cells were extracted on days 4, 6, 8, and 10 of culture, and CAT activities were determined. Experiments were repeated three times with similar results, and a representative result is shown.
(left) shows a diagram of deletion mutants used in the study. Stable transfectants were cultured with or without EGF (10 nM) on days 2-8 of culture. CAT activities in EGF-treated and untreated cells were compared (Fig. 2, right). Although transcriptional activities in deletion mutants up to Ϫ142 base pairs were promoted by EGF to an extent (2.6 -2.8-fold) similar to that in Ϫ290h␣CAT, Ϫ128h␣CAT, which contains only one CRE, showed a substantially decreased response to EGF.
Ϫ110h␣CAT, which does not contain CRE, did not respond to EGF. Ϫ290h␣CAT, and its deletion mutants up to Ϫ142 base pairs also responded to forskolin. Ϫ128h␣CAT showed an impaired response to forskolin, and Ϫ110h␣CAT did not respond to forskolin (data not shown), suggesting that CRE was functional in Rcho-1 cells. These results imply that CRE may be a response element for EGF.
To further confirm the involvement of CRE for the EGF responsiveness of hCG␣ gene, an 18-base pair region containing CRE was inserted into the tk-CAT vector, and the transcriptional activation by EGF was estimated. Rcho-1 cells were stably transfected with CRE-CRE-tk-CAT, CRE-tk-CAT, or tk-CAT gene and treated with or without EGF (10 nM) for days 2-8 of culture (Fig. 3A, left). CAT activities in EGF-treated and untreated cells were compared. EGF significantly promoted CRE-CRE-tk-CAT activity by 2.2-fold and substantially enhanced CRE-tk-CAT activity, whereas the tk-CAT gene did not respond to EGF (Fig. 3A, middle). The increase in CRE-CREtk-CAT activity by EGF was comparable with that in Ϫ290h␣CAT (Figs. 1A and 2). Forskolin also showed a similar effect on these genes. Forskolin significantly promoted CRE-CRE-tk-CAT activity, and a decreased response was observed in CRE-tk-CAT, whereas tk-CAT did not respond to forskolin (Fig. 3A, right). These results indicate that EGF confers a transcriptional activation even to a heterologous promoter through CRE in trophoblast cells.
Furthermore, Ϫ290h␣(mCRE)CAT, both CREs of which were mutated, did not respond to EGF (10 nM, 6 days of treatment) (Fig. 3B, middle) as well as to forskolin (Fig. 3B, right). These data strongly support the idea that the CRE on the hCG␣ promoter is essential for EGF responsiveness.
EGF Does Not Alter the CRE-binding Nuclear Proteins-To analyze the possible change by EGF in the transcriptional factors associating with CRE, we performed electrophoretic mobility shift assays using nuclear proteins from EGF-treated and untreated Rcho-1 cells (Fig. 4A). CRE formed several major DNA-protein complexes in both untreated (lane 2) and 10 nM EGF-treated (lane 5) Rcho-1 cells. These binding complexes seemed specific because these complexes were competed by excess amount of cold CRE (lanes 3 and 6) but not by excess amount of cold mutated CRE (lanes 4 and 7). No obvious differences were found in CRE-binding nuclear proteins between EGF-treated and untreated cells. These results suggest that EGF might not alter CRE-binding proteins.
To further analyze CRE-binding proteins, we performed Southwestern blot analysis using nuclear proteins from EGFtreated and untreated Rcho-1 cells. Southwestern blotting using 32 P-labeled CRE as a probe showed expression of a 43-kDa protein in nuclear protein samples from both EGF-treated and untreated Rcho-1 cells (Fig. 4B). The protein seemed to be the CREB because of its size (26). We also observed other bands in the Southwestern blotting, suggesting a possible involvement of other less defined CRE-binding proteins. The expression of CREB was further confirmed by Western blot analysis using anti-CREB antibody. A 43-kDa CREB was observed in nuclear protein samples from both EGF-treated and untreated cells (Fig. 4C). The band was not observed when the anti-CREB antiserum was substituted with a nonimmune serum (data not shown). The amounts of CREB were not different in each group.
EGF Phosphorylates CREB-It is known that CREB is mainly phosphorylated at Ser 133 and that the phosphorylation is essential for gene activation by CREB (27). To elucidate the mechanism of hCG␣ promoter activation by EGF through CRE, we determined whether or not EGF phosphorylates CREB protein. Nuclear extracts were obtained from EGF-treated (10 nM for 5 or 30 min) and untreated cells, and Western blot analysis was performed using anti-phosphorylated CREB antibody (Fig.  5A). Although phosphorylated CREB was not observed in nuclear proteins from untreated cells, it was present in cells treated for 5 min with EGF, and the amount of phosphorylated CREB decreased by 30 min of EGF treatment (Fig. 5A, upper  panel). Anti-CREB antibody detected a similar amount of CREB in nuclear protein samples from untreated and EGFtreated (5 and 30 min) cells (Fig. 5A, lower panel), showing that the changes in CREB phosphorylation by EGF were specific. We examined the time dependence of EGF effect on the CREB phosphorylation throughout 3 h (data not shown) and observed that the EGF effect was the most at 5 min. The results suggest that EGF promotes hCG␣ gene transcriptional activity through CRE, at least partly, by phosphorylating CREB.
EGF activates the mitogen-activated protein (MAP) kinase hCG␣ Transcriptional Activation by EGF cascade and induces transcription by phosphorylating transcriptional factors (28 -30). To investigate the mechanism of CREB phosphorylation by EGF, we tested the involvement of MAP kinase cascade. Rcho-1 cells were left untreated or were pretreated for 1 h with MEK1 inhibitor (PD98059 (50 M)) and then treated with EGF (10 nM, 5 min). Nuclear extracts were prepared, and Western blot analysis was performed using antiphosphorylated CREB antibody. Pretreatment with the MEK1 inhibitor did not inhibit EGF-induced phosphorylation of CREB as well as CREB expression (Fig. 5B). In MAP kinase assays using myelin basic protein as a substrate, we observed that EGF activated MAP kinase and that PD98059 at 50 M suppressed EGF-induced MAP kinase activation in Rcho-1 cells (data not shown). These results suggested that the MAP kinase cascade might not be a dominant pathway of CREB phosphorylation in Rcho-1 cells.
Induction of CREB Phosphorylation Depends on PKC-To further analyze the mechanism of CREB phosphorylation by EGF, we tested whether protein kinase A (PKA) and/or PKC pathways are involved. Rcho-1 cells were left untreated or pretreated for 30 min with PKC inhibitors (H7 (10 M), staurosporin (50 nM), chelerythrine (5 M)), or a PKA inhibitor (H8 (10 M)) and then stimulated with EGF (10 nM, 5 min) or with forskolin (10 M, 1 h). Nuclear extracts were prepared, and Western blot analysis was performed using anti-phosphorylated CREB antibody. Pretreatment with the PKC inhibitors inhibited EGF-induced phosphorylation of CREB (Fig. 6, upper  panel). Pretreatment with a PKA inhibitor H8 did not inhibit EGF-induced phosphorylation of CREB. H8 inhibited forskolininduced phosphorylation of CREB, showing the efficiency of the PKA inhibitor used in the study. Either PKC or PKA inhibitors did not affect CREB expression (Fig. 6, lower panel). These results suggest that EGF phosphorylates CREB mainly through the PKC-dependent pathway in trophoblast cells.
PKC Inhibitors Decrease EGF-promoted hCG␣ Transcriptional Activity-To study the effects of PKC and PKA inhibitors on the EGF-enhanced hCG␣ promoter activity, Rcho-1 cells stably transfected with Ϫ290h␣CAT were treated with EGF (10 nM) in the absence or the presence of PKC inhibitors (H7 (10 M), staurosporin (50 nM)), or a PKA inhibitor (H8 (10 M)) for days 2-8 of culture. Although H8 did not reduce the EGFpromoted hCG␣ promoter activity (2.6-fold), PKC inhibitors (H7 and staurosporin) significantly reduced the enhancement by EGF (Fig. 7). These PKC inhibitors alone did not reduce hCG␣ promoter activity (data not shown), indicating that they may specifically inhibit the EGF effect. In addition, H7 (10 M) did not affect P-450 side chain cleavage (P-450scc) gene transcription but rather promoted progesterone secretion by Rcho-1

FIG. 3. Functional analysis of CRE in EGF-induced promoter activation.
A, CRE confers transcriptional activation by EGF to a heterologous promoter. One or two copies of CRE were inserted in the native orientation directly upstream of tk-CAT. Left, schematics of CRE-tk-CAT constructs. Middle, EGF responsiveness of CRE-tk-CAT constructs. After a stable transfection, Rcho-1 cells were plated at 3.6 ϫ 10 5 cells/plate (60 mm) and maintained in 20% FBS/NCTC. On day 2, culture was shifted to 1% HS/NCTC without or with EGF (10 nM), and media containing or not containing EGF were changed every 2 days. On day 8, cells were extracted and CAT activities were determined. Fold induction is shown as a ratio in CAT activities of treated/untreated cells. Each value is the mean Ϯ S.E. of the mean of triplicate measurements. **, p Ͻ 0.005 versus tk-CAT. Right, forskolin responsiveness of CRE-tk-CAT constructs. Rcho-1 cells were incubated in 1% HS/NCTC medium with or without forskolin (10 M) for days 2-8. On day 8, cells were extracted and CAT activities were determined. Fold induction is shown as a ratio in CAT activities of treated/untreated cells. Each value is the mean Ϯ S.E. of the mean of triplicate measurements. **, p Ͻ 0.005 versus tk-CAT. B, absence of EGF effect on the Ϫ290h␣(mCRE)CAT. Left, schematics of Ϫ290h␣CAT, Ϫ290h␣(mCRE)CAT and promoterless CAT. Middle and right, Responsiveness of Ϫ290h␣CAT, Ϫ290h␣(mCRE) and promoterless CAT to 10 nM EGF (middle) and to 10 M forskolin (right). The method of culture and the treatment of EGF and forskolin were the same as for A. Fold induction is shown as a ratio in CAT activities of treated/untreated cells. Each value is the mean Ϯ S.E. of the mean of triplicate measurements. **, p Ͻ 0.005; ***, p Ͻ 0.001 versus promoterless CAT. cells (31). Therefore, the effect of the PKC inhibitors may not be due to cell toxicity. All these results suggest that phosphorylation of CREB through the PKC pathway may be involved in the hCG␣ promoter activation by EGF. DISCUSSION Many studies have shown the importance of EGF in the maintenance of pregnancy and in fetal development. EGF deficiency during pregnancy causes abortion in mice (4). Fetal mice lacking the EGF receptors are retarded in growth and die at midgestation in a 129/Sv genetic background (5). In humans, the amounts of placental EGF receptors are decreased in intrauterine growth-retarded pregnancy (32). The promotion of hCG gene expression by EGF might be a part of the pregnancysupporting system. It has been known that EGF stimulates hCG secretion by trophoblast cells (33). Cao et al. showed that EGF increased hCG␣ and hCG␤ mRNA level by stabilizing them in JEG-3 cells (16); however, it has not been known whether or not EGF promotes hCG gene transcription. In this report, we first showed that EGF increased the transcriptional activity of hCG␣ gene using Rcho-1 cells. Rcho-1 cells are derived from a rat choriocarcinoma. Although rodent placentas do not express chorionic gonadotropin, the hCG␣ gene promoter when expressed as part of a transgene is active in placentas and pituitary of transgenic mice (24). Also we observed differentiation-dependent transcriptional activation of hCG␣ gene (Fig. 1B). Therefore, the Rcho-1 cell line seems to possess the activation system of hCG␣ gene as well as human trophoblast cells. The EGF effect to promote the hCG␣ transcriptional activity was dose-and time-dependent; however, EGF was less effective on day 10 of culture than on day 8, suggesting that terminally differentiated cells have a reduced response to EGF.
EGF increases the transcriptional activities in various genes; EGF stimulation of gastrin transcription is mediated through a GC-rich gastrin EGF response element (34). Expression of the ovine P-450 side chain cleavage enzyme gene (CYP11A1) is stimulated by EGF through AP-1-like site (35). An AP-1-binding site in the c-fos gene can mediate the induction by EGF in HeLa cells (36). We showed that CRE was an EGF response element in the hCG␣ promoter in trophoblast cells. It has been known that CRE is essential for basal promoter activity and cAMP responsiveness of hCG␣ gene (37)(38)(39)(40). Some cAMP responsive genes also respond to EGF (35,(41)(42)(43). Although both cAMP and EGF activate the ovine CYP11A1 promoter in JEG-3 cells through distinguishable regions (35), both factors activated the hCG␣ gene via the same region (CRE) in this study. The mechanism of transcriptional activation by EGF is complicated and might be gene-specific.
Several proteins binding to CRE have been identified, which include CREB, cAMP response modulator, activating transcription factor-1, and cAMP response element-binding protein-1 (identical to activating transcription factor-2). These proteins are known as members of bZIP proteins (44). It has been known that these proteins form homodimers or heterodimers to bind to the CRE (45,46). We focused on CREB in this study because in Southwestern blot analysis we detected a 43-kDa CRE-binding protein, the migration of which in SDS-PAGE was the same as that of a 43-kDa immunoreactive CREB. However, we also observed other associating proteins (Fig. 4B). The possible involvement of other CRE-binding proteins remains to be investigated.
The activity of many transcription factors is regulated by posttranslational modification. Such modifications include phosphorylation and dephosphorylation of serine and threonine residues and oxidation and reduction of cysteines (47). We showed that CREB was phosphorylated by EGF. A most probable candidate for the phosphorylation site is Ser 133 (27). The importance of the phosphorylation of CREB in generating its transcriptional functions has been shown in transgenic mice expressing a CREB with a serine to alanine substitution mutation at Ser 133 (48). Phosphorylation can alter protein function by introducing an allosteric conformational change in the protein or by allowing (or blocking) specific electrostatic interactions with other molecules. These changes are thought to be involved in the regulation of transcription. The structural properties of CREB and phosphorylated CREB were analyzed by the method of CD, and it was shown that the phosphorylation at Ser 133 did not alter the secondary structure of CREB and the DNA binding affinity of CREB to CRE sequences (49). From these results the phosphorylation of CREB might induce the production of the specific interactions with proteins such as CREB-binding protein (CBP) rather than the conformational change or increased DNA binding affinity. Furthermore other studies suggest that although phosphorylation of CREB is re-

cells stably transfected with
Ϫ290h␣CAT were plated at 3.6 ϫ 10 5 cells/plate (60 mm) and maintained in 20% FBS/NCTC. On day 2, medium was changed to 1% HS/NCTC without or with EGF (10 nM) in the absence or the presence of H7 (10 M), staurosporin (50 nM), or H8 (10 M). On day 8, cells were extracted and CAT activities were determined. Fold induction is shown as a ratio in CAT activities of EGF-treated/untreated cells in each group. Each value is the mean Ϯ S.E. of the mean of triplicate measurements. *, p Ͻ 0.01 versus inhibitors.
hCG␣ Transcriptional Activation by EGF quired to form the CREB-CBP complexes, other events are also involved for activation of CREB-mediated transcription (50). Biophysical evaluation of phosphorylated CREB-CBP complexes will help to further understand the hCG␣ transcriptional activation.
Whereas much research on the regulation of CREB transactivation has been directed toward the mechanisms of phosphorylation, relatively little is known about the phosphatase-mediated inactivation of CREB. Hagiwara et al. provided evidence that protein phosphatase-1 selectively dephosphorylates Ser 133 in CREB and correspondingly attenuates the transcriptional activity of CREB (51). There is another study focused on the intracellular processes that regulate the phosphorylation state of CREB in the hippocampal neurons (52). In the study it is shown that synaptic activity simultaneously influences both phosphorylation and dephosphorylation of CREB. During a brief stimulus, simultaneous activation of both kinase and phosphatase ensures that pCREB elevation will be large but brief. Longer stimuli cause prolongation of nuclear pCREB, possibly by hampering the phosphatase pathway. A possible involvement of the CREB dephosphorylation mechanism in the EGF effect on the hCG␣ gene remains to be studied.
The phosphorylation of CREB at Ser 133 is mediated by PKA in pheochromocytoma cells (53) and by PKC in B lymphocytes (54). In this study, we showed that EGF-induced phosphorylation of CREB may be mainly mediated by PKC. In Rcho-1 cells forskolin also phosphorylated CREB (Fig. 6, upper panel), suggesting that there may be several pathways (at least two pathways, PKC and PKA) to mediate CREB phosphorylation. EGF has been shown to activate Ras (55,56) and to induce members of the MAP kinase family including the extracellular signalregulated kinases and the stress-activated protein kinases, also referred to as c-Jun N-terminal kinases (57)(58)(59). EGF and c-Jun act via a common DNA regulatory element to stimulate transcription of the ovine CYP11A1 (35), and induction of the CYP11A1 promoter by EGF involves a ras/MEK1/AP-1-dependent pathway (60). Several reports have shown that PKC activates Raf, suggesting cross-talk between MAP kinase and PKC pathways (61)(62)(63). We showed that MAP kinase pathway might not be essential for EGF-induced phosphorylation of CREB in trophoblast cells. Although our data do not exclude possible cross-talk between MAP kinase and PKC pathways, all these results suggests that EGF induces CREB phosphorylation mainly through the PKC pathway in trophoblast cells, resulting in transcriptional activation of the hCG␣ gene.