A novel AP-1 site is critical for maximal induction of the follicle-stimulating hormone beta gene by gonadotropin-releasing hormone

Abstract

Regulation of follicle-stimulating hormone (FSH) synthesis is a central point of convergence for signals controlling reproduction. The FSHbeta subunit is primarily regulated by gonadotropin-releasing hormone (GnRH), gonadal steroids, and activin. Here, we identify elements in the mouse FSHbeta promoter responsible for GnRH-mediated induction utilizing the LbetaT2 cell line that endogenously expresses FSH. The proximal 398 bp of the mouse FSHbeta promoter is sufficient for response to GnRH. This response localizes primarily to an AP-1 half-site (-72/-69) juxtaposed to a CCAAT box, which binds nuclear factor-Y. Both elements are required for AP-1 binding, creating a novel AP-1 site. Multimers of this site confer GnRH induction, and mutation or internal deletion of this site reduces GnRH induction by 35%. The same reduction was achieved using a dominant negative Fos protein. This is the only functional AP-1 site identified in the proximal 398 bp, since its mutation eliminates FSHbeta induction by c-Fos and c-Jun. GnRH regulation of the FSHbeta gene occurs through induction of multiple Fos and Jun isoforms, forming at least four different AP-1 molecules, all of which bind to this site. Mitogen-activated protein kinase activity is required for induction of FSHbeta and JunB protein. Finally, AP-1 interacts with nuclear factor-Y, which occupies its overlapping site in vivo.

FIG. 1. Localization of the GnRH-responsive region in the mouse FSHβ promoter

A, the 398-bp mouse FSHβ promoter was linked to a luciferase reporter gene (mFSHβ–luc) and transiently transfected into gonadotrope-derived LβT2 cells. The herpes thymidine kinase β-galactosidase reporter was co-transfected as an internal control. After overnight starvation in serum-free Dulbecco’s modified Eagle’s medium, the cells were treated with the indicated concentrations of GnRH for different lengths of time (indicated on the abscissa), after which the luciferase activity in the lysates was measured and normalized to β-galactosidase. Results represent the mean ± S.E. of at least three independent experiments, each performed in triplicate, and are presented as -fold induction from vehicle control. B, different lengths of the mouse regulatory region were transiently transfected into LβT2 cells and after overnight starvation, the cells were treated with 10 nm GnRH for 6 h. The results are represented as -fold induction from vehicle-treated cells for each truncation. Significantly different induction in the treated cells versus the control cells for each truncation is marked with an asterisk, whereas a significant drop in induction between the subsequent truncations is marked with a number sign. Results represent the mean ± S.E. of four independent experiments, each performed in triplicate.

FIG. 2. The −99/−65 region of the mouse FSHβ gene contains binding sites for NF-Y and AP-1

A, an NF-Y site and an adjacent AP-1 half-site were identified in the mouse FSHβ promoter using the Transfac® data base. Alignment of the sequence from −99 to −65 of the mouse FSHβ gene regulatory region reveals that the NF-Y site (underline) and the AP-1 half-site (dashed underline) are conserved in human and rodent species but are absent from the ovine and bovine promoters. Consensus binding sites are noted below the alignment. B, EMSA analysis of nuclear extracts from LβT2 cells treated with vehicle (0 h) or 10 nm GnRH (0.5, 2, or 6 h), using the −99/−65 sequence as a probe, are shown. The length of GnRH treatment in hours is indicated above each lane, and the antibodies used in supershift assay are marked above corresponding lanes. IgG represents a nonspecific antibody used as a control. The supershifted bands are indicated with ss, whereas 1 and 2 designate complexes that change following the treatment.

FIG. 3. Competition EMSA using oligonucleotides with scanning mutations as competitors reveals the base pairs required for NF-Y and AP-1 binding

A, alignment of wild-type sequence (WT) found in the mouse FSHβ promoter −99/−65 and oligonucleotides used as competitors (labeled A–K) is shown. Scanning mutations were introduced into oligonucleotides A–K, and these changes are underlined. These unlabeled oligonucleotides were used in a 100-fold excess in EMSA experiments in B–D, whereas wild-type sequence (WT) was used as a probe. The NF-Y binding site is underlined with a solid line in the wild-type sequence, whereas the AP-1 half-site is underlined with a dashed line. B, nuclear extracts from control cells were subjected to EMSA with radiolabeled wild-type probe. Mutated oligonucleotides were used as competitors in a 100-fold excess in the corresponding lanes. In the lane labeled AP-1, the AP-1 consensus sequence was used as a competitor in the same manner. C, nuclear extracts from cells treated with GnRH for 6 h were subjected to EMSA with the wild-type probe, and the same oligonucleotides with mutations as above were used as competitors. D, NF-Y antibody was added to nuclear extracts from the cells treated with 10 nm GnRH for 6 h, and competition EMSA was performed with the wild-type probe and competitor oligonucleotides indicated above the lanes.

FIG. 4. The AP-1 binding site is required for maximal induction with GnRH

A, mutations I and F (see Fig. 3) were introduced into the mFSHβ-luc vector, and transfections were performed using LβT2 cells. Cells were treated with 10 nm GnRH for 6 h, after which the luciferase activity was measured and normalized to β-galactosidase. The results are represented as -fold induction from the control cells transfected with the same reporter vector. The bar labeled AP-1 represents the induction of the reporter with a consensus AP-1 binding site introduced into the mFSHβ-luc reporter vector in place of the NF-Y/AP-1 site. Significantly different induction in the treated cells versus the control cells for each reporter is marked with an asterisk, whereas a significant difference in induction of the mutated reporter from induction of the wild-type reporter is marked with a number sign. Results represent the mean ± S.E. of four independent experiments, each performed in triplicate. B, a 9-bp deletion of the NF-Y/AP-1 site was created in mFSHβ-luc, and its induction following GnRH treatment was compared with the wild-type reporter. Significantly different induction in the treated cells versus the control cells for each reporter is marked with an asterisk, whereas a significant difference in induction of the mutated reporter from induction of the wild-type reporter is marked with a number sign. C, a reporter gene with four copies of the NF-Y/AP-1 element upstream of the thymidine kinase promoter was created, and its induction was compared with the induction of the control luciferase reporter driven by the thymidine kinase promoter. The activity from GnRH-treated cells was normalized to the activity from control cells, and results are represented as -fold induction. An asterisk marks that the induction of the reporter with multimerized AP-1 site is significantly different from the induction of the control plasmid.

FIG. 5. GnRH induces multiple AP-1 isoforms in LβT2 cells

A, in vitro transcribed and translated c-Fos and c-Jun were used in EMSA to compare their binding with complexes formed using LβT2 nuclear extract. Lane 1, reticulocyte lysate control; lane 2, control nuclear extracts; lane 3, nuclear extracts following 6 h of 10 nm GnRH treatment; lane 4, c-Jun and c-Fos in vitro transcribed and translated in reticulocyte lysate. B, LβT2 cells were treated with vehicle (0 h) or 10 nm GnRH for 1 or 3 h, after which the cells were lysed. Equal amounts of protein from whole cell lysates were run on the gel, and after transfer, the membranes were probed with antibodies specific for the indicated proteins. After secondary antibody, enhanced chemiluminescence was performed, and the blots were exposed to film. C, EMSA using the −99/−65 sequence from the mouse FSHβ promoter indicates that at least four different AP-1 isoforms bind the site. The length of GnRH treatment and the isoform specificity of the antibodies used for the supershift assay are indicated above the lanes. Lanes marked ns in the isoform lane included non-isoform-specific antibodies to Fos and Jun proteins, which therefore interacted with all isoforms of Fos and Jun families, respectively.

FIG. 6. Co-transfections using dominant negative Fos (A-Fos) indicate that AP-1 is necessary for maximal induction either by GnRH (A) or by overexpression of c-Jun and c-Fos (B)

A, cells transfected with wild-type mFSHβ-luc or mutation I introduced into mFSHβ-luc were treated for 6 h with vehicle (control) or 10 nm GnRH. Cells were co-transfected with an expression vector for dominant negative Fos (A-Fos) or its vector control to assess the necessity for AP-1 in the inductions by GnRH. Results represent the means of four independent experiments, each performed in triplicate. Results were analyzed by analysis of variance and Tukey-Kramer post hoc test, and an asterisk indicates a statistically significant difference from the control vehicle-treated cells, whereas a number sign indicates a significant difference from GnRH-treated cells co-transfected with wild-type reporter and empty vector control for A-Fos. B, cells transfected with wild-type mFSHβ-luc or mutation I introduced into mFSHβ-luc were co-transfected with c-Jun and c-Fos and with either dominant negative A-Fos or vector control. Twenty-four h after transfection, luciferase activity was measured and normalized to β-galactosidase. None of the vector controls for either c-Jun, c-Fos, or A-Fos had any effect on reporter expression (data not shown). Results represent the means of four independent experiments, each performed in triplicate. An asterisk indicates a statistically significant difference from the control cells, whereas a number sign indicates a significant difference from cells transfected only with c-Jun and c-Fos.

FIG. 7. Inhibition of the MAPK pathway during GnRH treatment prevents maximal induction of FSHβ and JunB

A, cells transfected with wild-type mFSHβ-luc were treated for 6 h with vehicle (control) or 10 nm GnRH. Cells were co-treated with the MEK inhibitor, UO126, at 1 µm to assess the necessity of the MAPK pathway in induction by GnRH. Results represent the mean of five independent experiments, each performed in triplicate. Results were analyzed by analysis of variance and Tukey-Kramer post hoc test, and an asterisk indicates a statistically significant difference from the control vehicle-treated cells, whereas a number sign indicates a significant difference from GnRH-treated cells without UO126. B, Western blot of JunB in whole cell lysates of LβT2 cells following 0, 1, or 3 h of GnRH treatment with or without the MEK inhibitor, UO126. The experiment was repeated three times, and a representative gel is shown.

FIG. 8. ChIP reveals that NF-Y binds DNA in the proximal mouse FSHβ promoter in both control and cells treated with GnRH for 3 h, whereas AP-1 binds DNA following GnRH treatment

A, chromatin was isolated from LβT2 cells treated with vehicle or 10 nm GnRH for 3 h and cross-linked with formaldehyde. After sonication, sheared chromatin was precipitated with the antibodies indicated above the lanes. The precipitated and purified DNA is then amplified in the PCR. The antibody specific for NF-Y precipitates the DNA specific for the sequence in the proximal mouse FSHβ promoter, in both control and GnRH-treated cells. Fos and Jun, on the other hand, bind DNA in vivo only following the GnRH treatment. In the first two lanes, chromatin was precipitated with protein A beads only serving as controls. B, chromatin prior to precipitation serves as the control for the amount of chromatin used for precipitation in the untreated and GnRH-treated samples. A serial dilution of the chromatin was performed and then used in PCR together with precipitated samples. C, four independent experiments were quantified using a PhosphorImager and then normalized to intensity in the control sample precipitated with protein A beads only to normalize for any difference in the activity of the [β-³²P]dATP used in PCRs. The solid bars represent chromatin immunoprecipitation from GnRH-treated cells, whereas open bars represent control samples.

FIG. 9. GST pull-down assays demonstrate that NF-YA can interact directly with c-Jun

³⁵S-Labeled proteins, indicated above each panel, were used in the binding assay with GST fusion proteins, labeled above each lane. GST fusion proteins were induced with isopropyl-β-d-thiogalactosidase overnight, and the bacterial pellets were sonicated. These proteins were bound to glutathione-Sepharose beads, and in vitro transcribed and translated labeled proteins were added. After extensive washing, the precipitates were run on a gel and subjected to autoradiography.