Is gonadotrope expression of the gonadotropin releasing hormone receptor gene mediated by autocrine/paracrine stimulation of an activin response element?

Expression of the FSHbeta subunit and GnRH receptor (GnRHR) genes in gonadotropes is stimulated by activin. We sought to identify the cis-acting element(s) in the murine GnRHR gene promoter which confer activin responsiveness. We established that 600 bp of 5'flanking sequence from the murine GnRHR gene were sufficient to confer activin responsiveness in the gonadotrope-derived alphaT3-1 cell line. Since alphaT3-1 cells, like gonadotropes, secrete activin, we examined the ability of follistatin, an activin binding protein, to block the activin response. Increasing concentrations of follistatin from 0 to 100 ng/ml resulted in a dose dependent decrease in activity of the -600 promoter. Contained within this region are three elements important for expression in alphaT3-1 cells: a Steroidogenic Factor-1 binding site (SF-1), an Activator Protein-1(AP-1) element, and an element termed the GnRH receptor activating sequence or GRAS. A block mutation of GRAS inhibited the ability of the promoter to respond to follistatin. A more refined analysis using a series of two-bp mutations which scan GRAS and flanking sequence revealed exact convergence of GRAS with activin/follistatin responsiveness. Finally, a construct consisting of 3 copies of GRAS placed upstream of a heterologous minimal promoter (3xGRAS-PRL-LUC) was responsive to both activin stimulation and follistatin inhibition in alphaT3-1 cells. Thus, autocrine/paracrine stimulation of gonadotropes by activin illustrates a unique mechanism for cell-specific gene expression.     

Differential expression of human gonadotropin-releasing hormone receptor gene in pituitary and ovarian cells

In terms of regulation of gene expression, gonadotropin-releasing hormone receptor (GnRHR) found in extrapituitary tissues has been suggested to be different from that in the pituitary. In the present study, we examined the molecular basis of this difference using the pituitary alphaT3-1 and ovarian carcinoma OVCAR-3 cells. As a first step, the different expression levels of GnRHR mRNA in the pituitary and ovarian cells were investigated using semi-quantitative RT-PCR. Quantitative analysis showed that the expression level of hGnRHR is a nine-fold higher in primary pituitary tissues than the primary culture of ovarian carcinomas (PCO). In pituitary alphaT3-1 cells, the expression level of hGnRHR was ten-fold higher than ovarian carcinoma OVCAR-3 cells. The possibility of the differential use of various cell-specific promoters in different cells was addressed by transiently transfecting cells with 3'-deletion clones of human GnRHR promoter. Sequential deletion of the 5'-flanking region of the gene revealed the use of two putative promoters, designated PR1 (-771 to -557) and PR2 (-1351 to -1022), and a negative control region (-1022 to -771), in the pituitary and ovarian cells. The same promoters appeared to be utilized for driving the basal promoter activities in both alphaT3-1 and OVCAR-3 cells, suggesting that there is no cell-specific promoter usage for the human GnRHR gene. Alternatively, the involvement of different regulatory protein factors was investigated using electrophoretic gel mobility shift assays. When end-labeled PR1 was used as a probe, two unique shifted complexes were identified in OVCAR-3 cells when compared to alphaT3-1 cells. One unique protein-DNA complex was observed in alphaT3-1 cells compared to OVCAR-3 cells when incubated with end-labeled PR2 as a probe. These DNA-protein complexes appeared to be specific, as they competed with excess amount of unlabelled competitor PR1 and PR2, respectively. In summary, it is unlikely that the use of a cell-specific promoter contributes to the different characteristics of ovarian GnRHR. Our study demonstrates that one mechanism by which cell-specific expression of human GnRHR is achieved is through the binding of distinct and/or combinations of cell-specific regulatory factors to various promoter elements in the 5'-flanking region of the gene.     

Identification of a negative regulatory element involved in tissue-specific expression of mouse renin genes.

The 5' flanking region of the mouse renin genes (Ren-1d and Ren-2d) contains two motifs that are homologous to known negative regulatory elements (NREs). Ren-2d has a 150-base-pair (bp) insertion 5' to the upstream putative NRE (NRE-1), which is lacking in Ren-1d. We tested the functionality of these sequences by using site-directed mutagenesis to delete individually each putative NRE from Ren-1d and to delete the 150-bp insertion from Ren-2d. We examined the effect of these mutations on the expression of the reporter gene chloramphenicol acetyltransferase, which was expressed from a truncated thymidine kinase promoter fused to the renin regulatory region. This plasmid was transfected into human choriocarcinoma JEG-3 cells. Only the upstream NRE (positions -619 to -597) was found to be functional in Ren-1d. The deletion of a 150-bp insertion from Ren-2d resulted in the suppression of chloramphenicol acetyltransferase activity to the level of Ren-1d expression. These data suggest that the upstream NRE that is functional in Ren-1d, but not in Ren-2d, may be partly responsible for differential expression of the renin genes in various tissues. The molecular mechanism of the NRE was examined by studying its interaction with nuclear proteins in submandibular gland and JEG-3 cells by gel-mobility-shift assays. Specific nuclear protein binding was observed only to the upstream NRE and the molecular mass of this protein was approximately 72 kDa as determined by Southwestern blot analysis. Thus our results suggest that both Ren-1d and Ren-2d conserve a cis-acting NRE in the 5' flanking region. In Ren-1d, this NRE could bind a specific nuclear protein resulting in the inhibition of Ren-1d expression in these tissues. On the other hand, the NRE in Ren-2d is nonfunctional due to interference by an adjacent 150-bp insertion.     

Oct-1, silencer sequence, and GC box regulate thyroid hormone receptor β1 promoter

Thyroid hormone, acting through thyroid hormone receptors (TRs), plays a crucial role in brain development and its insufficiency results in irreversible brain damage. TR mRNA is expressed continuously from early embryonic stages, but the level of TRβ1 mRNA in brain is more abundant in adult than in fetus. To identify important factors which regulate TRβ1 expression, we compared mouse fetal and adult brain nuclear extracts by DNase I footprinting and electrophoretic gel mobility shift assays (EMSA) of the TRβ1 promoter. We carried out transient transfection studies in COS 1 cells using the TRβ1 promoter fused to Luciferase gene, and used mutated promoter vectors and various expression vectors. In DNase I footprinting using the fragment -950 to -717, fetal brain nuclear extracts protected the areas -910 to -884 and -815 to -800 more than did adult extracts. In EMSA, proteins in fetal nuclear extracts bound to a silencer sequence (−924 to -916), GC box (−901 to -887), and E box (−810 to -805), more strongly than did proteins in adult brain extracts. The bands formed on GC box were not supershifted by Sp-1, Sp-2, Sp-3, Sp-4, EGR-1, or EGR-2 antibodies. Three bands were detected on the octamer binding site probe (−913 to -906) and one protein was supershifted by Oct-1 antibody. Adult brain extracts appear to contain more Oct-1 protein than do fetal extracts. The other two bands were more intense in fetal extracts than in adult extracts, but were not supershifted by either Oct-1 or Oct-2 antibodies. Mutation of the silencer response element, mutation of the GC box, and Oct-1 over expression in COS 1 cells increased TRβ1 promoter function as assayed by Luciferase reporter. Mutation of the octamer binding site, to which only Oct-1 bound in COS 1 cells, decreased Luciferase reporter activity. Thus the TRβ1 promoter was regulated negatively by the proteins bound to the silencer sequence and the GC box, and positively by Oct-1. Silencer and GC box binding proteins are more abundant in fetal brain, and Oct-1 is more abundant in adult brain. The results may be responsible for increased amounts of TRβ1 present in late fetal and adult brain.