No abstract

Members of the Oct family of transcription factors specifically interact with the octamer motif, ATGC-AAAT, a regulatory element important for tissue- and cell-specific transcription as well as for the expression of housekeeping genes. Except for Oct-1, all Oct factors are expressed in a temporally and spatially restricted mode during murine development and their number varies in a given cell type. Despite its ubiquitous expression pattern Oct-1 may play a role in murine development. As a first step towards elucidating the role of Oct-1 we report the complementary DNA cloning of the mouse Oct-1 gene. Two large transcripts of 5 and 14 kb are derived from a single gene. The expression patterns of three splicing products of Oct-1 are similar in a number of cells and tissues. In the POU region murine Oct-1 differs in four amino acids from the human homologue and these differences are restricted to helices 1 and 2. Interestingly, two of the four variant amino acids are identical to those in human and mouse Oct-2 and thus the murine Oct-1 homeodomain is intermediary in sequence between human Oct-1 and Oct-2. These two amino acids together with a third one have been shown to be relevant for the interaction between human Oct-1 and herpes simplex virus transactivator VP16. Nevertheless, VP16 interacts albeit weakly with murine Oct-1. We speculate that the differences in the human and mouse Oct-1 homeodomains reflect host-specific differences in protein-protein interactions.

The progesterone-binding protein uteroglobin has been expressed in Escherichia coli in an unfused, soluble form. Like mature uteroglobin from rabbit endometrium (UG), the E.coli produced uteroglobin (UG1) dimerizes in vitro, forms an antiparallel dimer with Cys3-Cys69' and Cys69-Cys3' disulfide bonds and binds progesterone under reducing conditions. In order to analyze the dimerization and the reduction dependence of progesterone binding in more detail, we separately replaced cysteine 3 and cysteine 69 by serines. Under reducing conditions, both uteroglobin variants (UG1-3Ser and UG1-69Ser) bind progesterone with the same affinity as the wild-type suggesting that both cysteine residues are not directly involved in progesterone binding. In contrast to the wild-type protein, both cysteine variants also bind progesterone with high affinity in the absence of reducing agents. In addition, UG1-3Ser and UG1-69Ser both form covalently linked homodimers. Thus, unnatural Cys69-69' and Cys3-3' disulfide bonds exist in UG1-3Ser and UG1-69Ser, respectively. These data together with computer models based on X-ray diffraction data strongly support the idea that progesterone reaches its binding site located in an internal hydrophobic cavity via a hydrophobic tunnel along helices 1 and 4. Under non-reducing conditions the tunnel is closed by two disulfide bridges (Cys3-Cys69' and Cys69-Cys3') that lie in the most flexible region of the dimer. Reduction or replacement of a cysteine residue enables conformational changes that open the channel allowing progesterone to enter.

No abstract

To understand the molecular mechanism of endometrial differentiation we have initiated an analysis of the uteroglobin promoter. Uteroglobin is normally expressed in endometrial tissues under the control of ovarian hormones. In gene transfer experiments with the Ishikawa cell line, derived from a human endometrial adenocarcinoma, we have identified several regions in the promoter of the uteroglobin gene that are responsible for its endometrium-specific expression. To evaluate the generality of these findings, we have begun cloning the promoter regions of potential endometrial markers, including the rat, mouse, and human uteroglobin gene. In the rat, expression of the uteroglobin-like gene, CC10, is dominant in the lung but is also observed in the endometrium of progesterone treated animals. A comparison of the 5'-flanking sequence of the rat and rabbit uteroglobin gene resulted in the detection of similarities and differences that could explain their differential expression in vivo. To substantiate these findings we have established several cell lines from rat endometrium using murine retroviral vectors containing a positive selection marker and various viral oncogenes, such as SV40 large T antigen, adenovirus E1A, and Ha-ras. Cell lines immortalized by SV40 T-antigen were subsequently transformed with the Ha-ras oncogene. Several cell lines exhibit properties of epithelial endometrial cells. Two cell lines generated with a temperature sensitive mutant of the SV40 large T-antigen grow as transformed cells at the permissive temperature, but differentiate upon shifting to the non-permissive temperature. These rat endometrial cell lines should be useful for the analysis of endometrium-specific gene expression and as model systems for endometrial carcinoma.