The focus of our research is the determinants of two transcriptional properties of steroid receptors: the dose-response curve of agonists and the partial agonist activity of antisteroids. The dose-response curve defines the EC50, or steroid concentration at which half-maximal response is seen, and is a crucial but poorly understood component of steroid hormone endocrinology. Differences in the EC50s of regulated genes provide a mechanism for differential expression by the common concentration of circulating steroid hormone in an organism. The partial agonist activity of antisteroids is an important consideration for limiting unwanted side effects during endocrine therapies by potentially allowing partial expression of genes other than the one targeted for suppression. While the EC50 and partial agonist activity were long considered to be invariant, mounting evidence indicates that they can be modulated by equilibrium interactions with a variety of factors. For glucocorticoid receptors (GRs), three modulatory factors are the homologous receptor, coactivators such as TIF2, and Ubc9, which is the human homolog of the E2 ubiquitin-conjugating enzymes of yeast. To understand the molecular mechanisms of this modulation, we first asked whether the modulatory activity of each factor requires the same or different domains of GRs. In all cases, neither the amino-terminal half of the receptor, which contains the AF-1 activation domain, nor the DNA binding domain is required. This contrasts with most steroid receptors where the AF-1 domain plays a major role in determining the amount of gene expression and partial agonist activity of antisteroids. The situation is more complicated with Ubc9, where GR N-terminal sequences interact with the GR C-terminal ligand binding domain (LBD) to block the actions of Ubc9, but not added GR or TIF2, at low GR concentrations. Increasing the concentration of GRs, or removing the N-terminal sequences, permits the modulatory actions of Ubc9. A shift from non-cooperative to cooperative steroid binding at high GR concentrations suggests that conformational changes reposition the inhibitory N-terminal sequence to allow Ubc9 interaction with the LBD. Thus, the LBD is necessary and sufficient for the activity of these modulatory factors. The above study was performed with only one antiglucocorticoid. Because the determinants of the partial agonist activity of antisteroids complexed with steroid receptors are not well understood, it is critical to know whether the effects of various parameters is the same for a variety of antisteroids. We therefore examined the role of the N-terminal half of GR (including AF-1), DNA binding site sequence, receptor contact with DNA, and coactivator binding on the expression of partial agonist activity in two cell lines for GRs bound by five antiglucocorticoids: dexamethasone mesylate (Dex-Mes), dexamethasone oxetanone (Dex-Ox), progesterone (Prog), deoxycorticosterone (DOC), and RU486. We find that the GR AF-1 domain is dispensable for the partial agonist activity of these antiglucocorticoids. This contrasts with the AF-1 domain being required for the partial agonist activity of antisteroids with most steroid receptors. DNA sequence (MMTV vs. a simple GRE enhancer) and cell-specific factors (CV-1 vs. Cos-7) exert minor effects on the level of partial agonist activity and suggest a contribution of DNA-induced conformational changes. The correlation of coactivator TIF2/GRIP1 binding to GR-steroid complexes with the increase in partial agonist activity upon addition of coactivator in CV-1 cells is consistent with the hypothesis that the partial agonist activity of a given GR-steroid complex in CV-1 cells is influenced by the binding of coactivators. Collectively, the ability to modify the levels of partial agonist activity through changes in steroid structure, DNA sequence, specific DNA-induced conformational changes, and coactivator binding suggests that useful variations in endocrine therapies may be possible by the judicious selection of these parameters to afford gene and tissue selective results. To guide our further studies on the modulation of GR transactivation properties, we have constructed a mathematical model that predicts the observed Michaelis-Menton curve for the many-step process of steroid regulated gene induction. This model suggests that altering steps downstream from receptor binding to DNA responsive elements can modulate the EC50 of agonist-complexes. The downstream steps that we selected (with their inhibitors in parentheses) are protein deacetylation (with TSA and VPA), DNA unwinding (with CPT), and CTD phosphorylation of RNA polymerase II (with DRB and H8). None of the inhibitors prevented the effects of added GRs. However, inhibitors of DNA unwinding and CTD phosphorylation did reverse the effects of Ubc9 with high concentrations of GRs. These results further support our earlier conclusion that different rate limiting steps are operative at low GR concentrations ? GRs or coactivators vs. high GR with Ubc9. The present data also suggest that a variety of processes can modulate the EC50 and partial agonist activity for GR-regulated gene induction and that, under appropriate conditions, two of these processes may be DNA unwinding and CTD phosphorylation of RNA pol II. Finally, in an effort to more easily examine the effects of GR mutations on the modulation of the dose-response curve and partial agonist activity, we developed a method by which a large number of alanine point mutations are introduced into a protein by two rounds of PCR, with the DNA products being directly transfected into cells. This PCR expression mutagenesis (PEM) technique is used to identify several new residues of the GR LBD that influence ligand binding and/or transactivation. PEM thus provides a new method for quickly screening large quantities of mutant proteins. As a result of the above studies, we have gained new molecular information about the modulation of the dose-response curve of agonists and the partial agonist activity of antisteroids. These results suggest new avenues for controlling these parameters, which are of general and widespread importance for the differential control of gene expression during development, differentiation, and homeostasis and for more selective endocrine therapies. These combined findings contribute to our long-term goal of defining the action of steroid hormones at a molecular level and of understanding their role in human physiology.
Showing the most recent 10 out of 31 publications
