Extreme position dependence of a canonical hormone response element.

Abstract

Hormone response elements (HREs) are considered enhancers, activating transcription in a relatively position- and orientation-independent fashion. Upon binding to an HRE, steroid receptors presumably contact coactivators and/or proteins associated with the transcription initiation complex. As a receptor target site is moved further from a fixed position such as the TATA box, not only will the spatial separation of the receptor with respect to its interaction partners change, so will the orientation due to the rotation of the DNA helix. Additional constraints may be imposed by the assembly of DNA into chromatin. Therefore, we have endeavored to test rigorously the assertion that HRE action is position independent. We have constructed a series of 42 chloramphenicol acetyltransferase expression vectors that contain a single progesterone/glucocorticoid receptor-binding site separated from a TATA box by 4 to 286 bp. The enhancer activity of the HRE was assessed after transient transfection of progesterone receptor-expressing fibroblasts. We find that the position of the HRE has a dramatic influence on induction by progestins. When closely juxtaposed to the TATA box, the HRE was unable to support a hormone response, perhaps due to direct steric hindrance with the transcription initiation complex. Full activity was gained by moving the HRE 10 bp further from the TATA sequence. As the HRE was moved incrementally further, activity remained near maximal over the next 26 bp. HRE activity then declined over the subsequent 26 bp and remained low for another 2.5 helical turns. Surprisingly, a narrow window of HRE activity occurred at an HRE-TATA box separation of 90-100 bp. Little or no hormone-induced transcriptional activity was observed when the HRE was positioned further from the TATA box. The addition of a second HRE or a basal (nuclear factor-1) element failed to relieve this constraint. A similar series of experiments was carried out in a mammary carcinoma cell line that expressed high levels of both glucocorticoid and progestin receptors. Data in these cells indicate that glucocorticoids and progestins supported a similar HRE position-activity profile, but this pattern of HRE activity was quite distinct from that seen in fibroblasts. This may be indicative of cell type-specific interactions between steroid receptors and adapter/coactivator proteins or cell type-specific activities such as acetylases or deacetylases participating in the steroid response.