Identification of a novel dexamethasone responsive enhancer in the human CYP3A5 gene and its activation in human and rat liver cells.

The human liver cytochromes P450 3A (CYP3As), orthologous to the rat glucocorticoid inducible forms, are composed of at least four differentially expressed members. To begin the study of the molecular events in the glucocorticoid regulation of CYP3A5, we fused 5' sequences of CYP3A5 to the chloramphenicol acetyltransferase gene in a vector that contains the herpes simplex virus thymidine kinase promoter. In HepG2 cells, the largest 5' CYP3A5 gene fragment (1.4 kb) suppressed the TK promoter. However, suppression was overcome by addition of 10 microM dexamethasone. A series of unidirectional deletions revealed a unique 219-bp fragment (-891 to -1109 bp upstream from the transcriptional start site) that conferred dexamethasone responsiveness on the TK promoter regardless of either the distance or orientation from the promoter and thus appears to be an enhancer. Nucleotide sequence analysis of this CYP3A5 enhancer revealed no consensus 15-bp glucocorticoid responsive element (GRE) (GGTACANNNTGTTCT); however, two GRE "half-sites" (TGTTCT) were found separated by 160 bp. Although dexamethasone stimulated the CYP3A5 enhancer only 3-4-fold in HepG2 cells, the CYP3A5 enhancer was stimulated 7- and 12-fold in immortalized primary human hepatocytes and primary rat hepatocyte cultures, respectively. The glucocorticoid receptor (GCR) seems to be indispensable to this process because 1) dexamethasone induction can be blocked by the antiglucocorticoid RU-486, 2) dexamethasone-dependent transcriptional activation of the CYP3A5 enhancer in HepG2 cells required cotransfection of an expression vector containing the intact GCR, yet 3) cotransfection with a plasmid that contains a mutation in the ligand binding domain of the GCR does not activate the CYP3A5 enhancer in the presence of dexamethasone. To further localize the dexamethasone responsive region of the 219-bp CYP3A5 enhancer, it was subdivided and fused to the TKCAT expression vector. Transfection analysis in HepG2 cells demonstrated that neither GRE half-site can independently confer dexamethasone responsiveness on the TK promoter. Block mutations of either of the two GRE half-sites or point mutations at specific GCR binding sites eliminates dexamethasone inducibility, demonstrating the half-sites need to interact. Electromobility shift assays indicate that the CYP3A5 5'-GRE half-site 1) specifically binds purified GCR, 2) can displace binding of the GCR to a consensus GRE, and 3) shifts a protein in HepG2 nuclear extracts that is supershifted by GCR antibody, demonstrating that this enhancer is an authentic GRE. This is the first study to demonstrate that a member of the human CYP3A gene family contains an enhancer that binds the GCR and that this binding is critical to transcriptional activation by dexamethasone.