Estrogen administration to postmenopausal women has been shown to increase plasma levels of apolipoprotein (apo) A-I. A human hepatoma cell line, Hep G2, was used to test the hypothesis that estrogen increases the hepatic production of apo A-I by modulating gene expression. When Hep G2 cells were treated fur 24 hours with E-2, the apo A-I content in the medium increased 4.3+/-1.0-fold at 10 mu mol/L E-2 and 1.8+/-0.4-fold at 1 mu mol/L E-2 compared with untreated cells. A time-course experiment indicated that there was no E-2-dependent (10 mu mol/L) increase in apo A-I medium content at 1 hour and 2 hours and that apo A-I was 165% of controls at 6 hours and 440% at 24 hours. Hep G2 cells were transfected, by the cationic lipid method, with constructs containing serial deletions of the 5' region of the apo A-I gene (-41/+397, -256/+397, and -2500/+397) cloned in front of the luciferase gene and with or without a 7-kb region spanning the apo C-III/A-IV intergenic region, which has been shown to contain regulatory elements for the expression of the apo A-I gene. With the exception of the construct containing only the basal promoter (-41/+397), the expression of all constructs was 2- to 3-fold greater in the presence of E-2. The smallest construct that maintained E-2 responsiveness,the -256/+397 construct, does not contain a typical estrogen-responsive element. In the same transfection experiments, the 4-fold increase in apo A-I in the culture medium was preserved. However, when the same set of transfections was performed by the calcium phosphate precipitation method, the E-2 effect on the apo A-I content in the culture medium and on transcription activation was nearly abolished. This effect was probably mediated by Ca2+ because incubation of cells with 20 mmol/L CaCl2 abolished the E,response. In conclusion, E-2 increases apo A-I production in hepatic cells by increasing the transcription of the apo A-I gene.