We determined optimum conditions for delivering DNA to transformed human bronchial epithelial cells expressing wild-type (BEAS) or abnormal (2CF) cystic fibrosis transmembrane conductance regulator (CFTR) using cationic liposomes (Lipofectin, [N-(N,N-dimethylaminoethane)carbamyl] cholesterol[DC-Chol]/dioleoylphosphatidylethanolamine[DOPE], or LipofectAMINE) and reporter genes which measured overall transgene expression (luciferase) or the fraction of cells transfected (heat-stable alkaline phosphatase). All liposomes showed dose-related toxicity. Optimal liposome and lipid: DNA ratios were different for BEAS than for 2CF cells. For all 3 liposome preparations, small particle size and net cationic charge related to transfection efficiency. Both LipofectAMINE and DC-Chol/DOPE transfected a maximum of 3% of BEAS cells, but luciferase expression could be increased without increasing the fraction of cells transfected. LipofectAMINE transfected a maximum of 6% of 2CF cells, and luciferase expression could be increased with no further increase in fraction of transfected cells. DC-Chol/DOPE transfected over 12% of 2CF cells with relatively small increases in luciferase expression. We conclude that an optimal cationic liposome and lipid: DNA ratio for transfecting bronchial epithelial cells depends on: (1) small particle size and net cationic charge, (2) whether the cells have the cystic fibrosis defect, and (3) whether the desired outcome is transfection of the maximum fraction of the cells or maximum total expression of the transgene.