We have tested the zeta potential (zeta, the surface charge density) of transfection complexes formed in serum-free medium as a rapid and reliable technique for screening transfection efficiency of a new reagent or formulation. The complexes of CAT plasmid DNA (1 mu g) and DC-chol/DOPE liposomes (3-20 nmol) were largely negatively charged (zeta= -15 to -21 mV), which became neutral or positive as 0.5 mu g or a higher amount of poly-L-lysine (PLL, MW 29 300 or MW 204 000) was added (-3.16 +/- 3,47 to +6.04 +/- 2.23 mV). However, the complexes of CAT plasmid DNA (1 mu g) and PLL MW 29 300 (0.5 mu g or higher) were neutral or positively charged (-3.22 +/- 2.3 to +5.55 +/- 0.64 mV), which remained the same as 6.6 nmol of the liposomes was added. The complexes formed between two positively charged compounds, PLL MW 29 300 (0.5 mu g) and the liposomes (3.20 nmol), were as closely positively charged as DNA/PLL or DNA/liposomes/PLL complexes (+3.31 +/- 0.41 to 7.16 +/- 1.0 mV). These results indicate that PLL determined the overall charge of the DNA/liposome/PLL ternary complexes. The complexes formed with histone (0.75 mu g or higher) were also positively charged, whose transfection activity was as high as PLL MW 39 300. However, the complexes formed with protamine or PLL MW 2400 remained negatively charged. These observations are in good agreement with the transfection activity of the formulation containing each polycationic polymer. The presence of PLL MW 29 300 did not change the hydrodynamic diameter of DNA/liposome/PLL complexes (d(H) = 275-312 nm). The complexes made of different sizes of PLL (MW 3400 and 204 000) also did not significantly change their size. This suggests that DNA condensation may not be critical. Therefore, zeta of the transfection complex can predict the transfection efficiency of a new formulation or reagent. (C) 2000 Elsevier Science B.V. All rights reserved.