Optimal in vitro gene delivery with cationic lipids requires an excess of cationic charges with respect to DNA phosphates. In these conditions, in vivo delivery will be hampered by interference from cationic lipid-binding macromolecules either circulating or in the extracellular matrix. To overcome this problem, we are developing a modular transfection system based on lipid-coated DNA particles reminiscent of enveloped viruses. The particle core consists of the lipopolyamine-condensed nucleic acid in an electrically neutral ratio to which other synthetic lipids with key viral properties are hydrophobically adsorbed. As a first result, we have found that a good transfection level can be achieved simply with the neutral core particle, provided a zwitterionic lipid (dioleoyl phosphatidylethanolamine) is added to completely coat the DNA. Addition of lipids bearing a fusogenic or a nuclear localization peptide head group to the particles does not significantly improve an already efficient system, in contrast to polylysine-based gene transfer methods that rely on lysosomotropic or fusogenic agents to be effective. This emphasizes the distinctive properties of the lipopolyamines, including cell membrane destabilization, endosome buffering capacity, and possibly nuclear tropism. Most importantly, addition of lipids with a triantennary galactosyl residue drives the neutral nucleolipidic particles to the asialoglycoprotein receptor of human hepatoma HepG2 cells: Transfection increases approximately 1000-fold with 25% galactolipid. This receptor-mediated process is saturable and slightly less efficient than receptor-independent transfection obtained in vitro with a large excess of cationic lipid alone. Yet, electrically silent particles may provide an attractive solution for gene transfer in vivo where their external saccharide coat should allow them to diffuse within the organism and reach their target cells.