Helical protein arrays on lipid tubules are valuable assemblies for studying protein structure and protein-lipid interactions through electron microscopy and crystallography. We describe conditions for forming such arrays from two proteins, IgG; and transducin, the photoreceptor Gr protein, using a variety of lipid surfaces. Anti-dinitrophenyl (DNP) IgG arrays formed on DNP-phosphatidylethanolamine (DNP-PE) mixed with either galactosylceramide lipids or phosphatidylcholine (PC) display different pH sensitivities and dimensions, yet have similar helical symmetries. DNP-PE/PC mixtures formed small crystals and large well-ordered flattened tubules. The peripheral membrane protein transducin (G(t)) formed helical arrays either on a mixture of cationic and neutral lipids or on residual photoreceptor lipids. Despite differences in lipid composition, the G(t) arrays have similar structures and show similar sensitivity to activation and variations in ionic environment. G(t) activation causes the helical assemblies to collapse to small vesicles, a process resembling the vesiculation of activated dynamin-lipid tubules. In a preliminary three-dimensional reconstruction, the hapten-bound IgG appears to make two contacts to the central lipid tubule, presumably via the F-ab domains. The ability to generate a three-dimensional reconstruction without tilts illustrates one advantage of helical structures for two-dimensional crystallography, especially for visualizing protein-lipid interactions.