Intracellular membrane docking and fusion requires the interplay between soluble factors and SNAREs. The SNARE hypothesis＇ postulates that pairing between a vesicular v-SNARE and a target membrane z-SNARE is the primary molecular interaction underlying the specificity of vesicle targeting as well as lipid bilayer fusion. This proposal is supported by recent studies using a minimal artificial system(2). However, several observations demonstrate that SNAREs function at multiple transport steps and can pair promiscuously, questioning the role of SNAREs in conveying vesicle targeting(3-6). Moreover, other proteins have been shown to be important in membrane docking or tethering(7-9). Therefore, if the minimal machinery is defined as the set of proteins sufficient to reproduce in vitro the fidelity of vesicle targeting, docking and fusion as in vivo, then SNAREs are not sufficient to specify vesicle targeting. Endosome fusion also requires cytosolic factors and is regulated by the small GTPase Rab5 (refs 10-20). Here we show that Rab5-interacting soluble proteins can completely substitute for cytosol in an in vivo endosome-fusion assay, and that the Rab5 effector EEA1 is the only factor necessary to confer minimal fusion activity. Rab5 and other associated proteins seem to act upstream of EEA1, implying that Ra65 effecters comprise both regulatory molecules and mechanical components of the membrane transport machinery. We further show that EEA1 mediates endosome docking and, together with SNAREs, leads to membrane fusion.