These patches may have originated as membrane tubules laid down on the mica surface during image acquisition, and subsequent enlargement might have been due to the diffusion of phospholipids from accumulation sites. Membrane tubules induced by MinE were coiled and bent , this differed from the smooth contour of those caused by the external force of buffer purposely blown over the SLBs . The images of fluorescently labeled MinE colocalized with membrane tubules indicated that tubule formation was associated with MinE . Replacing wild-type MinE with MinE1�C31 in the SLB experiments resulted in the formation of fluorescent foci, but no obvious membrane tubules were seen . Atto488-labeled anti-MinE antibody was used to identify MinE1�C31 on the fluorescent patches. MinE1�C31 was found at the vicinity of the lipid patches, but was not completely superimposed on them . We also identified arcs and enclosed rings of MinE1�C31 surrounding larger lipid patches. These data suggest that the association of MinE1�C31 with membranes resulted in the local accumulation of surrounding phospholipids. The number of phospholipids between the accumulation points significantly decreased and contributed to the reduction in background fluorescence . The differences between MinE1�C31 induced membrane deformation of giant vesicles and SLBs may reside in the continuity of the lipid supplies. Lipids were continuously drawn into the growing tubules in the giant vesicles until transformation was complete. The initiation points for tubule formation on SLBs were scattered and lipids were drawn independently into separate foci. This resulted in a shortage of lipids, which was not able to support tubule growth. These data indicate that MinE is able to cause membrane deformation and induce tubule formation in a flat membrane, which PR-957 further confirms our observation using the giant liposome system. We constructed a mutant MinE protein by substituting F6 with aspartic acid to weaken the amphipathicity of MinE2�C9. In the sedimentation assays, the purified mutant protein MinEF6D only retained 45% of the ability to co-sediment with liposomes , indicating the importance of this residue in supporting the protein-membrane interaction. The remaining Afatinib EGFR/HER2 inhibitor hydrophobic residues, A2 and L3, and the charged residues R10, K11, and K12 may have sustained part of the interaction. In addition, the large hydrophobic face might have allowed the mutant helix to rotate and associate with the membrane. Time-lapse fluorescence microscopy was used to examine liposome deformation induced by the mutant MinE proteins C1 and MinEF6D, which were defective in membrane association.