Lipids in biological membrane fusion

Conclusions

The results reviewed suggest that membrane fusion in diverse biological fusion reactions involves formation of some specific intermediates: stalks and pores. Energy of these intermediates and, consequently, the rate and extent of fusion depend on the propensity of the corresponding monolayers of membranes to bend in the required directions.

Proteins and peptides can control the bending energy of membrane monolayers in a number of ways. Monolayer lipid composition may be altered by different phospholipases [50, 85, 90], flipases and translocases [4, 50]. Proteins and peptides can change monolayer spontaneous curvature or hydrophobic void energy by direct interaction with membrane lipids [20, 32, 111]. Proteins may also provide some barriers for lipid diffusion in the plane of the monolayer [83, 141]. If diffusion of lipids at some specific membrane sites (e.g., in the vicinity of fusion protein) is somehow hindered, the energy of the bent fusion intermediates would reflect the elastic properties of these particular sites rather than the spontaneous curvature of the whole monolayers. Proteins may deform membranes while bringing them locally into close contact. The alteration of the geometric (external) curvature will certainly change the elastic energy of the initial state and, thus affect the energetic barriers of the formation of the intermediates [143]. In addition, the area and the energy of the stalk can be reduced by preliminary bending of the contacting membranes [111]. The possible effects of proteins and polymers on local elastic properties and local shapes of the membranes have been recently analyzed [22, 39, 45, 63]. These studies may provide a good basis for future development of theoretical models of protein-mediated fusion.

Various models for biological fusion have been presented as hypothetical sequences of intermediate conformations of proteins, with membrane lipids just covering the empty spaces between the proteins. Although the results discussed above do not allow us to draw an allexplaining cartoon of the fusion mechanism, they do indicate which properties of membrane lipid bilayers (if modified by fusion proteins) would get these bilayers to fuse. In addition, these data suggest a specific geometry to bent fusion intermediates (stalks and pores) and imply a contribution by lipids to the energy of these intermediates. We think that the synthesis of rapidly developing structural information on fusion proteins with the analysis of the physics of membrane rearrangement may soon yield a real understanding of the fascinating and fundamental phenomenon of membrane fusion.