Effect of single chain lipids on phospholipase C-promoted vesicle fusion. A test for the stalk hypothesis of membrane fusion.

The effect of low proportions (up to 5 mol %) of single-chain lipids on phospholipase C-promoted fusion of large unilamellar vesicles has been investigated with the aim of testing the so-called stalk model of membrane fusion. This model is known in two main versions, the one originally published by Kozlov and Markin [Kozlov, M. M. and Markin, V. S. (1983) Biofizika 28, 255-261] and what is known as the "modified stalk model" [Siegel, D. P. (1993) Biophys. J. 65, 2124-2140], that differ in a number of predictions. In the view of the latter author, hydrocarbons or other nonpolar lipids should help fusion by decreasing the interstitial energy of the stalk connecting the two apposed bilayers. We show that small amounts of hexadecane or squalene increase significantly the fusion rates in our system. Changes in monolayer curvature are the object of different predictions by the original and modified stalk theories. According to the original form, fusion would be promoted by lipids inducing a negative curvature in the closest (cis) monolayers of the fusing membranes and inhibited by the same lipids in the trans monolayers; the opposite would happen with lipids inducing a positive curvature. The modified stalk model predicts that fusion is helped by increasing the negative curvature of both monolayers. In our system, symmetrically distributed arachidonic acid, which increases the negative curvature, enhances lipid and content mixing, and the opposite is found with symmetrically distributed lysophosphatidylcholine or palmitoylcarnitine, which facilitate a positive monolayer curvature. In addition, fluorescence polarization and 31P NMR studies of the lamellar-to-isotropic (Q224 cubic) thermotropic transition of a lipid mixture corresponding to our liposomal composition reveal that all lipids that facilitate fusion decrease the transition temperature, while fusion inhibitors increase the transition temperature. Moreover, fusion (content mixing) rates show a maximum at the lamellar-to-isotropic transition temperature. These observations support the involvement of inverted lipid structures, as occurring in the inverted cubic phases, in membrane fusion. All these data are in full agreement with the stalk model of membrane fusion, particularly in its modified version.