Analysis of the Semliki Forest virus requirement for cholesterol during membrane fusion and infection

Abstract

Enveloped viruses infect cells by fusing their membrane with a cellular membrane. In vitro membrane fusion assays have shown that some enveloped viruses fused better with membranes that contain cholesterol, and one virus, Semliki Forest virus (SFV), required cholesterol for fusion with artificial membranes. Further analysis revealed that any 3{dollar}\beta{dollar}-hydroxy sterol could facilitate SFV mediated membrane fusion. Any alteration of cholesterol's 3{dollar}\beta{dollar}-hydroxyl group diminished fusion, suggesting that cholesterol's role in fusion is not an indirect effect of altered membrane fluidity. Fusion is mediated by the viral spike protein, a heterotrimer of two class 1 transmembrane peptides, E1 and E2, and a peripheral protein, E3. E1 must undergoes acid specific conformational changes before membrane fusion can occur. The truncated ectodomain of E1 also undergoes acid specific conformational changes, but only in the presence of cholesterol.;We decided to test the role of cholesterol in SFV mediated membrane fusion in vivo. We established two systems, partially cholesterol-depleted mammalian BHK cells and fully cholesterol-depleted mosquito C6/36 cells, to analyze the efficiency of SFV fusion with natural host cell membranes containing little or no cholesterol. Cholesterol in a cellular membrane is absolutely essential for SFV fusion. Cholesterol-depleted cells enriched with sterols lacking a 3{dollar}\beta{dollar}-hydroxyl group did not support fusion. Lack of cholesterol in a cell does not effect: SFV binding to cellular receptors, endocytosis of SFV by cells, and E1 acid conversion. SFV is {dollar}\sb\sim{dollar}2000 fold less efficient at infecting cholesterol-depleted C6/36 cells, and we used this selective pressure to isolate mutant virus that could grow on cholesterol-depleted cells that were enriched with chlorocholestene. Three mutant isolates, called srf-1, srf-2 and srf-3, were characterized. They fused 50-140 fold better than wt SFV with cholesterol-depleted cells, and 80-150 fold better than wt SFV with chlorocholestene enriched cells. wt SFV exhibits an inhibition of virus release from infected cholesterol-depleted cells, and the srf mutants seem less inhibited by the lack of cholesterol during virus release.