Specific Roles for Lipids in Virus Fusion and Exit Examples from the Alphaviruses - Fusion of Biological Membranes and Related Problems - page 431

these experiments is that the srf-3 mutation of SFV predicts a region of E l
involved in the cholesterol dependence of SIN. This region appears to act
by affecting the conformation of El, rather than by a simple role of specific
amino acid side chains.
5.2.2. Mechanism of the srf-3 Mutation
We hypothesized that the increased ability of the srf-3 spike protein to
fuse with cholesterol-depleted cell membranes is due to an alteration in
some aspect of the E l spike protein's interaction with cholesterol, result-
ing in the relative cholesterol-independence of a normally cholesterol-
requiring step in fusion. To test this hypothesis, the initial kinetics of the El
conformational changes were followed in the presence of either control or
cholesterol-depleted liposomes using either wt or srf-3 (Chatterjee and
Kielian, 1999b). srf-3 is strikingly less cholesterol-dependent than wt SFV
for both acid-specific epitope exposure and El homotrimer formation. At
time points when wt E l conversion is greatly reduced in the absence of cho-
lesterol, srf-3 showed comparable levels of El conversion in the presence
or absence of cholesterol. These E l conformational changes were unaf-
fected by the presence or absence of cholesterol in the virus membrane, as
demonstrated by using srf-3 stocks grown in either BHK cells or choles-
terol-depleted mosquito cells.
In contrast, the srf-3 El conformational changes are still dependent on
the presence of sphingolipid in the target bilayer (Chatterjee and Kielian,
1999b). Acid-specific epitope exposure and homotrimer formation were
evaluated using sterol-free liposomes, and neither wild type nor mutant
virus support efficient El conversion in the absence of sphingolipid. Studies
of virus infection in the presence of NH 4 Cl demonstrate very similar NH 4 Cl
sensitivities for wt and srf-3, indicating that the two viruses have a com-
parable pH-dependence for fusion (Chatterjee and Kielian, 1999b). Thus,
the srf-3 mutant appears unaltered in its sphingolipid dependence and pH
dependence, but is less dependent on cholesterol for the fusogenic confor-
mational changes in El.
The properties of wt and srf-3 fusion were assayed in vitro in lipid
mixing assays with pyrene-labeled wt or srf-3 virus (Chatterjee and Kielian,
1999b). The initial rates of fusion were followed at different temperatures
using complete liposomes (PC: PE : SPM : Chol. 1 : 1 : 1 : 1.5), and showed
similar kinetics and extent of fusion for the two viruses. An Arrhenius
diagram of initial fusion rates versus reciprocal temperature (Bron et al.,
1993) gave comparable plots for the two viruses, suggesting equivalent acti-
vation energies and general fusion characteristics between the two viruses.
In contrast, when fusion with cholesterol-free liposomes (PC: PE : SPM 1 :