Bluetongue virus outer capsid protein VP5 interacts with membrane lipid rafts via a SNARE domain

Abstract

Bluetongue virus (BTV) is a nonenveloped double-stranded RNA virus belonging to the family Reoviridae. The two outer capsid proteins, VP2 and VP5, are responsible for virus entry. However, little is known about the roles of these two proteins, particularly VP5, in virus trafficking and assembly. In this study, we used density gradient fractionation and methyl beta cyclodextrin, a cholesterol-sequestering drug, to demonstrate not only that VP5 copurifies with lipid raft domains in both transfected and infected cells, but also that raft domain integrity is required for BTV assembly. Previously, we showed that BTV nonstructural protein 3 (NS3) interacts with VP2 and also with cellular exocytosis and ESCRT pathway proteins, indicating its involvement in virus egress (A. R. Beaton, J. Rodriguez, Y. K. Reddy, and P. Roy, Proc. Natl. Acad. Sci. USA 99:13154-13159, 2002; C. Wirblich, B. Bhattacharya, and P. Roy J. Virol. 80:460-473, 2006). Here, we show by pull-down and confocal analysis that NS3 also interacts with VP5. Further, a conserved membrane-docking domain similar to the motif in synaptotagmin, a protein belonging to the SNARE (soluble N-ethylmaleimide-sensitive fusion attachment protein receptor) family was identified in the VP5 sequence. By site-directed mutagenesis, followed by flotation and confocal analyses, we demonstrated that raft association of VP5 depends on this domain. Together, these results indicate that VP5 possesses an autonomous signal for its membrane targeting and that the interaction of VP5 with membrane-associated NS3 might play an important role in virus assembly.

FIG. 1.

Comigration of BTV VP5 with raft fractions. (A) Wild-type VP5-transfected (top) and untransfected (bottom) HeLa cells were lysed with cold 1% Triton X-100, and the rafts were purified by density gradient fractionation. The lysates were run on SDS-PAGE and analyzed by Western blotting. Antibodies, molecular masses, and sucrose percentages are indicated on the right, left, and bottom, respectively. The pellet fraction is labeled P. The dotted arrows indicate localizations of different proteins in the raft fractions. (B) IFM colocalization of VP5 with caveolin (red) (a), actin (red) (b), tubulin (red) (c), calnexin (red) (d), and vimentin (red) (e). The expression of VP5 alone is shown in image f.

FIG. 2.

Effect of cholesterol sequestration on cells. Shown is the average percentage of dead cells at each concentration of mβcdx. The cell viability was assessed by staining the cells with trypan blue. The amounts of mβcdx and percentages of dead cells are indicated on the x and y axes, respectively. The experiment was repeated three times, and the average value was plotted. Standard errors are indicated by the error bars.

FIG. 3.

Cholesterol is required for BTV VP5 raft association. HeLa cells expressing VP5 were treated with 10 mM mβcdx (left) or untreated (right) and fractionated in a sucrose gradient. Molecular masses, antibodies, and sucrose percentages are indicated at the left, center, and bottom, respectively. The pellet fraction is labeled P. The dotted arrows indicate localizations of different proteins in the raft fractions.

FIG. 4.

VP5 expression in BTV-infected cells. VP5 in BTV-infected HeLa cells cofractionates with rafts. Cells infected with BTV-17 were analyzed by density gradient centrifugation 2, 4, 6, 8, 12, and 16 h p.i. Uninfected HeLa cells fractionated in parallel were used as controls. Antibodies, molecular masses, and sucrose percentages are indicated on the right, left, and bottom, respectively. The pellet fraction is labeled P. The dotted arrows indicate localizations of different proteins in the raft fractions.

FIG. 5.

Immunofluorescence analysis of VP5 expression in BTV-infected cells. (A) Localization of VP5 (green) in infected HeLa cells with cellular cytoskeleton; vimentin (red), tubulin (red), and actin (red) are indicated. The time p.i. is given in each image. The arrowheads indicate distributions of labeled proteins. (B) Localization of vimentin, tubulin, and actin in control uninfected cells.

FIG. 6.

Cholesterol depletion affects BTV-1 production. (A) HeLa cells infected with BTV-17 (left) and BSR cells infected with BTV-1 (right) were analyzed for relative virus titers. Virus was harvested at 8, 12, and 16 h p.i. in untreated and mβcdx (10 mM)-treated cells. Control untreated samples are labeled 8c, 12c, and 16c. The error bars indicate the standard errors of three replicates of the experiments. The total titer for each time p.i. was normalized to 100% for untreated cells. (B and C) Cells infected with BTV-17 (left) or BTV-1 (right) and analyzed for expression of NS2 (B) or tubulin (C). The presence and absence of mβcdx is shown as + and −, respectively.

FIG. 7.

Effect of cholesterol extraction on the migration of BTV proteins. Fractionation of BTV-infected BSR cells in the absence (left) and presence (right) of 10 mM mβcdx at 8 (top) and 16 (bottom) h p.i. Antibodies, molecular masses, and sucrose percentages are given at the center, left, and bottom, respectively. The pellet fraction is labeled P. The dotted arrows indicate localizations of different proteins in the raft fractions.

FIG. 8.

Interaction of VP5 with NS3. (A) Pull-down analysis of the interaction between VP5 and NS3. Lysates were pulled down with antibody against NS3 (left) or VP5 (right) and labeled with VP5 (left) or NS3 (right) antibodies. The plus and minus signs represent infected and uninfected cells. Lanes 1 and 2 of both blots represent cellular lysates of Sf9 cells that were either coinfected with recombinant baculoviruses (lane 1) or uninfected (lane 2). Lanes 3 to 6 are the cellular precipitates. The presence of beads or antibodies is indicated at the bottom, while the respective molecular masses are indicated on either side. (B) Immunofluorescence of cells expressing VP5 (green) and NS3 (red) in infected (left), uninfected control (center), and cotransfected (right) cells.

FIG. 9.

Mapping of a potential raft association domain in vp5. (A) Schematic showing the alignment of VP5 protein sequences belonging to serotypes 1, 10, and 17 of BTV and 1 and 2 of epizootic hemorrhagic disease virus. Amino acids that are completely conserved or conserved in charge are indicated by asterisks and colons, respectively. Amino acids targeted for mutagenesis are shown in boldface. (B) Distribution of wild-type VP5 (left) and M1 (right) in transiently transfected HeLa cells 48 h posttransfection.

FIG. 10.

Plasma membrane docking motif in VP5. (A) Fractionation of native (left) and mutated (right) VP5 in HeLa cells. Antibodies, molecular masses, and sucrose percentages are given at the center, left, and bottom, respectively. The pellet fraction is labeled P. The dotted arrows indicate localizations of different proteins in the raft fractions. (B) Distribution of native (left) and mutated (right) VP5 in HeLa cells.

FIG. 11.

Trimerization of full-length VP5 carrying the M1 mutation. Shown is a Western blot of nondenatured samples for native (left) and M1 (right) VP5 variants. Note that trimerization is unaffected in the M5 mutant.