|Home | About | Journals | Submit | Contact Us | Français|
Respiratory syncytial virus (RSV) is an important human pathogen that can cause severe and life-threatening respiratory infections in infants, the elderly, and immunocompromised adults. RSV infection of HEp-2 cells induces the activation of RhoA, a small GTPase. We therefore asked whether RhoA signaling is important for RSV replication or syncytium formation. The treatment of HEp-2 cells with Clostridium botulinum C3, an enzyme that ADP-ribosylates and specifically inactivates RhoA, inhibited RSV-induced syncytium formation and cell-to-cell fusion, although similar levels of PFU were released into the medium and viral protein expression levels were equivalent. Treatment with another inhibitor of RhoA signaling, the Rho kinase inhibitor Y-27632, yielded similar results. Scanning electron microscopy of C3-treated infected cells showed reduced numbers of single blunted filaments, in contrast to the large clumps of long filaments in untreated infected cells. These data suggest that RhoA signaling is associated with filamentous virus morphology, cell-to-cell fusion, and syncytium formation but is dispensable for the efficient infection and production of infectious virus in vitro. Next, we developed a semiquantitative method to measure spherical and filamentous virus particles by using sucrose gradient velocity sedimentation. Fluorescence and transmission electron microscopy confirmed the separation of spherical and filamentous forms of infectious virus into two identifiable peaks. The C3 treatment of RSV-infected cells resulted in a shift to relatively more spherical virions than those from untreated cells. These data suggest that viral filamentous protuberances characteristic of RSV infection are associated with RhoA signaling, are important for filamentous virion morphology, and may play a role in initiating cell-to-cell fusion.
Human Respiratory syncytial virus (RSV) belongs to the family Paramyxoviridae and is the leading viral cause of severe lower respiratory tract illness in infants and young children. The fusion (F) glycoprotein is necessary for cell-to-cell fusion and syncytium formation and is thought to be necessary for virion entry into cells, but the exact mechanisms of virus-induced membrane fusion have not been defined. RSV F1 is expressed on the virus envelope and on the surfaces of infected cells as a trimer (9, 53), similar to human immunodeficiency virus type 1 (HIV-1) gp41. Fusion proteins from several diverse enveloped viruses such as paramyxoviruses and lentiviruses have similar structural and functional domains and share similar fusion properties (7, 14, 24). Paramyxoviruses, including RSV, have a broad pH range for fusion and syncytium formation and directly fuse with the plasma membrane (41). Virus-mediated membrane fusion and entry are multistep processes that generally require attachment to the primary virus receptor, and in some cases, coreceptor binding. The fusion peptide is then inserted into the target cell membrane, followed by hemifusion, full fusion, the production of a fusion pore, and the release of the viral genome into the target cell cytoplasm (50). While the importance of virus-to-cell fusion during entry is clear, the teleological advantage to viruses of forming syncytia through cell-to-cell fusion is more uncertain. Viruses may use syncytium formation to spread quickly to neighboring cells or to evade host defense mechanisms. Cell-to-cell fusion mediated by some viral envelope proteins involves the cellular actin cytoskeleton and cell surface integrins (4, 12, 21, 23). Therefore, host cellular proteins that maintain cell membrane integrity, cell mobility, and adhesion might be expected to play a role in virus-induced fusion and syncytium formation since fusion involves direct cell-to-cell contact and the mixing of cell membranes, although there is currently no direct evidence for their involvement. Virus-induced membrane fusion mediated by the virus receptor and the fusion protein may occur similarly to intracellular vesicle fusion. Integral membrane proteins on the vesicle and target membrane known as v-snares and t-snares interact and undergo conformational changes which bring the target membranes close together to facilitate fusion (46, 47). Interestingly, a small GTPase, Rab5, is known to play a role in v-snare- and t-snare-mediated vesicle fusion (15, 45).
Many enveloped viruses cause characteristic changes in the surface morphology of infected cells. The surfaces of infected cells are covered by large clumps of filamentous protrusions, which can be visualized by light microscopy, immunofluorescence staining, and electron microscopy (2, 3, 35, 51). The morphology of budding virions depends on cellular determinants such as polarized cell phenotype and the integrity of the actin microfilament network (6, 39). The determinants of RSV's spherical and filamentous morphological forms and the roles of such particles in virus transmission and pathogenicity are not clearly defined. In RSV-infected cells, the filaments are coated with the viral envelope proteins F and G, suggesting a potential role for these proteins in forming cell-to-cell contacts that might initiate syncytium formation.
We have previously demonstrated that RhoA and its downstream signaling cascades are activated during RSV infection (16). RhoA is a small GTP binding protein in the Ras superfamily. RhoA is ubiquitously expressed in mammalian cells, and activated RhoA influences a variety of essential biological functions in eukaryotic cells, including gene transcription, cell cycle, vesicular transport, adhesion, cell shape, fusion, and motility, through its activation of signaling cascades (18, 30). RhoA affects the cytoskeleton by inducing the organization of actin stress fibers and the formation of focal adhesion plaques (37). Stress fiber formation is known to require RhoA activation (34). RhoA signaling pathways can also induce the production of interleukin-8, which is produced in abundance by RSV-infected cells (22). RhoA activation also leads to the formation of microvilli by the phosphorylation of moesin via Rho kinase (34, 44). Interestingly, viral filaments which are apparent during RSV infection resemble RhoA-induced microvilli. Based on these data, the goal of this study was to define the role of RhoA signaling in RSV infection. We used the following agents to determine if activated RhoA signaling is important for RSV replication and syncytium formation: the C3 exoenzyme from Clostridium botulinum, which specifically inactivates RhoA by ADP ribosylation of Asn 41 (40, 43); Y-27632, a Rho kinase inhibitor (49); and cytochalasin D, which inhibits actin polymerization (10). In this paper, we demonstrate that RhoA signaling is necessary for RSV-induced cell-to-cell fusion and for the formation of microvilli that promote the formation of filamentous virions. However, RhoA signaling is not required for an efficient infection or for the production of infectious but nonfilamentous virions. We also demonstrate an association between RSV-induced syncytium formation and the presence of RhoA-induced viral filaments. The data indicate that the requirements for the production of infectious RSV virions can be dissociated from the process of cell-to-cell fusion and that virus-induced RhoA activation and signaling are necessary for the filamentous virus structure and syncytium formation.
R. Chanock, National Institutes of Health, Bethesda, Md., provided the A2 strain of RSV. RSV stocks were prepared as previously described (17). Recombinant, green fluorescent protein-expressing RSV (rgRSV) was generated as previously described (19). The Long strain of RSV was obtained from the American Type Culture Collection. All cells were maintained in Eagle's minimal essential medium or Dulbecco's modified Eagle medium (DMEM) supplemented with glutamine, gentamicin, penicillin G, and 10% fetal bovine serum.
Two-day-old HEp-2 monolayers at 80% confluence in 12-well plates (Costar, Cambridge, Mass.) were used for plaque assays. The assay was performed as previously described (17).
HEp-2 monolayers at 80% confluence in 96-well plates were pretreated with medium containing 30 μg of C3 (CalBiochem, La Jolla, Calif.)/ml or 20 μM Y-27632 (a gift from Shuh Narumiya, Welfide Corporation, Iruma, Japan) beginning 24 h before RSV infection and continuing throughout the infection. Fifty microliters of RSV at a multiplicity of infection (MOI) of 0.1 was added to the cells and allowed to adsorb for 1 h at room temperature. After the adsorption of RSV, medium containing C3 or Y-27632 was added and the plates were incubated at 37°C. Untreated RSV-infected wells were used as controls. The dose levels and treatment effects of C3 and Y-27632 have been validated in previous publications (16, 27). Medium supernatants were collected daily, and virus growth was measured by a plaque assay for 8 consecutive days postinfection. Virus-induced syncytia and plaques were visualized at 4 days post-RSV infection.
HEp-2 cells were either left untreated or pretreated with 30 μg of C3/ml 24 h prior to RSV infection. At 72 h postinfection, the cells were harvested and 1-ml aliquots were centrifuged at 14,000 rpm for 10 min at 4°C in a Sorvall Surespin 630 centrifuge. The pellet was resuspended in mammalian protein extraction reagent (Pierce, Rockford, Ill.). The amount of protein was quantified by use of a BCA assay (Pierce). Equal amounts of proteins from C3-treated or untreated RSV-infected cells were resolved by sodium dodecyl sulfate-10% polyacrylamide gel electrophoresis and transferred to polyvinylidene difluoride membranes. The F protein was detected with an anti-F monoclonal antibody followed by a horseradish peroxidase (HRP)-conjugated anti-mouse antibody (Amersham Pharmacia Biotech, Bucks, United Kingdom). Proteins were visualized by ECL (Amersham Pharmacia Biotech).
Untreated HEp-2 cells or cells treated with 30 μg of C3/ml for 24 h were infected with RSV (MOI = 0.5) and fixed in 4% formaldehyde 48 h after infection. The fixed cells were stained with a 1:1,000 dilution of an anti-F monoclonal antibody (Chemicon, Temecula, Calif.) in 5% nonfat dry milk for 1 h. After three washes with phosphate-buffered saline (PBS)-Tween 20, the cells were stained with a 1:5,000 dilution of Alexa fluor 488 goat anti-mouse immunoglobulin G (IgG; Molecular Probes, Eugene, Oreg.) in 5% nonfat dry milk for 1 h followed by washing with PBS-Tween 20. The cells were analyzed with a FACSCaliber (Becton Dickinson, San Jose, Calif.) argon ion laser at 15 mW and 488 nm. Data were analyzed with FlowJo, version 6.0 (Tree Star, San Carlos, Calif.).
HEp-2 cells were grown on coverslips in six-well plates and infected with 200 μl of 103-PFU/ml rgRSV. These cells were analyzed by immunofluorescence microscopy at 24 h post-RSV infection. For the visualization of RSV-infected cells, cells were fixed on coverslips at room temperature in 3.7% formaldehyde for 10 min. The cells were observed under a Zeiss Axioplan fluorescence microscope, and photographs were taken with SPOT image capture software and a Zeiss MC80 microscope camera.
A cell-to-cell fusion assay was used to assess the requirement for RhoA signaling in RSV-induced cell-to-cell fusion. One population of HEp-2 cells (effector population) was infected with recombinant vaccinia virus vTF7-3, which encodes T7 polymerase, at a multiplicity of infection of 10 PFU per cell and then transfected with plasmids encoding RSV glycoproteins F, G, and SH under the control of the T7 promoter by the use of FuGene (Boehringer Mannheim, Indianapolis, Ind.). While F is the major determinant of cell-to-cell fusion, syncytium formation is optimized when all three surface proteins are expressed (20, 36). In addition, we have shown that coinfecting cells with recombinant vaccinia viruses expressing individual RSV proteins that combine the expression of F, G, and SH maximizes filament formation (unpublished observations). A second population of HEp-2 cells (target population) was infected with a recombinant vaccinia virus expressing β-galactosidase under the control of the T7 promoter (provided by E. A. Berger, National Institutes of Health, Bethesda, Md.). Four hours after transfection, the cells were trypsinized and suspended in DMEM containing 2.5% fetal bovine serum to a density of 2 × 107 cells per ml. The cells were divided evenly into aliquots, RhoA inhibitors were added to either effector or target cells, and the cells were incubated for 16 h at 37°C. The two sets of cells were then washed and suspended in DMEM at a concentration of 106 cells per ml. The two cell populations were then mixed in triplicate by adding 100 μl of each cell population to 96-well tissue culture plates, which were then incubated at 37°C for 4 h. The inhibitors were not present during the 4-h incubation for fusion. After 4 h, the cells were fixed in 2% glutaraldehyde-20% formaldehyde (Sigma, St. Louis, Mo.) in PBS for 10 min. One hundred fifty microliters of X-Gal solution (5 mM potassium ferricyanide, 5 mM potassium ferrocyanide, 2 mM magnesium chloride, 1 mg of X-Gal [5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside; Fisher, Springfield, N.J.]/ml, freshly diluted from a 40-mg/ml stock solution in dimethyl formamide) was then added. After 8 h, blue-stained fused cells were viewed with an inverted phase-contrast microscope.
HEp-2 cells on 12-mm-wide coverslips (Fisher, Pittsburgh, Pa.) were treated with 20 μM Y-27632 or 30 μg of C3/ml for 16 h and then infected with RSV (MOI = 1). Medium containing C3 or Y-27632 was added after adsorption to maintain the treatment throughout the course of infection. Untreated RSV-infected cells and uninfected cells treated with 20 μM lysophosphatidic acid (LPA) for 30 min were used as controls. LPA is a known activator of RhoA signaling. Infected samples were fixed 24 h after infection in 4% glutaraldehyde (Sigma) for 1 h, treated with 1% osmium for 15 min, and dehydrated through a series of 70 to 100% ethanol washes. Dehydrated cells were critical point dried, sputter coated with gold, and visualized by use of a Hitachi S4200 scanning electron microscope.
Vero cells or HEp-2 cells (1.5 × 105 cells) were plated in 24-well plates 1 day prior to infection with the Long strain of RSV at an MOI of 0.015 in medium containing 2% fetal calf serum for 1 h. The cells were then washed and incubated in fresh medium for 4 h.
For immuno-TEM, the medium was then replaced with fresh medium containing 10 μg of SB-209763, a humanized monoclonal antibody specific for the F protein of RSV that prevents syncytium formation, per milliliter for 24 to 48 h. The medium was aspirated and the cultures were fixed with 2.5% glutaraldehyde in 0.1 M phosphate buffer overnight. The cultures were washed and incubated for 45 min with HRP-conjugated donkey anti-human IgG (Jackson Immunoresearch, West Chester, Pa.). After the wash step, the peroxidase reaction product was developed with diaminobenzidene, and the cultures were postfixed with 1% osmium for 1 h and dehydrated in a graded series of ethanol.
For all samples, the cells were lifted off the wells with propylene oxide and embedded in EMBed 812 resin (Electron Microscopy Sciences, Ft. Washington, Pa.). Thin sections were cut and stained with uranyl acetate and lead citrate. For some sections, tannic acid staining was performed prior to the uranyl acetate stain to accentuate the glycoprotein spikes. The specimens were examined by use of a JEOL 100EX transmission electron microscope at 80 kV.
For evaluations of particles separated by velocity sedimentation, the bands associated with the peak fractions were concentrated by centrifugation onto a 60% sucrose cushion. A sample from the concentrated band was applied to a grid, stained with 0.5% uranyl acetate, and examined by use of a Hitachi H-7000 transmission electron microscope.
HEp-2 cells in 25-cm2 flasks were either left untreated or treated with 30 μg of C3/ml for 24 h and then infected with RSV (MOI of 0.5). After a 72-h infection, the infected cells were harvested and layered over a continuous 15 to 60% sucrose gradient. The virus was then separated by velocity centrifugation at 14,000 rpm for 10 min at 4°C in a Sorvall Sure Spin 630 centrifuge. Two-milliliter fractions were collected, and plaque assays were performed with the fractions as described previously (17).
Viruses were concentrated from peak individual fractions on a 60% sucrose cushion by ultracentrifugation at 14,000 rpm for 10 min at 4°C in a Sorvall Sure Spin 630 centrifuge. Virus particles were collected from the layer just above the sucrose cushion. Specimens for fluorescence microscopy were prepared on coverslips and stained. Briefly, 50 μl of the virus sample was smeared on a coverslip and allowed to air dry. The virus particles were fixed with 3.7% formaldehyde in PBS for 30 min. The samples were washed twice with PBS-Tween 20 and blocked with 5% nonfat dry milk in PBS for 30 min. They were then stained with anti-RSV antibody (Maine Biotechnology Services) in 1% nonfat dry milk for 1 h, followed by Alexafluor 488 anti-mouse IgG (Molecular Probes). After three washes with PBS-Tween 20, the coverslips were mounted on microscope slides. Specimens were viewed with a Zeiss AxioPlan 2 fluorescence microscope, and pictures were taken with a Hamamatsu ORCA-ER digital camera.
The C3 protein from C. botulinum is known to specifically ADP-ribosylate RhoA and irreversibly inactivates its ability to initiate signaling pathways. We used C3 to determine whether activated RhoA is required for RSV-induced syncytium formation. HEp-2 cells were treated with C3 beginning 24 h prior to virus infection. Untreated RSV-infected cells exhibited syncytium and plaque formation (Fig. (Fig.1A).1A). However, RSV-infected cells that had been pretreated and incubated with C3 did not have virus-induced syncytium or plaque formation (Fig. (Fig.1B)1B) and appeared similar to uninfected monolayers (Fig. (Fig.1C1C).
HEp-2 cells were treated with several doses of C3, up to 50 μg/ml, beginning 24 h prior to RSV infection (data not shown). RSV-induced syncytium formation was completely inhibited by 30 μg of C3/ml, while lower doses resulted in partial inhibition. Next, we asked whether RhoA activation is required for RSV infection and replication. HEp-2 cells in 96-well plates were either left untreated or pretreated with C3 beginning 24 h prior to RSV infection and throughout the course of infection. The contents of individual RSV-infected wells were transferred to HEp-2 cell monolayers in 12-well plates for plaque assays on eight consecutive days after RSV infection. Surprisingly, the number of PFU produced by C3-treated cells was equal to that produced by untreated RSV-infected cells, indicating that the production of infectious virions was not affected by the inactivation of RhoA (Fig. (Fig.1D).1D). In addition, RSV was produced from the C3-treated cells without producing syncytia for more than 8 days. To determine the potential effect of C3 treatment on the level of F protein expression, we analyzed the F protein by Western blotting and flow cytometry of RSV-infected cells, using a monoclonal anti-F antibody. Analyses of the RSV F protein in untreated or C3-treated RSV showed similar levels of protein production by Western blotting and similar levels of F expression on the cell surface by flow cytometry, indicating that an altered magnitude of RSV F expression is not the explanation for C3's inhibition of syncytium formation (Fig. (Fig.2).2). These data suggest that the presence of constitutive levels of inactive RhoA associated with the membrane is sufficient for virus entry and replication but that RhoA activation is required for virus-induced cell-to-cell fusion to form syncytia.
The Rho kinase inhibitor Y-27632 was used to further characterize the requirements for RhoA signaling in RSV infection and replication. First, we asked whether Y-27632 could interfere with single-cell infection by RSV. HEp-2 cells were treated with 20 μM Y-27632 beginning 24 h prior to infection with rgRSV (MOI = 0.1). The cells were fixed in 3.7% formaldehyde at 24 h postinfection. RSV-infected, green fluorescent protein-expressing cells were then visualized by fluorescence microscopy. Y-27632 had no effect on the number of RSV-infected cells (Fig. (Fig.3B)3B) compared to untreated controls (Fig. (Fig.3A3A).
Next, Y-27632 was used to create a virus growth curve to determine its effect on RSV replication. HEp-2 cells in 96-well plates were left untreated or treated with 20 μM Y-27632 beginning 24 h prior to RSV infection (MOI = 0.1). The contents of individual wells were transferred to HEp-2 cell monolayers in 12-well plates for plaque assays on eight consecutive days after RSV infection. The number of PFU of RSV produced by Y-27632-treated cells was equal to that produced by untreated RSV-infected cells, indicating that the production of infectious virions was not affected by the inactivation of RhoA signaling (Fig. (Fig.3C).3C). These data suggest that downstream signaling through Rho kinase is not required for virus entry or replication.
The role of RhoA in RSV-induced cell-to-cell fusion was then evaluated by use of a cell-to-cell fusion assay. As shown in Fig. Fig.4,4, one population of HEp-2 cells was infected with a recombinant vaccinia virus expressing T7 polymerase and transfected with plasmids expressing RSV F, G, and SH under the control of a T7 promoter (effectors). Another population of HEp-2 cells was infected with a recombinant vaccinia virus expressing lacZ under the control of a T7 promoter (targets). At 4 h postinfection, the effectors (hatched bars) or the targets (black bars) were treated with 20 μM Y-27632, 30 μg of C3/ml, or 10 μM cytochalasin D, an inhibitor of actin polymerization. The cells were incubated separately for 16 h and then mixed together. After 4 h, the cells were fixed and stained with X-Gal. Fused blue cells were then counted. Inhibiting RhoA signaling in effector cells (hatched bars) with Y-27632 or C3 dramatically reduced cell-to-cell fusion. A cytochalasin D treatment of effector cells also inhibited cell-to-cell fusion. In contrast, the treatment of target cells with Y-27632 or cytochalasin D did not have an effect on cell-to-cell fusion, although a treatment with C3 partially inhibited cell-to-cell fusion. These data indicate that RhoA signaling and actin polymerization are important for the effector cells in cell-to-cell fusion and that RhoA signaling through Rho kinase and actin polymerization are not required in the target cells for cell-to-cell fusion to occur. The inhibition of another RhoA signaling pathway or altered RhoA membrane localization may explain the partial effect on cell-to-cell fusion in C3-treated target cells.
To verify that the inhibition of cell-to-cell fusion was not due to an inhibition of vaccinia virus replication or the expression of T7 polymerase and lacZ, we infected HEp-2 cells with both the vaccinia virus expressing T7 polymerase and the vaccinia virus expressing lacZ, and after 4 h, added various inhibitors. Sixteen hours later, the cells were fixed and stained with X-Gal. All of the cells turned blue within 30 min, regardless of the treatment received, indicating that the inhibitors did not significantly affect vaccinia virus infection or the expression of T7 polymerase or lacZ (data not shown). Also, as indicated by trypan blue exclusion, none of the inhibitors caused toxicity to the cells during the 16 h of treatment compared to untreated cells (data not shown).
Taken together, as shown in Fig. Fig.4,4, there were distinct requirements for cell-to-cell fusion for the effector cells, which were transfected with plasmids encoding RSV envelope proteins, and for the target cells, which lacked RSV envelope proteins. RhoA signaling and actin polymerization were required for the infected cells to fuse with target, uninfected cells. On the other hand, RhoA signaling through Rho kinase and actin polymerization were not required in the uninfected target cells.
We next asked whether RhoA signaling and actin polymerization affected RSV-induced microvillus formation. We previously reported that the treatment of HEp-2 cells with 30 μg of C3/ml or 20 μM Y-27632, a Rho kinase inhibitor, beginning 24 h prior to infection and throughout RSV infection alters the pattern of F protein localization in infected cells (16). Untreated RSV-infected cells have punctate staining for F in the cytoplasm, and late in infection, have filamentous structures extending from the cell, which stain with an anti-F antibody (16). In C3- and Y-27632-treated cells, there were no F-staining filaments, and the cytoplasmic staining for F was more diffuse (16).
Therefore, we next determined the effect of RhoA activation and signaling on viral filament formation by SEM (Fig. (Fig.5).5). Untreated cells or cells treated with 30 μg of C3/ml for 24 h were infected with RSV and fixed in 4% glutaraldehyde 24 or 48 h after infection. When visualized by SEM, RSV-infected cells had large clumps of long filaments that protruded from several places across the cell surface (Fig. 5A and B) compared to uninfected HEp-2 cells (Fig. (Fig.5D).5D). These filaments also bridged across cell junctions to neighboring cells. By 48 h postinfection, the filaments covered the entire surface of the infected cell (Fig. (Fig.5B).5B). Interestingly, the viral filaments produced by C3-treated, RSV-infected cells were attenuated (as shown at 24 h in Fig. Fig.5C).5C). These cells produced blunted filaments that were not in clumps but were sparsely distributed and more disorganized across the surfaces of the cells. The filaments induced by RSV in the absence of C3 closely resembled the microvilli in uninfected cells induced by LPA, a known inducer of RhoA signaling (Fig. (Fig.5E)5E) (34, 44). These data suggest that RhoA-induced microvilli may play an important role in RSV filament formation and virus-induced cell-to-cell fusion, possibly by initiating cell-to-cell contacts.
RSV-induced filament formation was next evaluated by immunoperoxidase staining and TEM to confirm the presence of the F glycoprotein on the membranes of the filamentous particles. In Vero cells infected with the Long strain of RSV and processed for TEM, both viral particles and viral filaments were produced by infected cells (Fig. (Fig.6A).6A). At a higher magnification, viral glycoprotein spikes were evident on the surfaces of both viral particles and filaments (Fig. 6B and C), and nucleocapsid structures were evident within these structures. Immunoelectron microscopic localization of the anti-F antibody revealed intense staining of both viral particles and viral filaments (Fig. 6D, E, and F) overlying the locations of the glycoprotein spikes. These observations, along with those for Fig. Fig.5,5, suggest that the microvilli induced by RSV infection and RhoA activation are budding filamentous virus particles.
A scanning electron microscopic examination of C3-treated infected cells showed reduced quantities of single blunted filaments compared to the large clumps of long filaments in untreated infected cells (Fig. (Fig.5).5). Next, we developed a sucrose gradient velocity sedimentation technique to separate spherical and filamentous viruses and to attempt a semiquantitative measurement of the influence of C3 treatment on virus morphology. HEp-2 cells were either left untreated or treated with 30 μg of C3/ml for 24 h and then infected with RSV. After a 72-h infection, the infected cells and media were harvested and layered over a continuous 15 to 60% sucrose gradient. Viruses were then separated by velocity centrifugation in a Sorvall centrifuge. Two-milliliter fractions were collected and plaque assays were performed on fractions as described previously (17). RSV titers from sucrose gradient fractions showed a predominance of spherical forms for C3-treated RSV compared to equal amounts of filamentous and spherical virions for untreated RSV (Fig. (Fig.7A).7A). The peak fractions (2 and 6) were examined by fluorescence microscopy to confirm the separation of two morphological forms of the virus into two identifiable peaks. The virus particles were fixed on coverslips and stained with a rhodamine-conjugated anti-RSV antibody. The specimens were viewed with a Zeiss AxioPlan 2 fluorescence microscope. The peak fractions 2 and 6 showed spherical and filamentous morphologies, respectively (Fig. 7B and C). To confirm that the RSV F-expressing particles seen by fluorescence microscopy were virus particles and not membrane fragments, we performed TEM on the peak fractions. This confirmed that peak 1 consisted primarily of pleomorphic, somewhat spherical virus particles and peak 2 consisted of primarily filamentous particles.
We have previously shown that RhoA is activated during RSV infection (16). We now report data that suggest that RhoA-mediated signaling and actin polymerization are associated with the filamentous virion morphology and the syncytium-inducing phenotype. RhoA is an essential host cell protein with GTPase activity and is known to influence a variety of signaling pathways and basic cell functions (18, 30). The role of RhoA in virus-induced fusion and subsequent signaling events may have significance in virus infection to ensure the coordinated control of cellular activities required for virus replication, such as the stage of the cell cycle and reorganization of the actin cytoskeleton. There are many examples of viruses using the host cell machinery to complete their life cycles, so it is not surprising that RSV has adapted to utilize an essential GTPase to modify its morphological properties.
In C3-treated RSV-infected cells, there was no formation of syncytia (Fig. (Fig.1B).1B). Interestingly, blocking RhoA signaling with C3 or Y-27632 did not affect the efficiency with which RSV initiated infection or the production of infectious virions in cell culture (Fig. (Fig.1D1D and and3C).3C). Therefore, we have shown a distinction in the requirement for RhoA signaling events between the entry of cell-free virions (independent of RhoA signaling) and cell-to-cell fusion and syncytium formation (dependent on RhoA signaling). This distinction is not related to changes in the overall expression of RSV F, since Western blotting and flow cytometric analysis of C3-treated and untreated RSV-infected cells showed equivalent levels of F expression (Fig. (Fig.22).
The fusion assays whose results are shown in Fig. Fig.44 showed that there were distinct requirements for cell-to-cell fusion for the effector cells, which expressed the RSV envelope glycoproteins, and for the target cells, which did not express RSV envelope glycoproteins. This suggests that RhoA signaling and actin polymerization are required for an infected cell to fuse with an uninfected target cell. On the other hand, the data suggest that RhoA signaling through Rho kinase and actin polymerization are not required in uninfected cells for fusion to occur.
In Fig. Fig.4,4, we show that C3 partially inhibited cell-to-cell fusion in treated targets. A recent study showed that a C3 treatment does not remove RhoA from the membrane but may cause it to be redistributed into detergent-soluble regions of the membrane (29), thereby reducing its density within cholesterol-rich membrane microdomains (lipid rafts). Specific proteins, including RhoA and CD44, localize to these membrane microdomains (29, 31). It has been shown by our group and others that RSV proteins colocalize with cellular proteins associated with lipid microdomains, including caveolin-1 (5) and CD44, as well as with RhoA (27). The ADP-ribosylation of RhoA by C3 inactivates RhoA signaling, results in an overall decrease in both the number and length of viral filaments, and shifts the localization of F to nonlipid microdomain regions of the membrane (27). This suggests that the selective incorporation of RSV proteins into lipid microdomains may result in virus assembly in microvilli providing the filamentous viral morphology associated with syncytium formation. Viruses such as HIV-1, influenza virus, and simian virus 40 have also been shown to enter cells or to bud from lipid rafts (1, 8, 32, 52). The redistribution of RhoA into non-lipid-raft membrane domains of uninfected target cells by C3 does not influence replication or infection with individual virions, but it may diminish the ability of RSV-induced filaments to mediate cell-to-cell fusion. It is possible that there is a greater requirement for membrane order and the coalescence of factors within lipid rafts to support cell-to-cell fusion than for virus-to-cell membrane fusion. This may be related either to a required mixture of proteins and heparin binding domains which are needed to mediate cell-to-cell fusion or to the need for a threshold concentration.
RSV-infected cells form long filamentous protrusions expressing the F glycoprotein within 24 h after infection (16). Several viruses, including influenza virus and Ebola viruses, can also form long filamentous particles during infection, but it is unclear what role these filaments play in the spread of virus infection and in syncytium formation (11, 13). We were able to visualize these filamentous structures in RSV-infected cells by TEM (Fig. (Fig.6),6), SEM (Fig. (Fig.5),5), and F-specific indirect immunofluorescence (16) (Fig. (Fig.7).7). The treatment of cells with C3 or Y-27632 beginning 24 h before virus infection caused an altered pattern of F staining in infected cells by immunofluorescence (16). C3- and Y-27632-treated cells had diffuse staining for F throughout the cell, and there were no distinct filamentous structures present on or around the infected cells (16). By using SEM, we observed that viral filaments can bridge from the infected cell to a neighboring cell (Fig. (Fig.5A).5A). The viral filaments were blunted and more diffusely expressed on the surfaces of HEp-2 cells treated with C3 (Fig. (Fig.5C),5C), suggesting that filaments are needed to bridge cell junctions in order to initiate the cell-to-cell fusion process.
Interestingly, the RhoA-induced phosphorylation of moesin via Rho kinase can cause the production of microvilli which colocalize with CD44, a lipid raft protein, and resemble the RSV filaments shown by immunofluorescence (16) and SEM (Fig. (Fig.5)5) (34, 44). These microvilli are important for cell-to-cell adhesion in epithelial cells (48). The inhibition of RhoA signaling by C3 inhibits microvillus formation (Fig. (Fig.5C)5C) (44), suggesting that RhoA-induced microvilli may play an important role in the normal assembly of virus particles and the formation of filamentous virus structures. These data indicate a strong association between the ability of RSV to form filaments and its ability to undergo cell-to-cell fusion, further suggesting that viral filaments may be important for cell-to-cell fusion and syncytium formation. Immuno-TEM of microvilli and filamentous virions showed that the microvilli are coated with the RSV F glycoprotein (Fig. (Fig.6).6). These data suggest that the microvilli are sites of viral assembly and represent budding viral particles.
In order to confirm that C3 affects filamentous virus production, we developed a sucrose gradient velocity sedimentation technique to separate spherical and filamentous viruses. There were two peaks of RSV PFU in both C3-treated and untreated RSV-infected groups (Fig. (Fig.7A).7A). The relative separation of spherical and filamentous morphological forms into two identifiable peaks was further supported by immunofluorescence staining of virus particles with an anti-F monoclonal antibody and TEM (Fig. 7B and C). The results showed that the C3 treatment shifted the relative production of viral morphologies to more spherical particles than those obtained with untreated RSV (Fig. (Fig.7A).7A). A previous report has shown that cytochalasin D treatment enhances spherical influenza virion release and reduces the formation of filamentous influenza virus particles (39). These data are consistent with our findings and suggest that the assembly of viral filaments requires an intact actin microfilament network.
It has been reported that filamentous structures similar to those described for wild-type RSV formed at the cell surface, even when all three envelope glycoproteins were replaced by a single foreign viral glycoprotein (vesicular stomatitis virus G protein) carrying the RSV F cytoplasmic domain (33). The engineered recombinant virus induces filaments at the cell surface and causes cell-cell membrane fusion at pH 5.0 but not at pH 7.0. The requirement of viral genes for particle morphology differs among several viruses. For efficient particle assembly, vesicular stomatitis virus and rabies virus require the M protein (26, 28), influenza virus requires the M1 and M2 proteins (38), and simian virus 5 (another paramyxovirus) requires the coexpression of the N protein, the M protein, and one of the homologous transmembrane glycoproteins (42). For RSV, the interactions of the F protein cytoplasmic domain and the M protein with cellular proteins, RhoA, and the actin cytoskeleton may play a role in filamentous virus formation.
Our findings indicate that RhoA signaling is associated with RSV-induced filament formation and that the production of infectious virions in vitro does not require a syncytium-inducing phenotype. In addition, we report an association between RhoA-induced viral filaments and RSV-induced syncytium formation. A temperature-sensitive (ts) strain of RSV with a non-syncytium-inducing phenotype in cell culture has been reported (25). This virus produces equivalent numbers of progeny viruses as normal syncytium-inducing strains of RSV, but the virus is severely attenuated in vivo (25). In addition, the non-syncytium-inducing phenotype correlates with the lack of viral filaments seen by SEM (25). Thus, RSV infection in the presence of C3 and Y-27632 treatment may be analogous to ts RSV strains, which replicate with equal efficiencies in vitro as their counterpart syncytium-inducing strains, but without significant syncytium formation (25). However, the ts strains are attenuated in vivo, suggesting that viral filament formation and cell-to-cell fusion may represent virulence determinants. Learning how to produce non-syncytium-inducing viruses would be valuable for the development of live attenuated vaccines for two reasons. First, the virus may be less virulent, and second, if the virion morphology were more homogeneous, then the virus might be easier to purify and concentrate. More work is needed to define the precise steps at which RhoA activation is required for syncytium formation and to determine whether RhoA-induced signaling events are involved in RSV pathogenesis.
We thank James Wittig and Gary Olsen for their assistance with SEM (Vanderbilt University, Nashville, Tenn.), Beverly Maleeff and Sandra Griego for technical support with immuno-TEM (SmithKline Beecham, King of Prussia, Pa.), and Shuh Narumiya for the Y-27632 reagent (Welfide Corporation, Iruma, Japan).
This work was supported in part by grant RO1-AI-33933.