INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

Combinations of interconnected microtubule filaments, intermediate filaments, and actin microfilaments comprise the cytoskeleton of a living cell. The impacts of the different types of filaments on viral morphogenesis have been the focus of many studies. The microtubule filament network plays an important role in the trafficking of viral proteins from one cell compartment to another and in orchestrating the vectorial transport of these proteins in polarized cells (15, 28). Treatment of wild-type Sendai virus-infected cells with a microtubule-depolymerizing drug, for example, interferes with apical transport of the viral glycoproteins and consequently downregulates the polarized release of virus particles (15, 30, 31). In plants, the microtubule network has been shown to be essential for the cell-to-cell spread of tobacco mosaic tobamovirus (TMV) (12). The interaction of TMV with microtubules appears to be critical for the spread of this virus from the initial site of infection to adjacent cells and determines its host range. Multifunctional involvement of the actin microfilament network during viral infection has been documented as well. Vaccinia virus, for instance, utilizes actin microfilaments for its cell-to-cell spread (5). The intracellular enveloped form of this virus induces the nucleation of actin filament tails from the outer membrane surrounding the virus particles. The vaccinia virus particles extend outwards on actin projections to contact and infect adjacent cells. It has been recently reported that actin microfilaments contribute to the release of human immunodeficiency virus type 1 from the host cell and play a role in cell-to-cell transmission (20, 21). Actin and actin-associated proteins have also been found in released virus particles of rabies virus and measles virus, which may suggest the involvement of actin microfilaments in the release of these viruses as well (3, 18).

Although the involvement of the cytoskeleton during viral infection has been described for many members of different virus families, no information is available regarding members of the Bunyaviridae family. The Hantavirus genus consists of the Old and New World hantaviruses, whose genetic and morphologic organizations share significant similarity with those of the other members of the Bunyaviridae family (9, 17, 29). Hantaviruses consist of a lipid envelope with two incorporated glycoproteins, G1 and G2, that are proteolytic products of a glycoprotein precursor (GPC) (9, 23, 29). The core of the virus particle consists of nucleocapsid (N) structures which contain the viral genetic material encapsidated by N and RNA-dependent RNA polymerase L proteins. The viral genetic material consists of three segments of single-stranded, negative-sense RNA molecules. These segments separately encode the N, GPC, and L proteins (9, 23, 29). All the hantavirus genome-encoded proteins are structural proteins of the virions.

In this paper, we describe studies designed to ascertain the role of the cytoskeleton in the morphogenesis of Black Creek Canal virus (BCCV), a New World hantavirus (24, 26). The data obtained support the conclusion that the actin microfilament network is involved in the process of BCCV assembly and release.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Cell culture and virus strains. Vero and BHK21 cells were grown in Dulbecco's modified Eagle's medium (DMEM) containing 10% fetal bovine serum (FBS) and supplemented with penicillin-streptomycin antibiotic mix. Once confluence was reached, the BHK21 cells were maintained in DMEM with 5% FBS and the Vero cells were maintained in DMEM with 2% FBS. The stock of BCCV (3 × 10⁷ PFU) was grown and titered on Vero E6 cells. Vesicular stomatitis virus (VSV) (5 × 10⁹ PFU) was grown in BHK21 cells and plaque purified two times, and titers were determined by plaque assay on BHK21 cells. Punta Toro (PT) virus was grown on Vero E6 cells, and titers were determined by plaque assay.

Reagents. Monoclonal antibody (MAb) GB04-BF07 recognizing the BCCV N protein was provided by Michael D. Bowen (Centers for Disease Control and Prevention, Atlanta, Ga.). Anti-BCCV rabbit immune serum was a generous gift from Thomas Ksiazek (Centers for Disease Control and Prevention). Nocadozole, cytochalasin D, rhodamine-conjugated phalloidin, and anti--actin and anti--tubulin MAbs were purchased from Sigma (St. Louis, Mo.). [³⁵S]methionine-cysteine labeling mix was obtained from Amersham Corp. (Arlington Heights, Ill.). Fluorescein isothiocyanate-conjugated anti-mouse and anti-rabbit antibodies were purchased from Southern Biotechnology Associates (Birmingham, Ala.). Sindbis virus expression vectors SIN-rep5 and helper 5'-tRNA were provided as gifts by C. Rice. Hyperimmune mouse ascitic fluid to PT virus was a gift from J. F. Smith and D. Pifat (U.S. Army Medical Research Institute for Infectious Diseases, Fort Detrick, Md.).

Virus infectious-center assay. Since plaque assay of hantavirus infections is difficult, we developed an infectious-center assay to measure the virus titers. Tenfold dilutions of BCCV containing 100 µl of medium were placed on monolayers of Vero E6 cells grown in six-well plates, and virus was adsorbed for 2 h at 37°C. The cells were rinsed once in phosphate-buffered saline (PBS) and supplemented with 3 µl of liquefied 1% methylcellulose in DMEM with 2% FBS. After 6 days of incubation at 37°C, the methylcellulose-containing medium in the wells was liquefied by placing the plate on ice for 20 min and removed. The cells were rinsed three times in cold PBS to remove any residual methylcellulose and fixed in chilled 95% ethanol with 5% acidic acid for 20 min at 20°C. Nonspecific binding was blocked by preincubating the fixed cells in PBS containing 3% bovine serum albumin (BSA). The cells were then incubated with a 1:500 dilution of anti-N MAb. The bound antibody was detected with secondary antibody conjugated with horseradish peroxidase by using an immunostaining kit from Vector Laboratories (Burlingame, Calif.). The infectious centers were observed under a light microscope and counted.

Construction of recombinant plasmids. In our previous studies, we had cloned the entire sequences of BCCV S and M segments into a TA cloning vector, pCR 2.1, and used those plasmids (pCR-N, pCR-GPC) for nucleotide sequence analysis of the BCCV N and GPC genes (24). In the studies described here, the coding sequences of BCCV N and GPC genes were amplified by PCR from pCR-N and pCR-GPC plasmids with specific primers linked with XbaI adapters. The PCR fragments were digested with XbaI overnight at 37°C and purified through an agarose gel with glass beads (24). SIN-rep5 expression vector was linearized with XbaI and dephosphorylated by treatment with alkaline phosphatase according to the manufacturer's instructions. The ligation reaction was carried out at 12°C overnight, followed by transformation of Escherichia coli. The orientations of the cloned DNA inserts were determined by restriction digestion analysis, based on restriction maps of the vector and the DNA insert sequences (4, 24). Transcription reactions, electroporation, and handling of both the recombinant and the helper Sindbis virus RNA synthesized in vitro from linearized plasmid constructs were performed as described by Brebenbeek et al. (4). Expression of the recombinant proteins was carried out in BHK21 cells and examined by both indirect immunofluorescence analysis (IFA) and radioimmunoprecipitation analysis.

IFA. Cells grown on coverslips were rinsed three times in PBS and fixed in 3.7% paraformaldehyde for 15 min at room temperature. The fixed cells were then permeabilized in 0.1% Triton X-100 for 5 min. To eliminate nonspecific binding, the cells were preincubated with PBS with 3% BSA for 30 min at room temperature in a humidified chamber. Antibody dilutions of 1:100 for anti-BCCV sera, 1:500 for anti-N protein MAb, and 1:50 for phalloidin were prepared in PBS containing 3% BSA and added to the permeabilized cells for 30 min, with incubation at room temperature. After washing, anti-BCCV protein bound antibody was detected by incubation with anti-rabbit fluorescein isothiocyanate-conjugated secondary antibody for 30 min at room temperature. Actin microfilaments were detected with rhodamine-conjugated phalloidin added to the secondary antibody. Coverslips were washed three times with PBS, mounted with Vectashield (Vector Laboratories, Burlingame, Calif.), and examined with a Nikon Optiphot microscope.

Coprecipitation of BCCV N protein with actin. Three days postinfection, BCCV-infected Vero cells grown in 35-mm-diameter dishes were rinsed in cold PBS three times and lysed in 500 µl of lysis buffer (10 mM Tris HCl [pH 7.5], 1 mM EGTA, 100 mM NaCl, 0.5% Triton X-100) for 15 min on ice. In parallel, the same steps were carried out with BHK21 cells transfected with recombinant Sindbis virus RNAs expressing BCCV N protein two days posttransfection. The lysed proteins were separated from cell debris by centrifugation and then supplemented with either 1 µl of anti--actin MAb or 1 µl of anti--tubulin MAb and placed on ice for 1 h. Subsequently, 20 µl of Sepharose beads conjugated with protein A was added, and the tubes were left rotating overnight at 4°C. The Sepharose beads were collected by centrifugation and washed three times in the lysis buffer. The bound proteins were dissociated from the antibody by boiling in the Laemmli sample buffer for 5 min and separated on 10% polyacrylamide gels (14). Then the separated proteins were transferred to a nitrocellulose filter (Bio-Rad, Hercules, Calif.). The transferred BCCV N protein was detected by Western blotting with an ECL kit (Amersham, Little Chalfont, Buckinghamshire, England) and anti-N MAb.

Radioimmunoprecipitation analysis. Virus-infected cells were washed once in PBS supplemented with Eagle's medium without methionine and cysteine and incubated for 30 min at 37°C. Cells were then labeled with [³⁵S]methionine-cysteine labeling mix (100 µCi) for 3 h at 37°C and chased for 15 min. Cells were washed with ice-cold PBS and lysed with 600 µl of lysis buffer (10 mM Tris-HCl [pH 7.5], 0.15 M NaCl, 1% Triton X-100, 20 mM EDTA [pH 8.0]). Nuclei and cell debris were removed by centrifugation at 12,000 × g for 10 min. Cell lysate then was combined with 1 µl of anti-BCC virus rabbit sera by rotating for 1 h at 4°C. The antigen-antibody complexes were precipitated by incubation with 20 µl of protein A-Sepharose (Boehringer Mannheim, Indianapolis, Ind.). The precipitates were pelleted by centrifugation, washed three times with cold lysis buffer, and resuspended in Laemmli sample buffer (14). The samples were then analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Colocalization of BCCV N protein with actin filaments. Immunofluorescence staining of BCCV-infected Vero cells with a specific anti-BCCV rabbit serum elicits a characteristic filamentous pattern that is strikingly reminiscent of that of actin microfilaments. At 3 days postinfection, approximately 40% of the infected cells demonstrated this pattern (Fig. 1A). The fraction of infected cells showing the filamentous pattern increased, up to 80%, by 6 days postinfection (Fig. 1C). It seemed likely that the observed pattern was the result of interaction of the BCCV proteins with cytoskeletal structures. To examine this possibility, the BCCV-infected cells were subjected to treatment with cytoskeleton-disrupting drugs and examined by IFA. Treatment of the BCCV-infected cells with cytochalasin D, a microfilament-disrupting drug, resulted in a complete disintegration of the observed filamentous structures (Fig. 1B). In contrast, incubation of the infected cells in the presence of nocadozole, a microtubule-dissociating drug, did not have any effect on these structures. To gain further evidence of actin microfilament involvement, the BCCV-infected Vero cells were examined by double immunofluorescence staining with a specific anti-N protein MAb and phalloidin to stain actin microfilaments. As shown in Fig. 1E, the filamentous antigen pattern in infected Vero cells, observed predominantly at the edge of the cells, was found to be colocalized with the actin microfilaments stained with phalloidin (Fig. 1F). Thus, these results strongly indicate the association of the BCCV N protein with actin microfilaments.

View larger version (146K): [in this window] [in a new window]   FIG. 1.   Association of the BCCV N protein with actin filaments. Vero cells were infected with BCCV (at an MOI of 0.5) and examined by IFA with anti-BCCV rabbit serum either at 3 days (A and B) or at 6 days (C) postinfection. (A) Filamentous pattern of viral antigen in infected cells without treatment. (B) Absence of the filamentous pattern in infected Vero cells which were treated with cytochalasin D. (C) Vero cells infected with BCCV (at an MOI of 0.5) at 6 days postinfection. (D) Uninfected Vero cells. Colocalization of BCCV N protein with actin microfilaments was examined in Vero cells that were seeded on coverslips and infected with BCCV (at an MOI of 1). At 3 days postinfection, the cells were analyzed by double immunofluorescence with anti-N MAb and phalloidin. (E) Anti-N staining, some of which (arrows) coincides with anti-actin staining observed with phalloidin. (F) Anti-actin staining with phalloidin.

BCCV nucleocapsids and N proteins preferentially bind to the actin microfilaments. The BCCV genome encodes three major proteins: the G1 and G2 glycoproteins are derived from the GPC and are localized in both the Golgi complex, as inferred from studies with other members of the Bunyaviridae family (9, 16, 23), and the plasma membrane (8, 25); N protein is expressed as a cytoplasmic protein and is assembled into nucleocapsids. To examine the possible interaction of BCCV G1 and G2 proteins with actin microfilaments, the coding sequences for the GPC and N proteins were expressed separately with the Sindbis virus expression system. BHK21 cells were electroporated with the recombinant self-replicating RNAs, and the synthesis of the BCCV proteins was examined by radioimmunoprecipitation assay. Figure 2 shows that the G1 and G2 glycoproteins are synthesized at similar levels and processed to sizes which correspond to those of the viral proteins in infected cells. At 24 h posttransfection, BHK21 cells expressing BCCV proteins were fixed and examined by IFA (Fig. 3). The BCCV glycoproteins expressed in BHK21 cells can be observed both in the perinuclear region and in the endoplasmic reticulum (Fig. 3A). The proteins appear to be distributed throughout the cytoplasm, showing no structures reminiscent of those of actin microfilaments (Fig. 3B). In contrast, BHK21 cells expressing the N protein exhibit the filamentous pattern (Fig. 3C) like that observed in BCCV-infected cells and colocalized with actin filaments (Fig. 3D). However, the filamentous pattern in BHK21 cells is not as prominent as that observed in BCCV-infected Vero cells, and the number of cells showing this pattern was much lower (5 to 10% at 2 days postinfection). Taken together, these results indicate that the BCCV N protein, either expressed alone in transfected cells or presented in the form of a ribonucleoprotein complex (RNP), is capable of binding to the actin filaments.

View larger version (89K): [in this window] [in a new window]   FIG. 2.   Radioimmunoprecipitation analysis of Sindbis virus-expressed GPC and N proteins of BCCV in BHK21 cells with rabbit anti-BCCV sera. BHK21 cells were electroporated with recombinant Sindbis and helper virus RNAs, and at 24 h postinfection the cells were examined by radioimmunoprecipitation assay. In parallel, infected Vero cells were processed similarly at 3 days postinfection. Lane 1, three major BCCV proteins precipitated from the infected Vero cells. Lanes 2 to 4, BHK21 cells electroporated with helper Sindbis virus RNA alone, GPC containing Sindbis RNA expressing G1 and G2 glycoproteins, and N protein, respectively.

View larger version (147K): [in this window] [in a new window]   FIG. 3.   IFA of Sindbis virus-expressed GPC and N proteins of BCCV in BHK21 cells with rabbit anti-BCCV sera. BHK21 cells were electroporated with recombinant Sindbis and helper virus RNAs, and at 24 h postinfection the cells were examined by IFA. (A and B) Expression of GPC protein. (C and D) N-expressing BHK21 cells. The arrows indicate the areas of colocalization of the BCCV N protein with actin filaments.

BCCV N protein binds to actin monomers. In living cells, actin is present in two forms: G (monomeric) and F (filamentous). To confirm the immunofluorescence results and elucidate whether BCCV N protein and BCCV RNP are capable of binding to the G form of actin, we performed coimmunoprecipitation analysis of the BCCV N protein with anti--actin antibody (Fig. 4). In parallel, we carried out the same procedure with anti--tubulin antibody. BCCV-infected Vero cells at 3 days postinfection and BHK21 expressing the BCCV N protein alone at 2 days postelectroporation were compared. At selected time points, the levels of N protein in both cell types were found to be similar, as judged by IFA. During lysis at 4°C, only soluble G actin was extracted by the lysis buffer applied to the cells. Figure 4 shows that the N protein is coimmunoprecipitated with actin from both the infected Vero and the electroporated BHK21 cells. In contrast, no signal was obtained with anti--tubulin antibody. Coimmunoprecipitation of the N protein from BCCV-infected Vero cells was more efficient than coimmunoprecipitation from the electroporated BHK21 cells, which is consistent with the IFA observations showing less extensive association of N protein with actin filaments in BHK21 cells. In addition, we examined the Western blot of coimmunoprecipitated BCCV proteins with anti-BCCV rabbit immune serum, which recognizes all the viral proteins. No BCCV glycoproteins were detected on the blot; the only band seen on the gel was the one corresponding to the N protein, indicating that the BCCV glycoproteins are not coprecipitated with the actin antibody (not shown). From these data, we conclude that the interaction between the BCCV N protein and actin is not limited only to the filamentous form of actin, but that soluble G actin is also involved.

View larger version (78K): [in this window] [in a new window]   FIG. 4.   Coimmunoprecipitation of BCCV N protein with anti--actin antibody. Both BCCV-infected Vero cells at 3 days postinfection and BHK21 cells expressing BCCV N protein alone at 2 days postelectroporation were analyzed. Anti--tubulin antibody was used as a negative control. After immunoprecipitation with anti--actin MAb, the protein complexes were analyzed on a polyacrylamide gel with 0.1% sodium dodecyl sulfate, transferred to a nitrocellulose filter, and analyzed by Western blotting with a 1:500 dilution of anti-N MAb to detect BCCV N protein. Lane 1, coimmunoprecipitation of BCCV N protein from infected Vero cells with anti--actin; lane 2, absence of N protein with anti--tubulin antibodies; lane 3, no immunoprecipitation of N protein with anti--tubulin in BHK21 cells; lane 4, coimmunoprecipitation of N protein expressed alone in BHK21 cells with anti--actin; lane 5, N protein precipitated with anti-N antibody from infected Vero cells.

Depolymerization of actin microfilaments inhibits BCCV release. Although the association of viral proteins with actin microfilaments has been documented for many viruses, not all of the reported viruses exhibit dependence on this cytoskeletal structure. For example, VSV and influenza virus matrix proteins bind actin microfilaments efficiently (2). However, treatment of either VSV- or influenza virus-infected cells with cytochalasin D does not affect the yield or release of these viruses, indicating that actin is not critical for VSV or influenza virus morphogenesis (10, 11, 27). Therefore, we were interested in examining whether the interaction of the BCCV N protein with actin microfilaments is of any significance for viral morphogenesis or release. Vero cells were infected with BCCV (at a multiplicity of infection [MOI] of 1) and at 3 days postinfection supplemented with medium containing various concentrations of cytochalasin D. The medium was collected after 24 h, and the released-virus titers were quantitated by infectious-center assay. Figure 5A shows that the release of BCCV is significantly reduced in the cells with disrupted actin microfilaments. Less than 1 µg of cytochalasin D per ml was sufficient to induce the inhibitory effect. At this concentration, the released-BCCV titer decreased from 2.0 × 10⁶ to 5.9 × 10⁵ infectious-center units (ICU)/ml. In cells treated with higher cytochalasin D concentrations, progressively less BCCV was released. At 3 days postinfection, in the infected cells supplemented with 25 µg of cytochalasin D per ml, the virus titer decreased to 6% relative to the yield from uninfected cells. As expected, treatment of VSV-infected cells with cytochalasin D (Fig. 5B) did not affect the release of the virus, which is consistent with previous observations (27).

View larger version (18K): [in this window] [in a new window]   FIG. 5.   Depolymerization of actin microfilaments inhibits BCCV release. (A) Progressive decrease of the released-BCCV titers upon treatment of Vero cells for 24 h at 3 days postinfection with various concentrations of cytochalasin D (100% corresponds to 2.0 × 106 ICU/ml). (B) Control experiment with VSV which demonstrated no dependence of VSV release on actin microfilaments. VSV-infected Vero E6 cells (at an MOI of 0.1) were treated with cytochalasin D at 24 h postinfection for 24 h (100% corresponds to 4.2 × 108 PFU/ml).

View larger version (25K): [in this window] [in a new window]   FIG. 6.   Disruption of actin microfilaments does not affect BCCV entry. A monolayer of Vero cells was pretreated either without or in the presence of various concentrations of cytochalasin D. The medium was collected at 4 days postinfection, and BCCV titers were measured by infectious-center assay (100% corresponds to 2.8 × 104 ICU/ml).

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

In this study, we demonstrated that the BCCV N protein is capable of interacting with actin microfilaments and that intact actin filaments are important in BCCV morphogenesis. Immunofluorescence staining of virus-infected cells with a specific antibody showed a distinctive filamentous pattern that colocalized with actin microfilaments and disappeared upon treatment with an actin microfilament-depolymerizing drug. The observed filamentous pattern was not a peculiarity of a minority cell population within the culture; at 6 days postinfection, more than 80% of the infected cells exhibited these structures. However, the filamentous pattern was less prominent in the cells expressing BCCV N protein alone. In infected cells, N protein is present either as a free cytoplasmic protein or as a structural component of the BCCV nucleocapsid. Consequently, the affinities of these forms for actin microfilaments might differ as well. A similar theory was proposed to explain the interaction of influenza virus M1 protein with actin microfilaments (2). Normally, in influenza virus-infected cells, this protein is tightly associated with the cytoskeleton. However, when expressed either alone or with influenza virus nucleocapsid protein, M1 loses this property. Avalos et al. (2) concluded, on this basis, that viral RNP is the mediator of the M1 interaction with actin.

In all cell types, actin is present in two forms: filamentous (F) or monomeric (G). Using anti--actin, we demonstrated that the N protein can be coimmunoprecipitated with the G actin monomers. This indicates that the BCCV N protein interacts with both forms of actin, and depolymerization of the filaments per se does not eliminate the binding. However, it is the F form of actin that plays a critical role during BCCV infection, since its depolymerization by cytochalasin D leads to decreased BCCV production and virus release. Inhibition of the BCCV infection in Vero cells, however, was not complete, which could be due to several possible reasons. It is known that the barbed ends of the actin microfilaments associated with the plasma membrane remain resistant to cytochalasin D treatment, and it is conceivable that BCCV employs these resistant forms of actin microfilaments in order to continue a low level of assembly and release. Alternatively, the actin filaments may enhance BCCV assembly but may not be absolutely required for this process. Finally, as proposed for other members of the Bunyaviridae family (1), BCCV assembly may occur at the plasma membrane as well as in the intracellular membrane, and assembly processes at different sites may differ in their requirements for actin microfilaments.

Since pretreatment of Vero cells with various concentrations of cytochalasin D prior to virus adsorption did not interfere with the outcome of the BCCV infection, it seems unlikely that the actin microfilaments are involved at the stage of BCCV entry into a target cell. On the other hand, if the Vero cells are exposed to cytochalasin D during infection, the BCCV yield in the medium drops significantly. This suggests that BCCV interaction with actin microfilaments likely occurs during either virus assembly or release. Had actin microfilaments been involved only at the stage of release of mature virions, one would expect to observe the accumulation of levels of cell-associated BCCV particles corresponding to the decreased amounts of extracellular virus. However, in our experiments, the cell-associated-virus yields were also found to be reduced, which supports the hypothesis that actin microfilaments are involved in BCCV assembly. We propose that actin microfilaments, by interacting with BCCV nucleocapsids, transport the latter to the plasma membrane where the final step of virus assembly and release takes place. Recently, we were able to show that, in contrast with what occurs with many other members of the Bunyaviridae family (9, 16, 22, 23, 29), in polarized epithelial cells BCCV assembly and release occurs at the plasma membrane on the apical surface. In addition, Goldsmith et al. (8) showed that in Sin Nombre virus, another New World hantavirus (19), assembly occurred preferentially at the plasma membrane of infected cells. The possibility that the BCCV-RNP complex is transported to the plasma membrane via a process involving actin is also supported by the finding that, in contrast to that of BCCV, the release of PT virus does not exhibit dependence on actin microfilaments. In contrast to that of BCCV, the assembly of PT virus occurs at intracellular membranes (16).

The fact that a hantavirus protein binds to actin microfilaments may also contribute to an understanding of the molecular mechanisms of increased vascular permeability in hantavirus-infected individuals. The increased vascular permeability in the lung is a hallmark of hantavirus pulmonary syndrome, the human disease associated with New World hantaviruses (6, 13, 26, 32). Since actin microfilaments comprise a network which interconnects adjacent cells, it is conceivable that BCCV N protein binding might affect the integrity of these interconnections and, hence, the permeability of the endothelium.