Assembly and Intracellular Localization of the Bluetongue Virus Core Protein VP3

Abstract

The bluetongue virus (BTV) core protein VP3 plays a crucial role in the virion assembly and replication process. Although the structure of the protein is well characterized, much less is known about the intracellular processing and localization of the protein in the infected host cell. In BTV-infected cells, newly synthesized viral core particles accumulate in specific locations within the host cell in structures known as virus inclusion bodies (VIBs), which are composed predominantly of the nonstructural protein NS2. However, core protein location in the absence of VIBs remains unclear. In this study, we examined VP3 location and degradation both in the absence of any other viral protein and in the presence of NS2 or the VP3 natural associate protein, VP7. To enable real-time tracking and processing of VP3 within the host cell, a fully functional enhanced green fluorescent protein (EGFP)-VP3 chimera was synthesized, and distribution of the fusion protein was monitored in different cell types using specific markers and inhibitors. In the absence of other BTV proteins, EGFP-VP3 exhibited distinct cytoplasmic focus formation. Further evidence suggested that EGFP-VP3 was targeted to the proteasome of the host cells but was dispersed throughout the cytoplasm when MG132, a specific proteasome inhibitor, was added. However, the distribution of the chimeric EGFP-VP3 protein was altered dramatically when the protein was expressed in the presence of the BTV core protein VP7, a normal partner of VP3 during BTV assembly. Interaction of EGFP-VP3 and VP7 and subsequent assembly of core-like particles was further examined by visualizing fluorescent particles and was confirmed by biochemical analysis and by electron microscopy. These data indicated the correct assembly of EGFP-VP3 subcores, suggesting that core formation could be monitored in real time. When EGFP-VP3 was expressed in BTV-infected BSR cells, the protein was not associated with proteasomes but instead was distributed within the BTV inclusion bodies, where it colocalized with NS2. These findings expand our knowledge about VP3 localization and its fate within the host cell and illustrate the assembly capability of a VP3 molecule with a large amino-terminal extension. This also opens up the possibility of application as a delivery system.

Bluetongue virus (BTV) is an Orbivirus that has a nonenveloped icosahedral particle that contains an RNA genome of 10 double-stranded segments within three concentric protein shells. The outer shell is composed of two proteins, VP2 and VP5, and the inner shell, or core, consists of two other layers, each formed from a different viral protein, VP7 and VP3 (30). The VP3 layer, together with the transcription complex of three minor enzymatic proteins, VP1, VP4, and VP6, and the 10 double-stranded RNA segments, forms the subcore, which is the first assembled particle that acts as the foundation for virus assembly. Expression of VP3 (103 kDa) using a heterologous baculovirus expression system results in spontaneous assembly of a particulate structure that serves as a scaffold for the deposition of VP7 trimers, forming a stable core structure. The ability to form stable core-like particles (CLPs) of VP3 and VP7 proteins was exploited extensively in our previous studies to map the assembly pathway of these proteins and to determine the intra- and intermolecular interactions necessary for core assembly of BTV (14, 20, 34, 35).

In the mature core, the VP3 shell is fundamental to the structure of the particle and dictates the correct assembly of BTV capsids. The VP3 layer of the core is formed by 12 interconnected VP3 decamers; each decamer is composed of five A- and five B-type VP3 molecules (13). Although both A- and B-type VP3 molecules have an overall elongated triangular shape and each is divided into three domains (namely, apical, carapace, and dimerization), each type exhibits distinct folds. In the atomic structure, while the complete structure of the B molecule was clearly visible, the N terminus of the A molecule was not possible to resolve accurately, implying its conformational flexibility (13). The disordered, flexible amino terminus of a VP3 A molecule was hypothesized to be necessary to accommodate the minor rearrangement of the “apical” domain during assembly of the VP3 decamers. Recently, we have reported that the deletion of the first five residues did not significantly affect the efficiency of VP3 assembly and its interaction with VP7 trimers and subsequent core assembly. These data indicated that it may be possible to extend the amino terminus of VP3 without any adverse effect on protein and capsid assembly.

Although the precise molecular interactions that are necessary for the assembly of BTV are becoming increasingly well characterized, relatively little is known about where in the cell viral proteins assemble and the way in which the trafficking of virus proteins is controlled. In this study, the 27-kDa enhanced green fluorescent protein (EGFP) was fused to the amino terminus of VP3 in order to investigate intracellular localization of VP3 in the presence and absence of other viral proteins. EGFP has been extensively and selectively used as a fusion partner for different types of proteins to monitor their intracellular localizations and fates (16, 32, 39). The fusion construct was expressed successfully in both insect and mammalian cells, allowing examination of the intracellular trafficking of the protein in both cell types. In addition, the impact of extending the amino terminus of VP3 on its ability to oligomerize and to interact with the other core protein, VP7, in the formation of core-like particles was assessed. These studies demonstrate not only that VP3 will accommodate large extensions at the amino terminus without disrupting the formation of CLPs, but also that the presence of other viral proteins alters the intracellular fate of VP3 in both mammalian and insect cells.

MATERIALS AND METHODS

Virus and cells.

Autographa californica nuclear polyhedrosis virus and recombinant baculoviruses were propagated in Spodoptera frugiperda (Sf9) insect cells at 28°C in serum-free SF900 ll SFM medium (GIBCO) as described elsewhere (21).

DNA manipulation and construction of DNA clones.

Plasmid DNA manipulation was performed essentially as described previously (31). Restriction enzymes, T4 DNA ligase, and DNA polymerase were purchased from New England Biolabs. Calf intestinal alkaline phosphatase was obtained from Boehringer GMBH (Mannheim, Germany).

Construction of EGFP-VP3 plasmid for expression in insect and mammalian cell lines.

A modified pAcYM1 vector containing an SmaI restriction enzyme site was generated as described previously (20). The gene expressing EGFP was excised from pBI-EGFP (Clontech) and inserted at the amino terminus of the BTV type 17 (BTV-17) VP3 gene (10). Using SmaI, the EGFP-VP3 fragment was then cloned into the modified pAcYM1 vector for expression in the baculovirus expression system. For expression in the mammalian cell system, the pTARGET (pTARGET Mammalian Expression System; Promega Corporation) vector was used. The EGFP-L3 gene with the SmaI end was inserted into the multiple cloning sites downstream of the strong immediate-early enhancer-promoter region of human cytomegalovirus to promote transient expression of the inserted protein.

Generation of recombinant baculoviruses.

The lipofection technique was used to cotransfect monolayers of Sf9 cells with the recombinant transfer vectors, all generated as described above, and Bsu36I triple-cut Autographa californica nuclear polyhedrosis virus DNA (7). Recombinant baculoviruses were selected on the basis of their LacZ-negative phenotypes, plaque purified, and propagated as described elsewhere (22).

Protein analysis.

Confluent monolayers of Sf9 cells in 35-mm tissue culture dishes were infected with recombinant baculovirus at a multiplicity of infection (MOI) of 5. At 48 h postinfection, cells were harvested and lysed in 150 μl of lysis buffer (50 mM Tris-HCl, pH 8, 200 mM NaCl, 1% Triton X-100). Twenty microliters of the protein samples was boiled for 6 min in 5× sample buffer (2.3% sodium dodecyl sulfate [SDS], 2.5% [vol/vol] glycerol, 5% [vol/vol] β-mercaptoethanol, 62.5 mM Tris-HCl, pH 6.8, 0.01% bromophenol blue), and proteins were separated by SDS-10% polyacrylamide gel electrophoresis (PAGE) and stained with Coomassie blue.

Western blot analysis.

Western blot analysis of proteins was performed following the standard protocol as described previously (20).

Labeling of baculovirus-infected insect cells and immunoprecipitation.

Sf9 cells infected with one or two recombinant baculoviruses at an MOI of 5 were incubated in SF900 serum-free medium at 28°C. After 42 h postinfection, the cells were labeled with 100 μCi of [³⁵S]Met/Cys (NEN) per ml of medium for 2 h. Cells were then harvested, washed, and lysed with RIPA buffer (50 mM Tris-HCl [pH 7.5], 150 mM NaCl, 1% Triton X-100, 0.5% sodium deoxycholate, 1 mM EDTA), incubated in ice for 15 min, and centrifuged at 13,000 × g. The cell lysate was incubated with an anti-VP3 polyclonal antibody, and the complex was precipitated with protein A-agarose beads (Pierce), washed with RIPA buffer, and analyzed by SDS-10% PAGE and autoradiography.

Purification of core-like particles.

Baculovirus-expressed CLPs were purified as described previously (8). Briefly, Sf9 cells were coinfected with recombinant baculovirus in suspension culture expressing wild-type VP7 (27), together with either wild-type VP3 (25) or EGFP-VP3 fusion protein, using an MOI of 5. After incubation at 28°C for 48 h, infected cells were harvested and lysed by Dounce homogenization in TNN buffer (200 mM Tris-HCl, pH 8, 150 mm NaCl, and 0.5% Nonidet P-40). Assembled CLPs were purified from the soluble fraction by centrifugation on a 35% CsCl gradient as described previously (20). The presence of protein was analyzed by SDS-10% PAGE, followed by Coomassie blue staining and Western blotting.

Immunostaining and immunofluorescence microscopy.

For immunolabeling experiments, 1.5 × 10⁵ cells were seeded the day before infection or transfection in six-well plates (9.6 cm² per well) containing 18-mm round glass coverslips. A total of 1.8 μg of DNA was transfected into cells using 10 μl of Lipofectamine 2000 (Invitrogen) in OptiMEM (Invitrogen) and incubated for 4 h as suggested by the manufacturer. After incubation, fresh Dulbecco's modified Eagle's medium was added to the cells. Incubation was continued at 37°C in the presence of 5% CO₂ for 14 h before the cells were processed for immunofluorescence microscopy. For baculovirus infection, Sf9 cells were incubated with virus at an MOI of 5 for 42 h before being processed for further experiments. Infected or transfected cells were fixed by incubation at room temperature with 3% formaldehyde in phosphate-buffered saline (PBS). Fixed cells were permeabilized in 0.5% Triton X-100 in PBS, followed by blocking in blocking buffer (1% bovine serum albumin in PBS). The cells were incubated with the primary antibody, followed by incubation with appropriate secondary antibodies conjugated to Alexa Fluor 488 (green) or Alexa Fluor 594 (red) for 1 h. The immunostained cells were washed three times with PBS, followed by incubation with Hoechst 33342 for nuclear staining, and mounted on slides with Vectashield (Vector Laboratories). Immunolabeled samples were examined using a Leitz Orthoplan inverted microscope (Leitz Corporation) with fluorescence optics. Images were collected digitally using IP Lab software (Scananalytics, Fairfax, Va.) and edited in Adobe Photoshop (Adobe Systems, San Jose, Calif.). The monoclonal antibody against vimentin was purchased from Oncogene Research Products, CA. Lysotracker, Mitotracker, Hoechst 33342, and immunoglobulin G conjugated to Alexa Fluor 488 or Alexa Fluor 594 were obtained from Molecular Probes, OR. The rabbit polyclonal and monoclonal antibodies against proteasome S-20 and ubiquitin were obtained from Affiniti, Exeter, United Kingdom.

MG132 treatment of GFP-VP3-transfected BSR cells.

BSR cells were transfected with EGFP-VP3 plasmid as described above. After incubation, fresh Dulbecco's modified Eagle's medium was added to the cells with different amounts of MG132 (Calbiochem). As a control, dimethyl sulfoxide (DMSO) was used instead of MG132. The cells continued growing for 14 h at 37°C with 5% CO₂. After incubation, the cells were washed in PBS and fixed in 3% formaldehyde (Electron Microscopy Science). The nucleus was stained with Hoechst 33342, mounted with Vectashield (Vector Laboratories), and checked by confocal microscopy.

Negative staining and EM analysis.

Negative staining and electron microscopy (EM) analysis were performed following the standard protocol as described previously (20).

RESULTS

Synthesis of recombinant EGFP-VP3 in two different expression systems: characterization and identification of its intracellular localization.

In our previous studies, we reported that it was feasible to delete a few residues from both the amino and carboxyl termini of VP3 without perturbing VP3 assembly (20, 34). In this study, we investigated whether the VP3 N terminus could be extended with a large foreign protein, such as EGFP, that facilitates studies of intracellular tracking of the protein. EGFP was fused to the BTV L3 gene and expressed in insect cells (Sf9 cells) using the baculovirus protein expression system as described in Materials and Methods. Synthesis of a recombinant baculovirus expressing EGFP-VP3 was identified from infected Sf9 cells by SDS-PAGE and confirmed by immunoprecipitation using an anti-VP3 antibody (Fig. 1A). The apparent molecular size of the GFP-VP3 baculovirus-expressed protein observed in SDS-PAGE corresponded to the estimated size for the GFP-VP3 fusion protein (Fig. 1, lane 1) and migrated slower than the wild-type VP3 protein (Fig. 1A, lane 2). Expression of the fusion protein in the live Sf9 cells infected with the recombinant baculovirus was also visualized by GFP fluorescence using fluorescence microscopy (Fig. 1B).

FIG. 1.

Expression of the EGFP-VP3 fusion protein in insect cells. (A) Sf9 cells were infected with recombinant baculovirus expressing the EGFP-VP3 fusion protein or wild-type (WT) VP3 as a control at an MOI of 5, and the cells were labeled with ³⁵S for 2 hours. At 42 h postinfection, infected cells were lysed; proteins were immunoprecipitated using an anti-VP3 antibody and analyzed by SDS-PAGE and autoradiography. Lane 1, immunoprecipitated EGFP-VP3 protein; lane 2, control wild-type VP3. (B) Real-time monitoring of EGFP fluorescence by fluorescence microscopy. Sf9 cells were infected with a recombinant baculovirus expressing EGFP-VP3 chimeric proteins at an MOI of 5. At 40 h postinfection, the cells were examined by microscope. Pictures of the cells were taken in differential interference contrast mode, and GFP fluorescence was monitored with a fluorescence microscope.

To analyze the general trafficking behavior of newly synthesized EGFP-VP3, the distribution of the protein was analyzed in insect cells by fluorescence microscopy at various times of infection. Clear distribution of distinct punctate fluorescent structures, which were not present when EGFP alone was expressed (data not shown), was visualized in the cells infected with recombinant baculovirus (Fig. 2A, image a). With increased time of infection, the punctate structures were predominantly localized to one side of the nucleus (Fig. 2A, image c [60 h]).

FIG. 2.

Localization of transiently expressed EGFP fused to VP3 or VP3 alone. EGFP-VP3 was expressed in Sf9 cells infected with recombinant baculovirus or transiently expressed in BSR cells. Unmodified VP3 was also expressed in BSR cells. (A) Cells were fixed and visualized using a Leitz Orthoplan (Leitz Corporation) microscope. Pictures were taken with IP Lab software. (a to c) Localization of EGFP-VP3 in Sf9 cells at different time points. At different times postinfection (as indicated), cells were fixed and subjected to confocal microscopy. (d to f) Sf9 cells were infected with recombinant baculovirus expressing EGFP-VP3 and incubated for 48 h, fixed, labeled with an anti-VP3 antibody and anti-rabbit Alexa Fluor 594 as a secondary antibody, and subjected to confocal microscopy. Nuclei were stained with Hoechst 33342. (B) BSR cells transfected with plasmids expressing EGFP-VP3 (a) or VP3 (c) were fixed and observed by confocal microscopy (a) or, prior to microscopy, labeled with an anti-VP3 antibody followed by anti-rabbit Alexa Fluor 594 as a secondary antibody (c). Real-time monitoring of the EGFP fluorescence was done by fluorescence microscopy. BSR cells were transfected with 1.8 μg of EGFP-VP3 DNA and incubated for 14 h prior to examination by microscopy (b). Nuclei were stained with Hoechst 33342.

To make certain that these patterns did not represent any artifacts contributed due to the presence of EGFP, Sf9 cells infected with recombinant baculovirus expressing EGFP-VP3 were immunolabeled with an anti-VP3 polyclonal antibody and examined by confocal microscopy. For clarity, a single cell from a panel of transfected cells was selected (Fig. 2A, images d to f). The merged picture (f) exhibited reasonably good colocolization of the EGFP signal (Fig. 2A, image d) and the red stain (Fig. 2A, image e) of VP3. The sizes of VP3 foci in Fig (2B, image c) are a mixture of big and small foci, and some of them are definitely comparable to the size of EGFP-tagged VP3 foci. The reason for seeing only big foci in Fig. 2A, image d, might be the quenching of the smaller and weaker GFP fluorescence due to the fixative treatment. It is known that formaldehyde can quench substantial amounts of GFP fluorescence, and the quenching is proportional to the time of treatment. In the case of wild-type VP3, the red fluorescence is not impaired by the formaldehyde treatment. The colocalization thus confirmed that the punctate structures were EGFP-VP3 chimeras.

To ensure that the high expression level of EGFP-VP3 in the insect cells did not affect its normal localization, the EGFP-VP3 construct was expressed transiently in mammalian cells, as described in Materials and Methods. Expression in BSR cells transfected with a plasmid containing the EGFP-VP3 gene was first confirmed by Western blotting (data not shown), and the localization of the fusion protein in BSR cells was subsequently visualized by confocal microscopy. The fluorescence signals obtained from cells 12 h posttransfection clearly exhibited the presence of the same type of punctate structures (Fig. 2B, image a) as seen in insect cells. Similar fluorescence signals were also visible when the unfixed, live cells were examined by fluorescence microscopy (Fig. 2B, image b). The data confirmed that the distribution of EGFP-VP3 did not depend on the type of cells or expression system used.

To investigate whether VP3 itself, but not EGFP in the fusion protein, was driving focus formation in the cell, a plasmid was constructed for expression of untagged wild-type VP3 dependent on a cytomegalovirus promoter. The distribution of expressed VP3 in the BSR cells 14 h posttransfection was examined by immunolabeling of the fixed cells using anti-VP3 polyclonal antibodies (Fig. 2B, image c). Expression of native VP3 in BSR cells produced foci very similar to those observed on expression of the EGFP-VP3 fusion (Fig. 2B, images a and c). The data obtained confirmed that VP3 itself has the ability to form distinct punctate structures in the cytoplasm and that it was responsible for focus formation of the chimeric protein EGFP-VP3. Thus, addition of EGFP to the amino terminus of VP3 did not affect its normal targeting and subcellular localization.

Fate of expressed VP3 in cytoplasm.

The location of VP3 within the infected cells was determined by immunostaining of specific cellular proteins. A series of cellular markers for lysosomes, mitochondria, proteasome, and ubiquitin were used for the studies. The distinct colocalization pattern of GFP-VP3 and the proteasome in contrast to that in the lysosome and ubiquitin are shown in Fig. 3. Clearly, VP3 was not ubiquitinated (Fig. 3C). We used both monoclonal and polyclonal antibodies and repeated this experiment three or four times with each type of antibody. Unlike the polyclonal antibody, the monoclonal antibody we used recognizes only ubiquitinated proteins and not free ubiquitin. However, in our studies, we failed to visualize any colocalization of VP3 with ubiquitin, nor was VP3 identified in the lysosomal compartment (Fig. 3B) or in the mitochondria (data not shown). Only the polyclonal antibody to the 20S proteasome showed any signals of colocalization with EGFP-VP3 in the focal structure (Fig. 3A). The absence of colocalization signals with ubiquitin also indicated that the VP3 protein was not targeted by ubiquitin, implying that the degradation of the protein most likely does not follow the ubiquitin-proteasome pathway. Although it has been shown that most proteins that are degraded by the proteasome are multiubiquitinated, ubiquitin-independent proteasome degradation has also been demonstrated for some proteins (2, 9, 28, 29).

FIG. 3.

EGFP-VP3 targets the proteasome. Sf9 cells were infected with baculovirus expressing the EGFP-VP3 fusion protein for 42 h and were fixed as described in Materials and Methods. The cells were stained with specific antibodies and dyes as indicated below, followed by confocal fluorescence microscopy. The first column of each panel shows EGFP-VP3 localization in green, the second column shows the organelle position in red, and in the third column, green and red channels are merged. (A) Proteasome labeled with anti-proteasome S20 mouse monoclonal antibody at a concentration of 5 μg/ml, followed by anti-mouse Alexa Fluor 594 (2 μg/ml) as a secondary antibody. (B) Lysosomes were stained with Lysotracker Red dye at a concentration of 50 nM. (C) Ubiquitin was labeled with a specific mouse monoclonal antibody to ubiquitinylated protein (clone FK2) at a concentration of 5 μg/ml, followed by anti-mouse Alexa Fluor 594 (2 μg/ml) as a secondary antibody. Nuclei were stained with Hoechst 33342.

To characterize the proteasome targeting more precisely, EGFP-VP3-transfected BSR cells were treated with different concentrations (25 μM to 100 μM) of MG132 (in DMSO), a potent proteasome inhibitor, as described in Materials and Methods. After 14 h of treatment, the cells were washed and fixed, and the green fluorescence distribution in the cells was monitored by confocal microscopy (Fig. 4). As a control, only DMSO was added to the EGFP-VP3-transfected cells in order to confirm that it did not influence the formation of foci in the cells (Fig. 4a). It was evident that at higher doses of MG132 (25 μM or higher), the foci started to gradually disappear and eventually dispersed throughout the cell (Fig. 4b to d). In the presence of 100 μM of the inhibitor (Fig. 4d), instead of distinct focus formation, the green stain was dispersed all through the cell. There are also two distinct foci visible, and they are most probably the centrosome. It has been reported by Wigley et al. that the centrosome enlarges in response to inhibition of proteasome activity, and at high levels of misfolded protein, the structure not only expands, it also recruits cytoplasmic pools of 20S proteasome (40). Very similar effects of the inhibitor were also observed when insect cells were used (data not shown). These results, together with the colocalization data, confirmed that VP3, when expressed alone in the absence of other BTV proteins, is targeted specifically to the proteasome. When proteasome formation was inhibited, VP3 was dispersed throughout the cytoplasm.

FIG. 4.

Effect of proteasome inhibitor MG132 on EGFP-VP3 distribution. BSR cells were transfected with 1.8 μg of EGFP-VP3 DNA and incubated for 4 h prior to addition of MG132 (Calbiochem), as discussed in Materials and Methods. (a) Control BSR cell transfected with EGFP-VP3 and treated with DMSO. Transfected BSR cells incubated for 14 h in the presence of (b) 25 μM MG132, (c) 50 μM MG132, and (d) 100 μM MG132. After treatment, the cells were washed in PBS, fixed, and processed for immunofluorescence study.

It has been shown that cells respond to the synthesis of large amounts of misfolded, foreign proteins by transporting them to a perinuclear electron-dense inclusions called aggresomes (1, 9, 18). These aggresomes induce the redistribution of the normal filament protein, such as vimentin, around the aggregated protein, similar to a cage, and also recruit polyubiquitin and mitochondria to the site. To verify whether VP3 chimeras were not folded correctly and thus were deposited in an aggresome, the association of EGFP-VP3 and vimentin was investigated. BSR cells were transfected with the EGFP-VP3 plasmid and immunolabeled using a monoclonal antibody directed against vimentin filament. The examination of these cells by fluorescence microscopy demonstrated that vimentin did not structurally reorganize surrounding the foci (Fig. 5). We monitored vimentin structure and localization of mitochondria and ubiquitin up to 50 h posttransfection and found consistent results. These data and the absence of ubiquitin as described earlier suggest that the structures formed by EGFP-VP3 chimeras were not aggresomes.

FIG. 5.

EGFP-VP3 does not form an aggresome in transfected cells. BSR cells were transfected with EGFP-VP3 plasmid, and 14 h after transfection, the cells were immunolabeled with an anti-vimentin monoclonal antibody and visualized by confocal microscopy. The left panel shows the distribution of EGFP-VP3 (green), the middle panel shows vimentin (red), and the right panel is the merge of vimentin and EGFP-VP3. Note that GFP-VP3 did not induce the redistribution of vimentin.

VP3 could tolerate extensive extension at the N terminus without perturbing the core assembly.

BTV CLPs were formed efficiently in the cytoplasm of insect cells when VP3 and VP7 were cosynthesized by recombinant baculoviruses (8). This approach allowed determination of the effects of mutations introduced into viral structural proteins on protein-protein interaction and BTV particle assembly (11, 23, 24, 33, 36). We used this assay system to examine the assembly efficiency of chimeric VP3 protein. To test the assembly of CLPs made of EGFP-VP3, we coexpressed EGFP-VP3 with wild-type VP7 and purified CLPs through a cesium chloride (CsCl) gradient as described previously (20). The wild-type VP3 was used in place of EGFP-VP3 for the control experiment. In contrast to the native CLP band, a clear fluorescent band was observed at the position of native CLP in the gradient, indicating the efficient assembly of EGFP-VP3 and VP7. When analyzed by SDS-10% PAGE, the cesium chloride gradient band contained EGFP-VP3 and VP7 proteins as expected (Fig. 6A). However, surprisingly, some native VP3 was also present in the purified band (Fig. 6A). The presence of EGFP-VP3 was examined by Western blotting using both the anti-BTV antibody (Fig. 6B, lane 2) and an anti-EGFP polyclonal antibody (Fig. 6C, lane 2). Although EGFP-CLP was clearly assembled as expected, the production of chimeric CLPs was less efficient than the wild-type CLP production (Fig. 6A). The morphology of the purified EGFP-CLP was characterized by EM (Fig. 7, left). To assess more accurately the assembly efficiency of EGFP-VP3 with VP7 in comparison to the normal VP3, densitometric scanning of the Coomassie-stained protein bands of purified CLPs was performed to estimate the proportion of VP7 in each preparation (24). The intensity of the VP7 band in the chimeric CLPs was clearly less than the wild-type CLP VP7 band in the scan. In wild-type CLPs, 600 copies (200 trimers) of VP7 molecules per 120 copies of VP3 are considered the full complement (24). In EGFP-CLP, the number of VP7 molecules was approximately 360, nearly half of the VP7 content of wild-type CLPs. To detect the fluorescence in the EGFP-CLP, an aliquot (1 mg/ml) was examined directly by confocal microscopy, which revealed an almost homogeneous population of green spots of the same intensity (Fig. 7, right), most probably representing units of CLPs tagged with GFP as described for rotavirus CLPs (5).

FIG. 6.

Protein analysis of purified CLPs. Sf9 cells were coinfected with a recombinant baculovirus expressing BTV-10 VP7 and either EGFP-VP3 chimeric protein or BTV-17 VP3 as a control. Infected cells were harvested 48 h postinfection and lysed, and the expressed particles were purified by centrifugation on a CsCl gradient. Particulate fractions were analyzed either by SDS-10% PAGE followed by Coomassie blue staining (A) or by Western blotting using anti-BTV polyclonal antibodies (B) or anti-GFP polyclonal antibodies (C). (A) CLPs formed by VP7 with native VP3 (lane 1) or with EGFP-VP3 (lane 2). The positions of EGFP-VP3, VP3, and VP7 are indicated. (B) Western blot using an anti-BTV antibody showing the CLPs formed by VP7 with native VP3 (lane 1) or with EGFP-VP3 (lane 2). (C) Western blot using an anti-EGFP antibody showing CLPs formed by VP7 with native VP3 (lane 1) or with EGFP-VP3 (lane 2). The position of EGFP-VP3 is indicated. WT, wild type.

FIG. 7.

Visualization of individual CLPs by confocal or electron microscopy. Purified CLPs were examined by electron microscopy (left panel; bar size, 100 nm) or directly by confocal microscopy (right panel; bar size, 500 nm).

VP3 distribution pattern changes in the presence of VP7.

To investigate whether EGFP-VP3 and VP7 interaction makes any changes in the VP3 distribution pattern in infected cells, Sf9 cells were coinfected with baculoviruses expressing GFP-VP3 and VP7 and the distribution of the green fluorescence was analyzed in the cells by fluorescence microscopy at various times of infection. With increased time of infection, the punctate structures were predominantly localized around the nucleus (Fig. 8a to c). Since EGFP-VP3 interacts with VP7 as shown in Fig. 7, the redistribution of VP3 is likely to be due to its interaction with VP7 in transfected cells and the fact that the VP3-VP7 interaction is the driving force in the relocalization of EGFP-VP3 fluorescence.

FIG. 8.

EGFP-VP3 migrates toward nuclei when coexpressed with VP7. Shown are the patterns of EGFP-VP3 in the presence of VP7 in Sf9 cells (a to c) and in BSR cells (d). Sf9 cells were coinfected with recombinant baculovirus expressing EGFP-VP3 and VP7. At different times postinfection (as indicated), cells were fixed and subjected to confocal microscopy. Nuclei were stained with Hoechst 33342. (d) BSR cells transfected with plasmids expressing EGFP-VP3 and VP7 were fixed at 14 h and observed by confocal microscopy.

To obtain additional evidence of the redistribution of EGFP-VP3 in the presence of VP7, BSR cells were transfected with two plasmids, one expressing EGFP-VP3 and the other VP7. The presence of EGFP-VP3 was monitored by confocal microscopy 14 h posttransfection. Changes in the VP3 localization pattern were clearly visible by confocal analysis (Fig. 8d). As in insect cells, instead of focus formation, green fluorescence now relocated around the nucleus. However, there were also some distinct punctate fluorescent structures, indicating that not all EGFP-VP3 interacted with VP7. This was not surprising, since the expression of EGFP-VP3 and VP7 in BSR cells was considerably less than in Sf9 cells. It is therefore obvious that not all VP3 will interact with VP7 in BSR cells and that some EGFP-VP3 foci can be expected in the cells.

EGFP-VP3 is colocalized with the virus inclusion bodies in BTV-infected BSR cells.

It is believed that virus inclusion bodies (VIBs), in which BTV NS2 is a main constituent, are an assembly site for BTV components, including the structural core-associated proteins and virus RNA (37). Indeed, it was possible to detect the presence of VP3 by immunogold labeling within the VIBs in BTV-infected cells (3). These data indicated that assembly of VP3 is likely to occur within VIBs. Therefore, in the presence of VIBs, it is possible that EGFP-VP3 may migrate to VIBs but not to the proteasome. To investigate whether EGFP-VP3 could indeed be redirected to the VIBs, BSR cells were first infected with BTV for 14 h and subsequently transfected with the plasmid expressing EGFP-VP3. After 14 h of transfection, the BSR cells were washed and fixed, and the presence of EGFP-VP3 in the VIBs was examined. Since NS2 is the major constituent of VIBs, the presence of NS2 was used as a marker for VIBs (37). VIBs were examined by confocal microscopy by immunolabeling them with an anti-NS2 polyclonal antibody, and the presence of VP3 was determined by virtue of the fluorescence of EGFP-VP3. Confocal microscopic analysis demonstrated that the anti-NS2 antibody perfectly colocalizes with punctate foci of EGFP-VP3, thus indicating that VP3 was arrested within the VIBs in the BTV-infected cells (Fig. 9).

FIG. 9.

EGFP-VP3 targets NS2 inclusion bodies when expressed in BTV-infected cells. EGFP-VP3 plasmid was transfected into BTV-infected BSR cells, and at 14 h posttransfection, the cells were immunolabeled with an anti-NS2 polyclonal antibody and visualized by confocal microscopy. The top left panel shows the distribution pattern of EGFP-VP3 (green stain), and the top right panel demonstrates the distribution of NS2 (red stain). Nuclei were stained with Hoechst 33342 (blue stain). The lower two panels show the merged pictures. Merge RG, red and green stains; Merge RGB, red, green, and blue stains.

DISCUSSION

In this report, we have characterized the intracellular localization of the EGFP-tagged BTV core protein VP3 in both mammalian and insect cells. The 103-kDa VP3 protein was found to assemble into particulate structures in the absence of other BTV proteins (20) that are fundamental for the initiation of the assembly of more complex BTV particles (13).

The general trafficking behavior of newly synthesized EGFP-VP3 fusion protein and its intracellular distribution were analyzed by immunofluorescence microscopy at various times following infection. Expression of EGFP alone resulted in a diffuse distribution pattern throughout the cell, suggesting cytosolic and nuclear localization as reported earlier (4, 19). However, when EGFP was fused to VP3, the pattern of EGFP localization altered dramatically. At 42 h postinfection, the EGFP-VP3 fusion protein exhibited a distinct pattern of intense cytoplasmic foci. With increasing time of infection, these foci migrated to one side of the cell and were deposited near the nucleus. It is important to note that even though misfolding of EGFP-VP3 is likely to occur to some degree, the EGFP component of the chimera must have been properly folded, since green fluorescence was detected readily. To investigate whether EGFP-VP3 foci were cell type dependent, experiments were performed in both mammalian and insect cells with similar outcomes.

Overexpression of proteins had been shown to deposit protein into large aggregates, forming vimentin-caged aggresomes, as a general cellular response to the accumulation of undegraded protein aggregates (6, 9, 18). Aggresomes subsequently recruit proteasomes and ubiquitin for the degradation of aggregated proteins (9). In the case of VP3, our data suggested that VP3 aggregates did not rearrange the vimentin structures analogously to the aggresomes and were not associated with ubiquitin, as immunolabeling with anti-ubiquitin monoclonal antibody showed no colocalization with the VP3 foci. Therefore, EGFP-VP3 was not destined to follow the proteasome-ubiquitin pathway of degradation. However, 14 h after treatment of EGFP-VP3-expressing BSR cells with the proteasome inhibitor MG132, there was a marked change in the distribution of the fusion protein, with GFP fluorescence detectable throughout the cell. Therefore, normal cellular proteosome function is necessary for the redistribution of EGFP-VP3 into discrete foci. It is also interesting that EGFP-VP3 deposited in the proteasome was not ubiquitinated; hence, it can be hypothesized that VP3 was not degraded by the ubiquitin-proteasome pathway. Ubiquitin-independent proteasome degradation has been demonstrated for other proteins (2, 9, 15, 28, 29).

We recently reported that the N-terminal sequence of BTV-VP3 is crucial for assembly with the other major core protein, VP7. Although the loss of 5 N-terminal residues still allowed VP3-VP7 core assembly, deletion of 10 residues affected CLP assembly severely. Extension of the N terminus with EGFP had no apparent effect on the interaction between the chimeric protein and VP7 and subsequent assembly of particles. However, although the morphology of EGFP-containing particles as revealed by EM appeared to be similar to that of the normal VP3-VP7 particles, SDS-PAGE analysis showed the presence of both normal VP3 and chimeric VP3 in the population. This opens up two possibilities: the population either contained a mixture of normal CLPs and GFP-CLPs, or both native VP3 and EGFP-VP3 molecules were needed for core assembly. The mixed population of CLPs is possible due to the protease activity of insect cells on EGFP-VP3. However, it was not obvious when the fusion protein was expressed in the absence of VP7 (Fig. 1). The latter notion, on the other hand, could be more easily justified, as in native core structure, two types of VP3 molecules (A and B) are assembled to form a decamer prior to assembly of the VP3 shell. Therefore, it is possible that while one member of the AB dimer of VP3 molecules does indeed accommodate a large extension, the other does not, due to the inflexibility of the decameric structure. Since in the crystal structure, the amino terminus in molecule A, but not in B, is disordered, A is more likely to accommodate the extra sequences of EGFP while the stable folding of B would not permit any alteration. The presence of EGFP in the CLPs was confirmed by examining the purified CLP band in a confocal microscope. An effect of VP7 on the localization of VP3 chimeras also became evident from the confocal study by tracking the movement of green fluorescence, and it was confirmed by immunostaining with an anti-VP3 antibody.

BTV infection leads to the formation of perinuclear, non-membrane-bound cytoplasmic inclusions in the host cell (17). Moreover, the nonstructural protein NS2 is capable of forming viral inclusions when expressed independently of other BTV components in the cell (37). Current data suggest that these inclusions act as foci for the assembly of progeny BTV virions in infected cells by accumulating viral structural components and assembly intermediates. The data presented here demonstrate that, like the normal VP3 molecules, EGFP-VP3 also mostly localized in the VIBs (17, 26a), supporting the hypothesis that the EGFP tagging did not alter the biological properties of the VP3 molecule. Thus, an additional role for NS2 inclusions may be to sequester viral proteins like VP3, which would otherwise aggregate in the cytoplasm of infected cells. The assembly of the core proteins and RNA packaging in relation to NS2 will be an area for our future research.

Overall, our data suggest that it is possible to extend the VP3 amino terminus extensively without critically perturbing its ability to interact with the capsid protein VP7 and CLP assembly. These data open up the realistic possibility of using the BTV CLP particle as a delivery system for epitopes or enzymes to targeted cells (12, 26, 38).