/
Journal of General Virology(1992), 73, 2577-2584, Printedin Great Britain
2577
Interactions between bluetongue virus core and capsid proteins translated in vitro H. M. Liu, 1'2 T. F. Booth 2 and Polly Roy 1'2'3. 1Laboratory of Molecular Biophysics, Oxford University, South Parks Road, Oxford OX1 3QU, 2NERC Institute of Virology and Environmental Microbiology, Mansfield Road, Oxford OXI 3SR, U.K. and 3School of Public Health, University of Alabama at Birmingham, Alabama 35294, U.S.A.
To determine whether the two major core proteins (VP3 and VP7) of bluetongue virus can interact in vitro to form morphological structures, linearized VP3 and VP7 cDNA clones were transcribed using SP6 polymerase and the resultant transcripts were co-translated using rabbit reticulocyte lysates. The structures derived were isolated by sedimentation through a sucrose gradient and found to resemble V P 3 - V P 7 core-like particles (CLPs) expressed in vivo. Reacting CLPs
synthesized in vivo with outer capsid proteins translated in vitro (VP2 or VP5) indicated that each outer capsid protein has the capacity to bind to a preformed CLP. This was confirmed by in vivo expression of the appropriate genes using baculovirus vectors. The interaction of VP2 or VP5 with the C L P was analysed by electron microscopy and by using immunogoldlabelled monoclonal antibody.
Introduction
that VP2 can be selectively removed from virions (Huismans etal., 1987). Whether VP2 is attached directly to the core through a more labile association than that of VP5, or is attached to VP5, VP7 or VP3 is not known. Using the appropriate baculovirus expression vector it has been shown that virus-like particles (VLPs) can be formed by co-expression of the VP2, VP3, VP5 and VP7 genes (French et al., 1990). In addition, it has been shown that expression of these four genes and that of VP1, or the genes of VP3, VP7 and VP1, results in the encapsidation of the VP1 protein in vivo (Loudon & Roy, 1991). Similar data have been obtained for the encapsidation of the VP4 and VP6 proteins (Le Blois et al., 1992). When the VP2 and VP5 genes are co-expressed in the absence of the VP3 and VP7 core genes, the proteins fail to assemble (French et al., 1990). Such data indicate that the inner capsid isthe foundation for the assembly of the outer capsid. However, whether each outer capsid protein can interact independently with the core is not known. The results presented in this report conclusively demonstrate that VP5 or VP2 can bind to core-like particles (CLPs) independently of each other. In addition we have devised a rapid method to analyse experimentally the requirements for the interaction between VP3 and VP7 to form CLPs, involving the in vitro translation of m R N A transcripts and reaction of the products. By translation of VP2 or VP5 m R N A and reaction with CLPs the relationships between VP2 or VP5 and preformed cores can be investigated.
Bluetongue virus (BTV) is the prototype virus of the Orbivirus genus in the Reoviridae family. Like other members of the family, BTV is a double-shelled virus with a genome consisting of multiple R N A segments. The mature virion contains l0 dsRNA segments and three minor proteins (VP1, VP4 and VP6) surrounded by a core or inner shell consisting of two major proteins, VP3 and VP7. The core is encapsidated by the outer shell proteins VP2 and VP5. Recent cryo-electron microscopy data have indicated that the 69 nm diameter core structure exhibits icosahedral symmetry with a triangulation number of 13 (Prasad et al., 1992). Structural, biochemical and gene expression analyses have shown that the core is composed of two concentric layers of protein enclosing the nucleocapsid of R N A and minor proteins (see French & Roy, 1990; Loudon & Roy, 1991). The knobby outer layer of the core is made up solely of VP7. The smooth ring-like structure forming the inner layer is composed of the second major protein, VP3. The outer shell of negatively stained BTV lacks distinct morphological units, is loosely bound to the core, is composed of two proteins, VP2 and VP5 (Huismans & Van Dijk, 1990) and constitutes approximately 40% of the total protein content of the virus. Very little information is available on the topography and organization of VP2 and VP5 in the outer capsid of the virion, although the available biochemical evidence indicates 0001-0912 © 1992SGM
2578
H . M . Liu, T. F. Booth and P. R o y
Methods Viruses and cells. Recombinant baculoviruses were grown and assayed in Spodoptera frugiperda (St') cells in modified Grace (TC100) medium containing 10% (v/v) foetal bovine serum (FBS) according to the procedures described by Brown & Faulkner (1977). Prototype United States BTV-10 was propagated using monolayers of BHK-21 cells in Eagle's MEM containing 5% (v/v) FBS. The BTV particles were purified as described by Mertens et al. (1987). Synthesis of CLPs and other particles in insect cells. Sf cells were infected with a BTV-10 VP3-VP7 dual gene recombinant baculovirus at a multiplicity of 10 p.f.u./cell, or with the dual recombinant and a single gene recombinant baculovirus (expressing BTV-10 VP2 or BTV10 VP5) each at a multiplicity of 5 p.f.u./cell as described previously (French & Roy, 1990; French et al., 1990). The infected cells were harvested at 48 h post-infection, and the expressed particles (e.g. CLPs) were purified after lysing the cells and banding the extracts in either a CsCI gradient or a discontinuous sucrose gradient as described previously (French & Roy, 1990; French et al., 1990). The particles were recovered and analysed by 10% or 15% SDS-PAGE. Plasmid construction and in vitro transcription and translation. DNA representing the BTV-10 L2, M5 and $7 genes was recovered by BamHI digestion from the recombinant transfer vectors pAcYMIBTV 10.2, pAcYM1-BTV 10.5 (Marshall & Roy, 1989) and pAcYMIBTV 10.7 (Oldfield et al., 1990), respectively, and ligated into the BamHI site of the polylinker of the dual transcription vector pSPT-19. Similarly, DNA representing the BTV-17 L3 gene was recovered from plasmid pAcRP6 17.3 (Inumaru et al., 1987) and cloned into the SmaI site of pSPT-19. Clones determined by restriction enzyme mapping to contain the viral DNA sequences in the correct orientation for SP6 transcription were selected. For in vitro transcription, plasmid DNA was purified by CsCI gradient centrifugation and linearized downstream of the insert by treatment with a restriction enzyme. For pSPT19 BTV 10.2, DNA was digested with Sail, for pSPTI9 17.3 with HindIII, for pSPT19 10.5 with PstI, and for pSPT19 10.7 with H/ndIII. In each case the linearized DNA was recovered by phenolchloroform extraction and ethanol precipitation. Transcription of the linear DNA from the SP6 promoter was performed according to the procedures described by Master & Banerjee (1988). Typically, 1 pg of linear DNA template was incubated for 2 h at 40 °C with 15 units SP6 RNA polymerase (Pharmacia) in the presence of 1 n~-DTT, 40 mMTris-HCl pH 7.9, 6 m~-MgC12, 2 units RNAguard (Amersham) and 500 ItMeach nucleoside triphosphate (ATP, CTP, GTP and UTP). The RNA transcripts were recovered by phenol-chloroform extraction and ethanol precipitation. To prove that the transcripts were full-length, aliquots were resolved by electrophoresis on a 0-65% agarose gel. R N A samples were stored at - 7 0 °C in aliquots of 1 pg/pl. RNA transcripts were translated in vitro using a rabbit reticulocyte translation kit (Amersham) according to the manufacturer's instructions. The 25 gl reaction mixtures contained 20 ~tCi L-[3SS]methionine for each microgram of RNA template. After 1 h incubation at 30 °C the products were recovered and analysed either directly in the presence of CLPs, or after interaction (see below) in the absence of CLPs, for VP2 and/or VP5. Interaction assay and analyses of in vitro translation products. Posttranslation, the VP2 reaction products were incubated at 37 °C for a further 15 min with 10pg of purified baculovirus-expressed CLPs (French & Roy, 1990) in a final volume of 100 Ixl buffer (150 mMpotassium acetate, 25 mM-EDTA, 1 mM-DTT). Similar incubations were performed with the VP5 products and CLPs, VP2 plus VP5 products and CLPs (200 ~tl volume), or VP2 and VP3, VP5 and VP3, or VP7 and VP3 translation products in the absence of CLPs. The mixtures (or the individual translation products) were left for 30 min at
4 °C before layering onto a discontinuous sucrose gradient (10%, 40% and 60% w/v sucrose in 0.2 M-Tris-HCI pH 8.0) and centrifuging for 2h at 4°C and 85000g. The fractions forming the 40% to 60% interface were collected (0.5 ml volume) and analysed.
Immunoprecipitation analyses. The asS-labelled proteins of the material recovered from the sucrose gradient were incubated with rabbit anti-BTV-10 serum and immunoprecipitated using beads of Protein A-Sepharose CL-4B as described previously (French et al., 1989). The precipitated proteins were removed by boiling for 5 min in dissociation buffer (10% 2-mercaptoethanol, 10% SDS, 25% glycerol, 10 mM-Tris-HC1 pH 6.9, 0.02% bromophenoi blue) and resolved by 10 % or 15 % SDS-PAGE, and the radioactive products were located by fluorography. 14C-labelled marker proteins (Amersham) were employed as markers. Western immunoblot analyses. Proteins from the various morphological structures were dissociated by boiling for 5 min in dissociation buffer, and resolved on a 10% or 15% SDS-polyacrylamide gel. The proteins were transferred onto a Durapore filter and detected by probing with rabbit anti-BTV-10 or anti-BTV-17 sera as described previously (French et al., 1989). Electron microscopy. Particles were adsorbed to carbon-filmed 400mesh grids and negatively stained in 2 % uranyl acetate pH 4.4. Purified cores synthesized in vitro or in vivo were captured by using rabbit polyclonal anti-BTV-10 serum (1:500 dilution) preadsorbed onto copper 400-mesh Formvar carbon-coated grids. Grids were washed twice with 0-2 M-Tris-HCI pH 8.0 prior to staining with 2% uranyl acetate and were examined in a JEOL electron microscope. For immunogold studies, particles obtained from simultaneous expression of dual recombinant virus expressing CLPs and single recombinant virus expressing VP2 were adsorbed to carbon-filmed grids and incubated in a 1 : 100 dilution ofVP2 monoclonal antibody for 1 h. The unbound antibody was washed briefly at the end of the interaction and the grid was then incubated in goat anti-mouse immunogold reagent (5 nm gold particles; Bio Cell Research Laboratories). After another wash, grids were stained as described and observed in a JEOL 100 CX microscope at 100 kV. Low dose micrographs were recorded at a 50000 x magnification. Controls included omission of the primary antibody as well as the dual virus alone.
Results I n v i t r o synthesized V P 3 a n d V P 7 proteins assemble into CLPs We have reported previously that BTV CLPs are formed in S f cells i n f e c t e d b y a d u a l g e n e r e c o m b i n a n t baculovirus expressing the BTV VP3 and VP7 proteins. T h e r e f o r e it w a s o f i n t e r e s t to d e t e r m i n e w h e t h e r t h e t w o p r o t e i n s w o u l d a s s e m b l e in vitro. A n in vitro c e l l - f r e e a s s a y w o u l d b e s i m p l e r t h a n t h e in vivo s y s t e m f o r use in s t u d i e s d e s i g n e d to d e f i n e t h e sites o f V P 3 - V P 7 i n t e r a c t i o n n e c e s s a r y f o r c a p s i d a s s e m b l y b e c a u s e it w o u l d a v o i d t h e n e e d to s y n t h e s i z e r e c o m b i n a n t b a c u l o viruses expressing the relevant mutant proteins for each experiment. The SP6 transcription vector system was e m p l o y e d to o b t a i n f u l l - l e n g t h L3 a n d M 7 m R N A s p e c i e s as d e s c r i b e d in M e t h o d s . T h e m R N A s w e r e t r a n s l a t e d e i t h e r i n d i v i d u a l l y o r t o g e t h e r to p e r m i t i n t e r a c t i o n o f
B T V outer capsid
2579
(a) 1
2
3
4
5
1
2
VP3---~~96K
~i~ ! :
~ ~46K
~30K
~I4K
Fig. 1. (a) Analysis of the translation products of VP3 and VP7 mRNAs. Lane 1, translation products of VP3 mRNA; lane 2, translation products of VP7 mRNA; lane 3, co-translationproducts of VP3 and VP7 mRNAs; lane 4, translation products of rabbit reticulocyte lysate lacking exogenous mRNA; lane 5, 14C-labelled marker proteins. (b) Interaction of VP3 and VP7 in vitro. Lane 1, the assembled particles derived from co-translation of VP3 and VP7 mRNA were purified by banding on a sucrose gradient then immunoprecipitated by rabbit anti-BTV-10 sera, and separated by 10% SDS-PAGE and autoradiographed; lane 2, 14C-labelledprotein markers.
the de novo synthesized proteins. The protein products from each reaction were analysed by gel electrophoresis (Fig. 1). By comparison to Mr markers (Fig. 1 a, lane 5), the largest products of the individual m R N A s had the sizes expected for VP3 (lane 1) and VP7 (lane 2). From the VP3 m R N A , several products smaller than the fullsize VP3 were also obtained. Presumably these represented the products of incomplete VP3 m R N A transcripts or incompletely translated VP3 sequences. Reactions lacking m R N A failed to produce significant amounts of labelled protein (Fig. 1a, lane 4). The products of VP3 and VP7 m R N A co-translation (Fig. 1 a, lane 3) were resolved by centrifugation (see Methods). Material recovered from the position in the gradient expected to contain CLPs was immunoprecipitated and found to contain labelled VP3 and VP7 (Fig. 1 b, lane 1). Despite the fact that the crude translation products also contained incomplete VP3 translation products, the particulate material appeared to contain only full-size VP3, suggesting that incomplete VP3 was not incorporated into particles. To identify whether any morphological structures were formed, the materials were analysed by electron microscopy using anti-BTV serum to capture the particles (Fig. 2). Although very few morphological structures were identified under the electron microscope (Fig. 2b), it was evident that the VP3 and VP7 synthesized in vitro could assemble readily. Owing to
Fig. 2. Electron micrographs of particles assembled in vitro. (a) BTV CLPs synthesized in insect cells by a dual recombinant baculovirus expressing VP3 and VP7 (see French & Roy, 1990). (b) VP3 and VP7 products assembled in vitro purifiedby centrifugation and captured by anti-BTV-10 serum. Bar markers represent 100 nm.
lower levels of protein synthesis in vitro, the presence of fewer structures was anticipated. However, the particles assembled in vitro appeared to be morphologically different in comparison with baculovirus-expressed CLPs (Fig. 2a). They appeared to have a less well defined surface structure than CLPs, possibly due to fewer VP7 molecules being present on the VP3 framework. Nevertheless, the diameter of both types of CLPs was similar. No particles were identified in the translation products of the individual VP3 or VP7 m R N A species. However, when the VP3 and VP7 translation products were mixed post-translation, particles similar to those shown in Fig. 2(b) were identified and shown to contain VP3 and VP7 (data not shown). The results indicate that VP3 and VP7 proteins can associate in vitro to formparticles. The inability to detect particles in the single m R N A translation products may be due to the fact that such structures are not formed (i.e. VP3 requires VP7 for assembly), that any interactions which occurred were not ordered enough to be identifiable, or that any
2580
H. M. Liu, T. F. Booth and P. Roy
structures formed are unstable or present at a concentration too low to be detected by the methods employed. No attempt has been made to optimize the amount of particles formed or the abundance of VP7 relative to that of VP3. Interaction of VP2 and VP5 proteins synthesized in vitro with preformed CLPs The VP2 protein is the most variable protein of BTV and is the major, if not sole, determinant of type-specificity. This outer capsid protein is also responsible for virus neutralization and haemagglutination activities (Marshall & Roy, 1990). Available molecular and electron microscopic data indicate that the VP2 molecule is exposed on the surface of virus particles, whereas VP5 appears to be mainly unexposed (Roy et al., 1990; Eaton et al., 1990). It is not known whether VP2 binds to VP5 which in turn binds to the core, or whether both proteins interact directly with the core. Nor is it known whether VP2 or VP5 alone, or both proteins together, can form structures with VP3 in the absence of VP7. Initially we designed experiments to determine whether either VP2 or VP5 can assemble with CLPs in vitro. For this study, preformed CLPs synthesized from a dual gene baculovirus expression vector were employed (French & Roy, 1990). The CLPs were purified from Sf cells infected with the VP3-VP7 dual gene recombinant baculovirus as described in Methods. SP6 transcripts of VP2 and/or VP5 c D N A were synthesized using the SP6 polymerase as described in Methods and identified as full-length by agarose gel electrophoresis. Each transcript was then translated individually (Fig. 3 a, lanes 6 and 7) and subsequently incubated either alone or with an excess of unlabeUed purified CLPs (10 ktg) to allow interaction. The CLPs and assembled particles from each of the mixtures were then purified by sedimentation through sucrose gradients. To analyse the components of the derived particles, the peak fractions of the gradient were immunoprecipitated with anti-BTV-10 sera and products were analysed by 15 ~ S D S - P A G E . Autoradiographs of the 3sS-labelled protein components (VP2 and VP5) are shown in Fig. 3 in comparison with BTV proteins (Fig. 3a, b, lanes 1). As expected, when the translated products of the VP2 and VP5 genes were allowed to interact with unlabelled CLPs, both proteins assembled with the CLPs (Fig. 3a, lane 2). The results indicate that VP2 and VP5 synthesized in vitro can react post-translation with preformed CLPs. When the gel was stained (Fig. 3b, lane 2), in addition to antibody components the unlabelled VP3 and VP7 proteins of the carrier CLPs were identified, but VP2 and VP5 proteins were not. The low abundance o f the protein-CLP structures is not unexpected in view of the fact that an
(a) 1
2
3
4
5
6
7
vP2_ vP3--
VP5 - -
VP7 - -
(b) 1
2
3
4
5
VP3
~--VP7
Fig. 3. Interaction of VP2 and VP5 synthesizedin vitro with CLPs in insect cells by a dual recombinant baculovirus expressing VP3 and VP7. (a) Autoradiographof 15~ polyacrylamidegel containing: lane 1, BTV proteins in which the four major structural proteins were identified (VP2, VP3, VP5 and VP7); lane 2, 3sS-labelled individual translation productsof VP2 and VP5 reacted with unlabelledCLPs and purified through a discontinuous sucrose gradient; lanes 3 and 4, as above except only the VP5 (lane 3) or the VP2 0ane 4) mRNA translation product was reacted with unlabelled CLPs; lane 5, CLPs alone; lanes 6 and 7, individualtranslation products of VP5 (lane 6) or VP2 (lane 7) mRNA. Proteins of the assembled particles (lanes 2 to 5) were immunoprecipitatedusing anti-BTV-10sera prior to analysis by 15~ SDS-PAGE. (b) Stained gel (a), lanes 1 to 5.
excess of CLPs was employed to react with the limited amounts of VP2 and VP5 synthesized in vitro. The ability of VP5 protein to react with the core was demonstrated by the interaction of 35S-labelled VP5 alone in the presence of CLPs (Fig. 3a, b, lanes 3). When VP2 was
B T V outer capsid
1
2
3
(b)
(a) 1
2
Fig. 4. Interaction of VP2 with CLPs in vivo. Insect cells were coinfected with a recombinant baculovirus expressing BTV-10 VP2 and the dual recombinant baculovirusexpressingBTV-17 VP3 and BTV-10 VPT. The assembled particles were purified as described in Methods and analysed by 15% SDS-PAGE followed by Coomassie blue staining. (a) Lane 1, BTV-10proteins in which the four major proteins (VP2, VP3, VP5 and VP7) and the minor protein (VP1) are easily visible, and VP6 is scarcely visible. Minor protein VP4 was not detected in this preparation. Lane 2, CLPs containing VP3 and VPT; lane 3, CLPs associated with VP2. (b) Western blot analyses of CLPs (lane 1) and CLPs with VP2 (lane 2) using anti-BTV-10 sera.
used instead of VP5, it was found that radiolabelled VP2 synthesized in vitro was also attached to preformed CLPs (Fig. 3a, b, lanes 4). Some minor smaller protein bands (see Fig. 3, lanes 6 and 7) also appeared to be incorporated into the assembled particles. However, these bands were not detected in Western blot analysis using anti-BTV-10 sera (data not shown). The electron microscopic analysis of negatively stained C L P s together with VP2 or VP5 showed spherical, empty, virus-like particles lacking the knobby spike-like surface characteristics of the cores and CLPs (data not shown). In summary, the data demonstrate that when either VP2 or VP5, or both, are present they can assemble and interact with C L P s and form particles.
In vivo assembly o f VP2 or VP5 on to C L P s To confirm the above data and to determine whether VP2 or VP5 can assemble on to cores in an in vivo system,
2581
(b) 1
1
2
3
Fig. 5. Interaction of VP5 with CLPs in vivo. Insect ceils were coinfected with a recombinant baculovirus containing the BTV-10 VP5 gene and the dual recombinant baculovirus expressing BTV-17 VP3 and BTV-10 VP7. The assembled particles were purified as described in Methods and analysed by 10% SDS-PAGE. The gel was electroblotted on to an Immobilon membrane and probed with rabbit anti-BTV-17 serum (a) or an anti-BTV-10 serum (b). Lanes 1, CLPs containing VP3 and VP7; lanes 2, CLPs associated with VP5; lane 3, marker proteins.
we employed recombinant baculoviruses to infect Sf cells (Marshall & Roy, 1990; French & Roy, 1990). Cells were co-infected with a BTV-17 V P 3 - V P 7 dual gene recombinant baculovirus and single gene recombinant viruses expressing either BTV-10 VP2 or VP5. The assembled particles were recovered and purified through discontinuous sucrose gradients. The protein components of the particles were analysed by 15 % S D S - P A G E followed by either Coomassie blue staining or Western blot analyses. Shown in Fig. 4(a) are the proteins of BTV (lane 1), C L P s (lane 2) and VP2 assembled with C L P s (lane 3). The presence of VP2 in the derived particles was confirmed by Western blot analyses (Fig. 4b). The presence of VP5 associated with particles was not readily detected by staining procedures, as its expression in the absence of VP2 was considerably lower. However, the presence of VP5 was easily detectable by Western blot analysis using anti-BTV sera. The available anti-BTV-17 serum reacted with VP3, VP5 and VP7 (Fig. 5a) in the particles recovered, but not with VP2, although the available anti-BTV-10 serum reacted only with VP5 and VP7
2582
H. M. Liu, T. F. Booth and P. Roy
Fig. 6. E~ectr~nmicr~graphs~fpartic~esf~rmedbytheinteracti~n~fVP2~rVP5synthesizedinvie~withCLPs.Electron micrographs of (a) CLPs with VP2 and CLPs alone (inset), (b) CLPs with VP5. (c) CLPs with VP2 decorated with gold-labelled anti-VP2 monoclonal antibody. Bar marker represents 100 nm.
in the same samples (Fig. 5b). As with VP2 and & vitro analyses, the results demonstrate that VP5 binds to CLPs in vivo.
To analyse the morphology of these particles, both structures (one containing VP2, VP3 and VP7, the other VP5, VP3 and VP7) were examined by negative staining electron microscopy. In preparations from infections involving the VP3-VP7 dual recombinant vector, the only structures observed were CLPs (see Fig. 6a, b and ¢). In contrast, the co-expression of VP3 and VP7 with either VP2 or VP5 protein resulted in the formation of double-shelled particles resembling VLPs. However, these double-shelled particles had somewhat smaller diameters. VLPs with only one outer capsid protein (VP2
or VP5) had an approximate diameter of between 70 and 74nm, whereas VLPs containing both proteins measured approximately 82 to 85 nm. The data were substantiated by a study using gold-labelled monoclonal or monospecific antibody raised against VP2 or VP5. Conclusive results were obtained demonstrating that both VP2 (Fig. 6b) and VP5 (not shown) attach to the CLPs independently of each other.
Discussion We have demonstrated the ability of VP2 and VP5 of the outer capsid of BTV to interact separately with CLPs.
B T V outer capsid
Previously we have shown that CLPs have a knobby surface structure, similar to cores derived from virus by removal of VP2 and VP5 (French & Roy, 1990). The VP7 capsomers are easily removed from CLPs, resulting in a ring-like structure composed solely of VP3 (Loudon et aL, 1991). These findings, together with our present report, indicate that VP2 and VP5 probably interact only with the VP7 protein of the core. However, since the level of interaction of VP2 and VP5 with CLPs is much greater when the proteins are both present than when they are added individually, it is likely that during the morphogenesis of the virus the proteins stabilize each other in the virion structure. Site-directed mutagenesis of each protein will be required to investigate this point. Alignments o f the predicted amino acid sequences o f the VP2 proteins of different BTV serotypes have identified conserved and variable regions within the protein (Roy, 1989). However, comparisons of VP5 sequences have revealed that, in general, this protein is much more conserved, reflecting a degree of constraint against evolution of the three-dimensional structure of the molecule (Hirasawa et al., 1990). Perhaps little of VP5 is exposed on the virion surface. The lack o f recognition of BTV particles by anti-VP5 sera supports this view. The more variable VP2 protein and its role in inducing serotype-specific neutralizing antibodies and haemagglutination activity support the view that the VP2 protein is mostly exposed on the surface, despite its attachment to VP7 protein and any interaction with VP5. The conserved regions of VP2 are probably necessary for maintaining the conformation of the molecule, as well as for interaction with VP5 and VP7 proteins. Further studies involving specific deletions in these proteins are required to investigate this. Information obtained from this investigation and the data generated from various studies involving recombinant BTV proteins, plus the recent cryo-electron microscopic and image processing analyses of the BTV core, have given us a basic understanding of the assembly of the four major proteins in the mature virion particle. It appears that the assembly of BTV proteins occurs in various stages. Probably an unstable icosahedral particle composed only of VP3 is synthesized first, which interacts immediately with the three minor proteins. This subcore then forms a scaffold for the attachment of VP7 trimers. Either before or immediately after the CLPs are formed, the 10 R N A species become encapsidated. Eventually the VP2 and VP5 proteins simultaneously interact with the core to form the outer capsid of the virus particles. This hypothetical model and the proposed morphogenetic events will need to be confirmed by experimental approaches using both biochemical and ultrastructural analyses.
2583
We are grateful to Dr J. Mecham(USDA]ARS/ABADRL,Laramie, Wyo., U.S.A.) for providingus with VP2 monoclonalantibodies, Miss Stephanie Clarke for typing and Mr Christopher Hatton for photographic work. This work was partially supported by NIH grant A726879, Alabama State grant AE 89-40 and AFRC grant LRG 43535.
References BROWN, M. & FAULKNER,P. (1977). A plaque assay for nuclear polyhedrosisviruses using a solidoverlay.Journal of General Virology 36, 361-364. EATON,B. T., HYATT,A. D. & BROOKES,S. M. (1990). The replication of bluetongue virus. Current Topics in Microbiology and Immunology 162, 89-118. FRENCH,T. J. & ROY,P. (1990). Synthesis of bluetongu¢ virus (BTV) corelike particles by a recombinant baculovirus expressing the two major structural core proteins of BTV. Journal of Virology 64, 15301536. FRENCH,T. J., INUMARU,S. & ROY,P. (1989). Expressionof two related non-structural proteins of BTV-10 in insect cells by a recombinant baculovirus: production of polyclonalascitic fluid and characterization of the gene product in BTV infected BHK cells. Journal of Virology 63, 3270-3278. FaENCH, T. J., MARSHALL,J. J. A. & ROY, P. (1990). Assembly of double-shelled, virus-like particles of bluetongue virus by the simultaneous expression of four structural proteins. Journal of Virology 64, 5695-5700. HIRASAWA,T. & ROY,P. 0990). The complete nucleotide sequence of VP5 of a strain of bluetongue virus of serotype2 isolated in the USA reveals its close relationship with a virus of serotype 1 isolated in Australia. Virus Research 15, 107-112. HUISMANS,H. & VANDIJK,A. A. (1990). Bluetonguevirus components. Current Topics in Microbiology and Immunology 162, 21-41. HUXSMANS,H., VANDim, A. A. & ELS, H. J. (1987). Uncoating of parental bluetongue virus to core and subcore particles in infected L cells. Virology 157, 180-188. HYATr,A. D. & EATON,B. T. (1988). Ultrastructural distribution of the major capsid proteins within bluetongue virus and infected cells. Journal of General Virology 69, 805-815. INUMARU,S., GHIASI,H. & ROY,P. (1987). Expression of bluetongue virus group-specific antigen VP3 in insect cells by a baculovirus expression vector: its use for detection of bluetongue virus antibodies. Journal of General Virology 68, 1627-1635. LE BLOIS,H., FAYARD,B., URAKAWA,T. & ROY,P. (1992). Synthesis and characterization of chimeric particles between epizootic hemorrhagic disease virus and bluetonguevirus: functional domains are conserved on the VP3 protein. Journal of Virology (in press). LOUDON,P. T. & ROY, P. (1991). Assemblyof five bluetongue virus proteins expressed by recombinant baculoviruses: inclusion of the largest protein VPI in the core and virus-like particles. Virology 180, 798-802. LOUDON,P. T., OLDFIELD,S., HIRASAWA,T., MORIKAWA,S., MURPHY, M. & ROY,P. (1991). Expression of outer capsid protein VP5 of two bluetongue viruses and synthesis of chimeric double-capsid virus like particles using recombinant baculoviruses. Virology 182, 793802. MARSHALL,J. J. A. & ROY,P. (1990). High level expression of the two outer capsid proteins of bluetongue virus serotype 10: their relationship with the neutralization of virus infection. Virus Research 15, 189-195. MASTER, e. S. & BANERJEE,A. K. (1988). Resolution of multiple complexes of phosphoprotein NS with nucleocapsid protein N of vesicular stomatitis virus. Journal of Virology 62, 2651-2657. MERTENS, P. P. C., BURROUGHS, J. N. & ANDERSON, J. (1987). Purification properties of virus particles, infectious subviral particles, and cores of bluetongue virus serotypes 1 and 4. Virology 15, 189-195.
2584
H. M. Liu, T. F. Booth and P. Roy
OLDFIELD, S., ADACHI,A., URAKAWA,T., HIRASAWA,T. & RoY, P. 0990). Purification and characterization of the major group-specific core antigen VP7 of bluetongue virus synthesized by a recombinant baculovirus. Journal of General Virology 71, 2649-2656. PRASAD,B. V. V., YAMAGUCm,S. & ROY, P. (1992). Three-dimensional structure of single-shelled BTV. Journalof Virology 66, 2135-2142. RoY, P. (1989). Bluetongue virus genetics and genome structure. Virus Research 13, 179-206.
ROY, P., MARSHALL,J. J. A. & FRENCH, T. J. (1990). Structure of bluetongue virus genome and its encoded proteins. Current Topics in Microbiology and Immunology 162, 43-87.
(Received 11 February 1992; Accepted 16 June 1992)
Journal of General Virology(1992), 73, 2577-2584, Printedin Great Britain
2577
Interactions between bluetongue virus core and capsid proteins translated in vitro H. M. Liu, 1'2 T. F. Booth 2 and Polly Roy 1'2'3. 1Laboratory of Molecular Biophysics, Oxford University, South Parks Road, Oxford OX1 3QU, 2NERC Institute of Virology and Environmental Microbiology, Mansfield Road, Oxford OXI 3SR, U.K. and 3School of Public Health, University of Alabama at Birmingham, Alabama 35294, U.S.A.
To determine whether the two major core proteins (VP3 and VP7) of bluetongue virus can interact in vitro to form morphological structures, linearized VP3 and VP7 cDNA clones were transcribed using SP6 polymerase and the resultant transcripts were co-translated using rabbit reticulocyte lysates. The structures derived were isolated by sedimentation through a sucrose gradient and found to resemble V P 3 - V P 7 core-like particles (CLPs) expressed in vivo. Reacting CLPs
synthesized in vivo with outer capsid proteins translated in vitro (VP2 or VP5) indicated that each outer capsid protein has the capacity to bind to a preformed CLP. This was confirmed by in vivo expression of the appropriate genes using baculovirus vectors. The interaction of VP2 or VP5 with the C L P was analysed by electron microscopy and by using immunogoldlabelled monoclonal antibody.
Introduction
that VP2 can be selectively removed from virions (Huismans etal., 1987). Whether VP2 is attached directly to the core through a more labile association than that of VP5, or is attached to VP5, VP7 or VP3 is not known. Using the appropriate baculovirus expression vector it has been shown that virus-like particles (VLPs) can be formed by co-expression of the VP2, VP3, VP5 and VP7 genes (French et al., 1990). In addition, it has been shown that expression of these four genes and that of VP1, or the genes of VP3, VP7 and VP1, results in the encapsidation of the VP1 protein in vivo (Loudon & Roy, 1991). Similar data have been obtained for the encapsidation of the VP4 and VP6 proteins (Le Blois et al., 1992). When the VP2 and VP5 genes are co-expressed in the absence of the VP3 and VP7 core genes, the proteins fail to assemble (French et al., 1990). Such data indicate that the inner capsid isthe foundation for the assembly of the outer capsid. However, whether each outer capsid protein can interact independently with the core is not known. The results presented in this report conclusively demonstrate that VP5 or VP2 can bind to core-like particles (CLPs) independently of each other. In addition we have devised a rapid method to analyse experimentally the requirements for the interaction between VP3 and VP7 to form CLPs, involving the in vitro translation of m R N A transcripts and reaction of the products. By translation of VP2 or VP5 m R N A and reaction with CLPs the relationships between VP2 or VP5 and preformed cores can be investigated.
Bluetongue virus (BTV) is the prototype virus of the Orbivirus genus in the Reoviridae family. Like other members of the family, BTV is a double-shelled virus with a genome consisting of multiple R N A segments. The mature virion contains l0 dsRNA segments and three minor proteins (VP1, VP4 and VP6) surrounded by a core or inner shell consisting of two major proteins, VP3 and VP7. The core is encapsidated by the outer shell proteins VP2 and VP5. Recent cryo-electron microscopy data have indicated that the 69 nm diameter core structure exhibits icosahedral symmetry with a triangulation number of 13 (Prasad et al., 1992). Structural, biochemical and gene expression analyses have shown that the core is composed of two concentric layers of protein enclosing the nucleocapsid of R N A and minor proteins (see French & Roy, 1990; Loudon & Roy, 1991). The knobby outer layer of the core is made up solely of VP7. The smooth ring-like structure forming the inner layer is composed of the second major protein, VP3. The outer shell of negatively stained BTV lacks distinct morphological units, is loosely bound to the core, is composed of two proteins, VP2 and VP5 (Huismans & Van Dijk, 1990) and constitutes approximately 40% of the total protein content of the virus. Very little information is available on the topography and organization of VP2 and VP5 in the outer capsid of the virion, although the available biochemical evidence indicates 0001-0912 © 1992SGM
2578
H . M . Liu, T. F. Booth and P. R o y
Methods Viruses and cells. Recombinant baculoviruses were grown and assayed in Spodoptera frugiperda (St') cells in modified Grace (TC100) medium containing 10% (v/v) foetal bovine serum (FBS) according to the procedures described by Brown & Faulkner (1977). Prototype United States BTV-10 was propagated using monolayers of BHK-21 cells in Eagle's MEM containing 5% (v/v) FBS. The BTV particles were purified as described by Mertens et al. (1987). Synthesis of CLPs and other particles in insect cells. Sf cells were infected with a BTV-10 VP3-VP7 dual gene recombinant baculovirus at a multiplicity of 10 p.f.u./cell, or with the dual recombinant and a single gene recombinant baculovirus (expressing BTV-10 VP2 or BTV10 VP5) each at a multiplicity of 5 p.f.u./cell as described previously (French & Roy, 1990; French et al., 1990). The infected cells were harvested at 48 h post-infection, and the expressed particles (e.g. CLPs) were purified after lysing the cells and banding the extracts in either a CsCI gradient or a discontinuous sucrose gradient as described previously (French & Roy, 1990; French et al., 1990). The particles were recovered and analysed by 10% or 15% SDS-PAGE. Plasmid construction and in vitro transcription and translation. DNA representing the BTV-10 L2, M5 and $7 genes was recovered by BamHI digestion from the recombinant transfer vectors pAcYMIBTV 10.2, pAcYM1-BTV 10.5 (Marshall & Roy, 1989) and pAcYMIBTV 10.7 (Oldfield et al., 1990), respectively, and ligated into the BamHI site of the polylinker of the dual transcription vector pSPT-19. Similarly, DNA representing the BTV-17 L3 gene was recovered from plasmid pAcRP6 17.3 (Inumaru et al., 1987) and cloned into the SmaI site of pSPT-19. Clones determined by restriction enzyme mapping to contain the viral DNA sequences in the correct orientation for SP6 transcription were selected. For in vitro transcription, plasmid DNA was purified by CsCI gradient centrifugation and linearized downstream of the insert by treatment with a restriction enzyme. For pSPT19 BTV 10.2, DNA was digested with Sail, for pSPTI9 17.3 with HindIII, for pSPT19 10.5 with PstI, and for pSPT19 10.7 with H/ndIII. In each case the linearized DNA was recovered by phenolchloroform extraction and ethanol precipitation. Transcription of the linear DNA from the SP6 promoter was performed according to the procedures described by Master & Banerjee (1988). Typically, 1 pg of linear DNA template was incubated for 2 h at 40 °C with 15 units SP6 RNA polymerase (Pharmacia) in the presence of 1 n~-DTT, 40 mMTris-HCl pH 7.9, 6 m~-MgC12, 2 units RNAguard (Amersham) and 500 ItMeach nucleoside triphosphate (ATP, CTP, GTP and UTP). The RNA transcripts were recovered by phenol-chloroform extraction and ethanol precipitation. To prove that the transcripts were full-length, aliquots were resolved by electrophoresis on a 0-65% agarose gel. R N A samples were stored at - 7 0 °C in aliquots of 1 pg/pl. RNA transcripts were translated in vitro using a rabbit reticulocyte translation kit (Amersham) according to the manufacturer's instructions. The 25 gl reaction mixtures contained 20 ~tCi L-[3SS]methionine for each microgram of RNA template. After 1 h incubation at 30 °C the products were recovered and analysed either directly in the presence of CLPs, or after interaction (see below) in the absence of CLPs, for VP2 and/or VP5. Interaction assay and analyses of in vitro translation products. Posttranslation, the VP2 reaction products were incubated at 37 °C for a further 15 min with 10pg of purified baculovirus-expressed CLPs (French & Roy, 1990) in a final volume of 100 Ixl buffer (150 mMpotassium acetate, 25 mM-EDTA, 1 mM-DTT). Similar incubations were performed with the VP5 products and CLPs, VP2 plus VP5 products and CLPs (200 ~tl volume), or VP2 and VP3, VP5 and VP3, or VP7 and VP3 translation products in the absence of CLPs. The mixtures (or the individual translation products) were left for 30 min at
4 °C before layering onto a discontinuous sucrose gradient (10%, 40% and 60% w/v sucrose in 0.2 M-Tris-HCI pH 8.0) and centrifuging for 2h at 4°C and 85000g. The fractions forming the 40% to 60% interface were collected (0.5 ml volume) and analysed.
Immunoprecipitation analyses. The asS-labelled proteins of the material recovered from the sucrose gradient were incubated with rabbit anti-BTV-10 serum and immunoprecipitated using beads of Protein A-Sepharose CL-4B as described previously (French et al., 1989). The precipitated proteins were removed by boiling for 5 min in dissociation buffer (10% 2-mercaptoethanol, 10% SDS, 25% glycerol, 10 mM-Tris-HC1 pH 6.9, 0.02% bromophenoi blue) and resolved by 10 % or 15 % SDS-PAGE, and the radioactive products were located by fluorography. 14C-labelled marker proteins (Amersham) were employed as markers. Western immunoblot analyses. Proteins from the various morphological structures were dissociated by boiling for 5 min in dissociation buffer, and resolved on a 10% or 15% SDS-polyacrylamide gel. The proteins were transferred onto a Durapore filter and detected by probing with rabbit anti-BTV-10 or anti-BTV-17 sera as described previously (French et al., 1989). Electron microscopy. Particles were adsorbed to carbon-filmed 400mesh grids and negatively stained in 2 % uranyl acetate pH 4.4. Purified cores synthesized in vitro or in vivo were captured by using rabbit polyclonal anti-BTV-10 serum (1:500 dilution) preadsorbed onto copper 400-mesh Formvar carbon-coated grids. Grids were washed twice with 0-2 M-Tris-HCI pH 8.0 prior to staining with 2% uranyl acetate and were examined in a JEOL electron microscope. For immunogold studies, particles obtained from simultaneous expression of dual recombinant virus expressing CLPs and single recombinant virus expressing VP2 were adsorbed to carbon-filmed grids and incubated in a 1 : 100 dilution ofVP2 monoclonal antibody for 1 h. The unbound antibody was washed briefly at the end of the interaction and the grid was then incubated in goat anti-mouse immunogold reagent (5 nm gold particles; Bio Cell Research Laboratories). After another wash, grids were stained as described and observed in a JEOL 100 CX microscope at 100 kV. Low dose micrographs were recorded at a 50000 x magnification. Controls included omission of the primary antibody as well as the dual virus alone.
Results I n v i t r o synthesized V P 3 a n d V P 7 proteins assemble into CLPs We have reported previously that BTV CLPs are formed in S f cells i n f e c t e d b y a d u a l g e n e r e c o m b i n a n t baculovirus expressing the BTV VP3 and VP7 proteins. T h e r e f o r e it w a s o f i n t e r e s t to d e t e r m i n e w h e t h e r t h e t w o p r o t e i n s w o u l d a s s e m b l e in vitro. A n in vitro c e l l - f r e e a s s a y w o u l d b e s i m p l e r t h a n t h e in vivo s y s t e m f o r use in s t u d i e s d e s i g n e d to d e f i n e t h e sites o f V P 3 - V P 7 i n t e r a c t i o n n e c e s s a r y f o r c a p s i d a s s e m b l y b e c a u s e it w o u l d a v o i d t h e n e e d to s y n t h e s i z e r e c o m b i n a n t b a c u l o viruses expressing the relevant mutant proteins for each experiment. The SP6 transcription vector system was e m p l o y e d to o b t a i n f u l l - l e n g t h L3 a n d M 7 m R N A s p e c i e s as d e s c r i b e d in M e t h o d s . T h e m R N A s w e r e t r a n s l a t e d e i t h e r i n d i v i d u a l l y o r t o g e t h e r to p e r m i t i n t e r a c t i o n o f
B T V outer capsid
2579
(a) 1
2
3
4
5
1
2
VP3---~~96K
~i~ ! :
~ ~46K
~30K
~I4K
Fig. 1. (a) Analysis of the translation products of VP3 and VP7 mRNAs. Lane 1, translation products of VP3 mRNA; lane 2, translation products of VP7 mRNA; lane 3, co-translationproducts of VP3 and VP7 mRNAs; lane 4, translation products of rabbit reticulocyte lysate lacking exogenous mRNA; lane 5, 14C-labelled marker proteins. (b) Interaction of VP3 and VP7 in vitro. Lane 1, the assembled particles derived from co-translation of VP3 and VP7 mRNA were purified by banding on a sucrose gradient then immunoprecipitated by rabbit anti-BTV-10 sera, and separated by 10% SDS-PAGE and autoradiographed; lane 2, 14C-labelledprotein markers.
the de novo synthesized proteins. The protein products from each reaction were analysed by gel electrophoresis (Fig. 1). By comparison to Mr markers (Fig. 1 a, lane 5), the largest products of the individual m R N A s had the sizes expected for VP3 (lane 1) and VP7 (lane 2). From the VP3 m R N A , several products smaller than the fullsize VP3 were also obtained. Presumably these represented the products of incomplete VP3 m R N A transcripts or incompletely translated VP3 sequences. Reactions lacking m R N A failed to produce significant amounts of labelled protein (Fig. 1a, lane 4). The products of VP3 and VP7 m R N A co-translation (Fig. 1 a, lane 3) were resolved by centrifugation (see Methods). Material recovered from the position in the gradient expected to contain CLPs was immunoprecipitated and found to contain labelled VP3 and VP7 (Fig. 1 b, lane 1). Despite the fact that the crude translation products also contained incomplete VP3 translation products, the particulate material appeared to contain only full-size VP3, suggesting that incomplete VP3 was not incorporated into particles. To identify whether any morphological structures were formed, the materials were analysed by electron microscopy using anti-BTV serum to capture the particles (Fig. 2). Although very few morphological structures were identified under the electron microscope (Fig. 2b), it was evident that the VP3 and VP7 synthesized in vitro could assemble readily. Owing to
Fig. 2. Electron micrographs of particles assembled in vitro. (a) BTV CLPs synthesized in insect cells by a dual recombinant baculovirus expressing VP3 and VP7 (see French & Roy, 1990). (b) VP3 and VP7 products assembled in vitro purifiedby centrifugation and captured by anti-BTV-10 serum. Bar markers represent 100 nm.
lower levels of protein synthesis in vitro, the presence of fewer structures was anticipated. However, the particles assembled in vitro appeared to be morphologically different in comparison with baculovirus-expressed CLPs (Fig. 2a). They appeared to have a less well defined surface structure than CLPs, possibly due to fewer VP7 molecules being present on the VP3 framework. Nevertheless, the diameter of both types of CLPs was similar. No particles were identified in the translation products of the individual VP3 or VP7 m R N A species. However, when the VP3 and VP7 translation products were mixed post-translation, particles similar to those shown in Fig. 2(b) were identified and shown to contain VP3 and VP7 (data not shown). The results indicate that VP3 and VP7 proteins can associate in vitro to formparticles. The inability to detect particles in the single m R N A translation products may be due to the fact that such structures are not formed (i.e. VP3 requires VP7 for assembly), that any interactions which occurred were not ordered enough to be identifiable, or that any
2580
H. M. Liu, T. F. Booth and P. Roy
structures formed are unstable or present at a concentration too low to be detected by the methods employed. No attempt has been made to optimize the amount of particles formed or the abundance of VP7 relative to that of VP3. Interaction of VP2 and VP5 proteins synthesized in vitro with preformed CLPs The VP2 protein is the most variable protein of BTV and is the major, if not sole, determinant of type-specificity. This outer capsid protein is also responsible for virus neutralization and haemagglutination activities (Marshall & Roy, 1990). Available molecular and electron microscopic data indicate that the VP2 molecule is exposed on the surface of virus particles, whereas VP5 appears to be mainly unexposed (Roy et al., 1990; Eaton et al., 1990). It is not known whether VP2 binds to VP5 which in turn binds to the core, or whether both proteins interact directly with the core. Nor is it known whether VP2 or VP5 alone, or both proteins together, can form structures with VP3 in the absence of VP7. Initially we designed experiments to determine whether either VP2 or VP5 can assemble with CLPs in vitro. For this study, preformed CLPs synthesized from a dual gene baculovirus expression vector were employed (French & Roy, 1990). The CLPs were purified from Sf cells infected with the VP3-VP7 dual gene recombinant baculovirus as described in Methods. SP6 transcripts of VP2 and/or VP5 c D N A were synthesized using the SP6 polymerase as described in Methods and identified as full-length by agarose gel electrophoresis. Each transcript was then translated individually (Fig. 3 a, lanes 6 and 7) and subsequently incubated either alone or with an excess of unlabeUed purified CLPs (10 ktg) to allow interaction. The CLPs and assembled particles from each of the mixtures were then purified by sedimentation through sucrose gradients. To analyse the components of the derived particles, the peak fractions of the gradient were immunoprecipitated with anti-BTV-10 sera and products were analysed by 15 ~ S D S - P A G E . Autoradiographs of the 3sS-labelled protein components (VP2 and VP5) are shown in Fig. 3 in comparison with BTV proteins (Fig. 3a, b, lanes 1). As expected, when the translated products of the VP2 and VP5 genes were allowed to interact with unlabelled CLPs, both proteins assembled with the CLPs (Fig. 3a, lane 2). The results indicate that VP2 and VP5 synthesized in vitro can react post-translation with preformed CLPs. When the gel was stained (Fig. 3b, lane 2), in addition to antibody components the unlabelled VP3 and VP7 proteins of the carrier CLPs were identified, but VP2 and VP5 proteins were not. The low abundance o f the protein-CLP structures is not unexpected in view of the fact that an
(a) 1
2
3
4
5
6
7
vP2_ vP3--
VP5 - -
VP7 - -
(b) 1
2
3
4
5
VP3
~--VP7
Fig. 3. Interaction of VP2 and VP5 synthesizedin vitro with CLPs in insect cells by a dual recombinant baculovirus expressing VP3 and VP7. (a) Autoradiographof 15~ polyacrylamidegel containing: lane 1, BTV proteins in which the four major structural proteins were identified (VP2, VP3, VP5 and VP7); lane 2, 3sS-labelled individual translation productsof VP2 and VP5 reacted with unlabelledCLPs and purified through a discontinuous sucrose gradient; lanes 3 and 4, as above except only the VP5 (lane 3) or the VP2 0ane 4) mRNA translation product was reacted with unlabelled CLPs; lane 5, CLPs alone; lanes 6 and 7, individualtranslation products of VP5 (lane 6) or VP2 (lane 7) mRNA. Proteins of the assembled particles (lanes 2 to 5) were immunoprecipitatedusing anti-BTV-10sera prior to analysis by 15~ SDS-PAGE. (b) Stained gel (a), lanes 1 to 5.
excess of CLPs was employed to react with the limited amounts of VP2 and VP5 synthesized in vitro. The ability of VP5 protein to react with the core was demonstrated by the interaction of 35S-labelled VP5 alone in the presence of CLPs (Fig. 3a, b, lanes 3). When VP2 was
B T V outer capsid
1
2
3
(b)
(a) 1
2
Fig. 4. Interaction of VP2 with CLPs in vivo. Insect cells were coinfected with a recombinant baculovirus expressing BTV-10 VP2 and the dual recombinant baculovirusexpressingBTV-17 VP3 and BTV-10 VPT. The assembled particles were purified as described in Methods and analysed by 15% SDS-PAGE followed by Coomassie blue staining. (a) Lane 1, BTV-10proteins in which the four major proteins (VP2, VP3, VP5 and VP7) and the minor protein (VP1) are easily visible, and VP6 is scarcely visible. Minor protein VP4 was not detected in this preparation. Lane 2, CLPs containing VP3 and VPT; lane 3, CLPs associated with VP2. (b) Western blot analyses of CLPs (lane 1) and CLPs with VP2 (lane 2) using anti-BTV-10 sera.
used instead of VP5, it was found that radiolabelled VP2 synthesized in vitro was also attached to preformed CLPs (Fig. 3a, b, lanes 4). Some minor smaller protein bands (see Fig. 3, lanes 6 and 7) also appeared to be incorporated into the assembled particles. However, these bands were not detected in Western blot analysis using anti-BTV-10 sera (data not shown). The electron microscopic analysis of negatively stained C L P s together with VP2 or VP5 showed spherical, empty, virus-like particles lacking the knobby spike-like surface characteristics of the cores and CLPs (data not shown). In summary, the data demonstrate that when either VP2 or VP5, or both, are present they can assemble and interact with C L P s and form particles.
In vivo assembly o f VP2 or VP5 on to C L P s To confirm the above data and to determine whether VP2 or VP5 can assemble on to cores in an in vivo system,
2581
(b) 1
1
2
3
Fig. 5. Interaction of VP5 with CLPs in vivo. Insect ceils were coinfected with a recombinant baculovirus containing the BTV-10 VP5 gene and the dual recombinant baculovirus expressing BTV-17 VP3 and BTV-10 VP7. The assembled particles were purified as described in Methods and analysed by 10% SDS-PAGE. The gel was electroblotted on to an Immobilon membrane and probed with rabbit anti-BTV-17 serum (a) or an anti-BTV-10 serum (b). Lanes 1, CLPs containing VP3 and VP7; lanes 2, CLPs associated with VP5; lane 3, marker proteins.
we employed recombinant baculoviruses to infect Sf cells (Marshall & Roy, 1990; French & Roy, 1990). Cells were co-infected with a BTV-17 V P 3 - V P 7 dual gene recombinant baculovirus and single gene recombinant viruses expressing either BTV-10 VP2 or VP5. The assembled particles were recovered and purified through discontinuous sucrose gradients. The protein components of the particles were analysed by 15 % S D S - P A G E followed by either Coomassie blue staining or Western blot analyses. Shown in Fig. 4(a) are the proteins of BTV (lane 1), C L P s (lane 2) and VP2 assembled with C L P s (lane 3). The presence of VP2 in the derived particles was confirmed by Western blot analyses (Fig. 4b). The presence of VP5 associated with particles was not readily detected by staining procedures, as its expression in the absence of VP2 was considerably lower. However, the presence of VP5 was easily detectable by Western blot analysis using anti-BTV sera. The available anti-BTV-17 serum reacted with VP3, VP5 and VP7 (Fig. 5a) in the particles recovered, but not with VP2, although the available anti-BTV-10 serum reacted only with VP5 and VP7
2582
H. M. Liu, T. F. Booth and P. Roy
Fig. 6. E~ectr~nmicr~graphs~fpartic~esf~rmedbytheinteracti~n~fVP2~rVP5synthesizedinvie~withCLPs.Electron micrographs of (a) CLPs with VP2 and CLPs alone (inset), (b) CLPs with VP5. (c) CLPs with VP2 decorated with gold-labelled anti-VP2 monoclonal antibody. Bar marker represents 100 nm.
in the same samples (Fig. 5b). As with VP2 and & vitro analyses, the results demonstrate that VP5 binds to CLPs in vivo.
To analyse the morphology of these particles, both structures (one containing VP2, VP3 and VP7, the other VP5, VP3 and VP7) were examined by negative staining electron microscopy. In preparations from infections involving the VP3-VP7 dual recombinant vector, the only structures observed were CLPs (see Fig. 6a, b and ¢). In contrast, the co-expression of VP3 and VP7 with either VP2 or VP5 protein resulted in the formation of double-shelled particles resembling VLPs. However, these double-shelled particles had somewhat smaller diameters. VLPs with only one outer capsid protein (VP2
or VP5) had an approximate diameter of between 70 and 74nm, whereas VLPs containing both proteins measured approximately 82 to 85 nm. The data were substantiated by a study using gold-labelled monoclonal or monospecific antibody raised against VP2 or VP5. Conclusive results were obtained demonstrating that both VP2 (Fig. 6b) and VP5 (not shown) attach to the CLPs independently of each other.
Discussion We have demonstrated the ability of VP2 and VP5 of the outer capsid of BTV to interact separately with CLPs.
B T V outer capsid
Previously we have shown that CLPs have a knobby surface structure, similar to cores derived from virus by removal of VP2 and VP5 (French & Roy, 1990). The VP7 capsomers are easily removed from CLPs, resulting in a ring-like structure composed solely of VP3 (Loudon et aL, 1991). These findings, together with our present report, indicate that VP2 and VP5 probably interact only with the VP7 protein of the core. However, since the level of interaction of VP2 and VP5 with CLPs is much greater when the proteins are both present than when they are added individually, it is likely that during the morphogenesis of the virus the proteins stabilize each other in the virion structure. Site-directed mutagenesis of each protein will be required to investigate this point. Alignments o f the predicted amino acid sequences o f the VP2 proteins of different BTV serotypes have identified conserved and variable regions within the protein (Roy, 1989). However, comparisons of VP5 sequences have revealed that, in general, this protein is much more conserved, reflecting a degree of constraint against evolution of the three-dimensional structure of the molecule (Hirasawa et al., 1990). Perhaps little of VP5 is exposed on the virion surface. The lack o f recognition of BTV particles by anti-VP5 sera supports this view. The more variable VP2 protein and its role in inducing serotype-specific neutralizing antibodies and haemagglutination activity support the view that the VP2 protein is mostly exposed on the surface, despite its attachment to VP7 protein and any interaction with VP5. The conserved regions of VP2 are probably necessary for maintaining the conformation of the molecule, as well as for interaction with VP5 and VP7 proteins. Further studies involving specific deletions in these proteins are required to investigate this. Information obtained from this investigation and the data generated from various studies involving recombinant BTV proteins, plus the recent cryo-electron microscopic and image processing analyses of the BTV core, have given us a basic understanding of the assembly of the four major proteins in the mature virion particle. It appears that the assembly of BTV proteins occurs in various stages. Probably an unstable icosahedral particle composed only of VP3 is synthesized first, which interacts immediately with the three minor proteins. This subcore then forms a scaffold for the attachment of VP7 trimers. Either before or immediately after the CLPs are formed, the 10 R N A species become encapsidated. Eventually the VP2 and VP5 proteins simultaneously interact with the core to form the outer capsid of the virus particles. This hypothetical model and the proposed morphogenetic events will need to be confirmed by experimental approaches using both biochemical and ultrastructural analyses.
2583
We are grateful to Dr J. Mecham(USDA]ARS/ABADRL,Laramie, Wyo., U.S.A.) for providingus with VP2 monoclonalantibodies, Miss Stephanie Clarke for typing and Mr Christopher Hatton for photographic work. This work was partially supported by NIH grant A726879, Alabama State grant AE 89-40 and AFRC grant LRG 43535.
References BROWN, M. & FAULKNER,P. (1977). A plaque assay for nuclear polyhedrosisviruses using a solidoverlay.Journal of General Virology 36, 361-364. EATON,B. T., HYATT,A. D. & BROOKES,S. M. (1990). The replication of bluetongue virus. Current Topics in Microbiology and Immunology 162, 89-118. FRENCH,T. J. & ROY,P. (1990). Synthesis of bluetongu¢ virus (BTV) corelike particles by a recombinant baculovirus expressing the two major structural core proteins of BTV. Journal of Virology 64, 15301536. FRENCH,T. J., INUMARU,S. & ROY,P. (1989). Expressionof two related non-structural proteins of BTV-10 in insect cells by a recombinant baculovirus: production of polyclonalascitic fluid and characterization of the gene product in BTV infected BHK cells. Journal of Virology 63, 3270-3278. FaENCH, T. J., MARSHALL,J. J. A. & ROY, P. (1990). Assembly of double-shelled, virus-like particles of bluetongue virus by the simultaneous expression of four structural proteins. Journal of Virology 64, 5695-5700. HIRASAWA,T. & ROY,P. 0990). The complete nucleotide sequence of VP5 of a strain of bluetongue virus of serotype2 isolated in the USA reveals its close relationship with a virus of serotype 1 isolated in Australia. Virus Research 15, 107-112. HUISMANS,H. & VANDIJK,A. A. (1990). Bluetonguevirus components. Current Topics in Microbiology and Immunology 162, 21-41. HUXSMANS,H., VANDim, A. A. & ELS, H. J. (1987). Uncoating of parental bluetongue virus to core and subcore particles in infected L cells. Virology 157, 180-188. HYATr,A. D. & EATON,B. T. (1988). Ultrastructural distribution of the major capsid proteins within bluetongue virus and infected cells. Journal of General Virology 69, 805-815. INUMARU,S., GHIASI,H. & ROY,P. (1987). Expression of bluetongue virus group-specific antigen VP3 in insect cells by a baculovirus expression vector: its use for detection of bluetongue virus antibodies. Journal of General Virology 68, 1627-1635. LE BLOIS,H., FAYARD,B., URAKAWA,T. & ROY,P. (1992). Synthesis and characterization of chimeric particles between epizootic hemorrhagic disease virus and bluetonguevirus: functional domains are conserved on the VP3 protein. Journal of Virology (in press). LOUDON,P. T. & ROY, P. (1991). Assemblyof five bluetongue virus proteins expressed by recombinant baculoviruses: inclusion of the largest protein VPI in the core and virus-like particles. Virology 180, 798-802. LOUDON,P. T., OLDFIELD,S., HIRASAWA,T., MORIKAWA,S., MURPHY, M. & ROY,P. (1991). Expression of outer capsid protein VP5 of two bluetongue viruses and synthesis of chimeric double-capsid virus like particles using recombinant baculoviruses. Virology 182, 793802. MARSHALL,J. J. A. & ROY,P. (1990). High level expression of the two outer capsid proteins of bluetongue virus serotype 10: their relationship with the neutralization of virus infection. Virus Research 15, 189-195. MASTER, e. S. & BANERJEE,A. K. (1988). Resolution of multiple complexes of phosphoprotein NS with nucleocapsid protein N of vesicular stomatitis virus. Journal of Virology 62, 2651-2657. MERTENS, P. P. C., BURROUGHS, J. N. & ANDERSON, J. (1987). Purification properties of virus particles, infectious subviral particles, and cores of bluetongue virus serotypes 1 and 4. Virology 15, 189-195.
2584
H. M. Liu, T. F. Booth and P. Roy
OLDFIELD, S., ADACHI,A., URAKAWA,T., HIRASAWA,T. & RoY, P. 0990). Purification and characterization of the major group-specific core antigen VP7 of bluetongue virus synthesized by a recombinant baculovirus. Journal of General Virology 71, 2649-2656. PRASAD,B. V. V., YAMAGUCm,S. & ROY, P. (1992). Three-dimensional structure of single-shelled BTV. Journalof Virology 66, 2135-2142. RoY, P. (1989). Bluetongue virus genetics and genome structure. Virus Research 13, 179-206.
ROY, P., MARSHALL,J. J. A. & FRENCH, T. J. (1990). Structure of bluetongue virus genome and its encoded proteins. Current Topics in Microbiology and Immunology 162, 43-87.
(Received 11 February 1992; Accepted 16 June 1992)
