Antigenic structure of the capsid protein of rabbit ... - Semantic Scholar
Journal of General Virology (1998), 79, 1901–1909. Printed in Great Britain ...................................................................................................................................................................................................................................................................................
Antigenic structure of the capsid protein of rabbit haemorrhagic disease virus Jorge L. Martı! nez-Torrecuadrada,1 Elena Corte! s,1 Carmen Vela,1 Jan P. M. Langeveld,2 Rob H. Meloen,2 Kristian Dalsgaard,3 William D. O. Hamilton4 and J. Ignacio Casal1 Inmunologı! a y Gene! tica Aplicada SA (INGENASA), Hnos Garcia Noblejas 41, 28037 Madrid, Spain ID-DLO, PO Box 65, NL-8200 Lelystad, The Netherlands 3 Danish Veterinary Institute for Virus Research, DK-4771 Kalvehave, Denmark 4 Axis Genetics plc, Babraham, Cambridge CB2 4AZ, UK 1 2
Rabbit haemorrhagic disease virus (RHDV) causes an important disease in rabbits. The virus capsid is composed of a single 60 kDa protein. The capsid protein gene was cloned in Escherichia coli using the pET3 system, and the antigenic structure of RHDV VP60 was dissected using 11 monoclonal antibodies (MAbs) and 12 overlapping fragments of the protein expressed in E. coli. Two antigenic regions were found. Ten out of the 11 MAbs recognized different discontinuous epitopes in the most immunodominant region of the viral capsid. This
Introduction Rabbit haemorrhagic disease (RHD) is a viral disease of the species Oryctolagus cuniculus and was first described by Liu et al. (1984). The disease, characterized by high mortality in adult rabbits (Liu et al., 1984 ; Ohlinger et al., 1989), is caused by rabbit haemorrhagic disease virus (RHDV), a member of the family Caliciviridae (Ohlinger et al., 1990). The family Caliciviridae comprises several groups of nonenveloped viruses which share many biological properties (for a review see Clarke & Lambden, 1997). Among these properties is the presence of cup-like depressions (calyces) in the surface of the virus. However, this property can no longer be regarded as a general feature of caliciviruses as many of them present an ill-defined surface morphology. Thus, antigenically wellcharacterized caliciviruses such as feline calicivirus (FCV) and human Norwalk virus (NV) present a fuzzy, non-distinctive surface morphology. In contrast, RHDV shows a classic ‘ Star of David ’ calicivirus morphology. The 3D structure of RHDV has not been determined. Analysis of the recombinant NV particles by electron cryoAuthor for correspondence : Ignacio Casal. Fax 34 1 408 75 98. e-mail INGENASA!alc.es
0001-5335 # 1998 SGM
domain was located between residues 31 and 250 of the VP60 N terminus. The other MAb revealed the presence of an antigenic site within 102 aa of the C terminus. This MAb did not recognize the major cleavage product of the full-length 60 kDa protein. These results indicate that, in contrast to other caliciviruses such as Norwalk virus (NV), the 36 kDa cleavage product probably forms the N-terminal region of VP60. However, as in NV, the cleavage region appears to be the most immunodominant region.
microscopy and computer image processing techniques has shown that the capsid is composed of 90 dimers of the capsid protein that form a shell domain from which arch-like capsomers protrude (Prasad et al., 1994). Based on sequence similarities with other T ¯ 3 RNA viruses, it has been predicted that the N-terminal amino acids (aa) 1–250 form the shell domain and the remaining C-terminal residues make up the arches (Prasad et al., 1994). The RHDV genome consists of a single-stranded RNA molecule of positive polarity, approximately 7±5 kb in size. The genome organization shows one long ORF that represents 86 % of the viral genome, with the VPg located downstream of the 3C-like protease gene, in the 5« region (Meyers et al., 1991). The capsid protein of RHDV has an apparent molecular mass of 60 kDa. Expression of a cDNA that encodes RHDV VP60 in insect cells infected with a recombinant baculovirus results in spontaneous assembly of the protein into empty recombinant RHDV (rRHDV) virus-like particles (VLPs) (Laurent et al., 1994 ; Plana-Dura! n et al., 1996). These VLPs are morphologically and antigenically similar to native RHDV particles. The immunogenicity of these particles is extremely high ; doses as low as 0±5 µg are able to fully protect rabbits by the systemic route. Since RHDV cannot be propagated in vitro, the availability of large amounts of rRHDV VLPs has facilitated
BJAB
J. L. Martı! nez-Torrecuadrada and others
the production of monoclonal antibodies (MAbs) useful for antigenic characterization of the virus. Capucci et al. (1995) and Laurent et al. (1997) described the production and characterization of a panel of anti-RHDV MAbs. However, the MAb binding sites were not physically mapped in the capsid protein, and so the epitopes and their position on the RHDV capsid remain unknown. Antigenic mapping studies carried out in other caliciviruses such as NV (Hardy et al., 1996) and FCV (Milton et al., 1992 ; Seal et al., 1993 ; Tohya et al., 1997) have identified the presence of several epitopes in the extreme C-terminal regions of the capsid protein. The purpose of this study was to begin to dissect the antigenic and structural topography of the RHDV capsid using MAbs generated to rRHDV particles and VP60-truncated fragments expressed in E. coli. The antigenic map was compared to that obtained for prototype caliciviruses such as FCV and NV.
Methods + Viruses and cells. The RHDV Olot}89 strain used for these studies was isolated and propagated as described previously (PlanaDura! n et al., 1996). The Spodoptera frugiperda cell line Sf9 (ATCC g CRL 1711) was used to propagate recombinant baculovirus AcRHDV-710 and to produce rRHDV VP60 particles (Plana-Dura! n et al., 1996). Sf9 cells were grown in suspension at 27 °C in Grace’s medium supplemented with yeastolate and lactalbumin hydrolysate (Summers & Smith, 1987), containing 5 % foetal calf serum and antibiotics. + Production of antibodies. MAbs specific for rRHDV VP60 were prepared from BALB}c mice immunized with rRHDV particles, purified as previously described (Plana-Dura! n et al., 1996). Protocols for immunization and the preparation of MAbs have been described previously (Sanz et al., 1985 ; Lo! pez de Turiso et al., 1991). All hybridomas were cloned by limiting dilution at least three times. Hybridoma supernatants were screened by ELISA and immunoblotting analysis. The isotype of the MAbs was determined by ELISA using specific anti-mouse subtype antisera (Sigma). Polyclonal rabbit anti-RHDV sera were obtained from rabbits experimentally infected or vaccinated with an inactivated vaccine or recombinant VP60. + Cloning of VP60 gene in pET expression vectors : preparation of truncated forms. The pET3 and pET3x vectors (Studier et al., 1990) were used to construct fusion proteins between gene 10 protein and VP60 fragments. The expression vector pET3 was mainly used ; pET3x was chosen for those fragments smaller than 400 bp. The fusion protein consisted of 11 aa of the gene 10 protein for pET3 and 260 residues for pETx. Standard cloning techniques (Sambrook et al., 1989) using convenient restriction enzyme sites within the VP60 gene were applied to obtain in-frame fusion constructs between the gene 10–VP60. Fig. 1 shows the VP60 fragments obtained from the recombinant plasmid pAcRHDV-710 (Plana-Dura! n et al., 1996) by endonuclease digestion : A (nt 1–90, BamHI fragment) ; B (nt 91–1744, BamHI fragment) ; C (nt 91–527, BamHI to AflIII) ; D (nt 528–1744, AflIII to BamHI) ; E (nt 91–1429, BamHI to PvuII) ; F (nt 91–567, BamHI to SpeI) ; G (nt 568–1429, SpeI to PvuII) and H (nt 1429–1744, PvuII to BamHI).
BJAC
If appropriate restriction sites were not available, VP60 fragments were generated by PCR with Vent DNA polymerase (New England Biolabs). PCR comprised 25 cycles of denaturation at 94 °C for 1 min, primer annealing at 50 °C for 1 min, and extension at 72 °C for 1±5 min. Primers used were as follows (BamHI sites shown in italics) : RHDV-1, 5« CATGGATCCTGGCGTTGTGG 3« ; RHDV-2 : 5« GAAGGATCCATTCACAGCCGTGCTG 3« ; RHDV-3, 5« TGCGGATCCCAGTGAGGACTGGGGTC 3« ; RHDV-4, 5« AAAGGATCCTGGTGCAACCTGGGAG 3« ; RHDV-5, 5« GGTGGATCCAGACGGCTTTCCTGAT 3« ; and RHDV-6, 5« TGGGGATCCTGTGGCGTTGACGTCA 3«. The VP60 fragments generated by PCR were : I (nt 91–999, from primer RHDV-1 to RHDV-4) ; J (nt 301–1429, from RHDV-2 to RHDV-6) ; K (nt 997–1429, from RHDV-5 to RHDV-6) ; and L (nt 91–750, from RHDV1 to RHDV-3). Fragments with BamHI-compatible ends were ligated into BamHI-digested phosphatase-treated pET3 or pET3x (fragments A and H). Fragments lacking BamHI-compatible ends were first subcloned in a pMTL plasmid in order to generate BamHI-compatible ends. The ligation mixtures were used to transform XL-1 Blue or HMS174 competent cells. The resulting colonies were characterized by restriction mapping. The orientation of the insert and the junction sequences of the recombinant plasmids were sequenced by the dideoxynucleotide method. + Induction and purification of E. coli-derived VP60 fragments. The expression and purification of fusion proteins were performed as described previously (Martı! nez-Torrecuadrada & Casal, 1995). Briefly, purification of the different truncated forms of VP60 was carried out by solubilization of the cell pellet with 4 M guanidinium hydrochloride after lysis by sonication. This simple treatment allowed a quick purification of the fusion proteins (greater than 85 %) in a single step. These partially purified fusion proteins were used for epitope mapping. + SDS–polyacrylamide gels and immunoblotting assays. Purity of the fusion protein and protein concentration were estimated from Coomassie-stained SDS–polyacrylamide gels. Proteins were diluted in loading buffer and boiled for 5 min, or not boiled where indicated, before analysis by 11 % SDS–PAGE. The resolved proteins were transferred to a Hybond-C membrane (Amersham) using semi-dry equipment (Bio-Rad). The membranes were cut into strips and saturated by incubation for 1 h at room temperature with blocking buffer (3 % skimmed milk–0±05 % Tween 20 in PBS). After blocking, filters were incubated for more than 2 h at room temperature with the purified MAbs diluted at 1 µg}ml in blocking buffer. After several washes with 0±05 % Tween 20 in PBS, bound antibodies were detected by alkaline phosphatase-conjugated anti-mouse IgG (Sigma) diluted 1 : 2500 in blocking buffer. For colour development, nitroblue tetrazolium chloride (Gibco–BRL) and bromochloroindoyl phosphate (Pierce) were used as substrates. When RHDV-infected and VP60-immunized rabbit sera, diluted 1 : 1000, were used for immunoblotting analyses, the binding of the antibodies was detected by incubation with horseradish peroxidase (HRP)-labelled protein A (Sigma) at a dilution of 1 : 500 in blocking buffer. 4-Chloro-1-naphthol (0±5 mg}ml ; Sigma), 17 % (v}v) methanol and 0±015 % hydrogen peroxide in PBS were used for colour development. + ELISA. The reactivity of the rabbit sera and the mouse MAbs with VP60 and truncated fragments was determined by indirect ELISA. Polystyrene microtitre plates (Labsystem) were coated with 0±5 µg of the different antigens, RHDV virions, VLPs or E. coli-derived VP60 fragments overnight at 4 °C in 0±05 M carbonate buffer (pH 9±6). Washes between consecutive steps were performed with 0±05 % Tween 20 in PBS. Plates were incubated with the rabbit sera or MAbs diluted in PBS containing 0±35 M NaCl, 0±05 % Tween 20 for 1 h at 37 °C. HRP-labelled protein A (Sigma) or anti-mouse IgG (Pierce) diluted 1 : 1000 in blocking buffer were
Antigenic structure of RHDV
Fig. 1. Schematic diagram of VP60 fragments expressed in E. coli. VP60 fragments were constructed as described in Methods using convenient restriction enzyme sites and primers for PCR. The RHDV VP60 gene is depicted by a darker line, with the initiation site (ATG) and the restriction enzyme sites used in the fragmentation process. The lines with the restriction enzyme sites (italic) or primers for PCR (bold) at the ends represent the location and length of fragments. Names (A–L) and sizes are indicated above each fragment.
used as conjugates. After incubation for 1 h at 37 °C, bound antibodies were detected by adding ABTS [2,2«-azino-di(ethyl-benzothiazoline-6sulfonic acid) ; Sigma] as substrate. The reaction was stopped with 2 % SDS for 15 min and the absorbance was measured at 405 nm in an ELISA reader (Bio-Tek Instruments). Double antibody sandwich (DAS)–ELISA was used to determine the accessibility of the epitopes. Briefly, 96-well microtitre plates were coated with 0±5 µg purified rabbit IgG anti-rRHDV. The plates were washed four times with 0±05 % Tween 20 in PBS and incubated with 0±5 µg rRHDV particles in PBS containing 0±35 M NaCl, 0±05 % Tween 20 (dilution buffer) for 1 h at 37 °C. Serial dilutions of each MAb starting at 10 µg purified IgG in dilution buffer were added to each well and incubated for 1 h at 37 °C. After washing, plates were incubated for 1 h at room temperature with peroxidase-labelled goat anti-mouse Ig (Pierce) diluted 1 : 2000 in dilution buffer. Peroxidase activity was detected as above. + Competitive antibody binding. Microtitre plates were coated overnight at 4 °C with 0±25 µg per well purified rRHDV in 0±05 M carbonate buffer (pH 9±6). After washing with 0±05 % Tween 20 in PBS, plates were incubated for 1 h at 37 °C with the corresponding HRPconjugated MAb at a predetermined optimal concentration and mixed with twofold serial dilutions of each of the unlabelled purified MAbs. MAbs were labelled with peroxidase according to the method of Nakane & Kawaoi (1974). Peroxidase activity was detected by adding ABTS as substrate. The reaction was stopped with 2 % SDS and the absorbance of each well was determined at 405 nm. Grouping of MAbs 1E12 and 2A4 was based on additional data described below.
Results Production and general properties of RHDV-specific MAbs
A total of 11 anti-RHDV MAbs were obtained from two different fusions. The specificity, isotype and ELISA titres are shown in Table 1. Except for MAb 2A7, which had isotype IgG2b, the most frequent isotype was IgG1. The MAb titre determined by ELISA ranged between 10& and 10(, except for MAbs 1E12 and 2A4 which gave titres of 10%. The specificities were determined by ELISA with empty RHDV capsids and semi-purified RHDV virions, and by immunoblot analysis. In the ELISA, all of the MAbs reacted with RHDV virions and VLPs. In immunoblotting, most of the MAbs reacted vigorously with RHDV VP60 and RHDV virions, except for MAbs 1H6, 2F5, 1E12 and 2A4, which reacted weakly, and 2C1, which failed to react (Table 1). MAb 2C1 recognized only the proteins not previously denatured by boiling. To test the accessibility of the epitopes on the surface of the virus, DAS–ELISA was carried out using a rabbit IgG as capture antibody. The results indicated that all of the MAbs reacted with the intact particles (data not shown), suggesting that the binding epitopes are exposed on the surface of the particles.
BJAD
J. L. Martı! nez-Torrecuadrada and others
Table 1. Characterization of MAbs to RHDV VP60 Reactivity by Western blot† Hybridoma 3D6 1H6 2E11 2F5 1A2 1H7 2D6 2A7 2C1 1E12 2A4
Isotype
IgG titre*
VP60
36 kDa fragment
IgG " IgG " IgG " IgG " IgG " IgG " IgG " IgG b # IgG " IgG " IgG "
10( 10( 10( 10' 10& 10( 10( 10' 10' 10% 10%
®
® ®
* Calculated at a concentration of 1 mg}ml. ELISA titre is expressed as the reciprocal of the dilution that gave an absorbance greater than the cut-off value [cut-off value was taken as the mean³(2¬SD of the negative control MAb values)]. † ®, Negative reaction ; , positive reaction (, strongest ; , weakest).
Fig. 2. Antigenic sites on RHDV VP60 defined by MAbs in competition ELISAs. Competition between peroxidase-labelled MAb and the competitor unlabelled MAb is indicated by filled squares. MAbs 1E12 and 2A4 lost their reactivities after peroxidase labelling and could not be included in these experiments as tracers. The proposed antigenic groups are indicated on the right.
A truncated form of 36 kDa always appears in RHDV preparations. Most of the MAbs that were positive by immunoblotting reacted with both forms, VP60 and the 36 kDa fragment, except for MAb 2A7, which recognized exclusively the VP60 form. Competitive binding assay
To study the interaction between these MAbs and their ability to compete for binding to RHDV VP60, a competitive binding assay was carried out. The observed competition values were used to construct a tentative epitope map of the protein (Fig. 2). Based on reciprocal competition, MAbs could
BJAE
Fig. 3. Expression of the VP60 fragments in E. coli. Bacteria containing the recombinant pET3- or pET3x-derived expression vectors were induced for 3 h with 0±4 mM IPTG. The resultant fusion proteins were partially purified by solubilization with 4 M guanidinium hydrochloride and boiled for 5 min with loading buffer. Proteins were resolved on 11 % SDS–polyacrylamide gels and visualized by staining with Coomassie blue. Each lane represents bacterial lysates overexpressing the VP60 fragments indicated at the top of the figure. Molecular mass marker sizes are given in kDa on the lefthand side, and positions of fusion proteins are also indicated with arrowheads.
be placed in five groups, defining at least five different antigenic sites. Group I consisted of MAbs 1A2, 1H7, 2E11, 2F5, 3D6 and 1H6, which showed a complete reciprocal competition, suggesting that the corresponding epitopes were very close or overlapping. No competition was observed among MAbs 2D6, 2A7 and 2C1 or between these MAbs and group I MAbs ; thus, each one defined a different antigenic site (II, III and IV, respectively). MAbs 1E12 and 2A7 lost their binding ability after labelling. However, they did not compete with any other MAb when used as competitors, and formed a separate group, site V, based on the same reactivity pattern shown by these MAbs in different assays. In summary, the results indicated the existence of at least five different antigenic sites in RHDV VP60. The next step was to physically map the sequences recognized by these MAbs. Construction of RHDV VP60 truncated forms
Twelve pET-derived recombinant plasmids were constructed containing overlapping fragments of the RHDV VP60 with various lengths (87–1654 bp), as shown in Fig. 1. They were designated pET3xc-∆A, pET3b-∆B, pET3b-∆C, pET3b∆D, pET3b-∆E, pET3b-∆F, pET3c-∆G, pET3xa-∆H, pET3b-∆I, pET3b-∆J, pET3b-∆K and pET3b-∆L. Production of the RHDV VP60 fragments in E. coli was analysed by Coomassie blue staining of SDS–polyacrylamide gels (Fig. 3). The levels of expression were high in every case, ranging from 0±2 mg}ml of culture for fragments A, G, H, I and K to 0±05 mg}ml of culture for fragment J, as estimated by visual comparison with known quantities of BSA in SDS– polyacrylamide gels. The expression of large VP60 fragments B and D in E. coli led to the production of abundant proteolytic fragments. The identity of the proteins as RHDV-derived
Antigenic structure of RHDV
Table 2. Reactivity of RHDV-specific antibodies with VP60 fragments by immunoblotting ®, No reaction ; , positive reaction (, maximum reaction ; , slight reaction). MAb group* Fragment A B C D E F G H I J K L
Rabbit serum
I
II
III
IV
V
α-RHDV†
α-VP60‡
® ® ® ® ® ® ® ®
® ® ® ® ® ® ® ®
® ® ® ® ® ® ® ® ®
® ® ® ® ® ® ® ® ®
® ® ® ® ® ® ® ® ®
® ® ®
* MAbs were used as purified immunoglobulins at a concentration of 1 µg}ml. † Serum from RHDV-infected rabbit at a 1 : 500 dilution. ‡ Serum raised against baculovirus-derived VP60 at a 1 : 500 dilution.
(a)
(b)
Fig. 4. Reactivity of fragments L (a) and H (b) with anti-VP60 MAbs by Western blot analysis. Bacterial lysates overexpressing fragments L or H were prepared, mixed with protein sample buffer and loaded onto 11 % SDS–polyacrylamide gels. Fractionated proteins were transferred to nitrocellulose filter, cut into strips and incubated with every MAb as described in Methods. Molecular mass markers are in kDa on the left-hand side of each panel. Arrows indicate migration of fragments L and H.
fragments was confirmed by immunoblotting analysis using specific rabbit sera. All the fragments reacted with a polyclonal rabbit sera anti-recombinant VP60 (Table 2). Fragments A, C and F of the VP60 N terminus were not recognized by sera from RHDV-infected rabbits. Reactivity of RHDV-specific antibodies with the capsid protein fragments by immunoblotting
To accomplish the epitope mapping, the reactivity of different fragments of the VP60 was probed with a panel of
rabbit sera and mouse MAbs by immunoblotting (Table 2). Group I and II MAbs showed a similar recognition pattern, reacting strongly with fragments B, E, I and L, except for MAbs 1H6 and 2F5, which showed a weak reactivity. MAbs 1H6 and 2F5 could be recognizing the same epitope in a different orientation that is more sensitive to denaturation. Fragment L (659 bp) was the shortest fragment recognized (Fig. 4 a) ; it is located between residues 31 and 250 of VP60. Fragments shorter than L, such as F (aa 31–189) or C (aa 31–176), were not recognized, hindering a better definition of the epitope. MAb
BJAF
J. L. Martı! nez-Torrecuadrada and others
Table 3. Reactivity of RHDV VP60 fragments determined by ELISA VP60-specific MAbs, RHDV-infected rabbit antiserum and VP60-specific rabbit antiserum were screened by ELISA for antibodies against RHDV VP60 fragments at a 1 : 200 dilution of serum or 50 ng purified IgG. The strength of the reaction was expressed according to absorbance values at 405 nm (®, below cut-off ; , between cut-off value and 0±5 ; , 0±5–1±5 ; , 1±5–2±5 ; , " 2±5). Baculovirus-derived VP60 was used as a positive control. MAb group Fragment A B C D E F G H I J K L VP60
Rabbit serum
I
II
III
IV
V
α-RHDV
α-VP60
® ® ® ® ® ® ® ®
® ® ® ® ® ® ® ®
® ® ® ® ® ® ® ® ®
® ® ® ® ® ® ® ®
® ® ® ® ® ® ® ®
® ® ®
2A7 (group III) reacted with fragments B, D and H. Fragment H comprises the last 101 residues of the C terminus of VP60, between residues 477 and 579 (Fig. 4 b). Surprisingly, MAb 2C1 (site IV), which did not recognize the complete VP60, was able to react, albeit at low levels, with fragments E, I and L. Group V MAbs displayed a similar reactivity pattern to group IV by immunoblotting. They reacted better with shorter fragments such as I and L than with the larger fragments. Characterization of RHDV epitopes by ELISA
To study whether the reactivity of the different groups of MAbs with the different fragments was sensitive to denaturation, an indirect ELISA was carried out (Table 3). ELISA results showed that the reactivity of group I and II MAbs with the fragments was similar to that of 2C1 (group IV), indicating that the binding of MAb 2C1 occurs within the same region of RHDV VP60, but the recognition site was more susceptible to denaturation than the epitope(s) recognized by groups I and II. Group V MAbs were characterized by a generally weaker reaction and did not react (MAb 2A4), or did so very weakly (MAb 1E12), with the same fragments recognized by groups I, II and IV. Results obtained with MAb 2A7 (group III) confirmed the same recognition site previously found by immunoblotting. The results obtained with the MAbs are in good agreement with those obtained with a specific rabbit anti-RHDV sera (Table 3), indicating the biological relevance of these regions in the recognition of the virus. In summary, RHDV epitopes seem to be concentrated in two regions, defined by fragments L and H.
BJAG
Discussion Based on the reactivity of rRHDV VP60 and truncated fragments with a collection of mouse MAbs, the antigenic structure of the RHDV VP60 capsid protein appears to be dominated by a large discontinuous antigenic domain (Fig. 5). Most of the MAbs recognized this site (fragment L) comprising 219 residues close to the N terminus of the molecule. This immunodominant region probably includes several stretches of residues brought together by folding of the polypeptide chain. There are several subsites inside this large recognition site, as several groups of MAbs exhibit different recognition properties and different degrees of competition. These characteristics, contiguity in the binding site and different degrees of competition, were already shown by group II RHDV-specific MAbs reported by Capucci et al. (1995). In addition, the epitopes are conformation-dependent, as the MAbs recognize fragments I and L, but not fragments C, F, J and K. Based on this reactivity, the recognition site in fragment L would involve residues 31–100 and residues 189–250, although we cannot discount the possibility that one of the regions is necessary only for folding of the protein in the required conformation. The fact that some MAbs are able to react by immunoblotting, despite the fact that they recognize discontinuous epitopes, is not unusual. Similar observations have been made, for example, in other viruses (Lo! pez de Turiso et al., 1991) and in bacterial toxins (Bartoloni et al., 1988). A puzzling observation is the fact that, in Western blot analysis, antibodies from groups IV and V recognize fragments
Antigenic structure of RHDV
Fig. 5. Diagram of RHDV antigenic domains. The full length of VP60 capsid protein is represented by the black box at the top. The putative position of the 36 kDa cleavage product is shown below as a stippled box, since the exact location of this fragment has not been determined. The antigenic regions L and H are indicated as open boxes. Reacting MAbs are shown below the corresponding fragment (corresponding antigenic groups are indicated between parentheses). Numbers on the top of VP60 and the antigenic fragments indicate the first and last residues.
E, I and L, but not B, which includes the former three fragments. The explanation for this is probably in the observed ELISA reactivity : the MAbs react with fragment B by ELISA and not by immunoblotting, indicating that the epitopes are susceptible to denaturation. Thus, we can speculate that correct folding of larger polypeptide fragments expressed in E. coli is more difficult than folding of smaller fragments. Therefore, if the epitope is conformation-dependent, it is probably easier to regain a proper conformation in a small fragment than in a large fragment. The observation that rabbit polyclonal sera do not react, or react very weakly, with the N-terminal fragments A, C and F confirmed the lack of epitopes in this area of VP60. This coincidence between the mouse MAbs and the rabbit sera confirms the relevance of these results in reflecting the situation in the natural host, despite the known limitations of the inbred mouse system. Notwithstanding the considerable homology between RHDV and European brown hare syndrome virus (EBHSV), comparative sequence analysis of residues 37–50 has revealed that this N-terminal region is quite variable among RHDV and EBHSV (Wirblich et al., 1994). Some variability also occurs among residues 220–250. These two regions are involved in epitope recognition by the collection of MAbs. Phylogenetic analysis of the caliciviruses shows that RHDV and EBHSV are located on a branch apart from the rest of the caliciviruses (Berke et al., 1997), being equally distant from the FCV group and Norwalk-like viruses. Recent reports have proposed classifying RHDV and EBHSV as two serotypes of a single serogroup within the Caliciviridae due to the antigenic crossreactivity among them (Laurent et al., 1997). In this context, it is not surprising that the antigenic structure of RHDV should be quite different to that of FCV and NV. So, while in FCV and NV the antigenic sites are concentrated in the C-terminal region of the protein (Milton et al., 1992 ; Hardy et al., 1996 ; Tohya et al., 1997), the most immunogenic domain of RHDV is close to the N-terminal region. It should be noted that no MAbs were obtained against the most variable region of the capsid protein (residues 407–434), although MAb 2A7
partially overlaps the hypervariable E region described for FCV (residues 426–520) (Seal et al., 1993). Previous reports grouped RHDV epitopes in three different types, based on Western blot analysis and}or ELISA reactivity : surface linear, surface conformational and internal linear epitopes (Laurent et al., 1997 ; Capucci et al., 1995). Our MAbs could fit in any of the first two types. By DAS–ELISA, they seem to be localized on the surface of the particles, although they also react with the VP60 by immunoblotting. In Western blot analysis, only MAb 2C1 does not react with the RHDV capsids. Since the 2C1 epitope was finally mapped in the same domain as the other MAbs, this result suggests that the reactivity by Western blot or ELISA is not enough to determine whether the epitopes are linear or conformation-dependent. It might be that some of the MAbs previously assigned as reacting with linear epitopes (Laurent et al., 1997) could be reacting with conformation-dependent or discontinuous epitopes. Since most of the RHDV-neutralizing MAbs bind to the surface of the particle (Laurent et al., 1997 ; Thouvenin et al., 1997), while non-neutralizing cross-reactive epitopes are buried inside the virion, there exists a large probability that some of our MAbs may be neutralizing. This possibility will be explored in further experiments in vivo. Fragment L, the smallest fragment recognized by most of the MAbs, would correspond to the proposed shell (S) domain of the capsid for NV (Prasad et al., 1994). Prasad et al. suggest that the N-terminal 250 residues of the capsid protein form the S domain, which is buried and should therefore be less antigenic. However, the antigenic regions described in this report appear to be on the surface of the particle, as the MAbs are able to react with intact RHDV particles, either by indirect ELISA or DAS–ELISA. It should be noted, however, that good exposure on the surface of the capsids does not always correlate with higher immunogenicity. For instance, in African horsesickness virus, the VP7 protein of the inner shell of the virion is the most immunogenic protein of the virus (Chuma et al., 1992). The same situation occurs in several retroviruses, where the internal protein complex gag is the most immuno-
BJAH
J. L. Martı! nez-Torrecuadrada and others
dominant antigen (Reid et al., 1991). Therefore, both possibilities could be feasible. In any case, the real situation will not be clarified until the 3D structure of RHDV and the precise position of the residues can be determined. An important characteristic of caliciviruses is the presence of large amounts of soluble protein derived from the proteolytic digestion of the capsid protein (i.e. a 32 kDa fragment in NV ; Hardy et al., 1995). In the case of RHDV, this soluble protein is also abundant. It is the only component of the core-like particles (25–27 nm), which are detected in a number of rabbits that have died with typical RHD symptoms (Granzow et al., 1996). This truncated fragment was recognized by the MAbs with different efficiency. Most of the MAbs reacted with the 36 kDa band, confirming that it is a C-terminal truncated fragment of VP60. In contrast, MAb 2A7 (group III) reacted preferentially with the VP60 and weakly with a less abundant proteolytic fragment, which migrates slightly slower than the 36 kDa fragment and would represent the C-terminal half of the molecule. Thus, (i) in contrast to NV, the RHDV major proteolytic fragment appears to be the N-terminal part ; and (ii) similarly to NV, the cleavage fragment appears to contain the most antigenic region of the capsid. Remarkably, the presence of core-like particles is associated with the occurrence of sub-acute cases of the disease, where rabbit mortality is much lower. Whether this reduced mortality may be due to the higher immunogenicity of the core-like particles inducing more antibodies is not known and further studies are necessary. We thank Juan Plana-Dura! n for providing anti-RHDV rabbit antisera from vaccinated and infected animals. This report is the result of an integrated European collaboration supported in part by the EU, FAIR programme, contract no. FAIR-CT95-0720.
References Bartoloni, A., Pizza, M., Bigio, M., Nucci, D., Ashworth, L. A., Irons, L. I., Robinson, A., Burns, D., Manclark, C., Sato, H. & Rappuoli, R. (1988). Mapping of a protective epitope of Pertussis toxin by in vitro
refolding of recombinant fragments. Bio}Technology 6, 709–712. Berke, T., Golding, B., Jiang, X., Cubitt, D., Wolfaardt, M., Smith, A. & Matson, D. O. (1997). Phylogenetic analysis of the caliciviruses. Journal
of Medical Virology 52, 419–424. Capucci, L., Frigoli, G., Ronshold, L., Lavazza, A., Brocchi, E. & Rossi, C. (1995). Antigenicity of the rabbit hemorrhagic disease virus studied
by its reactivity with monoclonal antibodies. Virus Research 37, 221–238. Chuma, T., Le Blois, H., Sa! nchez-Vizcaı! no, J. M., Dı! az-Laviada, M. & Roy, P. (1992). Expression of the major core antigen VP7 of African
horsesickness virus by a recombinant baculovirus and its use as a groupspecific diagnostic reagent. Journal of General Virology 73, 925–931. Clarke, I. N. & Lambden, P. R. (1997). The molecular biology of caliciviruses. Journal of General Virology 78, 291–301. Granzow, H., Weiland, F., Strebelow, H.-G., Liu, C. M. & Schirrmeier, H. (1996). Rabbit hemorrhagic disease virus (RHDV) : ultrastructure and
biochemical studies of typical and core-like particles present in liver homogenates. Virus Research 41, 163–172.
BJAI
Hardy, M. E., White, L. J., Ball, J. M. & Estes, M. K. (1995). Specific
proteolytic cleavage of recombinant Norwalk virus capsid protein. Journal of Virology 69, 1693–1698. Hardy, M. E., Tanaka, T. N., Kitamoto, N., White, L., Ball, J. M., Jiang, X. & Estes, M. (1996). Antigenic mapping of the recombinant Norwalk
virus capsid protein using monoclonal antibodies. Virology 217, 252–261. Laurent, S., Vautherot, J. F., Madelaine, M.-F., Le Gall, G. & Rasschaert, D. (1994). Recombinant rabbit hemorrhagic disease virus capsid protein
expressed in baculovirus self-assembles into virus-like particles and induces protection. Journal of Virology 68, 6794–6798. Laurent, S., Vautherot, J. F., Le Gall, G., Madelaine, M.-F. & Rasschaert, D. (1997). Structural, antigenic and immunogenic relationships between
European brown hare syndrome virus and rabbit haemorrhagic disease virus. Journal of General Virology 78, 2803–2811. Liu, S. J., Xue, H. P., Pu, B. Q. & Quian, N. H. (1984). A new viral disease in rabbits. Animal Husbandry and Veterinary Medicine 16, 253–255. Lo! pez de Turiso, J. A., Corte! s, E., Garcı! a, J., Ranz, A., Sanz, A., Vela, C. & Casal, I. (1991). Fine mapping of canine parvovirus B cell epitopes. Journal of General Virology 72, 2445–2456. Martı! nez-Torrecuadrada, J. L. & Casal, J. I. (1995). Identification of a linear neutralization domain in the protein VP2 of African horsesickness virus. Virology 210, 391–399. Meyers, G., Wirblich, C. & Thiel, H.-J. (1991). Rabbit hemorrhagic disease virus : molecular cloning and nucleotide sequence of a calicivirus genome. Virology 184, 664–676. Milton, I. D., Turner, J., Teelan, A., Gaskell, R., Turner, P. C. & Carter, M. J. (1992). Location of monoclonal antibody binding sites in the
capsid protein of feline calicivirus. Journal of General Virology 73, 2435–2439. Nakane, P. K. & Kawaoi, A. (1974). Peroxidase-labeled antibody : a new method of conjugation. Journal of Histochemistry and Cytochemistry 22, 1084–1091. Ohlinger, V. F., Haas, B., Ahl, R. & Weiland, F. (1989). Die infektio$ se ha$ morrhagische Krankheit des Kaninchen – eine durch ein Calicivirus verursachte Tierseuche. Tieraerztliche Umschau 44, 284–294. Ohlinger, V. F., Haas, B., Meyer, G., Weiland, F. & Thiel, H.-J. (1990).
Identification and characterization of the virus causing rabbit hemorrhagic disease. Journal of Virology 64, 3331–3336. Plana-Dura! n, J., Bastons, M., Rodriguez, M. J., Climent, I., Corte! s, E., Vela, C. & Casal, I. (1996). Oral immunization of rabbits with VP60 particles confers protection against rabbit hemorrhagic disease. Archives of Virology 141, 1423–1436. Prasad, B. V. V., Rothnagel, R., Jiang, X. & Estes, M. K. (1994). Three dimensional structure of baculovirus-expressed Norwalk virus capsids. Journal of Virology 68, 5117–5125. Reid, G., Rigby, M. A., McDonald, M., Hosie, M. J., Neil, J. C. & Jarret, O. (1991). Immunodiagnosis of feline immunodeficiency virus infection
using recombinant viral p17 and p24. AIDS 5, 1477–1483. Sambrook, J., Fritsch, T. J. & Maniatis, T. (1989). Molecular Cloning : A Laboratory Manual, 2nd edn. Cold Spring Harbor, NY : Cold Spring Harbor Laboratory. Sanz, A., Garcia-Barreno, B., Nogal, M. L., Vin4 uela, E. & Enjuanes, L. (1985). Monoclonal antibodies specific for African swine fever virus proteins. Journal of Virology 54, 199–206. Seal, B. S., Ridpath, J. F. & Mengeling, W. L. (1993). Analysis of feline calicivirus capsid protein genes : identification of variable antigenic determinant regions of the protein. Journal of General Virology 74, 2519–2524. Studier, F. W., Rosenberg, A. H., Dunn, J. J. & Dubendorff, J. W.
Antigenic structure of RHDV (1990). Use of T7 RNA polymerase to direct expression of cloned genes.
(1997). Mapping of antigenic sites involved in neutralization on the
Methods in Enzymology 185, 60–89. Summers, M. D. & Smith, G. E. (1987). A Manual of Methods for
capsid protein of feline calicivirus. Journal of General Virology 78, 303–305.
Baculovirus Vectors and Insect Cell Culture Procedures. Texas Agricultural Experiment Station Bulletin no. 1555.
Wirblich, C., Meyers, G., Ohlinger, V. F., Capucci, L., Eskens, U., Haas, B. & Thiel, H.-J. (1994). European brown hare syndrome virus :
Thouvenin, E., Laurent, S., Madelaine, M.-F., Rasschaert, D., Vautherot, J.-F. & Hewat, E. A. (1997). Bivalent binding of a neutralising
relationship to rabbit hemorrhagic disease virus and other caliciviruses. Journal of Virology 68, 5164–5173.
antibody to a calicivirus involves the tortional flexibility of the antibody hinge. Journal of Molecular Biology 270, 238–246. Tohya, Y., Yokohama, N., Maeda, K., Kawaguchi, Y. & Mikami, T.
Received 19 November 1997 ; Accepted 17 March 1998
BJAJ
Antigenic structure of the capsid protein of rabbit haemorrhagic disease virus Jorge L. Martı! nez-Torrecuadrada,1 Elena Corte! s,1 Carmen Vela,1 Jan P. M. Langeveld,2 Rob H. Meloen,2 Kristian Dalsgaard,3 William D. O. Hamilton4 and J. Ignacio Casal1 Inmunologı! a y Gene! tica Aplicada SA (INGENASA), Hnos Garcia Noblejas 41, 28037 Madrid, Spain ID-DLO, PO Box 65, NL-8200 Lelystad, The Netherlands 3 Danish Veterinary Institute for Virus Research, DK-4771 Kalvehave, Denmark 4 Axis Genetics plc, Babraham, Cambridge CB2 4AZ, UK 1 2
Rabbit haemorrhagic disease virus (RHDV) causes an important disease in rabbits. The virus capsid is composed of a single 60 kDa protein. The capsid protein gene was cloned in Escherichia coli using the pET3 system, and the antigenic structure of RHDV VP60 was dissected using 11 monoclonal antibodies (MAbs) and 12 overlapping fragments of the protein expressed in E. coli. Two antigenic regions were found. Ten out of the 11 MAbs recognized different discontinuous epitopes in the most immunodominant region of the viral capsid. This
Introduction Rabbit haemorrhagic disease (RHD) is a viral disease of the species Oryctolagus cuniculus and was first described by Liu et al. (1984). The disease, characterized by high mortality in adult rabbits (Liu et al., 1984 ; Ohlinger et al., 1989), is caused by rabbit haemorrhagic disease virus (RHDV), a member of the family Caliciviridae (Ohlinger et al., 1990). The family Caliciviridae comprises several groups of nonenveloped viruses which share many biological properties (for a review see Clarke & Lambden, 1997). Among these properties is the presence of cup-like depressions (calyces) in the surface of the virus. However, this property can no longer be regarded as a general feature of caliciviruses as many of them present an ill-defined surface morphology. Thus, antigenically wellcharacterized caliciviruses such as feline calicivirus (FCV) and human Norwalk virus (NV) present a fuzzy, non-distinctive surface morphology. In contrast, RHDV shows a classic ‘ Star of David ’ calicivirus morphology. The 3D structure of RHDV has not been determined. Analysis of the recombinant NV particles by electron cryoAuthor for correspondence : Ignacio Casal. Fax 34 1 408 75 98. e-mail INGENASA!alc.es
0001-5335 # 1998 SGM
domain was located between residues 31 and 250 of the VP60 N terminus. The other MAb revealed the presence of an antigenic site within 102 aa of the C terminus. This MAb did not recognize the major cleavage product of the full-length 60 kDa protein. These results indicate that, in contrast to other caliciviruses such as Norwalk virus (NV), the 36 kDa cleavage product probably forms the N-terminal region of VP60. However, as in NV, the cleavage region appears to be the most immunodominant region.
microscopy and computer image processing techniques has shown that the capsid is composed of 90 dimers of the capsid protein that form a shell domain from which arch-like capsomers protrude (Prasad et al., 1994). Based on sequence similarities with other T ¯ 3 RNA viruses, it has been predicted that the N-terminal amino acids (aa) 1–250 form the shell domain and the remaining C-terminal residues make up the arches (Prasad et al., 1994). The RHDV genome consists of a single-stranded RNA molecule of positive polarity, approximately 7±5 kb in size. The genome organization shows one long ORF that represents 86 % of the viral genome, with the VPg located downstream of the 3C-like protease gene, in the 5« region (Meyers et al., 1991). The capsid protein of RHDV has an apparent molecular mass of 60 kDa. Expression of a cDNA that encodes RHDV VP60 in insect cells infected with a recombinant baculovirus results in spontaneous assembly of the protein into empty recombinant RHDV (rRHDV) virus-like particles (VLPs) (Laurent et al., 1994 ; Plana-Dura! n et al., 1996). These VLPs are morphologically and antigenically similar to native RHDV particles. The immunogenicity of these particles is extremely high ; doses as low as 0±5 µg are able to fully protect rabbits by the systemic route. Since RHDV cannot be propagated in vitro, the availability of large amounts of rRHDV VLPs has facilitated
BJAB
J. L. Martı! nez-Torrecuadrada and others
the production of monoclonal antibodies (MAbs) useful for antigenic characterization of the virus. Capucci et al. (1995) and Laurent et al. (1997) described the production and characterization of a panel of anti-RHDV MAbs. However, the MAb binding sites were not physically mapped in the capsid protein, and so the epitopes and their position on the RHDV capsid remain unknown. Antigenic mapping studies carried out in other caliciviruses such as NV (Hardy et al., 1996) and FCV (Milton et al., 1992 ; Seal et al., 1993 ; Tohya et al., 1997) have identified the presence of several epitopes in the extreme C-terminal regions of the capsid protein. The purpose of this study was to begin to dissect the antigenic and structural topography of the RHDV capsid using MAbs generated to rRHDV particles and VP60-truncated fragments expressed in E. coli. The antigenic map was compared to that obtained for prototype caliciviruses such as FCV and NV.
Methods + Viruses and cells. The RHDV Olot}89 strain used for these studies was isolated and propagated as described previously (PlanaDura! n et al., 1996). The Spodoptera frugiperda cell line Sf9 (ATCC g CRL 1711) was used to propagate recombinant baculovirus AcRHDV-710 and to produce rRHDV VP60 particles (Plana-Dura! n et al., 1996). Sf9 cells were grown in suspension at 27 °C in Grace’s medium supplemented with yeastolate and lactalbumin hydrolysate (Summers & Smith, 1987), containing 5 % foetal calf serum and antibiotics. + Production of antibodies. MAbs specific for rRHDV VP60 were prepared from BALB}c mice immunized with rRHDV particles, purified as previously described (Plana-Dura! n et al., 1996). Protocols for immunization and the preparation of MAbs have been described previously (Sanz et al., 1985 ; Lo! pez de Turiso et al., 1991). All hybridomas were cloned by limiting dilution at least three times. Hybridoma supernatants were screened by ELISA and immunoblotting analysis. The isotype of the MAbs was determined by ELISA using specific anti-mouse subtype antisera (Sigma). Polyclonal rabbit anti-RHDV sera were obtained from rabbits experimentally infected or vaccinated with an inactivated vaccine or recombinant VP60. + Cloning of VP60 gene in pET expression vectors : preparation of truncated forms. The pET3 and pET3x vectors (Studier et al., 1990) were used to construct fusion proteins between gene 10 protein and VP60 fragments. The expression vector pET3 was mainly used ; pET3x was chosen for those fragments smaller than 400 bp. The fusion protein consisted of 11 aa of the gene 10 protein for pET3 and 260 residues for pETx. Standard cloning techniques (Sambrook et al., 1989) using convenient restriction enzyme sites within the VP60 gene were applied to obtain in-frame fusion constructs between the gene 10–VP60. Fig. 1 shows the VP60 fragments obtained from the recombinant plasmid pAcRHDV-710 (Plana-Dura! n et al., 1996) by endonuclease digestion : A (nt 1–90, BamHI fragment) ; B (nt 91–1744, BamHI fragment) ; C (nt 91–527, BamHI to AflIII) ; D (nt 528–1744, AflIII to BamHI) ; E (nt 91–1429, BamHI to PvuII) ; F (nt 91–567, BamHI to SpeI) ; G (nt 568–1429, SpeI to PvuII) and H (nt 1429–1744, PvuII to BamHI).
BJAC
If appropriate restriction sites were not available, VP60 fragments were generated by PCR with Vent DNA polymerase (New England Biolabs). PCR comprised 25 cycles of denaturation at 94 °C for 1 min, primer annealing at 50 °C for 1 min, and extension at 72 °C for 1±5 min. Primers used were as follows (BamHI sites shown in italics) : RHDV-1, 5« CATGGATCCTGGCGTTGTGG 3« ; RHDV-2 : 5« GAAGGATCCATTCACAGCCGTGCTG 3« ; RHDV-3, 5« TGCGGATCCCAGTGAGGACTGGGGTC 3« ; RHDV-4, 5« AAAGGATCCTGGTGCAACCTGGGAG 3« ; RHDV-5, 5« GGTGGATCCAGACGGCTTTCCTGAT 3« ; and RHDV-6, 5« TGGGGATCCTGTGGCGTTGACGTCA 3«. The VP60 fragments generated by PCR were : I (nt 91–999, from primer RHDV-1 to RHDV-4) ; J (nt 301–1429, from RHDV-2 to RHDV-6) ; K (nt 997–1429, from RHDV-5 to RHDV-6) ; and L (nt 91–750, from RHDV1 to RHDV-3). Fragments with BamHI-compatible ends were ligated into BamHI-digested phosphatase-treated pET3 or pET3x (fragments A and H). Fragments lacking BamHI-compatible ends were first subcloned in a pMTL plasmid in order to generate BamHI-compatible ends. The ligation mixtures were used to transform XL-1 Blue or HMS174 competent cells. The resulting colonies were characterized by restriction mapping. The orientation of the insert and the junction sequences of the recombinant plasmids were sequenced by the dideoxynucleotide method. + Induction and purification of E. coli-derived VP60 fragments. The expression and purification of fusion proteins were performed as described previously (Martı! nez-Torrecuadrada & Casal, 1995). Briefly, purification of the different truncated forms of VP60 was carried out by solubilization of the cell pellet with 4 M guanidinium hydrochloride after lysis by sonication. This simple treatment allowed a quick purification of the fusion proteins (greater than 85 %) in a single step. These partially purified fusion proteins were used for epitope mapping. + SDS–polyacrylamide gels and immunoblotting assays. Purity of the fusion protein and protein concentration were estimated from Coomassie-stained SDS–polyacrylamide gels. Proteins were diluted in loading buffer and boiled for 5 min, or not boiled where indicated, before analysis by 11 % SDS–PAGE. The resolved proteins were transferred to a Hybond-C membrane (Amersham) using semi-dry equipment (Bio-Rad). The membranes were cut into strips and saturated by incubation for 1 h at room temperature with blocking buffer (3 % skimmed milk–0±05 % Tween 20 in PBS). After blocking, filters were incubated for more than 2 h at room temperature with the purified MAbs diluted at 1 µg}ml in blocking buffer. After several washes with 0±05 % Tween 20 in PBS, bound antibodies were detected by alkaline phosphatase-conjugated anti-mouse IgG (Sigma) diluted 1 : 2500 in blocking buffer. For colour development, nitroblue tetrazolium chloride (Gibco–BRL) and bromochloroindoyl phosphate (Pierce) were used as substrates. When RHDV-infected and VP60-immunized rabbit sera, diluted 1 : 1000, were used for immunoblotting analyses, the binding of the antibodies was detected by incubation with horseradish peroxidase (HRP)-labelled protein A (Sigma) at a dilution of 1 : 500 in blocking buffer. 4-Chloro-1-naphthol (0±5 mg}ml ; Sigma), 17 % (v}v) methanol and 0±015 % hydrogen peroxide in PBS were used for colour development. + ELISA. The reactivity of the rabbit sera and the mouse MAbs with VP60 and truncated fragments was determined by indirect ELISA. Polystyrene microtitre plates (Labsystem) were coated with 0±5 µg of the different antigens, RHDV virions, VLPs or E. coli-derived VP60 fragments overnight at 4 °C in 0±05 M carbonate buffer (pH 9±6). Washes between consecutive steps were performed with 0±05 % Tween 20 in PBS. Plates were incubated with the rabbit sera or MAbs diluted in PBS containing 0±35 M NaCl, 0±05 % Tween 20 for 1 h at 37 °C. HRP-labelled protein A (Sigma) or anti-mouse IgG (Pierce) diluted 1 : 1000 in blocking buffer were
Antigenic structure of RHDV
Fig. 1. Schematic diagram of VP60 fragments expressed in E. coli. VP60 fragments were constructed as described in Methods using convenient restriction enzyme sites and primers for PCR. The RHDV VP60 gene is depicted by a darker line, with the initiation site (ATG) and the restriction enzyme sites used in the fragmentation process. The lines with the restriction enzyme sites (italic) or primers for PCR (bold) at the ends represent the location and length of fragments. Names (A–L) and sizes are indicated above each fragment.
used as conjugates. After incubation for 1 h at 37 °C, bound antibodies were detected by adding ABTS [2,2«-azino-di(ethyl-benzothiazoline-6sulfonic acid) ; Sigma] as substrate. The reaction was stopped with 2 % SDS for 15 min and the absorbance was measured at 405 nm in an ELISA reader (Bio-Tek Instruments). Double antibody sandwich (DAS)–ELISA was used to determine the accessibility of the epitopes. Briefly, 96-well microtitre plates were coated with 0±5 µg purified rabbit IgG anti-rRHDV. The plates were washed four times with 0±05 % Tween 20 in PBS and incubated with 0±5 µg rRHDV particles in PBS containing 0±35 M NaCl, 0±05 % Tween 20 (dilution buffer) for 1 h at 37 °C. Serial dilutions of each MAb starting at 10 µg purified IgG in dilution buffer were added to each well and incubated for 1 h at 37 °C. After washing, plates were incubated for 1 h at room temperature with peroxidase-labelled goat anti-mouse Ig (Pierce) diluted 1 : 2000 in dilution buffer. Peroxidase activity was detected as above. + Competitive antibody binding. Microtitre plates were coated overnight at 4 °C with 0±25 µg per well purified rRHDV in 0±05 M carbonate buffer (pH 9±6). After washing with 0±05 % Tween 20 in PBS, plates were incubated for 1 h at 37 °C with the corresponding HRPconjugated MAb at a predetermined optimal concentration and mixed with twofold serial dilutions of each of the unlabelled purified MAbs. MAbs were labelled with peroxidase according to the method of Nakane & Kawaoi (1974). Peroxidase activity was detected by adding ABTS as substrate. The reaction was stopped with 2 % SDS and the absorbance of each well was determined at 405 nm. Grouping of MAbs 1E12 and 2A4 was based on additional data described below.
Results Production and general properties of RHDV-specific MAbs
A total of 11 anti-RHDV MAbs were obtained from two different fusions. The specificity, isotype and ELISA titres are shown in Table 1. Except for MAb 2A7, which had isotype IgG2b, the most frequent isotype was IgG1. The MAb titre determined by ELISA ranged between 10& and 10(, except for MAbs 1E12 and 2A4 which gave titres of 10%. The specificities were determined by ELISA with empty RHDV capsids and semi-purified RHDV virions, and by immunoblot analysis. In the ELISA, all of the MAbs reacted with RHDV virions and VLPs. In immunoblotting, most of the MAbs reacted vigorously with RHDV VP60 and RHDV virions, except for MAbs 1H6, 2F5, 1E12 and 2A4, which reacted weakly, and 2C1, which failed to react (Table 1). MAb 2C1 recognized only the proteins not previously denatured by boiling. To test the accessibility of the epitopes on the surface of the virus, DAS–ELISA was carried out using a rabbit IgG as capture antibody. The results indicated that all of the MAbs reacted with the intact particles (data not shown), suggesting that the binding epitopes are exposed on the surface of the particles.
BJAD
J. L. Martı! nez-Torrecuadrada and others
Table 1. Characterization of MAbs to RHDV VP60 Reactivity by Western blot† Hybridoma 3D6 1H6 2E11 2F5 1A2 1H7 2D6 2A7 2C1 1E12 2A4
Isotype
IgG titre*
VP60
36 kDa fragment
IgG " IgG " IgG " IgG " IgG " IgG " IgG " IgG b # IgG " IgG " IgG "
10( 10( 10( 10' 10& 10( 10( 10' 10' 10% 10%
®
® ®
* Calculated at a concentration of 1 mg}ml. ELISA titre is expressed as the reciprocal of the dilution that gave an absorbance greater than the cut-off value [cut-off value was taken as the mean³(2¬SD of the negative control MAb values)]. † ®, Negative reaction ; , positive reaction (, strongest ; , weakest).
Fig. 2. Antigenic sites on RHDV VP60 defined by MAbs in competition ELISAs. Competition between peroxidase-labelled MAb and the competitor unlabelled MAb is indicated by filled squares. MAbs 1E12 and 2A4 lost their reactivities after peroxidase labelling and could not be included in these experiments as tracers. The proposed antigenic groups are indicated on the right.
A truncated form of 36 kDa always appears in RHDV preparations. Most of the MAbs that were positive by immunoblotting reacted with both forms, VP60 and the 36 kDa fragment, except for MAb 2A7, which recognized exclusively the VP60 form. Competitive binding assay
To study the interaction between these MAbs and their ability to compete for binding to RHDV VP60, a competitive binding assay was carried out. The observed competition values were used to construct a tentative epitope map of the protein (Fig. 2). Based on reciprocal competition, MAbs could
BJAE
Fig. 3. Expression of the VP60 fragments in E. coli. Bacteria containing the recombinant pET3- or pET3x-derived expression vectors were induced for 3 h with 0±4 mM IPTG. The resultant fusion proteins were partially purified by solubilization with 4 M guanidinium hydrochloride and boiled for 5 min with loading buffer. Proteins were resolved on 11 % SDS–polyacrylamide gels and visualized by staining with Coomassie blue. Each lane represents bacterial lysates overexpressing the VP60 fragments indicated at the top of the figure. Molecular mass marker sizes are given in kDa on the lefthand side, and positions of fusion proteins are also indicated with arrowheads.
be placed in five groups, defining at least five different antigenic sites. Group I consisted of MAbs 1A2, 1H7, 2E11, 2F5, 3D6 and 1H6, which showed a complete reciprocal competition, suggesting that the corresponding epitopes were very close or overlapping. No competition was observed among MAbs 2D6, 2A7 and 2C1 or between these MAbs and group I MAbs ; thus, each one defined a different antigenic site (II, III and IV, respectively). MAbs 1E12 and 2A7 lost their binding ability after labelling. However, they did not compete with any other MAb when used as competitors, and formed a separate group, site V, based on the same reactivity pattern shown by these MAbs in different assays. In summary, the results indicated the existence of at least five different antigenic sites in RHDV VP60. The next step was to physically map the sequences recognized by these MAbs. Construction of RHDV VP60 truncated forms
Twelve pET-derived recombinant plasmids were constructed containing overlapping fragments of the RHDV VP60 with various lengths (87–1654 bp), as shown in Fig. 1. They were designated pET3xc-∆A, pET3b-∆B, pET3b-∆C, pET3b∆D, pET3b-∆E, pET3b-∆F, pET3c-∆G, pET3xa-∆H, pET3b-∆I, pET3b-∆J, pET3b-∆K and pET3b-∆L. Production of the RHDV VP60 fragments in E. coli was analysed by Coomassie blue staining of SDS–polyacrylamide gels (Fig. 3). The levels of expression were high in every case, ranging from 0±2 mg}ml of culture for fragments A, G, H, I and K to 0±05 mg}ml of culture for fragment J, as estimated by visual comparison with known quantities of BSA in SDS– polyacrylamide gels. The expression of large VP60 fragments B and D in E. coli led to the production of abundant proteolytic fragments. The identity of the proteins as RHDV-derived
Antigenic structure of RHDV
Table 2. Reactivity of RHDV-specific antibodies with VP60 fragments by immunoblotting ®, No reaction ; , positive reaction (, maximum reaction ; , slight reaction). MAb group* Fragment A B C D E F G H I J K L
Rabbit serum
I
II
III
IV
V
α-RHDV†
α-VP60‡
® ® ® ® ® ® ® ®
® ® ® ® ® ® ® ®
® ® ® ® ® ® ® ® ®
® ® ® ® ® ® ® ® ®
® ® ® ® ® ® ® ® ®
® ® ®
* MAbs were used as purified immunoglobulins at a concentration of 1 µg}ml. † Serum from RHDV-infected rabbit at a 1 : 500 dilution. ‡ Serum raised against baculovirus-derived VP60 at a 1 : 500 dilution.
(a)
(b)
Fig. 4. Reactivity of fragments L (a) and H (b) with anti-VP60 MAbs by Western blot analysis. Bacterial lysates overexpressing fragments L or H were prepared, mixed with protein sample buffer and loaded onto 11 % SDS–polyacrylamide gels. Fractionated proteins were transferred to nitrocellulose filter, cut into strips and incubated with every MAb as described in Methods. Molecular mass markers are in kDa on the left-hand side of each panel. Arrows indicate migration of fragments L and H.
fragments was confirmed by immunoblotting analysis using specific rabbit sera. All the fragments reacted with a polyclonal rabbit sera anti-recombinant VP60 (Table 2). Fragments A, C and F of the VP60 N terminus were not recognized by sera from RHDV-infected rabbits. Reactivity of RHDV-specific antibodies with the capsid protein fragments by immunoblotting
To accomplish the epitope mapping, the reactivity of different fragments of the VP60 was probed with a panel of
rabbit sera and mouse MAbs by immunoblotting (Table 2). Group I and II MAbs showed a similar recognition pattern, reacting strongly with fragments B, E, I and L, except for MAbs 1H6 and 2F5, which showed a weak reactivity. MAbs 1H6 and 2F5 could be recognizing the same epitope in a different orientation that is more sensitive to denaturation. Fragment L (659 bp) was the shortest fragment recognized (Fig. 4 a) ; it is located between residues 31 and 250 of VP60. Fragments shorter than L, such as F (aa 31–189) or C (aa 31–176), were not recognized, hindering a better definition of the epitope. MAb
BJAF
J. L. Martı! nez-Torrecuadrada and others
Table 3. Reactivity of RHDV VP60 fragments determined by ELISA VP60-specific MAbs, RHDV-infected rabbit antiserum and VP60-specific rabbit antiserum were screened by ELISA for antibodies against RHDV VP60 fragments at a 1 : 200 dilution of serum or 50 ng purified IgG. The strength of the reaction was expressed according to absorbance values at 405 nm (®, below cut-off ; , between cut-off value and 0±5 ; , 0±5–1±5 ; , 1±5–2±5 ; , " 2±5). Baculovirus-derived VP60 was used as a positive control. MAb group Fragment A B C D E F G H I J K L VP60
Rabbit serum
I
II
III
IV
V
α-RHDV
α-VP60
® ® ® ® ® ® ® ®
® ® ® ® ® ® ® ®
® ® ® ® ® ® ® ® ®
® ® ® ® ® ® ® ®
® ® ® ® ® ® ® ®
® ® ®
2A7 (group III) reacted with fragments B, D and H. Fragment H comprises the last 101 residues of the C terminus of VP60, between residues 477 and 579 (Fig. 4 b). Surprisingly, MAb 2C1 (site IV), which did not recognize the complete VP60, was able to react, albeit at low levels, with fragments E, I and L. Group V MAbs displayed a similar reactivity pattern to group IV by immunoblotting. They reacted better with shorter fragments such as I and L than with the larger fragments. Characterization of RHDV epitopes by ELISA
To study whether the reactivity of the different groups of MAbs with the different fragments was sensitive to denaturation, an indirect ELISA was carried out (Table 3). ELISA results showed that the reactivity of group I and II MAbs with the fragments was similar to that of 2C1 (group IV), indicating that the binding of MAb 2C1 occurs within the same region of RHDV VP60, but the recognition site was more susceptible to denaturation than the epitope(s) recognized by groups I and II. Group V MAbs were characterized by a generally weaker reaction and did not react (MAb 2A4), or did so very weakly (MAb 1E12), with the same fragments recognized by groups I, II and IV. Results obtained with MAb 2A7 (group III) confirmed the same recognition site previously found by immunoblotting. The results obtained with the MAbs are in good agreement with those obtained with a specific rabbit anti-RHDV sera (Table 3), indicating the biological relevance of these regions in the recognition of the virus. In summary, RHDV epitopes seem to be concentrated in two regions, defined by fragments L and H.
BJAG
Discussion Based on the reactivity of rRHDV VP60 and truncated fragments with a collection of mouse MAbs, the antigenic structure of the RHDV VP60 capsid protein appears to be dominated by a large discontinuous antigenic domain (Fig. 5). Most of the MAbs recognized this site (fragment L) comprising 219 residues close to the N terminus of the molecule. This immunodominant region probably includes several stretches of residues brought together by folding of the polypeptide chain. There are several subsites inside this large recognition site, as several groups of MAbs exhibit different recognition properties and different degrees of competition. These characteristics, contiguity in the binding site and different degrees of competition, were already shown by group II RHDV-specific MAbs reported by Capucci et al. (1995). In addition, the epitopes are conformation-dependent, as the MAbs recognize fragments I and L, but not fragments C, F, J and K. Based on this reactivity, the recognition site in fragment L would involve residues 31–100 and residues 189–250, although we cannot discount the possibility that one of the regions is necessary only for folding of the protein in the required conformation. The fact that some MAbs are able to react by immunoblotting, despite the fact that they recognize discontinuous epitopes, is not unusual. Similar observations have been made, for example, in other viruses (Lo! pez de Turiso et al., 1991) and in bacterial toxins (Bartoloni et al., 1988). A puzzling observation is the fact that, in Western blot analysis, antibodies from groups IV and V recognize fragments
Antigenic structure of RHDV
Fig. 5. Diagram of RHDV antigenic domains. The full length of VP60 capsid protein is represented by the black box at the top. The putative position of the 36 kDa cleavage product is shown below as a stippled box, since the exact location of this fragment has not been determined. The antigenic regions L and H are indicated as open boxes. Reacting MAbs are shown below the corresponding fragment (corresponding antigenic groups are indicated between parentheses). Numbers on the top of VP60 and the antigenic fragments indicate the first and last residues.
E, I and L, but not B, which includes the former three fragments. The explanation for this is probably in the observed ELISA reactivity : the MAbs react with fragment B by ELISA and not by immunoblotting, indicating that the epitopes are susceptible to denaturation. Thus, we can speculate that correct folding of larger polypeptide fragments expressed in E. coli is more difficult than folding of smaller fragments. Therefore, if the epitope is conformation-dependent, it is probably easier to regain a proper conformation in a small fragment than in a large fragment. The observation that rabbit polyclonal sera do not react, or react very weakly, with the N-terminal fragments A, C and F confirmed the lack of epitopes in this area of VP60. This coincidence between the mouse MAbs and the rabbit sera confirms the relevance of these results in reflecting the situation in the natural host, despite the known limitations of the inbred mouse system. Notwithstanding the considerable homology between RHDV and European brown hare syndrome virus (EBHSV), comparative sequence analysis of residues 37–50 has revealed that this N-terminal region is quite variable among RHDV and EBHSV (Wirblich et al., 1994). Some variability also occurs among residues 220–250. These two regions are involved in epitope recognition by the collection of MAbs. Phylogenetic analysis of the caliciviruses shows that RHDV and EBHSV are located on a branch apart from the rest of the caliciviruses (Berke et al., 1997), being equally distant from the FCV group and Norwalk-like viruses. Recent reports have proposed classifying RHDV and EBHSV as two serotypes of a single serogroup within the Caliciviridae due to the antigenic crossreactivity among them (Laurent et al., 1997). In this context, it is not surprising that the antigenic structure of RHDV should be quite different to that of FCV and NV. So, while in FCV and NV the antigenic sites are concentrated in the C-terminal region of the protein (Milton et al., 1992 ; Hardy et al., 1996 ; Tohya et al., 1997), the most immunogenic domain of RHDV is close to the N-terminal region. It should be noted that no MAbs were obtained against the most variable region of the capsid protein (residues 407–434), although MAb 2A7
partially overlaps the hypervariable E region described for FCV (residues 426–520) (Seal et al., 1993). Previous reports grouped RHDV epitopes in three different types, based on Western blot analysis and}or ELISA reactivity : surface linear, surface conformational and internal linear epitopes (Laurent et al., 1997 ; Capucci et al., 1995). Our MAbs could fit in any of the first two types. By DAS–ELISA, they seem to be localized on the surface of the particles, although they also react with the VP60 by immunoblotting. In Western blot analysis, only MAb 2C1 does not react with the RHDV capsids. Since the 2C1 epitope was finally mapped in the same domain as the other MAbs, this result suggests that the reactivity by Western blot or ELISA is not enough to determine whether the epitopes are linear or conformation-dependent. It might be that some of the MAbs previously assigned as reacting with linear epitopes (Laurent et al., 1997) could be reacting with conformation-dependent or discontinuous epitopes. Since most of the RHDV-neutralizing MAbs bind to the surface of the particle (Laurent et al., 1997 ; Thouvenin et al., 1997), while non-neutralizing cross-reactive epitopes are buried inside the virion, there exists a large probability that some of our MAbs may be neutralizing. This possibility will be explored in further experiments in vivo. Fragment L, the smallest fragment recognized by most of the MAbs, would correspond to the proposed shell (S) domain of the capsid for NV (Prasad et al., 1994). Prasad et al. suggest that the N-terminal 250 residues of the capsid protein form the S domain, which is buried and should therefore be less antigenic. However, the antigenic regions described in this report appear to be on the surface of the particle, as the MAbs are able to react with intact RHDV particles, either by indirect ELISA or DAS–ELISA. It should be noted, however, that good exposure on the surface of the capsids does not always correlate with higher immunogenicity. For instance, in African horsesickness virus, the VP7 protein of the inner shell of the virion is the most immunogenic protein of the virus (Chuma et al., 1992). The same situation occurs in several retroviruses, where the internal protein complex gag is the most immuno-
BJAH
J. L. Martı! nez-Torrecuadrada and others
dominant antigen (Reid et al., 1991). Therefore, both possibilities could be feasible. In any case, the real situation will not be clarified until the 3D structure of RHDV and the precise position of the residues can be determined. An important characteristic of caliciviruses is the presence of large amounts of soluble protein derived from the proteolytic digestion of the capsid protein (i.e. a 32 kDa fragment in NV ; Hardy et al., 1995). In the case of RHDV, this soluble protein is also abundant. It is the only component of the core-like particles (25–27 nm), which are detected in a number of rabbits that have died with typical RHD symptoms (Granzow et al., 1996). This truncated fragment was recognized by the MAbs with different efficiency. Most of the MAbs reacted with the 36 kDa band, confirming that it is a C-terminal truncated fragment of VP60. In contrast, MAb 2A7 (group III) reacted preferentially with the VP60 and weakly with a less abundant proteolytic fragment, which migrates slightly slower than the 36 kDa fragment and would represent the C-terminal half of the molecule. Thus, (i) in contrast to NV, the RHDV major proteolytic fragment appears to be the N-terminal part ; and (ii) similarly to NV, the cleavage fragment appears to contain the most antigenic region of the capsid. Remarkably, the presence of core-like particles is associated with the occurrence of sub-acute cases of the disease, where rabbit mortality is much lower. Whether this reduced mortality may be due to the higher immunogenicity of the core-like particles inducing more antibodies is not known and further studies are necessary. We thank Juan Plana-Dura! n for providing anti-RHDV rabbit antisera from vaccinated and infected animals. This report is the result of an integrated European collaboration supported in part by the EU, FAIR programme, contract no. FAIR-CT95-0720.
References Bartoloni, A., Pizza, M., Bigio, M., Nucci, D., Ashworth, L. A., Irons, L. I., Robinson, A., Burns, D., Manclark, C., Sato, H. & Rappuoli, R. (1988). Mapping of a protective epitope of Pertussis toxin by in vitro
refolding of recombinant fragments. Bio}Technology 6, 709–712. Berke, T., Golding, B., Jiang, X., Cubitt, D., Wolfaardt, M., Smith, A. & Matson, D. O. (1997). Phylogenetic analysis of the caliciviruses. Journal
of Medical Virology 52, 419–424. Capucci, L., Frigoli, G., Ronshold, L., Lavazza, A., Brocchi, E. & Rossi, C. (1995). Antigenicity of the rabbit hemorrhagic disease virus studied
by its reactivity with monoclonal antibodies. Virus Research 37, 221–238. Chuma, T., Le Blois, H., Sa! nchez-Vizcaı! no, J. M., Dı! az-Laviada, M. & Roy, P. (1992). Expression of the major core antigen VP7 of African
horsesickness virus by a recombinant baculovirus and its use as a groupspecific diagnostic reagent. Journal of General Virology 73, 925–931. Clarke, I. N. & Lambden, P. R. (1997). The molecular biology of caliciviruses. Journal of General Virology 78, 291–301. Granzow, H., Weiland, F., Strebelow, H.-G., Liu, C. M. & Schirrmeier, H. (1996). Rabbit hemorrhagic disease virus (RHDV) : ultrastructure and
biochemical studies of typical and core-like particles present in liver homogenates. Virus Research 41, 163–172.
BJAI
Hardy, M. E., White, L. J., Ball, J. M. & Estes, M. K. (1995). Specific
proteolytic cleavage of recombinant Norwalk virus capsid protein. Journal of Virology 69, 1693–1698. Hardy, M. E., Tanaka, T. N., Kitamoto, N., White, L., Ball, J. M., Jiang, X. & Estes, M. (1996). Antigenic mapping of the recombinant Norwalk
virus capsid protein using monoclonal antibodies. Virology 217, 252–261. Laurent, S., Vautherot, J. F., Madelaine, M.-F., Le Gall, G. & Rasschaert, D. (1994). Recombinant rabbit hemorrhagic disease virus capsid protein
expressed in baculovirus self-assembles into virus-like particles and induces protection. Journal of Virology 68, 6794–6798. Laurent, S., Vautherot, J. F., Le Gall, G., Madelaine, M.-F. & Rasschaert, D. (1997). Structural, antigenic and immunogenic relationships between
European brown hare syndrome virus and rabbit haemorrhagic disease virus. Journal of General Virology 78, 2803–2811. Liu, S. J., Xue, H. P., Pu, B. Q. & Quian, N. H. (1984). A new viral disease in rabbits. Animal Husbandry and Veterinary Medicine 16, 253–255. Lo! pez de Turiso, J. A., Corte! s, E., Garcı! a, J., Ranz, A., Sanz, A., Vela, C. & Casal, I. (1991). Fine mapping of canine parvovirus B cell epitopes. Journal of General Virology 72, 2445–2456. Martı! nez-Torrecuadrada, J. L. & Casal, J. I. (1995). Identification of a linear neutralization domain in the protein VP2 of African horsesickness virus. Virology 210, 391–399. Meyers, G., Wirblich, C. & Thiel, H.-J. (1991). Rabbit hemorrhagic disease virus : molecular cloning and nucleotide sequence of a calicivirus genome. Virology 184, 664–676. Milton, I. D., Turner, J., Teelan, A., Gaskell, R., Turner, P. C. & Carter, M. J. (1992). Location of monoclonal antibody binding sites in the
capsid protein of feline calicivirus. Journal of General Virology 73, 2435–2439. Nakane, P. K. & Kawaoi, A. (1974). Peroxidase-labeled antibody : a new method of conjugation. Journal of Histochemistry and Cytochemistry 22, 1084–1091. Ohlinger, V. F., Haas, B., Ahl, R. & Weiland, F. (1989). Die infektio$ se ha$ morrhagische Krankheit des Kaninchen – eine durch ein Calicivirus verursachte Tierseuche. Tieraerztliche Umschau 44, 284–294. Ohlinger, V. F., Haas, B., Meyer, G., Weiland, F. & Thiel, H.-J. (1990).
Identification and characterization of the virus causing rabbit hemorrhagic disease. Journal of Virology 64, 3331–3336. Plana-Dura! n, J., Bastons, M., Rodriguez, M. J., Climent, I., Corte! s, E., Vela, C. & Casal, I. (1996). Oral immunization of rabbits with VP60 particles confers protection against rabbit hemorrhagic disease. Archives of Virology 141, 1423–1436. Prasad, B. V. V., Rothnagel, R., Jiang, X. & Estes, M. K. (1994). Three dimensional structure of baculovirus-expressed Norwalk virus capsids. Journal of Virology 68, 5117–5125. Reid, G., Rigby, M. A., McDonald, M., Hosie, M. J., Neil, J. C. & Jarret, O. (1991). Immunodiagnosis of feline immunodeficiency virus infection
using recombinant viral p17 and p24. AIDS 5, 1477–1483. Sambrook, J., Fritsch, T. J. & Maniatis, T. (1989). Molecular Cloning : A Laboratory Manual, 2nd edn. Cold Spring Harbor, NY : Cold Spring Harbor Laboratory. Sanz, A., Garcia-Barreno, B., Nogal, M. L., Vin4 uela, E. & Enjuanes, L. (1985). Monoclonal antibodies specific for African swine fever virus proteins. Journal of Virology 54, 199–206. Seal, B. S., Ridpath, J. F. & Mengeling, W. L. (1993). Analysis of feline calicivirus capsid protein genes : identification of variable antigenic determinant regions of the protein. Journal of General Virology 74, 2519–2524. Studier, F. W., Rosenberg, A. H., Dunn, J. J. & Dubendorff, J. W.
Antigenic structure of RHDV (1990). Use of T7 RNA polymerase to direct expression of cloned genes.
(1997). Mapping of antigenic sites involved in neutralization on the
Methods in Enzymology 185, 60–89. Summers, M. D. & Smith, G. E. (1987). A Manual of Methods for
capsid protein of feline calicivirus. Journal of General Virology 78, 303–305.
Baculovirus Vectors and Insect Cell Culture Procedures. Texas Agricultural Experiment Station Bulletin no. 1555.
Wirblich, C., Meyers, G., Ohlinger, V. F., Capucci, L., Eskens, U., Haas, B. & Thiel, H.-J. (1994). European brown hare syndrome virus :
Thouvenin, E., Laurent, S., Madelaine, M.-F., Rasschaert, D., Vautherot, J.-F. & Hewat, E. A. (1997). Bivalent binding of a neutralising
relationship to rabbit hemorrhagic disease virus and other caliciviruses. Journal of Virology 68, 5164–5173.
antibody to a calicivirus involves the tortional flexibility of the antibody hinge. Journal of Molecular Biology 270, 238–246. Tohya, Y., Yokohama, N., Maeda, K., Kawaguchi, Y. & Mikami, T.
Received 19 November 1997 ; Accepted 17 March 1998
BJAJ
