key: cord-0747473-7ro72sv3
authors: Lenstra, Johannes A.; Kusters, Johannes G.; Koch, Guus; van der Zeijst, Bernard A.M.
title: Antigenicity of the peplomer protein of infectious bronchitis virus
date: 1989-01-31
journal: Molecular Immunology
DOI: 10.1016/0161-5890(89)90014-x
sha: fbe77b2a78477062a2b4a5d7101c94c97bf94d0b
doc_id: 747473
cord_uid: 7ro72sv3

Abstract To study the antigenic structure of the peplomer protein of the avian coronavirus infectious bronchitis virus, fragments from the peplomer gene were generated by restriction-enzyme cleavage or by limited DNase digestion and inserted in the Escherichia coli expression plasrnid pEX (Stanley and Luzio, 1984). The antigenicity of the expression products was tested using a number of polyclonal antisera and monoclonal antibodies. The polyclonal antisera recognized different sets of epitopes in the 1162-residue sequence. The N-terminal region of one of the two subunits, S2, was recognized by all polyclonal sera and by two monoclonal antibodies. This clearly immunodominant region contains at least two adjacent or overlapping epitopes, one of which has been localized within 18 residues. The epitopes found as antigenic pEX expression products do not coincide with the regions in the S1 subunit that have been found to contain hypervariable sequences. We suggest that these regions constitute conformation dependent neutralization epitopes that cannot be detected in the pEX system. The relevance of our finclings for vaccine development is discussed.

Coronaviruses act as causative agents of a number of infectious diseases in domesticated animals (for a review about these positive stranded RNA viruses, see Siddell et al., 1983; Sturman and Holmes, 1983) . Conventional vaccines do not exist or are rendered ineffective by antigenic variation. This latter mechanism is thought to be relevant for the epidemiology of infectious bronchitis virus (IBV), which causes a highly contagious respiratory disease in chickens. After recent outbreaks of the disease, new serotypes have been isolated that may have originated from the live attenuated strains used for vaccination (Kusters et al., 1987) . Obviously, recombinant-DNA expression products or peptide vaccines would offer a valuable alternative for conventional vaccines. The antigenic properties of IBV or other coronaviruses most likely reside in the peplomer (Cavanagh et al., 1984) , a club-shaped spike projecting from the viral membrane and consisting of a dimer or trimer of the E2 peplomer protein (Cavanagh, 1983) . Analysis of the amino acid sequences of the peplomer proteins from representatives of the three main coronavirus serotypes (De Groot et a[., 1987) has led to a model, in which two §Author to whom correspondence should be addressed. Abbreviations:

ELISA, enzyme-linked immunosorbent assay; IBV, infectious bronchitis virus; MAb, monoclonal antibody;

PBS, phosphate-buffered saline.

long cc-helices are located in the C-terminal half of the monomer subunit. In the oligomer, those helices intercalate and form a coiled-coil structure, which constitutes the stalk of the peplomer. The transmembrane segment at the C-terminal end serves as membrane anchor. The N-terminal half is thought to form the spherical outer part of the peplomer.

In IBV, proteolytic cleavage results in an N-and a C-terminal subunit, designated Sl and S2, respectively. Sl is thought to be the most relevant part for the antigenic properties of IBV (Cavanagh et al., 1986) . It is recognized by strongly neutralizing monoclonal antibodies (Mockett et al., 1984; Koch et al., unpublished) .

Furthermore, isolated Sl is capable of eliciting neutralizing antisera (Cavanagh et al., 1986) . By comparing the amino acid sequences of different strains, it was found that most amino acid substitutions, likely to reflect antigenic variation, have occurred in the Sl subunit (Niesters et ul., 1986; Binns et al., 1986; Kusters et al., unpubljshed) .

In this study, we have localized epitopes by expressing parts of the peplomer sequence in a prokaryotic expression system, and measuring the antigenicity of the expression products (Stanley and Luzio, 1984) . Most probably, this method detects epitopes that are essentially conformation independent. Antigenic regions were almost exclusively found in the S2 subunit, suggesting that the antigenic variation in Sl reflects nonformation dependent epitopes. One immunodominant region has been localized near the N-terminus of S2. This region was recognized by all polyclonal antisera tested as well as by two neutralizing monoclonal antibodies, and contains at least two different epitopes. The sequence of this region may be the starting point for the development of peptide vaccines.

Schuell). The Hinfl ends of hl were made blunt by treatment with the Klenow fragment of DNA polymerase. Depending on the desired reading frame, the fragments were ligated in compatible sites of pEX I, pEX2 or pEX3 (Stanley and Luzio, 1984) . which had been treated with calf intestine phosphatase if linearized by a single digestion.

The cDNA clones p39 and p42 of the peplomer gene of IBV strain M41 in the vector pUC9 have been described previously (Niesters et al., 1986) .

Rabbit antisera against M41 were raised by intramuscular injection of purified virus (1 mg protein), lysed by Triton X-100, in Freund's complete adjuvant, followed by the same dose in Freund's incomplete adjuvant after two and three weeks, respectively.

Mouse antiserum was raised by intraperitoneal injection of 100 pg virus protein, followed by a second dose after 1 month. Chicken antisera were collected from convalescent (chickens 1 and 3) or hyperimmune (chickens 2 and 4) animals, infected with IBV strain M41 (chickens 1 and 2) or strain D207 (chickens 3 and 4).

Recombinant plasmids were introduced into E. coli strain pop 2136 via the CaClz transformation procedure. Recombinants expressing the IBV gene fragments were selected by immunoscreening with a rabbit antiserum or, when the expression product was not recognized antigenically, by hybridization followed by gel electrophoresis of the protein expression products.

As hybridization probe, a mixture of HindIIIiHinfI fragments of the insertion from plasmid p39 was used, which could be isolated free from the 2.7-kb pUC fragment by gel electrophoresis.

The preparation of the monoclonal antibodies 26.1 and 31.5 against IBV has been described previously (Koch et al., 1986) .

Alkaline phosphatase-conjugated goat anti-(mouse IgG, H + L) and goat anti-(rabbit IgG, Fc) were from Promega (Madison); anti-(chicken IgG, H + L) conjugates were from Nordic, Tilburg, The Netherlands (horseradish peroxidase) or from Miles (alkaline phosphatase).

As pEX host strain, E. coli pop 2136 was used, which was constructed by Dr Raibaud (Institut Pasteur, Paris).

Unless mentioned otherwise, all DNA manipulations used were done essentially as described by Maniatis et al. (1982) or Davis et al. (1986) .

Genomic cDNA clones p39 and p42 were cut with restriction enzymes as indicated in Fig. 1 Tris-HC1 (pH 7.5) 1.5 mM MnCl, and 0.1 mg/ml BSA at 22 C for 6 min (first 5 /tl of the reaction mixture) or 10 min (remaining 5 ~1). As judged by gel electrophoresis and autoradiography after end-labelling with the Klenow fragment of DNA polymerase, fragments of 2&500 bp were generated.

These fragments were made blunt-ended by treatment with T4 DNA polymerase, followed by treatment with the Klenow fragment. After sequential extractions with phenol/chloroform and ether, the fragments were ligated directly in pEX2 DNA, which had been cut with SmaI and dephosphorylated with calf intestine phosphatase.

The reaction was carried out overnight This rapid cloning procedure obviates the use of linkers (Haymerle ez al., 1986 ) but generates nevertheless a sufficient number of clones from a low amount of insert DNA.

Immunoscreening of bacterial colonies transferred to nitrocellulose membranes (BA85 from Schleicher and Schuell) was performed as described (Stanley, 1983) , using a gelatin/Triton X-1OOjPBS buffer and an E. coli lysate to quench aspecific reactions. The dilutions of the primary antisera were 1: 1000 for polyclonal sera and 1: 100 or 1:500 for ascites fluids containing monoclonal antibodies. Binding of antisera to colonies was visualized using alkaline phosphatase-conjugated secondary antisera (diluted 1:5000) and nitro blue tetrazolium and 5-bromo-4-chloro-3-indolyl phosphate as substrates (Blake et al., 1984) .

To localize DNase fragments from pEX clones that were positive after immunoscreening or to check the correct reading frames for the expression of restriction fragments, DNA from pEX/IBV recombinants was isolated by the alkaline-lysis method, cut with two restriction enzymes flanking the insertion and with Hind111 to cleave the pEX vector sequence. The resulting mixture of fragments was ligated in M13mp9 DNA cut with the appropriate enzymes. Subsequently, the nucleotide sequence of the insertion was determined by the dideoxynucleotide termination method and compared to the sequence of the IBV M41 peplomer gene.

Expression of the pEX fusion gene, leading to the synthesis of a hybrid b-galactosidase protein, was induced in a 5 ml exponential culture (ODMM ca. 0.25) by incubation at 42°C for 90 min. Hybrid proteins were purified as Triton X-100 pellet (Stanley and Luzio, 1984) . Cells were spun down (10 min at 5OOOg), resuspended in 100 ~1 15% (w/v) sucrose, 50mM

Tris-HCI (pH 8.0), 50mM EDTA, and treated with lysozyme (1 mg/ml) for 10 min. After addition of 140 ~10.2% (v/v) Triton X-100 in 10 mM Tris-HCl (pH 8.0) 1 mM EDTA, the suspension was sonicated in a bath sonicator for 15 min. The lysozyme digestion as well as use of the sucrose buffer appeared essential for the purity of the final preparation. After spinning down the insoluble hybrid protein (10 min at 10,OOOg) and removing the viscous supernatant as completely as possible, the pellet was resuspended in Triton X-100 buffer. Sonification, centrifugation and removal of the supernatant were repeated (l-2 times) until no viscosity remained. Finally, the pellet was resuspended in 250~1 PBS.

After adding concentrated SDS-containing lysis buffer (Laemmli and Favre, 1973) , hybrid proteins were spotted via a manifold on nitrocellulose membranes or fractionated by SDSpolyacrylamide gel electrophoresis (7.5%) followed by Western blotting. Hybrid proteins thus immobilized were analysed as described for the filters of the immunoscreening (see above). Binding of chicken antisera was visualized using peroxidase-conjugated secondary antisera (diluted 1: 1000) and 3'-3' diaminobenzidine and H,Oz as substrates.

In the pEX expression system, foreign gene fragments are inserted at the 3' end of a cro-IacZ expression vector (Stanley and Luzio, 1984) . Expression of this gene is under control of the lambda P, promotor and induced by inactivation of the temperature sensitive cl repressor at 42°C. The resulting expression products precipitate inside the cell and are thus protected against proteolysis.

After solubilization, the product of the foreign gene fragment, linked to the C-terminal residue of the /?-galactosidase part, may be recognized by appropriate antisera. Figure 1 shows the Hinfl, Pst I, Sau3AI and Sau3AI/PstI fragments that have been excised from the genomic cDNA plasmids p39 or p42 (Niesters et al., 1984) and expressed in pEX plasmids. The Hinfl site in the region encoding the N-terminal signal sequence of the peplomer protein allowed the cloning of the 5' part of the coding sequence without the stop codons just upstream of (and in frame with) the ATG start codon.

As judged from an SDS-polyacrylamide gel [ Fig. 2(A) ], the hybrid proteins have sizes predicted by the lengths of the insertions. Exceptions were ps4, which contains the stop codon of the peplomer gene, and the 144 bp fragment p4, which was inserted twice, both fragments expressed in tandem and in the same reading frame. Analysis of the sequence showed that only in the case of hl and p4, hybrid proteins of (almost) the same size could be obtained after translation in a wrong reading frame; the proper reading frames of these clones have been checked by sequencing. Fragment pl starts with a stretch of 12 dG residues (originating from ~42); apparently, this did not disturb the reading frame by ribosome slippage along the homopolymer track (Stanley, 1983) . The IBV peplomer sequences contained in the hybrid proteins are listed in Table 1 . Only residue 1064 has not been overlapped by an expression product.

Further, a reactive region exactly on the junction of two fragments or in the gap between p4 and p5 can only be detected in an overlapping fragment if this fragment does not contain other reactive regions. In the theoretically worst case, the reactivity of 8% of the total peplomer sequence would not be detected, assuming an epitope size of 10 residues. 

In Fig. 2(B) , a Western blot of the IBV/pEX hybrid proteins, incubated with a rabbit antiserum, is shown. Only three regions were recognized strongly, one contained in p2, ~4, psl and ps2, the second in p3 and ~4, and the third in ps4. Weaker signals were observed for the regions contained in p4, in p5 while not in ps3, and in the overlap of s5 and p6, respectively.

have been carried out with another rabbit antiserum, with a mouse antiserum and with four antisera from chickens, infected with IBV strain M41 (chickens 1 and 2) or with strain D207 (chickens 3 and 4) . The binding patterns are summarized in Fig. 3 . Because of high background signals of the chicken antisera, only strong signals cquld be scored. Apparently, each antiserum recognized a different set of epitopes.

The number of The positions of the restriction-enzyme fragments of the peplomer gene are shown in Fig. I . The fragments were ligated to linearized pEX 1, -2 or -3 as indicated. The last column indicates which parts of the amino acid sequence of the peplomer protein are synthesized by the recombinants after induction of the expression. antigenic regions as defined by the restriction fragments varies from 1 (mouse) to 6 (rabbit 1). One region, corresponding to fragment ps2, was recognized by all homologous and heterologous antisera. (Koch et al., unpublished results) . None of these sera bound any of the pEX products, most likely because the epitopes of these antisera are conformation dependent. However, the lack of binding of other, non-neutralizing monoclonal antibodies recognizing conformation independent epitopes on the peplomer protein suggests that other factors than the conformation dependence may affect the reactivity of monoclonal antibodies with pEX hybrid proteins.

Koch et al. (1986) described a number of monoclonal antisera with weak neutralizing activity that bound to the S2 subunit on Western blots. One of these, MAb 26.1, recognized all IBV strains tested and has a neutralization titre of about 10'. Figure  2 (C) shows a Western blot of hybrid proteins screened with this MAb. The strong binding to fragments p2, s4, psl and ps2 localizes the epitope of MAb 26.1 between residues 390 and 612, the same region that is recognized by all polyclonal sera. Since subunit S2 starts at residue 538, the epitope must lie in the N-terminal region (residues 538 to 612) of S2. The same localization has been found for the epitopes of another S2-specific MAb 31.5.

To further localize the epitope of MAb 26.1, a library of small random DNase fragments in pEX has been constructed.

After immunoscreening, 0.2% of the colonies were found to react positively with MAb 26.1. Six clones were selected for sequence analysis after subcloning in Ml3 replicative-form DNA. The results are summarized in Fig. 4 . All clones contain nucleotides 1620-1673, restricting the localization of the epitope of MAb 26.1 to residues 541-558. Figure 5 shows the antigenicity of four expression products of the DNase fragments to MAb 26. antisera and fragments are listed in Table 2 . The antiserum from chicken 1 binds to the expression products of the same fragments as MAb 26.1, indicating the recognition of the same or an adjoining epitope. The results with antiserum of chicken 2 were clouded by a particularly high background reaction, but the clear positive reactions with two of the fragments indicate a binding to the same region. Interestingly, the negative reactions of fragments (15841682) and (1617-1673) demonstrate that the epitope(s) of MAb 31.5 and of the scra of rabbit I, rabbit 2 and mouse are located partially or completely in the region 562-574.

The positions of the DNase defined regions reactive with the polyclonal sera are graphically represented in Fig. 3 by short horizontal lines. The delineations of the individual epitopes are indicated in Fig. 4 by heavy lines, the dashed line indicating the possibility that the region 541-561 contributes to the epitopes recognized by the MAb 31.5 and by the rabbit and mouse antisera.

In indirect ELISAs on Western blots with intact viral proteins, MAbs 26.1 and 31.5 have different specificities (Koch et al., 1986. unpublished results). MAb 26.1 recognizes the $2 subunit of all strains tested, while MAb 31.5 only binds to S2 of the H120/B222/M41 serotype. The same specificity was found with pEX expression products: MAb 26.1 binds to the hybrid proteins of strains M41, D207 and D1466; MAb 31.5 only to M41 products. An accurate definition of the epitope of MAb 26.1 and other epitopes in the same region, based on the sequences of several IBV strains, on PEPSCAN peptide synthesis (Geysen et al., 1984) and on competition experiments will be published elsewhere (Kusters et al.; Koch et al., in preparation) .

In the pEX system, the expression product of an inserted DNA fragment constitutes the C-terminus of a hybrid /l-galactosidase protein, which precipitates in the cell. Prior to the assay of its antigenicity, the hybrid protein is solubilized by SDS and transferred to a nitrocellulose filter. Presumably, epitopes found by cross-reaction with antisera are essentially conformation independent.

The localization of epitopes within stretches of 10-20 residues in several coronaviruses (Fig. 4 , unpublished results) demonstrates that the binding of antibody does not depend on particular flanking sequences, and that any nativelike folding of the epitope in the hybrid protein is confined to the same small region.

Classically, continuous epitopes are localized by testing the antigenicity of peptides (Atassi, 1984) . Prokaryotic expression of subgenomic fragments in pEX or in i.gtll (Nunberg et ul., 1984; Mehra et al., 1986 ) offers a convenient alternative and is not limited to small proteins.

We have combined the expression of large (13&1400 bp) restriction fragments to localize antigenic regions, and of small DNase fragments to delineate epitopes more accurately. Another strategy is the subcloning of small Kusters et al., Luytjes et al., unpublished results) . This obviates the sequence analysis of DNase fragments, but depends on the presence of suitable restriction sites or on the availability of homologous sequences.

Another new and complementary approach to the localization of continuous epitopes is the synthesis of a large number of peptides coupled to a solid phase, the so-called PEPSCAN (Geysen et al., 1984) . This method has been used in conjunction with the pEX expression system for the localization of a neutralization epitope of the peplomer protein of mouse hepatitis virus (Luytjes et al., unpublished results). However, not all monoclonal antibodies that cross-react with pEX expression products bind to PEPSCAN peptides spanning the antigenic region (Posthumus and Meloen, unpublished results). For instance, MAb 26.1 did not recognize any of the overlapping nonapeptides spanning the N-terminal region of the S2 subunit. Whether this should be explained by the lengths of the peptides or by local conformations in the hybrid proteins will be the subject of further study.

The reactivity patterns of the different polycionaf antisera (Fig. 3 ) outline the variability of the humoral immune response against the IBV peplomer protein at the molecular level. There are also a number of other noteworthy features.

Firstly, all antisera recognize in the 1162-residue sequence a relatively small (l-6) number of regions, predominantly in the S2 subunit; the actual number of epitopes per antiserum depends on the degree of clustering of epitopes within the antigenic regions. So, if the whole surface of the proteins is potentially antigenic (Berzofsky, 1985; Geysen et al., 1987) , this is clearly not the case for the determinants detected as antigenic pEX hybrid proteins.

Secondly, the N-terminal part of the S2 subunit is recognized by all polyclonal antisera tested and by a number of monocional antibodies. Within this region, the monoclonal and polyclonal antibodies recognized different epitopes ( Table 2 ). The MAbs have neutraIization titres ranging from 10 to 1000 (Koch et al., 1986 ) suggesting a role of this region in the neutralization of the infection.

Thirdly, the regions found to be antigenic do not coincide with the hypervariable regions in the Sl subunit (Niesters et al., 1986; Binns et al., 1986) . These regions are supposed to be part of neutralization epitopes that are the targets of antigenie variation. In fact, although the Sl subunit is thought to induce the protective immunity against IBV (Cavanagh et al., 1986) we found it to be relatively, if not completely, devoid of antigenicity (Fig. 3) .

As the most obvious explanation, we propose that the neutralization epitopes in St are conformation dependent and cannot be mimicked by pEX hybrid proteins. This would also explain the lack of reaction of the neutralizing MAbs that were directed against epitopes in SI. Apparently, these epitopes are of the same category as the conformation dependent neutralization epitopes of the haemagglutinin of influenza virus (Wiley et al., 1981) . In contrast, the epitopes localized in the N-terminal region of S2 may be comparable to the epitopes of foot-and-mouth disease virus, detectable by cross-reaction with peptides (Geysen et al.. 1984; Rossman et al., 1985) .

respectively. The rabbit antisera were generously provided by MS N.M.C. Bleumink-Pluvm and Dr H. G. M. Niesters (Veterinary Faculty). The experimental contributions of

Antigenic structures of proteins

Intrinsic and extrinsic factors in protein antigenic structures

Comparison of the spike precursor sequences of coronavirus IBV strains M41 and 6/82 with that of IBV Beaudette

A rapid, sensitive method for detection of alkaline phosphatase-conjugated antiantibodies on Western blots

Coronavirus IBV: structural characterization of the spike protein

By virtue of their Continuous epitopes often correspond to relatively inherent cross-reactivity with short peptides, these coronaviruses

Use of peptide synthesis to probe viral antigens for epitopes to a resolution of a single amino acid

Chemistry of antibody binding to a protein

Efficient construction of cDNA libraries in plasmid expression vectors using an adaptor strategy

Antigenic differentiation of avian bronchitis virus variant strains employing monoclonal antibodies

Molecular epidemiology of infectious branchitis virus in the Netherlands

Mokectttar Cloning, a Laboratory Manual. Cold Spring Harbor Laboratories

Efficient mapping of protein antigenic determinants

Monoclonal antibodies to the Sl spike and membrane proteins of avian infectious bronchitis virus strain Massachusetts M41

The pcplomer protein sequence of the M41 strain of coronavirus IBV and its comparison with Beaudette strains

Epitopes on the peplomer protein of infectious bronchitis virus strain M4I as defined by monoclonal antibodies

Method to map antigenic determinants recognized by monoclonal antibodies: localization of a determinant of virus neutralization on the feline leukemia virus envelope protein ~~70

The biology of coronaviruses

Solubilization and immune-detection of p-galactosidase hybrid proteins carrying foreign antigenic determinants

~onst~ction of a new family of high efficiency bacterial expression vectors: identification of cDNA clones coding for human liver proteins

Antigenic crossreactivity between proteins and peptides: new insights and applications

Correlation between segmental mobility and the location of antigenic determinants in proteins

Frankenberger identification of the antibodv-combining site of Hone E

mobile regions in the protein structure (Westhof et regions are the prime candidates for the development of peptide vaccines (Van Regenmortel, 1987) . Thus, for the N-terminal region of the IBV S2 subunit. structural considerations may outweigh the relatively weak neutralization, if compared with the epitopes in Sl (Mockett et al., 1984; Koch et al., Further, M41 sequences cross-react with chicken antisera against D207, belonging to a different serotype, (Fig. 3, Table 2 ) and vice versa (not shown). Apparently, this region is not subject to antigenic variation. This would be another advantage of the epitopes found in 52.Thirdly, the recognition of the N-terminal S2 sequence by several polyclonal antisera indicates that this region is strongly immunodominant.This would favour a response in all vaccinated animals.These considerations may become relevant as soon as the technology for designing effective peptide vaccines becomes available.