The Epitope Recognized by Monoclonal Antibody 2B6 in the B/C Domains of Classical Swine Fever Virus Glycoprotein E2 Affects Viral Binding to Hyperimmune Sera and Replication

Abstract

Classical swine fever (CSF) is a highly contagious disease of pigs caused by CSF virus (CSFV). E2 is the major viral envelope protein of immune dominance that induces neutralizing antibodies and confers protection against CSFV infection. The B/C domains of E2 are variable among CSFV isolates, which could affect immunogenicity and binding to antibodies. We attempted to characterize the epitope recognized by a monoclonal antibody 2B6 (mAb-2B6) raised against the E2 B/C domains of the vaccine C-strain and to examine if mutations in the epitope region would affect antibody binding and viral neutralization. The epitope specific for mAb-2B6 recognition is linear, spanning five residues ⁷⁷⁴DGXNP⁷⁷⁸ in the B/C domains. The residue N777 is indispensable for the specificity. The epitope exists only in group 1 strains, but not in those of group 2. The recombinant viruses containing individual mutations on the epitope region lost the reactivity to mAb-2B6. The mutant virus RecC-N777S had low replication potential, about 10-fold decrease in the yield of progeny virus particles, whereas the mutant virus RecC-P778A reverted to proline upon continuous passaging. The mutations on the mAb-2B6 epitope region did not affect neutralization by anti-C-strain polyclonal sera from pigs. Deletion from aa774 covering the mAb-2B6 epitope, but not that from aa781, also affected binding with the polyclonal antibodies from vaccinated pigs, although the major binding region for the vaccinated antibodies is aa690-773.

Introduction

Classical swine fever virus (CSFV), a member of the genus Pestivirus within the family Flaviviridae, is the causative agent of classical swine fever that is highly contagious and causes significant economic losses to the swine industry [1,15,19]. CSFV is an enveloped RNA virus with the genome size of approximately 12.3 kb, flanked by 5´- and 3´-nontranslated regions (NTRs). The viral genome encodes a polyprotein of 3,898 amino acids that is processed by cellular and viral proteases to generate four structural (C, Erns, E1, and E2) and eight nonstructural proteins (Npro, P7, NS2, NS3, NS4A, NS4B, NS5A, and NS5B)[14,29].

Phylogenetic analysis of the 5’-NTR, E2, or NS5B divides CSFV strains into three groups with 11 subgroups [22,26]. Typical historical strains, including the vaccine strains, belong to group 1, whereas currently circulating strains and those that have caused infections since the 1980s form group 2, and those distributed in Taiwan and some other regions belong to group 3 [8,23,25,33]. Several reports indicate that subgroup 2.1 strains have branched away from the vaccine C-strain and become dominant in China [4,17,31].

E2 is the major viral envelope protein of high immunogenicity that induces neutralizing antibodies and confers protection against infection. Four major antigenic domains A, B, C, and D, all located in the N-terminal region of E2, have been identified using a panel of monoclonal antibodies (mAbs) [35,36]. They form two independent antigenic domains. The B/C domains are linked by a putative disulfide bond between 693C and 737C, whereas the D/A domains are connected by two separate disulfide bonds [2,35,36]. Independent studies also report conformational and linear epitopes located in the B/C and A domains [3,13,24,34,38]. A highly conserved linear epitope 829TAVSPTTLR837 located in the A domain has been utilized in epitope vaccine development and considered for specific serodiagnostic assays [16,27,28,38]. Another conserved motif, 772LFDGTNP778, constitutes a linear epitope between domains B/C and A [24]. In addition, the antigenic motif 771LLFD774 in the B/C domains is essential for the integrity of the antigenic structure of E2 [2].

The B/C domains are much more variable than the D/A domains, and the antigenic variations in this region form two distinct clusters, AR1 and AR2 [4-6]. Variations of E2 from field strains influence the reaction patterns to and cross-neutralization ability with both anti-CSFV sera and mAbs [6]. Antigenic analysis using different mAbs reveals the antigenic diversity of E2 from different CSFV isolates [11,18,20,21,39]. Therefore, CSFV field strains could be distinguished from the vaccine strain by different reaction patterns with some mAbs [18,20]. Exchanges of single residues in the B/C domains between the group 1 vaccine strain lapinized Philippines Coronel (LPC) and the group 3 field isolate could change the mAb reaction patterns [3]. Substitution of the residue 710 on the E2 protein of different strains affected viral binding and neutralization by a number of mAbs [34]. These studies suggest that alterations of amino acids between different CSFVs may change the antigenicity.

In this study, a series of recombinant C-strain-based mutant E2 proteins, with single substitutions based on amino acid differences in the N-terminus between the C-strain and subgroup 2.1 isolates, were used to define specific residues critical for recognition by mAb-2B6 that was derived from the E2 B/C domains of the C-strain and reacted only with a linear epitope of group 1 strains. Site-directed mutations of critical residues were also introduced into the infectious clone of the C-strain to examine their effects on the reactivity and replication of the recombinant viruses.

 

Materials and Methods

Cells and Viruses

Swine testicle (ST) cells were grown in Dulbecco's minimal essential medium (Hyclone, Thermo Scientific, USA) with 10% fetal calf serum (Hyclone). Four CSFV strains were used: two subgroup 1.1 strains (vaccine C-strain and Shimen strain) and two subgroup 2.1 clinical isolates (strains QZ-07 and HZ1-08). The viral strains were propagated and titrated in ST cells. Virus stocks were prepared by passages in ST cells and stored at -80℃. All stocks were tested for the absence of bovine viral diarrhea virus by PCR using specific primers.

Plasmids

The plasmids containing the full-length E2 gene of the vaccine C-strain and seven subgroup 2.1 strains were constructed according to established procedures in our laboratory. Prokaryotic expression plasmids were constructed (as previously described [6]) to generate two recombinant proteins that contain the His-tag in both N-and C-termini: rE2-BC (spanning aa 690-814) and rE2-AD (spanning aa690-865) covering both B/C and A/D domains [38,39]. For analysis of the reaction patterns of truncated rE2-BC with monoclonal and polyclonal antibodies, a panel of deletion constructs in pET30a (Novagen, USA) fused with a C-terminal GST-tag were derived by overlapping PCR from C-strain-based pET30-rE2-BC: E2-690-814, E2-768-814, E2-690-791, E2-690-780, E2-690-773, E2-768-791, and E2-773-780. For eukaryotic expression plasmid construction, a 1,212 bp cDNA fragment spanning the signal sequence to the end of the full-length C-strain E2 was amplified and ligated into pcDNA3.1 following BamHI and XhoI digestion.

Antibodies

Generation of the murine mAbs 2B6 and 6B8 was described in our previous study [6]. The rabbit antiserum to the C-strain virus was generated by intravenous injection of 103 TCID50 of C-strain into New Zealand white rabbits, which were boosted after 21 days. The serum samples collected at 14 days after boosting were used for indirect immunofluorescence assay (IFA). Acquisition of the pig hyperimmune sera (#467 and #516) against CSFV vaccine C-strain was as previously described [6]. These two sera were used for binding ELISA, western blotting, and viral neutralization assays. The animal experiments were approved by the Laboratory Animal Management Committee of Zhejiang University (Approval No. 20101015).

Expression of rE2 Proteins, SDS-PAGE, and Western Blotting

Expression of rE2 proteins, both full-length and truncated, was performed by routine methods used in our laboratory [6]. Briefly, E. coli Rosetta (DE3) cells harboring the expression plasmids were induced with 1 mM isopropylthio-β-D-galactoside (Sangon Shanghai, China). Bacterial cells were disrupted by sonication and proteins were purified on a Ni-NTA affinity column (Novagen, USA). The purified rE2 proteins were confirmed by western blotting with mouse monoclonal anti-His-tag antibody (Sigma-Aldrich, USA) and quantified by the Bradford assay. The antigenic reactivity of different purified rE2 proteins with mAb-2B6 or pig anti-C sera was performed by ELISA as previously described [6].

Site-Directed Mutagenesis of the C-Strain rE2-BC

Multiple alignment of E2 sequences identified 16 major variable residues in the B/C domains. To substitute such C-strain residues with those found in group 2 isolates, plasmids containing individual mutations were generated by site-directed mutagenesis on C-strain-based pET30-rE2-BC. All substitutions were performed using the Quick-change Site-Directed Mutagenesis Kit (Stratagene, USA) according to the manufacturer’s instructions. The primers were designed via the Quick-change Primer Design program (http://www.stratagene.com). The expected nucleotide changes in each mutant were verified by sequencing. The rE2-BC proteins included substitutions at A692S, D705N, E706K, L709P, G713E, N723S, D725G, N729D, S736I, V738T, T745I, N777S, S779A, T780I, R788G, and S789F.

Monoclonal Antibody Binding ELISA for Mutated rE2-BC Proteins

The binding ELISA was conducted in triplicate under stringent conditions to avoid nonspecific reactions [6]. The binding efficiency was expressed as the ratio of mAb-2B6 bound to individual mutant protein to that bound to native C-strain version rE-BC, which was set as 100%. The average binding efficiency was calculated for three independent ELISA assays. Relative binding of lower than 50% was considered as significant.

Construction and In Vitro Rescue of C-Strain-Based Infectious Clones Containing Major Substitutions on the mAb-2B6 Epitope

Plasmid pA-RecC, a full-length cDNA clone of the C-strain viral genome, was used as the template in which each residue of the mAb-2B6 epitope was mutated by the Quick Change XL Site-Directed Mutagenesis Kit (Stratagene, USA). Each of the fulllength genomic clones was linearized by XhoI and transcribed in vitro using the T7 Megascript system (Ambion, Austin, USA). After TURBO DNase I (Ambion, USA) digestion, the RNA transcripts were precipitated with LiCl and quantified by a spectrophotometer, ND-1000 (Nano Drop Technologies, USA). The RNA transcripts were individually transfected into ST cells by electroporation at 150 V with a BTX 630 electroporator (BTX, San Diego, USA). Cells were plated in 25 cm2 flasks and incubated for 4 days at 37℃ and 5% CO2. Virus particles were detected by immunostaining the E2 protein as described below. Stocks of rescued viruses were stored at -80℃.

Indirect Immunofluorescence Assays

Cells seeded on 24-well plates (Costar; Corning, USA) were fixed with ethanol-acetone (1:1 by volume) for 20 min at -20℃, washed twice in PBS (10 mM), and blocked in PBS containing 1% BSA. Different primary antibodies diluted in PBS containing 0.5% non-fat milk powder were added for 2 h at 37℃ and then washed with PBS. The cells were probed with Alexa-Fluor-labeled antimouse or anti-rabbit secondary antibodies (488 or 568; Invitrogen, USA) for 1 h, followed by two rounds of washing, and examined under the IX71 inverted fluorescence microscope (Olympus, Japan).

Single-Step Growth Curve

Growth characteristics of rescued mutant viruses RecC-D775A and RecC-N777S were evaluated relative to the wild-type recombinant C-strain virus in a single-step curve. Monolayer ST cells seeded in 24-well plates were infected with 150 TCID50 viruses and adsorbed for 1 h at 37℃ (set as time 0h). The supernatant samples were then removed and a volume of 0.7 ml of fresh medium was added to the cells in each well. Cell samples were collected at 12, 24, 48, 72, and 96 h post-infection and tested for virus titration and RNA copies.

In Vitro Virus Neutralization

Virus neutralization was assessed as described previously [6]. Briefly, each of the stock viruses (500 TCID50) was mixed with an equal volume of 2-fold serially diluted pig hyperimmune sera and incubated at 37℃ for 1 h before being inoculated onto the preseeded ST cells in 96-well plates. The starting dilution of each serum was 1:50. At 96 h post inoculation, cells were fixed and stained for glycoprotein E2 by IFA. The neutralization index (NI) is the log10 of the reciprocal of the antibody dilution factor where 50% of the wells were protected from infection.

 

Results

Reaction Patterns of mAb-2B6 with Different Genotypes of CSFV Strains

Immunofluorescence was performed on ST monolayers infected with CSFV strains from different groups and probed with the mAb-2B6 to identify their binding. Fluorescent signals were observed only on cells infected with group 1 strains (Shimen and C-strain) but not those of group 2 (Fig. 1A). This is consistent with the observation that only prokaryotic-derived E2 proteins (rE2-AD) of group 1 strains reacted to mAb-2B6 (Fig. 1B). Substitution of any putative cysteine residues involved in the disulfide bond in the eukaryotic E2 protein had no impact on recognition of mAb-2B6 (data not shown). We speculate that mAb-2B6 may recognize a linear epitope that is present only on E2 of group 1 strains.

Fig. 1.The monoclonal antibody 2B6 only recognizes a linear epitope existing on the glycoprotein E2 of group 1 classical swine fever virus strains, as shown by immunofluorescence and immunoblotting. (A) ST cells infected with individual CSFV isolates were co-immunostained by rabbit anti-C-strain serum and mAb-2B6 and then probed with Alexa-Fluor-labeled secondary antibodies (488 for anti-Rabbit and 568 for anti-Mouse). Red fluorescent signals were observed only on cells infected with group 1 strains (Shimen and C-strain) but not on those infected with group 2 strains. (B) The mAb-2B6 reacted only with purified rE2-AD proteins derived from group 1 CSFV strains (C-strain and Shimen; lanes 1 and 2 on the top panel). The anti His-tag mAb was used as the control antibody to show similar loading of different rE2-AD proteins (bottom panel).

Fine Mapping of Residues Responsible for mAb 2B6 Binding

To help identify the residues responsible for the observed 2B6 reaction patterns (Fig. 1A), C-strain-based mutant rE2-BC proteins were generated by substituting each of the 16 major residues in the corresponding region of subgroup 2.1 strains showing difference from the vaccine strain [6]. Binding efficiency of mAb-2B6 to the mutant proteins was tested by binding ELISA. Fig. 2A shows that the substitution N777S located at the C-terminus of domains B/C caused significant reduction of binding to mAb-2B6 (about 30%, as compared with wild type rE2-BC). This implies that the residues around N777 could be responsible for mAb-2B6 binding.

Fig. 2.Fine mapping of the E2 epitope recognized by the monoclonal antibody 2B6. (A) Site-directed mutagenesis was used to substitute individual amino acids in C-strain rE2-BC protein with those found at the same positions in subgroup 2.1 strains. The substituted amino acids are depicted on the X-axis. The Y-axis shows the binding efficiency of individual rE2 proteins to mAb-2B6 and anti-His mAb, respectively. The binding efficiency is relative to C-strain rE2-BC. (B) Site-directed mutagenesis was used for alanine substitution of amino acids in C-strain rE2-BC protein around N777. (C) Western blotting of alanine-substituted rE2-BC proteins probed with mAb-2B6 and anti-His mAb. (D) Eukaryotically derived C-strain rE2-BC and alanine-substituted mutant proteins were examined by immunofluorescence probed with mAb-2B6 and rabbit anti-C-strain serum (R-anti-C). The group 1 C-strain and group 2 strain HZ1-08 were used as controls. The mAb-2B6 clearly recognizes an epitope consisting of four critical residues, 774D-G-X-N-P778.

To further define the minimal region for recognition by mAb-2B6, residues around N777 (768-782) in the C-strain rE2-BC protein were mutated one by one based on Alascanning. Substitutions D774A, G775A, N777S, and P778A caused remarkable decrease in binding efficiency, whereas other substitutions did not show significant changes in binding ELISA (Fig. 2B). Western blotting further shows that all the above four substitutions did not have apparent reactivity to mAb-2B6 (Fig. 2C). A panel of alanine-mutated eukaryotic-derived E2 proteins of the C-strain were also used to mimic the natural E2 protein. IFA demonstrated that these four substitutions were not recognizable by the mAb-2B6, although all were positive to rabbit anti-C strain sera (Fig. 2D). The above results reveal that mAb-2B6 recognized an epitope consisting of four critical residues,774DGXNP778.

Effects of Single Residue Substitutions of E2 in C-Strain-Based Recombinant Viruses on Reactivity to Two mAbs

We rescued three recombinant viruses (RecC-G775A, RecC-N777S, and RecC-P778A), but failed to obtain the D774A mutant. We further examined the reaction patterns of mAb-2B6 with three rescued viruses by IFA. Rabbit anti-C-strain sera and mAb-6B8 (another E2 mAb targeted on a conformational epitope in the A/D domains) were used for comparison. All three mutant viruses did not show any reactivity to 2B6, but they exhibited similar binding with positive rabbit sera and mAb-6B8 (Fig. 3). These findings indicate that each of the three residues in the epitope774DGXNP778of the C-strain is equally important for binding to mAb-2B6.

Fig. 3.Reaction patterns of the monoclonal antibodies 2B6 and 6B8 with recombinant viruses RecC-G775, RecC-N777S, RecC-P778A, and Rec-C. ST cell monolayers were infected with 200 TCID50 virus and incubated at 37℃ for 2 days. The plates were fixed with ethanol-aceton(1:1 by volume) and differentially probed with mAb-2B6, mAb-6B8, and rabbit anti-C-strain serum.

In Vitro Replication of Mutant Viruses

Prior to evaluation of the in vitro growth characteristics of the three mutant viruses, viral stocks were prepared using repeated infections of fresh ST cells. RecC-P778A reverted to the wild-type sequence during subculture and was not used in the in vitro growth assay. RecC-G775A exhibited almost the same titers and replication ability as the wild-type C-strain, whereas RecC-N777S mutant showed significant reduction in replication (shown as TCID50, p< 0.05 by two-tailed Student’s t-test) and reduced genomic RNA copies (though not statistically significant, p > 0.05) at 96 h post infection (Fig. 4).

Fig. 4.In vitro growth characteristics of the recombinant viruses. ST cell monolayers were infected (MOI = 0.001) with RecC-G775A, RecC-N777S, or RecC. Virus samples were collected at indicated times post infection for titers (A) and RNA copies (B). Data represent means and standard deviations from three independent experiments. The difference of TCID50 between RecC and RecC-N777S was significant (p< 0.05) at 96h post infection, whereas the RNA copies between these two strains did not differ significantly (p > 0.05).

Neutralization of Mutant Viruses by Anti-C-Strain Sera

In a previous study, we found that substitution of N777S in the C-strain-derived E2 increased binding to sera from group 2 infected pigs [6]. We, therefore, wondered if mutations in the 2B6 epitope774DGXNP778 would have a similar impact on the neutralization by anti-C-strain sera. Surprisingly, serum neutralization results indicated that mutant viruses were inhibited to the same level as the wild-type C-strain (Fig. 5), suggesting that the 2B6 epitope may not be involved in viral neutralization.

Fig. 5.Neutralization analysis of the monoclonal antibody 2B6 epitope mutants RecC-G775 and RecC-N777S by pig anti-C-strain sera. Two-fold serial dilutions of two heat-inactivated sera (#467 and #516) were mixed in equal volumes of 500 TCID50 of the viruses, incubated at 37℃ for 1 h, and then transferred to confluent ST cell monolayers in 96-well plates. The starting dilution of each serum was 1:50. At 72 h post infection, the cells were fixed and immunostained for E2 glycoprotein. The wild-type recombinant virus RecC was used for comparison.

To determine whether this group 1 specific epitope could affect binding to polyclonal antibodies from C-strain vaccinated pigs, a total of seven truncated C-strain E2 proteins were generated to identify the minimal regions essential for binding to pig anti-C-strain sera and mAb-2B6 by binding ELISA. The minimal region for mAb-2B6 binding was aa773-780 on the B/C domains since deletion of the region from aa774, but not from aa781, abolished the reactivity (Fig. 6). Deletion from aa774 covering the mAb-2B6 epitope, but not that from aa781, also affected binding to the polyclonal antibodies from vaccinated pigs, although the major binding region for the vaccinated antibodies is aa690-773. The GST fusion protein containing the minimal region covering the 2B6 epitope (aa773-780) also showed no binding to polyclonal sera (Fig. 6).

Fig. 6.Definition of the minimal domain of the C-strain E2 protein critical for binding by the monoclonal antibody 2B6 and pig anti-C-strain serum. Truncated E2 proteins fused with GST were generated as described in Materials and Methods. Reactivity of the fusion proteins with mAb-2B6 or anti-C-strain sera was analyzed by ELISA and compared with the un-truncated fragment aa690-814 set as 100%.

 

Discussion

The E2 glycoprotein of CSFV induces neutralizing antibodies and protects swine from lethal infection [7,12,16,32,36]. Genetic analysis has revealed that the N-terminal half of E2 is more variable and may be responsible for its antigenic difference among three groups of CSFV strains [4,6]. The mAbs that could differentially recognize the N-terminal B/C domains from different groups of CSFV strains have the potential for distinguishing the vaccine C-strain and field isolates [2,3]. Thus, fine mapping of the epitopes in the N-terminal region could provide useful information not only for a better understanding of the relationship between the E2 structure and antigenic variability of CSFV isolates, but also for development of serological differential assays [9,10,16,24,34].

Our previous study showed that the mAb-2B6 generated from the B/C domains of C-strain E2 exhibited strong binding affinity only to E2 of group 1 strains, but not those of group 2 isolates, as shown by ELISA and western blotting [6]. Since 2B6 was generated from mice immunization with the prokaryotically expressed E2 protein, its target epitope may not have post-translational modifications as commonly seen in eukaryotic cells [24]. Further mutation of either one of the cysteine residues in the B/C domains did not affect its reactivity to E2 expressed in eukaryotic cells, confirming that 2B6 recognizes a linear epitope unique in E2 of group 1 strains.

Site-directed mutagenesis was used to replace 16 major amino acids of the E2 that are different between the C-strain and subgroup 2.1 strains. We found that one residue replacement, N777S, remarkably reduced the binding of 2B6 to the C-strain B/C domains. Further alanine-scanning mutations around N777 demonstrated that residues 774, 775, 777, and 778 significantly reduced the reactivity of 2B6 with the mutant E2 proteins expressed in eukaryotic cells (shown by immunofluorescence) and in prokaryotic cells (shown by western blotting), whereas substitution of the residue 776 did not alter the reaction pattern. Thus, we fine-mapped an epitope,774DGXNP778, located at the border between the B/C domains and domain A.

Alignment of the sequences aa690–800 of E2 from multiple CSFV isolates revealed that the mAb-2B6 epitope exists only on E2 of group 1 strains. The residue N777 is highly conserved within group 1 strains but different in those of group 2 (Fig. S1), implying its potential role as a major determinant of the antigenic variations. This is in general agreement with an earlier report that the residue 777 in the E2 epitope recognized by the mAb HQ06 did significantly affect its reactivity to different CFSV isolates [24], although they found that mutations of other residues such as aa773, 775, and 778 completely abolished the reactivity. Earlier reports suggested that different CSFV isolates could be typed by a panel of mAbs based on their binding profiles [11], although this approach might be impractical because of the need of quite a number of mAbs. A common conformational epitope was mapped at the N-terminal 90 residues between aa690 and aa779 on the B/C domains by using mAbs against the group 2.1 and group 1 vaccine strains, and the region aa770 to 779 was found essential not only for the structural integrity of conformational recognition (771LLFD774) but also for specific binding and possibly for differentiation between group 1 and group 2 strains [2]. Interestingly, our mAb-2B6 linear epitope falls within this region. These results suggest that group 1 and group 2 strains might be differentiated with only two mAbs.

To understand if these mutations affect the binding of the mAb-2B6 to E2 of different CSFV isolates, site-directed mutagenesis of the 2B6 epitope was performed in the C-strain-based infectious clone. Of the four mutations, only three recombinant mutant viruses were rescued (RecC-G775A, RecC-N777S, and RecC-P778A). Reactivity to mAB-2B6 was lost in all mutant viruses, as evidenced by IFA. To our surprise, mutation P778A affected not only the immune reactivity but also the yield of the mutant progeny virus. Continuous passaging of the mutant virus P778A led to reversion to proline at the same position, thus rendering us unable to obtain enough virus stock for in vitro growth analysis. The mutant virus N777S also had poor replication potential, as compared with the wild-type virus. This is probably due to changes of viral E2 in infectivity because E2 is an important protein involved in viral attachment and/or entry [37]. Suggestive of this were less populous fluorescence plaques with smaller sizes during early passages of the mutant virus N777S, compared with wild-type C-strain (Fig. 3). Similar results were seen as reduced plaque sizes in continuous in vitro passage of the CSFV strain Brescia, due to a single amino acid mutation in the Erns protein [30].

Viral neutralization showed that there were virtually no apparent changes of neutralization activity of the hyper-immune swine sera on the mutant viruses. This may indicate that the mAb-2B6 epitope is not immune-dominant during natural infection, and sera from vaccinated or infected pigs contain less IgG targeting this epitope. Thus, substitution of those residues in the epitope did not have any impact on virus neutralization, because viral isolates might be overwhelmingly neutralized by IgG targeting other immunodominant epitopes in E2.

Minimal peptide binding tests showed that the C-strain-positive serum did not bind to individual GST-fused 773-780 proteins but exhibited reactivity with N-terminal 690-780 of E2 fused with GST-tag. This is consistent with an early study that showed that the N-terminal residues 690-812 contain several essential sites for serum binding, and deletion of either end of the peptide 690-779 abolishes the reactivity [13]. However, it should be noted that deletion of the 2B6 epitope within N-terminal 690-780 of E2 apparently reduces the binding of the anti-C-strain pig serum to the GST-fusion peptide 690-773. These data suggest that the mAb-2B6 epitope774DGXNP778, although not involved in neutralization, may serve as a critical element to maintain the conformational structure of the B/C domains that affect binding to polyclonal sera in a way similar to a previously reported motif (771LLFD774) that was found to affect the reactivity of a panel of mAbs specific to the E2 B/C domains [2].

In conclusion, we identified a specific epitope,774DGXNP778, critical for mAb-2B6 recognition at the protein (prokaryotically and eukaryotically expressed) or whole-virus level. The mutant virus RecC-N777S exhibited low replication potential. Since the epitope is conserved in group 1 CSFV strains, it may be utilized as a marker for differentiation between group 1 and group 2 CSFV isolates.