The insect cell lines, Spodoptera frugiperda 9 (Sf9) and High Five were cultured at 27°C in medium IB905 (Zhejiang Yishengke Biotechnology Co., Ltd., China). Murine SP2/0-Ag14 (SP2/0) myeloma cells were maintained in RPMI-1640 medium (Thermo Fisher Scientific Inc.) at 37°C with 10% fetal bovine serum (FBS, Thermo Fisher Scientific Inc.) and 5% CO₂.

Horseradish peroxidase (HRP) conjugated mouse anti-His IgG or sheep anti-mouse IgG, and ×2 High-Fidelity PCR Master Mix were purchased from Sangon Biotech Co. Ltd. (Shanghai, China); SuperScript™ IV CellsDirect™ cDNA Synthesis Kit was purchased from Thermo Fisher Scientific Inc.

According to the instruction manual of the “Bac-to-Bac^® Baculovirus Expression System” (Thermo Fisher Scientific Inc.), the genes of the recombinant proteins, E2ab-2.1 and E2ab-C (Figure 1A) were inserted respectively into the plasmid, pFastBac™1 through the restriction endonuclease sites BamHI and XhoI. These plasmids were used to transfect Sf9 cells for generation of the baculoviruses, which were used to infect High Five cells for expression of the recombinant proteins. The High Five cells were lysed and centrifuged to obtained expressed proteins, E2ab-2.1 or E2ab-C, in the pellets. The pellets were washed using a buffer solution containing 50 mM Tris, 150 mM NaCl, 3 M guanidine hydrochloride, 1 mM EDTA, and 1% Triton X-100. The washed pellets were dissolved with denaturing buffer containing 6 M guanidine hydrochloride. The proteins were recovered and concentrated by gradually removing guanidine hydrochloride with 10KD ultrafiltration tubes (Amicon Ultra-15 Centrifugal Filter, Millipore).

The 6 weeks-old female BALB/c mice were immunized by subcutaneous injection of the purified protein E2ab-2.1 (10 μg per mouse), which was emulsified with Freund’s complete adjuvant. After 4 weeks, the mice were injected with the same dose of the protein emulsified with Freund’s incomplete adjuvant, followed by an additional intravenous injection of the same dose after additional 3 weeks. Subsequently, another week later, hybridoma cells were produced by sacrificing the mice and fusing their spleen cells with the SP2/0 cells with the conventional method (Cho et al., 2005).

To obtain monoclonal antibodies, the hybridoma cells were cultured in 96-well plate until single cell clones were obtained. The culture supernatants were tested for the presence of antibodies against E2 using indirect ELISA. The clones expressing E2 specific antibodies were further subcloned and expanded. The mAbs were purified from the culture supernatant using protein A/G agarose beads (Beyotime Biotechnology Co., Ltd., China).

To identify the binding sites of the mAbs on E2 protein, indirect ELISA was performed by using the different E2 proteins or peptides. Briefly, 96-well plates were coated with the protein E2ab-2.1 or E2ab-C (10 μg/mL), or chemically synthesized peptides (20 μg/mL) at 4°C overnight. The coated plates were blocked with 5% skimmed milk for 2 h at 37°C followed by incubation with the hybridoma cell culture supernatant, or purified mAbs (1 μg/mL), or pig serum (diluted 1:100) for 1 h at 37°C. The plates were washed with PBS and incubated with HRP-conjugated goat anti-mouse or anti-pig IgG secondary antibodies for 30 min at 37°C. Finally, the TMB substrate solution was added to the plates and the absorbance was measured at 450 nm. The statistic analysis of the binding affinity was performed using the OD450 values.

The classes of the mouse mAbs (isotypes of heavy chain and light chain) were identified using a commercial indirect ELISA kit (Wuhan Chundu Biotechnology Co., Ltd., China). The assay was performed according to the manufacturer’s instructions.

To check the purified recombinant proteins, the samples were separated in 10% SDS-PAGE gels and transferred to polyvinylidene fluoride (PVDF) membranes (Bio-Rad, United States). The membranes were blocked with 5% skimmed milk, probed with mouse anti-His antibody (1:5,000 dilution) (Sangon Biotech Co. Ltd., Shanghai, China), mAb MM1 (1 μg/mL), or mAb MM5 (1 μg/mL) followed by HRP-conjugated goat anti-mouse IgG secondary antibody (Sungene Biotechnology Co., Ltd., Shanghai, China). The enhanced chemiluminescent (ECL) reagent (NCM Biotechnology Co., Ltd., Suzhou, China) was added to capture the images with MiniChemi 610 chemiluminescent imaging system (SinSage Technology Co., Ltd., Beijing, China).

To prepare cDNA that encodes the variable regions of the mAbs, hybridoma cells were collected at the logarithmic growth phase by removing the culture medium, washing cells once with pre-cooled PBS, and resuspending with pre-cooled PBS. The cell concentration was adjusted to 2 × 10⁶ cells/mL for mRNA isolation and cDNA synthesis according to the instruction manual of the SuperScript™ IV CellsDirect™ cDNA Synthesis Kit.

To design the amplification primers, multiple mouse immunoglobulin gene sequences were downloaded from the GenBank. The gene sequences of the variable regions were confirmed as shown in Figure 4A by aligning these sequences with each other. The GenBank codes of the sequences used for design of the heavy chain primers were X70423.1, BC092295.1, BC092271.1, BC057672.1, BC018280.1, and BC003878.1, and for light chain primers were DQ078272.1, LC522515.1, MH208237.1, LC199874.1, LC026057.1, BC080787.1, BC094013.1, BC091750.1, BC028540.1, BC019474., and BC002112.1.

The primer locations on the gene sequences were as shown in Figure 4A. Since the nucleotide sequences of the leader region within the variable regions were different between the mouse immunoglobulin gene sequences, multiple upstream primers were included along with a downstream primer in each PCR reaction. The upstream primers for the heavy chain included three oligonucleotides, Fh-1: 5′-ATG​GGA​TGG​AGC​TGT​ATC​ATC​C-3′, Fh-2: 5′-AGG​AAC​TGC​AGG​TGT​CC-3′, and Fh-3: 5′-CAG​CTA​CAG​GTG​TCC​ACT​CC-3′, with a downstream primer, Rh: 5′-AGA​AGG​TGT​GCA​CAC​CGC​TGG​AC-3′. The upstream primers for the light chain included two oligonucleotides, Fκ-1: 5′-CAGTTCCTGTTTCTGTTARTGCTCTGG-3′ and Fκ-2: 5′-TGG​GTG​CTG​CTG​CTC​TGG​GT-3′, with a downstream primer, Rh: 5′-AGA​AGG​TGT​GCA​CAC​CGC​TGG​AC-3′. The resulting PCR products were sequenced, and corresponding amino acid sequences were analyzed to confirm the framework regions (FR) and the complementary determining regions (CDR).

Similar to the indirect ELISA, 96-well plates were coated with polypeptides 2.1-1 or 2.1-7 listed in Table 1 (7 μg/mL, 50 μL/well) at 4°C overnight. The plates were washed and incubated with 80 μL/well pig serum (1:100 dilution) followed by 100 μL/well MM5 antibody (1 μg/mL) or 100 μL/well MM12 antibody (1 μg/mL) for 1 h at 37°C. Subsequently, the plates were incubated with HRP-conjugated goat anti-mouse secondary antibody for 30 min at 37°C. The TMB substrate solution was added to the plates, and the absorbance was measured at 450 nm. The data were analyzed same as indirect ELISA.

The animal studies were approved by Ethics Committee of Guangdong Haid Group Co., Ltd. (HD-211110-1, 07 December 2021).

As shown in Figures 1A–C, the E2 AB domains from the 2.1 strain were selected as the immunogen. To keep a natural E2-like structure, the third cysteine was substituted with serine, which would remove possibilities of forming disulfide bonds with other cysteines in the recombinant protein (Chang et al., 2012a; Chang et al., 2012b). The immunogen was designated as E2ab-2.1, while the control protein was designated as E2ab-C, which contained E2 AB domains from the C strain without a mutation. The purified recombinant proteins E2ab-2.1 and E2ab-C were verified by Western blotting, the results indicated that these two proteins had similar molecular weight, about 20 kDa (Figure 1D).

Hybridoma cells were generated by fusion of spleen cells from E2ab-2.1 immunized mice with SP2/0 cells. Twenty hybridoma cell lines were screened after subcloning, which expressed mAbs that were named as MM1-MM6, MM8-MM15, and MM17-MM22. Indirect ELISA results showed that four mAbs, MM1, MM5, MM18, and MM20 only recognized E2ab-2.1, but not E2ab-C (Figure 1E). The class testing for these mAbs revealed that the heavy chain of MM1, MM18, and MM20 were IgG2a, while the heavy chain of MM5 was IgG1. The light chain was the κ chain for all four antibodies (Figures 1F, G). Remaining 16 mAbs recognized both the E2ab-2.1 and E2ab-C, which suggested that they were unable to distinguish whether E2-containing samples came from the 2.1 strain or the C strain of CSFV.

To determine the antigenic epitopes recognized by the mAbs, seven overlapped polypeptides for the 2.1 strain and five for the C strain were designed and synthesized (GL Biochem Ltd., Shanghai, China) based on the amino acid sequences of the respective E2 AB domains (Figure 2A). The amino acids sequences for these peptides are indicated in Table 1.

The data from the indirect ELISA confirmed that the four 2.1 strain-specific mAbs only bound to 2.1 strain-derived polypeptides. Specifically, MM1, MM18, and MM20 only recognized peptide 2.1-2, and MM5 only recognized peptide 2.1-1 (Figure 2B). The rest of the 16 mAbs, which bound to both the 2.1 strain and C strain derived E2 recombinant proteins, recognized peptides (2.1-7 and C5) from both the strains (Figures 2C, D).

Among of the 2.1 strain-specific mAbs, MM1 displayed the lowest binding affinity to C strain, and MM5 was the only mAb with the heavy chain that belonged to the IgG1 subtype (Figures 1E, F), therefore MM1 and MM5 were selected for further epitope analysis.

To determine the epitopes of MM1 and MM5, the non-conservative residues in the peptides 2.1-1 and 2.1-2 were individually mutated to the same residues as those in the C strain (Figures 3A–D). The data from ELISA demonstrated that the substitutions of the residues “709P” to “L” in peptide 2.1-1-A, or “713E” to “G” in peptide 2.1-1-B, significantly reduced the binding affinity to MM5, but no reduction in affinity was observed with the substitution of “705N” to “D” in peptide 2.1-1-D. Therefore, residues “709P” and “713E” play critical roles in the epitope of MM5, especially the residue “713E”, as the binding affinity was reduced by 95% with the E713G substitution (Figures 3A, B).

Similarly, three substitution mutations (R720K, G725D, and D729N) in the variant peptide 2.1-2-B resulted in a complete loss of binding affinity to MM1 antibody. The substitutions in the last two non-conservative residues “734R” and “738T” to “K” and “V”, respectively, in the variant peptide 2.1-2-A, did not affect the binding affinity (Figure 3C).

As shown in Figure 3D, peptide 2.1-2-C was a shorter version of the peptide 2.1-2. To detect the critical residues which were responsible for MM1 recognition, four non-conservative residues in the peptide 2.1-2-C were mutated one by one to the respective residues in the C strain. The data from ELISA indicated that both the residues “723S” and “725G” were very critical in the binding of MM1 to the peptide 2.1-2-C, and replacing any of them, such as peptides 2.1-2-F or 2.1-2-F, could lead to disruption of the interaction between MM1 and peptide 2.1-2-C. Additionally, the replacement of residue “729D” with “N” (peptide 2.1-2-G) also slightly reduced the MM1 binding. These data suggest that the critical binding site for MM1 should be residues 723S, 725G and 729D on the E2 protein of the 2.1 strain.

The Western blotting results also indicated that MM1 and MM5 were able to specifically detect E2 protein derived from the 2.1 strain, but not from the C strain (Figure 3E).

The 3D model of the BVDV E2 was used to simulate structure of the CSFV E2 protein, which indicated that the locations of the MM1 and MM5 epitopes are at two different loops in the domain A (Figure 3F).

The immunoglobulin variable region genes of the mAbs, MM1 and MM5 were amplified from the hybridoma cells (Figure 4A). Agarose gel electrophoresis results indicated that the PCR products of the heavy chains were about 550 bp and of the light chains were about 400 bp (Figure 4B). The gene fragments were sequenced and translated into amino acid sequences, from which three complementary determining regions of each the heavy chains and the light chains were identified (Figure 4C). The detail of the gene sequences and the amino acid sequences can be found in the Supplementary Data S1 (Monoclonal antibody MM1) and Supplementary Data S2 (Monoclonal antibody MM5).

To further prove the recognition specificity between the mAbs and their epitopes, the pig serum was generated by immunization of pigs with the 2.1 strain derived E2 protein (Figure 5A). As shown in Figure 5B, the pig serum presented much higher binding affinities to the peptides 2.1-1 and 2.1-7 than the peptide 2.1-2. Since the peptide 2.1-1 contained the epitope of MM5, and the peptide 2.1-7 exhibited the best binding affinity to MM12 (Figures 2C, D), the binding specificity of these peptides were tested by the competitive ELISA using pig serum followed by MM5 and MM12 antibodies. The results indicated that the binding affinity of MM5 or MM12 to the peptides were significantly reduced by pre-incubation with the pig serum (Figure 5C).

As shown in Table 2, the Supplementary Data S3 E2 sequences downloaded from GeneBank at NCBI, there were total 100 sequences for E2 protein in the genotype 1 collected from 1945 to 2021, and only one of them contained the MM5 epitope, which was isolated 30 years ago. For the genotype 2, there were in total 353 sequences for E2 protein, which belonged to seven sub-genotypes (2.1–2.7). Interestingly, 226 of 353 sequences were from sub-genotype 2.1, out of which, 215 (95.1%) sequences contained MM5 epitope. The sub-genotypes 2.2, 2.3, 2.4, 2.5, and 2.7 also had the MM5 epitope in some sequences (15.2%, 5.6%, 100%, 66.7%, and 100%, respectively), but not in sub-genotype 2.6. In addition, there were total 16 isolates in genotype 3, where only 3 of the 9 sequences from sub-genotype 3.2 had MM5 epitope.

In addition, beside the genotype 2, none of the E2 sequences from genotype 1 and genotype 3 contained the MM1 epitope. The MM1 epitope tended to exist in the sub-genotypes 2.2 and 2.3, rather than 2.1. However, if the E2 sequences from genotype 2 that contain either MM1 or MM5 epitopes counted together, the percentage of these sequences reached up to 90.8%.

The sub-genotype 2.1 predominated in world, especially in China since last 20 years (Table 2, Figure 6, Supplementary Table 1). There were total 212 sequences for E2 protein uploaded to GenBank from China from all the genotypes, where 163 (>76%) sequences belonged to sub-genotype 2.1. Out of these 163 sequences, 158 sequences (96.9%) contained the MM5 epitope. The remaining 49 sequences out of 212 were from the sub-genotypes 1.1, 2.2, 2.5, and 3.4. Notably, all the seven sequences from the sub-genotype 2.5 also contained the MM5 epitope.

Moreover, it has been reported that closely related viruses, BVDV and BDV share the same epitopes with CSFV (Huang et al., 2021). However, there are only nine E2 sequences that were derived from the pig originated BVDV, and four E2 sequences from pig originated BDV in GeneBank, and the MM5 epitope was absent in the BVDV E2 sequences, but present in two of the BDV sequences. In addition, there was no MM1 epitope (SHGLQLD) present in the BDV or BVDV E2 sequences, (Supplementary Data S3 E2 sequences downloaded from GeneBank at NCBI).