Escherichia coli DH5α and BL21(DE3) were purchased from Takara Bio Inc (Beijing, China) and grown in Luria–Bertani medium. H. parasuis DT3 and ZS7 strains (Laboratory collection) were isolated from the lung tissue and nasal cavity of different diseased pigs in southern China, respectively; and were grown at 37°C in an aeroculture in tryptic soy broth and tryptic soy agar, supplemented with 50 μg/mL nicotinamide adenine dinucleotide (NAD) and 5% inactivated equine serum.

PAMs (3D4/21, ATCC, Manassas, VA, USA) and SP2/0 myeloma cells were purchased from Otwo Biotech Inc. (Guangdong, China). PAMs were cultured in Dulbecco’s Modified Eagle’s medium (DMEM) containing 10% fetal bovine serum (FBS) and 1 × penicillin–streptomycin solution (PS). SP2/0 cells were grown in RPMI 1640 medium supplemented with 10% FBS and 1 × PS. All cells were grown at 37°C in a 5% CO₂ atmosphere. All the cell culture media and supplements were purchased from Thermo Fisher Scientific (Waltham, MA, USA). The cells were subcultured until they reached an 80%–90% confluence.

Recombinant truncated OmpP2 (rOmpP2) without the signal peptide was expressed using the prokaryotic expression vector pET-28a. Different rOmpP2 genes were amplified from purified genomic DNA of clinical samples (nasal swabs, lungs, lymph nodes, joint effusion, etc.) from healthy or diseased pigs, with the primer set 5′-CATGCCATGGTAACAGTTTATGAAAATGAAGGT-3′ and 5′-CCGCTCGAGCCATAATACACGTAAACC-3′, to ensure that a 6 × His-tag was attached to the C-terminal. The expressed fusion proteins were located within inclusion bodies. The inclusion bodies were washed, the protein solubilized, and then refolded by dialysis against a decreasing concentration gradient (6 M, 4 M, 2 M, 0 M) of urea at 4°C. The recombinant proteins were separated by 10% sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) after concentration by ultrafiltration (Millipore, Billerica, MA, USA). They were identified by Coomassie bright-blue staining or western blotting (WB) using an anti-histidine tag antibody (Sangon Biotech, Shanghai, China). Protein concentration was determined using the BCA Protein Assay Kit (Sangon Biotech), and protein purity was automatically calculated by the image tools of GelDoc XR+ IMAGELAB (BioRad, Hercules, CA, USA). rOmpP2 proteins were used as antigens for mouse immunization and serological detection.

Native OmpP2 (nOmpP2) proteins were extracted from the H. parasuis DT3 and ZS7 strains as previously described (Zhang et al., 2012b; Zhou et al., 2014), and were confirmed to be genotype I by DNA sequencing.

To locate the B cell epitopes of different genotypes of OmpP2, three groups of overlapping peptides ( Table 1 ) were designed and synthesized by Sangon Biotech. In the first group, twenty-one 21–24-mer peptides with 8-mer overlaps were designated as Pt1 to Pt21, based on the full-length gene of genotype I rOmpP2 (the synthesis of Pt10 failed). In addition, three peptides (Pt22, Pt23, and Pt24) were designed based on the insertion sequence of genotype II rOmpP2. In the second group, eight peptides were designed based on the predicted linear B-cell epitopes and loops (epitope prediction: IEDB software, http://tools.iedb.org/bcell/; loop prediction: PRED-TMBB software, http://bioinformatics.biol.uoa.gr/PRED-TMBB/input.jsp) of the genotype I rOmpP2 protein, to complement the first group and to cover all the predicted epitopes and loops. In the third group, based on the peptide–enzyme-linked immunosorbent assay (ELISA) results from the first and second groups, nine peptides were designed to confirm the position or genotype-specificity of the epitope. All peptides were 90% pure and were dissolved in sterile water (poorly soluble peptides were completely dissolved by ultrasound treatment then filtered through 0.22 μm filters).

To obtain mAbs against different genotypes of the OmpP2 protein, two mouse immunization groups were set up and immunized with rP2-I2 and rP2-II. Three 6-week-old female BALB/c mice were used in each group. The first subcutaneous injection contained 100 μg rOmpP2 protein emulsified in equal proportions with Freund’s complete adjuvant (Sigma–Aldrich, St Louis, MO, United States). Two weeks later, the second subcutaneous injection, containing 50 μg rOmpP2 protein emulsified in equal proportions with Freund’s incomplete adjuvant (Sigma–Aldrich) was administered. Immunizations were performed two or more times at 2-week intervals, depending on the antibody titer, which was assessed by indirect ELISA using rOmpP2 as the antigen. If the antibody titer exceeded 12,800, the final boost immunization, of 50 μg of rOmpP2 in phosphate-buffered saline (PBS), was intraperitoneally injected.

Three days after the final immunization, spleen cells were aseptically harvested and fused with SP2/0 cells at a ratio of 3:1 using polyethylene glycol 1450. Supernatants from the antibody-producing hybridomas were screened by indirect ELISA using wells coated with the antigens injected into each immunization group. Positive hybridomas were selected and subcloned three times using limited dilution. All the clones from the same fusion cell well were treated as positive hybridoma cell lines.

To screen for positive hybridoma cell clones and linear B cell epitopes, indirect ELISA was performed as previously described (Qiu et al., 2012), with some modifications. Briefly, immuno-microtiter plates were coated with 5 μg/mL of peptides or 3 μg/mL of proteins in carbonate-buffered saline (pH = 9.6) at 4°C overnight. Plates were washed three times with PBST (0.01 M PBS, pH 7.2; 0.05% Tween 20) and blocked with 4% bovine serum albumin at 37°C for 2 h. Hybridoma cell supernatant (1:1) and serum (1:50 to 1:102,400) were diluted and added to the plate in a volume of 100 μL/well and incubated at 37°C for 1 h. After washing five times with PBST, horseradish peroxidase (HRP)-conjugated rabbit anti-mouse IgG (1:10,000 diluted) was added as the secondary antibody and plates were incubated at 37°C for 45 min. After washing five times with PBST, 100 μL chromogenic solution (TMB, Sangon Biotech) was added and incubated at 37°C for 20 min. The absorbance was measured using a microplate reader at an absorption of 450 nm. An mAb to the P30 protein of African swine fever virus (ASFV) was used as a negative control, and test results were calculated as the sample to cut-off (S/CO) ratio: S/CO value = sample detection value/cutoff (CO) value, where the CO value = 2.1 × negative-control detection value. An S/CO ratio ≥ 1.0 was considered as positive.

rOmpP2 and nOmpP2 were separated by SDS-PAGE and transferred to polyvinylidene fluoride membranes. After blocking overnight at 4°C in 5% skim milk, the protein-bound membranes were incubated with the diluted supernatants of mAbs for 1 h at room temperature. After washing, the membranes were incubated with HRP-conjugated rabbit anti-mouse IgG (1:10000 diluted) for 1 h at room temperature. Finally, the proteins were detected by adding an enhanced chemiluminescence reagent (Thermo Fisher Scientific). The P30 protein of ASFV was used as a negative control.

The nOmpP2 genes of the DS3 and ZS7 strains and the rOmpP2 genes were amplified and sequenced using the primers described above. Thirty OmpP2 protein sequences of 15 serovars from H. parasuis reference strains (15 international reference strains and 15 Chinese reference strains) were obtained from Genbank (Additional file 2). Sequence alignment was performed using MEGA 5.1 software, and the sequence consistency, mutation sites of each epitope locus, and the relationships among the epitope, genotype, and pathogenicity were analyzed. If an epitope existed only in a single genotype, it was defined as a GS epitope, and the mAb that interacted with it was defined as the GS-antibody (GS-Ab). If the epitope was present in both genotypes, it was defined as the GC epitope, and the mAb that interacted with it was defined as the genotype-common antibody (GC-Ab).

A total of 187 swine serum samples were collected from nine farms and epitope peptide-ELISA was performed to analyze the clinical positivity rate of antibodies to each epitope. The ELISA protocol was similar to that previously described. Swine serum (diluted 1:400) was used as the primary antibody, and HRP-conjugated rabbit anti-pig IgG (Sangon Biotech; diluted 1:10,000) was used as the secondary antibody. The incubation time with the secondary antibody was reduced to 30 min, the incubation time with TMB was reduced to 10 min, and the other procedures remained unchanged. A mock control, without an antigen coating, was set up for each serum sample, and serum from newborn swine born to negative sows was used as the negative control. The S/CO value was calculated as follows: sample OD₄₅₀ value (S) = OD₄₅₀ for serum with peptides - OD₄₅₀ for serum with mock; CO value = 2.1 × (OD₄₅₀ for negative control with peptides - OD₄₅₀ for negative control with mock). An S/CO ratio ≥ 1.0 was considered as positive. The 187 clinical serum samples were detected in parallel using a commercial H. parasuis antibody ELISA kit (Shenzhen Ziki Biotech Co., Shenzhen, China), which used a one-step assay with the serum diluted 1:15. We then compared the results of the commercial kit with those of the epitope peptide-ELISA to analyze the potential of the epitopes in clinical diagnosis. Finally, 20 double-positive serum samples (positive by commercial ELISA and epitope peptide-ELISA) were randomly selected and used as primary antibodies to perform an indirect ELISA with the remaining peptides to screen for missed epitopes in the mouse experiments by using polyclonal antibodies in pig serum.

Log-phase growing PAMs were seeded in 24-well plates at a density of 2 × 10⁵ cells/well and cultured overnight. When the cultures reached confluence, the cells were stimulated in triplicate with 130 nmol/mL peptides (diluted with serum-free DMEM) for 12 h. Untreated cells (cultured in serum-free DMEM) were used as mock controls in each assay. Following treatment, the cells were lysed and total RNA purified using the TRIzol method. cDNA was synthesized using the Evo M-MLV RT Kit with clean gDNA (Accurate Biology, Hunan, China). Real-time polymerase chain reaction (PCR) was performed to determine the relative quantities of IL-1α, IL-1β, IL-6, IL-8, TNF-α, and GAPDH mRNA, with SYBR Green Premix Pro Taq HS qPCR Kit (Accurate Biology) and ABI 7500 Real-Time PCR System (Applied Biosystems, CA, USA). The primers ( Table S1 ) used for RT-PCR were synthesized by Sangon Biotech, Inc. Each sample was assessed in triplicate, in at least three independent experiments. Relative expression of the cytokine genes was normalized to the internal control gene (GAPDH), analyzed using the 2^-ΔΔCT method, and was presented as the fold-change relative to the mock control.

Statistical analysis of the data was conducted by one-way analysis of variance (ANOVA) using SPSS v16 software (IBM SPSS Inc., Armonk, NY, USA). The post-hoc Duncan test was used to identify differences between group means. p < 0.05 were considered significant.

Three different ompP2 genes of genotype I were amplified from clinical samples of diseased pigs, and one ompP2 gene of genotype II was amplified from a nasal swab of a healthy pig. An expression vector containing any one of the four ompP2 genes was used to express heterologous proteins in E. coli BL21 (DE3) cells. Western blotting with an anti-His tag antibody confirmed that all four heterologous proteins contained a His tag ( Figure 1A ), indicating that the proteins were, as expected, rOmpP2 proteins. We named the three proteins of genotype I rP2-I1, rP2-I2, and rP2-I3 and the protein of genotype II as rP2-II. As expected, the apparent molecular weight of rOmpP2 proteins was 43–44.3 kDa, and the masses of the three genotype I proteins were similar and slightly smaller than that of the rP2-II protein. After being concentrated and verified by SDS-PAGE ( Figure 1B ), the concentrations of rP2-I1, rP2-I2, rP2-I3, and rP2-II were 2.9 mg/mL, 5.8 mg/mL, 5 mg/mL, and 1.9 mg/mL, respectively, with a purity of 90%, 94.6%, 93.2%, and 88.4% respectively. These results indicated that they recombinant proteins were adequate for use as immunogenic antigens. rP2-I2 (the highest and purest OmpP2 content among genotype I samples) and rP2-II were used for mouse immunization.

After the second immunization, the antibody titers of mice immunized with rP2-I2 and rP2-II were 51,200 and 25,600, respectively. The mice were then sacrificed and their spleens were harvested for cell fusion. After three subcloning steps, 19 and 13 mAbs (32 in total) against OmpP2 were established by indirect ELISA and were selected from the rP2-I2 and rP2-II immunization groups, respectively. The reactivity and specificity of each mAb were confirmed by indirect ELISA and WB analysis with the four rOmpP2 proteins ( Figure 2 ). Indirect ELISA revealed that 81.25% of the mAbs (rP2-I2, 15/19; rP2-II, 11/13) reacted with the four rOmpP2 proteins, showing that most of the mAbs could recognize genotype I and II OmpP2 proteins in the tertiary structure, which were defined as GC-Abs. Additionally, mAbs 2C5, 1A10, 4H3, and 4E4 from the rP2-I2 immunization group only recognized one or more proteins from genotype I, which were defined as GS-Abs of genotype I, accounting for 21.05% (4/19) of the mAbs from the rP2-I2 immunization group. In contrast, mAbs 2F1 and 5F1 from the rP2-II immunization group only recognized the rP2-II protein of genotype II, which was defined as a GS-Abs of genotype II, and accounted for 15.38% (2/13) of the mAbs from the rP2-II immunization group. The genotypes of the proteins recognized by the antibody used in the WB assay were consistent with the ELISA results. WB analysis indicated that 84.38% (27/32) of the mAbs could recognize the primary (linear) structure of the OmpP2 protein, suggesting that these mAbs reacted with linear B-cell epitopes. In addition, four mAbs (4E4, 1F8, 3B1, and 6G5) from the rP2-I2 immunization group and mAb 3D9 from the rP2-II immunization group recognized only the tertiary structure, accounting for 21.05% (4/19) and 7.69% (1/13) of the mAbs in the rP2-I2 and rP2-II immunization groups, respectively, indicating that these mAbs reacted with conformational B-cell epitopes.

To confirm mAb recognition of the epitopes, 27 mAbs were detected for reactivity with 29 overlapping peptides in groups A and B. The results ( Figure 3 ) indicated that 10 mAbs were positive for one or two peptides, including mAbs 1A10, 2C5, 1E6, 2C11, 6E8, and 2B8 from the rP2-I2 immunization group, and 2E2, 1D2, 3A3, and 5F9 from the rP2-II immunization group.

Among the GC-mAbs, mAbs 2E2 and 1D2 reacted with Pt1 and Pt1a, and the epitope recognized was the overlapping region Pt1a: VTVYENEGTKVDFDGQ (rP2-I2, 1–16 aa). mAb 3A3 reacted with Pt7 and Pt7a, and the recognized epitope was the overlapping region GKQAVIGDSIGQAGFDKVYG (rP2-I2, 101–120 aa). mAb 5F9 reacted with Pt9. We then further truncated Pt9 to Pt9a and Pt9b, and found that mAb 5F9 reacted with Pt9a but not with Pt9b; thus, the key region of epitope of 5F9 was GFDILTA (non-overlapping region of Pt9a and Pt9b, rP2-I2, 137–143 aa). Our results confirmed that the epitope recognized by 5F9 resides in Pt9a: GFDILTASSDSAINYT (rP2-I2, 137–152 aa). mAb 2B8 reacted with Pt17, and the recognized epitope was IDFVRTGARFDVTPKSGVYGNYSY (rP2-I2, 257–280 aa). mAbs 1E6, 2C11, and 6E8 reacted with Pt19 and Pt19a, and the epitope recognized was the overlapping region YKLHKQVVTFVE (rP2-I2, 301–312 aa).

The GS-mAbs of genotype I 1A10 reacted with Pt5, but not with Pt5-I3 and Pt5-II, which are peptides in the same region (65–88 aa) of the rP2-I3 and rP2-II proteins, respectively. However, the GS-mAbs of genotype I 2F1 reacted with Pt5-II, but not with Pt5-I3 and Pt5. We then synthesized the truncated peptides, Pt5a and Pt5b, based on the mutation sites in Pt5, and found that neither 1A10 nor 2F1 reacted with these peptides. These results confirmed that the epitope recognized by 1A10 resides in Pt5: YETRLGRNSKNDAGWGDVTTDEAY, while the epitope recognized by 2F1 resides in Pt5-II: YETRLGSGSKNAAKWGDVTTDEAY.

In addition, The GS-mAbs of genotype I 2C5 reacted with Pt11 and Pt11a but not with the truncated peptides Pt11b and Pt11c; therefore, the epitope recognized was the overlapping region of Pt11 and Pt11a, NYNVANERDNKGEVKVDSTKS (rP2-I2, 164–184 aa). Then, Pt11a-II was synthesized (the homologous region of the rP2-II protein). We found that the GS-mAbs of genotype II 5F1 reacted with this peptide, indicating that the epitope recognized by 5F1 resided in the Pt11-II, NYNVANEREKADVKVDSIKS (rP2-II, 180–199 aa).

We compared the sequences of the 36 OmpP2 proteins and showed that the consistency of all sequences was 85.62%, and there was significant heterogeneity between the two genotypes. The epitope sequence alignment results were analyzed in combination with the ELISA results ( Figure 4 ).

In terms of the GC-mAbs, the epitopes VTVYENEGTKVDFDGQ (rP2-I2, 1–16 aa) and YKLHKQVVTFVE (rP2-I2, 301–312 aa), recognized by mAbs 1D2/2E2 and 2C11/1E6/6E8, were completely conserved in all OmpP2 sequences. Thus, these two epitopes were GC epitopes. Epitope GFDILTASSDSAINYT (rP2-I2, 137–152 aa), recognized by mAb 5F9, had several amino acid mutations, but these mutations were present within the antigens that reacted with mAb 5F9, including rP2-I1, rP2-I3, rP2-II, and OmpP2 of the DT3 and ZS7 strains. These results showed that the mutations were not at the pivotal amino acids, that all mutation sequences were linear epitopes recognized by mAb 5F9, and that they were also GC epitopes. Similarly, the epitope IDFVRTGARFDVTPKSGVYGNYSY (rP2-I2, 257–280 aa) recognized by mAb 2B8, which had a mutation at N-258 within genotype II proteins, which did not affect its reactivity with mAb 2B8. Thus, these epitopes were GC epitopes. The sequence polymorphism of the epitope GKQAVIGDSIGQAGFDKVYG (rP2-I2, 101–120 aa) recognized by mAb 3A3 was more complicated. Although mutation at D-120 did not affect binding of this peptide to mAbs, we failed to rule out the effect of mutations at aa 104 and 111 on epitope integrity. Multiple alignment revealed that 83.33% (30/36) OmpP2 proteins contained epitopes in this region that could react with mAb 3A3, including genotype I and II proteins; therefore, this epitope was likely to be a GC epitope.

In contrast, the sequences of Pt5 (65–88 aa) and Pt11/Pt11a (164–184 aa), which reacted with GS-mAb, could be divided into three clusters. Except for the Pt5 region of strain no. 4, the two regions of all OmpP2 proteins had obvious genotype specificity; clusters 1 and 2 were genotype I, and cluster 3 was genotype II. mAb 1A10 reacted specifically with the epitopes of rP2-I1, rP2-I2, DT3, and ZS7 in cluster 1, indicating that aa 85 and 86, which had high-frequency mutations, were not pivotal amino acids, and that mutations at these residues did not change the integrity of the epitopes. Therefore, mAb 1A10 interacted specifically with and identifies all OmpP2 proteins in cluster 1. In addition, mAb 1A10 could not recognize rP2-I3 and rP2-II, indicating that mutations at aa 70, 71, 72, 74, 76, 78, and 83 changed the integrity of the epitope, and that these may be pivotal amino acids in the epitope. Therefore, mAb 1A10 could not recognize the sequences of cluster 2 and cluster 3. The epitope YETRLGRNSKNDAGWGDVTTXXAY was the GS epitope of cluster 1 of genotype I. Similarly, mAb 2F1 only recognized the proteins of genotype II in cluster 3, but not those in clusters 1 and 2; thus, epitope YETRLGSGSKNAAKWGDVTTXXAY was the GS epitope of genotype II. Multiple alignments showed that the epitope Pt11/Pt11a NYNVANERDNKGEVKVDSTKS, recognized by mAb 2C5, was not unique to rP2-I2, and four OmpP2 proteins from representative strains, such as Hs-DY09, also had the same epitope. This epitope was thus a GS epitope belonging to cluster 1 and was significantly different from those in clusters 2 and 3. In contrast, the epitope Pt11a-II NYNVANEREKADVKVDSIKS, recognized by mAb 5F1, was a GS epitope that belonged only to cluster 3 and genotype II. These results illustrated that the pivotal amino acids of the Pt11/Pt11a epitope region contain abundant polymorphisms.

In the peptide-ELISA analysis of mouse serum ( Table 2 ), Pt1/Pt1a and Pt9/Pt9a epitopes reacted with mouse serum from both the rP2-I2 and rP2-II immune groups, whereas other epitopes were only exempted in rP2-I2 or rP2-II. The antibody titers of different epitopes showed visible differences in mouse serum; the highest antibody titer of 51,200 of anti-Pt1/P1a was observed in the rP2-II immunization group. This was higher than that of the injected antigen rP2-II (25,600), and 50 in the rP2-I2 group. The antibody titer of anti-Pt5 was 25,600 (rP2-I2 immunization group), whereas the antibody titers of the remaining epitopes ranged between 50 and 3,200. ELISA results showed that anti-Pt14 (THDDYKSGSVNKKDKDGVYFGLKY, rP2-I2, 209–232 aa) antibody was present in the serum of rP2-I2 immunized mice, with a titer of 800, while anti-Pt22 (TYKVDESITVNNTQGTFKYSAPQE, rP2-II, 129–152 aa) antibodies were present in the serum of rP2-II immunized mice, and with a titer of 3,200, indicating that Pt14 and Pt22 may also be epitopes.

In the peptide-ELISA analysis of sera from naturally infected pigs ( Table 2 ), positive swine sera could be screened for all epitopes, but the seroprevalence differed markedly among the epitopes. The highest positivity rates were 41.18%, 34.22%, 45.99%, 44.39%, and 43.42% for Pt1/Pt1a, Pt7a, Pt11a-II, Pt17, and Pt19/Pt19a, respectively. The lowest positivity rate was 1.6% for Pt5, and that for the other epitopes was between 5.88% and 22.99%. The overall positivity rate of peptide ELISA for all epitopes and swine serum was 67.38%, similar to the positivity rate (68.98%) for the two recombinant proteins (rP2-I2, rP2-II). Comparative analysis with commercial kits showed that the agreement of each epitope exceeded 65%, with highest being 73.8% for Pt14. The overall agreement between all the epitopes was 67.91%, which was slightly higher than that between the two recombinant proteins (rP2-I2 and rP2-II).

In addition, ELISA for the 20 double-positive serum samples and the remaining peptides showed that the positivity rate of Pt4, Pt15, and Pt21 reached 60%–80%, indicating that these were highly likely to be epitopes. However, the titers of the three peptides were below 50 in mouse serum.

We compared the mRNA expression levels of IL-1α, IL-1β, IL-6, IL-8, and TNF-α between peptide-stimulated cells (experimental group) and unstimulated cells (mock group). The results are shown in Figure 5 . The mRNA expression levels of most cytokines were upregulated (p < 0.05) after peptide stimulation, except for that of IL-6, which was downregulated by Pt5 and Pt17 stimulation (0.51- and 0.47-fold, respectively, compared to the mock control). The extent of significantly upregulated ranges (compared with mock control, p < 0.05) in mRNA expression levels of proinflammatory cytokines were 1.76- to 3.42-fold for IL-1α, 1.56- to 5.15-fold for IL-1β, 1.45- to 3.71-fold for IL-6, 1.68- to 3.69-fold for IL-8, and 2.13- to 7.59-fold for TNF-α. The highest values were observed upon stimulation with Pt20 (IL-1α), Pt1a (IL-1β), Pt20 (IL-6), Pt20 (IL-8), and Pt9 (TNF-α) stimulation. Among the peptides producing significant differences in proinflammatory cytokines, 72.73% (8/11, IL-1α), 66.67% (10/15, IL-1β), 52.94% (8/17, IL-6), 72.73% (8/11, IL-8), and 64.71% (11/17, TNF-α) were epitopes; 63.64% (7/11, IL-1α); 53.33% (8/15, IL-1β), 70.59% (12/17, IL-6), 54.55% (6/11, IL-8), 58.82% (10/17, TNF-α) were loops; and 36.36% (4//11, IL-1α), 33.33% (5/15, IL-1β), 35.29% (6/17, IL-6), 36.36% (4/11, IL-8), 35.29% (6/17, TNF-α) were peptides with overlapping epitopes and loops (epitope & loop). Combined analysis of the five proinflammatory cytokines showed that 64.79% (46/71) of the peptides with significant differences were epitopes, 60.56% (43/71) were loops, and 35.21% (25/71) were both epitopes and loops. Although most of the peptides with significant differences in proinflammatory cytokines were epitopes and/or loops, there was no significant difference in the average mRNA expression levels of various cytokines between the epitope and non-epitope groups, or the loop and non-loop groups ( Figure S1 ).

The average mRNA expression levels were calculated from five proinflammatory cytokines in each peptide, and a line diagram was drawn with the mean of the minimum upregulation level (significant upregulation compared to the mock control, p < 0.05) of each cytokine as the threshold ( Figure 5 ). The results showed that the mean values of Pt1, Pt1a, Pt7, Pt9, Pt11, Pt11a, Pt13, Pt18, Pt19, Pt20, and Pt21 exceeded the threshold, and the corresponding amino acid sites were 1–24 aa, 97–120 aa, 129–152 aa, 161–190 aa, 193–216 aa, and 273–341 aa in rP2-I2, which were localized within linear B cell epitopes and loops. These peptides could induce mRNA expression of most inflammatory cytokines. Among them, Pt1, Pt9, and Pt20 significantly induced the mRNA expression of all five inflammatory cytokines (p < 0.05) most pronouncedly. In addition, although Pt17 did not exceed the threshold, it affected the expression of five inflammatory cytokines: while it stimulated PAMs to upregulate the expression of IL-1α, IL-1β, IL-8 and TNF-α, it downregulated the expression of IL-6 (p < 0.05).

Finally, we analyzed the effects of different homologous GS epitopes between the two genotypes on proinflammatory cytokines ( Figure 6 ). After Pt5-I3 stimulation, the mRNA expression levels of IL-1α, IL-1β, IL-8, and TNF-α were significantly upregulated (p < 0.05), while IL-6 levels remained unchanged (p > 0.05). Pt5-II could significantly upregulate the levels of IL-1α, IL-8, TNF-α, and significantly downregulate IL-6 levels (p < 0.05). Pt5-I2 could significantly upregulate TNF-α and significantly downregulate IL-6 levels (p < 0.05). The comparison among homologous GS epitopes showed that there was no significant difference between Pt5 and Pt5-II (p > 0.05), but the induced expression level of IL-1β was significantly lower than that of Pt5-I3 (p < 0.05). Furthermore, Pt5-II can significantly inhibit the expression of IL-6 compared with Pt5-I3 (p < 0.05). Taken together, these results illustrate that the proinflammatory activities of Pt5-I3 were stronger than those of Pt5 and Pt5-II. Similarly, the expression levels of IL-1β, IL-6, IL-8, and TNF-α were significantly upregulated after Pt11 and Pt11a stimulation (p < 0.05), while only those of IL-1α, IL-8, and TNF-α were significantly upregulated after Pt11a-II stimulation (p < 0.05). Comparison among the three homologous GS epitopes showed that the mRNA expression levels of IL-1β and IL-6 after Pt11 and Pt11a stimulation were significantly higher than those after Pt11a-II stimulation, but the level of IL-1α was lower than that after Pt11a-II stimulation (p < 0.05).