Carboxyl-Terminal Decoy Epitopes in the Capsid Protein of Porcine Circovirus Type 2 Are Immunogenicity-Enhancers That Elicit Predominantly Specific Antibodies in Non-Vaccinated Pigs

In the context of the carboxyl-terminus (C-terminus) of the capsid protein of porcine circovirus type 2a (PCV2a) and PCV2a vaccines, this study aimed to explore its unrevealing cryptic epitope and its relation to PCV2-infected herd immunity. To discover the C-terminus of the capsid protein of PCV2a, monoclonal antibodies (mAbs) were generated in this work. Two mAbs bound the two minimal linear epitopes (229PPLKP233 and 228DPPLNP233 (or 229PPLNP233)), which were located at the C-terminus of the capsid proteins of PCV2a and PCV2b, respectively. One mAb bound to the minimal linear epitope (220QFREFNLK227, peptide P82), but it neither bound the virus-like particle (VLP) of PCV2a nor produced positive staining in PCV2a-infected cells by immunofluorescence assay. Further, the residues 220–227 were not accessible on the surface of the VLP on the three-dimensional model, but the residues 228–231 extend toward the VLP exterior. Immunoassays were conducted in this study to screen anti-viral peptide-specific IgGs, which could differentiate vaccinated pigs from non-vaccinated ones. The data show two 220QFREFNLKDPPLKP233-containing peptides had a significantly higher binding reactivity with sera from PCV2-infected pigs in the control group than with sera from the VLP-vaccine group, particularly seen in sera from swine aged 15 weeks to 24 weeks. However, the peptide P82 had not this phenomenon in that test. This study confirmed that C-terminal epitopes play an important role in PCV2-induced decoy of swine humoral immunity.

Although vaccination against PCV2 infection has been widely utilized for disease control and effectively reduces PCV2-associated diseases [11,13,43], low levels of PCV2 infection were still detected in vaccinated pigs under field conditions [13]. It was difficult to evaluate the immune response of the PCV2a vaccine in field farms since the prevalence of seropositive animals has grown steadily by vaccination or infection [11][12][13]44,45]. Therefore some researchers developed the Rep' protein antibody detection method to distinguish natural infection with PCV2 from subunit vaccine immunization [46]. Some researchers found that high levels of anti-capsid peptide (residues 169-180) activity are associated with naturally PCV2-infected animals [35,47,48]. They indicated this decoy epitope (residues 169-180), which is exposed on the surface of the capsid subunit, but, in the structural context of the virus-like particle (VLP), this epitope is buried and inaccessible to antibody [35]. The fundamental characteristics of decoy epitopes are directing the humoral response away from protective epitopes and inducing non-neutralizing antibodies against these epitopes [35,[49][50][51][52]. Therefore, the specific antibodies which against decoy epitopes or VLP, provide opportunities to devise diagnostic tests for monitoring the immunological effects of vaccination [47]. The other view is that modification of decoy epitopes and other epitopes of PCV2 in vaccines could improve immunogenicity for PCV2 control since the shifting prevalence of circulating PCV2 subtype strains, which could be increasingly inadequate for current vaccines.
The following critical amino acid substitutions (at residue 206, 210, 215, 232, and 234) in the C-terminus of the capsid protein of various PCV2 genotypes were noticed in molecular genetic analysis of PCV2 sequences [19,20]. Particularly, the extra amino acid (residue 234 K) was in the C-terminus of the capsid protein of PCV2d strains that might influence host immune response, then these strains might favor survival, virulence, and circulating among vaccinated swine herds. For this reason, immunorelevant epitopes of the C-terminus of the capsid protein of PCV2 should be studied. According to previous reports, the antigenic domain (residues193-207) [30], conformational epitopes (within residues 165-200, and 230-233 of PCV2a [31], and residues 231-233 of PCV2b [32]), linear B-cell epitopes (residues 195-202, and 231-233) [32] were found by epitope mapping with mAbs against PCV2 virion or recombinant capsid protein. Likewise, another study demonstrated that the chemical synthesis of the peptide C3 (residues 195-233) can mimic the C-terminus of the capsid protein of PCV2b, particularly the immunoreactivity of the peptide C3 with sera from PCV2infected pigs [42]. These anti-C3 mAbs recognized two core motifs (P62 ( 228 DPPLNP 233 ), and P67 ( 228 DPPLNPK 234 )), and each motif was located at the C-terminus of the capsid protein of PCV2b and PCV2d, respectively [33]. The residues ( 225 NLKDPPLNP 233 ) of the Cterminus of the capsid protein of PCV2b have been reported to be critical to VLP assembly, cell entry, and propagation [53]. However, there is little information about the connection between the C-terminus of the capsid protein of PCV2 and PCV2-infected herd immunity. Based on previous studies [30][31][32][33], there should be epitopes on the sequence of the capsid protein of PCV2a between residues 205 and 233. If the synthetic capsid peptide (residues 205-233) can mimic the epitopes of the native capsid protein, it could elicit antibodies that could bind to the native capsid protein.
Therefore, this paper presents a new approach to verify the chemical synthesis of the peptide can also mimic the native capsid protein of PCV2a. This study generated six mAbs against the C-terminus of the capsid protein of PCV2a (peptide P64, residues 205-233) and defined their minimal binding regions. This study also used these mAbs and other antibodies to label the capsid protein and other viral proteins to determine conformational epitopes in PCV2 virus-infected porcine kidney cells and VLP. Further, this study used the UCSF Chimera version to figure out the locations of binding residues of mAbs on the 3-D Model of the VLP of PCV2. Furthermore, more than 60 viral peptides (including capsid peptides and other ORF peptides) were used as coating antigens to screen anti-viral peptide IgGs that could significantly differentiate vaccinated piglets from the non-vaccinated. The relationship among the C-terminus of the capsid protein, epitopes, and their antibody responses in the PCV2-infected herd with or without vaccination was also discovered.
The method is essentially the same as that used by the author with some adjustments. More detail can be found in the previous paper [33]. Briefly, four five-week-old, female, BALB/cByJNarl mice were purchased from a specific pathogen-free (SPF) colony (the National Applied Research Laboratories, Taiwan). The mice were maintained in isolation rooms in filtertop cages. Mice were inoculated with 50 µg of immunogen (the peptide P64 conjugated with KLH) emulsified in Freund's adjuvant (Sigma-Aldrich, St. Louis, MO, USA). Subsequently, mice were boosted fortnightly with the same amount of immunogen. Three days after the final booster, the spleens of the mice were harvested for hybridoma generation. Hybridomas were screened for the secretion of anti-P64-specific mAbs by indirect enzyme-linked immunosorbent assay (iELISA) using the peptide P64 as a coating antigen. That secondary antibody solution (Peroxidase-conjugated goat anti-mouse IgG (subclasses 1 + 2a + 2b + 3, Fcγ, Jackson ImmunoResearch, West Grove, PA, USA)) was used at 1:2500 dilution for hybridoma screening. The hybridomas that secreted anti-P64 mAbs were subcloned by limited dilution of the cells. Then, ascites containing mAbs were collected as previously proposed.

Isotyping of MAbs
The heavy chain and the light chain of the six mAbs secreted by each hybridoma were determined by using an SBA Clonotyping System/HRP (Southern Biotechnology, Birmingham, AL, USA). It was performed according to the previously published method [33]. Briefly, the MaxiSorp microtiter plate (Thermo Fisher Scientific, Inc., Waltham, MA, USA) was coated with the capture antibody (goat anti-mouse Ig was diluted to 5 µg/mL in 0.05-M bicarbonate buffer (Sigma-Aldrich, St. Louis, MO, USA)) at 4 • C overnight. After three washes with PBS containing 0.05% Tween 20 (PBST), the plate was blocked with the blocking buffer (PBST containing 5% casein hydrolysate) at 37 • C for 30 min. After washing, Viruses 2022, 14, 2373 6 of 19 a volume of 100 µL of each culture supernatant of hybridoma was added sequentially, and the plate was incubated at 37 • C for 2 h. After washing, 100 µL of HRP labeled goat anti-mouse IgG1, -IgG2a, -IgG2b, -IgG3, -IgA, -IgM, -κ light chain, and -λ light chain were added to appropriate wells of the plate. PBST served as the background. The plate was incubated at 37 • C for 1 h and afterward washed with PBST. Then, 100 µL of 2,2 -azinobis(3-ethylbenzothiazoline-6-sulfonic acid) diammonium salt (ABTS, Sigma-Aldrich, St. Louis, MO, USA) buffer was added to each well. After 30 min, a SpectraMax M5 microplate reader (Molecular Devices, San Jose, CA, USA) was used to measure the optical density (OD) of each well at 405 nm.

Peptide Array-Based Epitope Mapping
The peptide array-based epitope mapping was performed by iELISA using truncated peptides (as shown in Table 1) as coating antigens. Briefly, peptides contained the Cterminus of the capsid peptide between residues 205 and 233, associated 10-mer peptides, and truncated derivatives. Two commercial virus-like particles (VLP) of PCV2 vaccines (VLP-A (Porcilis ® PCV, Intervet International B.V., An Boxmeer, The Netherlands), and VLP-B (Ingelvac CircoFLEX ® , Boehringer Ingelheim, St Joseph, MO, USA)) and one anti-PCV2 mAb 36A9 (Ingenasa, Madrid, Spain) were also used in this experiment. Maxisorp plates were coated with each peptide (5 µg/mL) or vaccine (15 µL/mL) in bicarbonate buffer and incubated at 4 • C overnight. After three washes with PBST, the plates were blocked with the blocking buffer at 37 • C for 30 min. After washing, the culture supernatant of hybridoma was added, and plates were again incubated at 37 • C for 2 h. The anti-peptide P64 mouse serum (at a 1:1000 dilution) was used as the positive control, and the anti-PCV2 mAb 36A9 served as the control. After rinsing three times with PBST, the secondary antibody solution was applied to the appropriate wells. Peroxidase-conjugated goat anti-mouse IgG (subclasses 1 + 2a + 2b + 3, Fcγ, Jackson ImmunoResearch, West Grove, PA, USA) was used at a 1:2500 dilution for binding to the mAbs. After 1 h, the plates were washed three times. The colorimetric reaction was developed by using the ABTS solution. Following incubation for 30 min, the OD was measured at 405 nm.

Generation of Structural Images for Capsid Protein of PCV2a
The three-dimensional (3-D) structures of the capsid protein were used as approaches to figure out the binding interaction of peptides and antibodies. The structural model of the capsid protein and the VLP of PCV2a coordinates were retrieved from the Protein Data Bank (PDB) entries for the capsid protein of PCV2a (PDB code: 3JCI) [54], and images were generated using UCSF Chimera version 1.14 from the Resource for Biocomputing, Visualization, and Informatics at the University of California, San Francisco, USA [55].

Archival Pig Sera Collected from the Piglets following Vaccination or Unvaccinated Controls
The previous study herd had PCV2 infection without getting the PCV2 vaccine in the conventional pig farm with the farrow-to-finish operation [45]. All pigs were notched in their ear with a unique identification number on the day of birth. They were weaned at 4 weeks of age and transferred into nursery houses. Then, those weaning piglets were moved to the fattening house at 8 weeks of age. This study was previously approved by the IACUC of the Livestock Research Institute (approval no. LRIIACUC100-33) and complied with all of the requirements and regulations for animal experimentation, following current regulations in Taiwan.
Two piglets of the same sex and similar body weight (close to the mean body weight of the entire newborn litter) were selected from each litter. This study involved 20 newborn piglets (10 male and 10 female) of TLRI Black Pig No.1 (TBP), delivered from 10 sows between January and March 2011. One piglet from each litter was injected intramuscularly with the VLP-B vaccine (Ingelvac CircoFLEX ® , 1.0 mL, one dose at 3 weeks of age). The other piglet from each litter was injected intramuscularly with 2.0 mL of saline at 3 weeks of age and served as a control. Blood samples from each pig were collected at 3, 6, 9, 12, 15, 18, 21, and 24 weeks of age. Subsequently, sera separated from the aforementioned blood samples and stored at −20 • C. At last, there were only three piglets from the control group that died during this experiment. One pig died before 18 weeks of age, and two died before 21 weeks of age.
All sera (n = 153) had been examined with the immunofluorescence assay and polymerase chain reaction. The result showed that both vaccinated pigs and the controls had suffered from PCV2 infection [45] (as Figure S1).

Detection of the Specific Antibodies against the Capsid Peptides or Other ORF Peptides in Sera among Vaccinated Piglets and Control Ones
First, primary screening was used to select anti-viral peptide-specific antibodies, which could significantly differentiate vaccinated piglets from non-vaccinated ones. Six sera were selected from six animals (three litters) at three different ages (18, 21, and 24 weeks of age, respectively). Each group contained three samples from three individuals of different ages. Two VLPs, two inactive PCV2 viruses (VC (chimeric PCV1/2, Fostera™, Zoetis, Charles City, IA, USA), and VD (PCV2a virus, Circovac ® , Merial, Saint-Priest, France)), and 56 peptides (capsid peptides and other ORFs peptides) were used as coating antigens in iELISA test.
They performed as previously described for pig anti-peptide ELISA [56], with the exception that the capsid peptides and other ORFs peptides (as Table 2) were used as the coating antigen, and pig sera at a 1:200 dilution were used. The subsequent procedures of blocking, archival pig sera, washing, secondary antibody, substrate, and the OD reading, were as previously described [56].

Statistical Analysis
The data were analyzed by using a one-way analysis of variance (ANOVA) and Tukey's Studentized Range (HSD, honestly significant difference) multiple comparisons test using the SAS Enterprise Guide 7.1 ® software (SAS Institute Inc., Cary, NC, USA). A p-value < 0.05 was considered significant. Following the fusions, hybridoma supernatants were screened for the presence of anticapsid peptide (P64) specific antibodies by iELISA. For choosing hybridomas that only produce IgG mAb, the peroxidase-conjugated goat anti-mouse IgG (subclasses 1 + 2a + 2b + 3, Fcγ) secondary antibody was used for hybridoma screening. Among them, 41 hybridomas supernatants reacted with the peptide P64 at the first screening. After repeated subcloning by limiting dilution and selection, six stable hybridomas secreting anti-P64 mAbs were obtained. Three hybridomas (clone 4C9, 4F6, and 4H9) produced IgG1 mAb, two hybridomas (clone 6E3 and 6H11) produced IgG2b mAb, and one hybridoma (clone 3H11) produced IgG2a mAb. Those hybridomas produced mAbs with kappa-light chains. The results are presented in Table 3.
that of the mAb 4F6 and other linear peptides (P39, P79, P80, P83, P107, and P108). Two mAbs (4C9 and 4H9) not only reacted with the peptide of immunogen (peptide P64) but also reacted mildly with the peptide P108 (210DYNIRVTMYVQFREFNLKDPPLKP233). These data indicated that the mAb 6E3 reacted only with the peptide of immunogen (P64). The positive control (mAb 36A9) not only reacted with VLPs of PCV2 but also showed minor reactions against some linear peptides (P64, P86, and P108). Mapping of the linear and minimal epitopes for anti-capsid peptide (peptide P64) mAbs binding. Anti-P64 mAbs bound the linear peptide spanning from residues 205 to 233. The mAb binding was tested by using an iELISA. Peptides contained the sequence of the capsid protein between residues 205 and 233, associated 10-mer peptides, and truncated derivatives (as shown in Table 1). The experiments were performed three times. Data represent the mean ± standard error (SE). Mapping of the linear and minimal epitopes for anti-capsid peptide (peptide P64) mAbs binding. Anti-P64 mAbs bound the linear peptide spanning from residues 205 to 233. The mAb binding was tested by using an iELISA. Peptides contained the sequence of the capsid protein between residues 205 and 233, associated 10-mer peptides, and truncated derivatives (as shown in Table 1). The experiments were performed three times. Data represent the mean ± standard error (SE).

Immunoreactivity of Anti-P64 MAbs with PCV2 Virus
The specificity of mouse immune serum, pig sera, and mAbs was primarily determined by IFA with the PCV2 FA substrate slides according to the manufacturer's manual. Slides were PCV2 virus (strain: T657)-infected porcine kidney cells fixed on the surface of Teflon-masked slides. The data showed that the PCV2-infected pig antiserum produced strong intranuclear and cytoplasmic staining in PCV2-infected PK cells ( Figure 2C, D). This staining was used to identify all viral proteins of PCV2. When PCV2 FA substrate slides were stained with anti-P64 mouse sera, positive signals were displayed in a local distribution in the nucleus only ( Figure 2E,F). This step was used to identify the C-terminus of the capsid protein of PCV2a. In the same way, the anti-VLP of PCV2 mouse sera produced dispersing granular and cytoplasmic staining in PCV2-infected PK cells ( Figure 2G,H). These findings were similar to the previous study [34]. The previous data indicated that the anti-peptide C3 (PCV2b) mAbs could not specifically detect this PCV2 virus (strain: T657) [34]. Surprisingly, this study showed that not only the mAb 3H11 produced strong specific intranuclear staining and slightly cytoplasmic staining in PCV2-infected PK cells ( Figure 2I,J), but also the mAb 6H11 produced strong specific intranuclear staining in PCV2-infected PK cells ( Figure 2S,T). Based on these data, this strain (T657) was assumed to belong to PCV2a. Inexplicably, the mAb 4C9 produced unspecific cytoplasmic staining in PCV2-infected PK cells ( Figure 2K,L), and the mAb 4H9 produced intranuclear staining and slightly unspecific cytoplasmic staining in PCV2-infected PK cells ( Figure 2O,P). The mAbs (4F6 and 6E3) did not recognize this PCV2 virus (strain: T657) in PCV2-infected PK cells ( Figure 2M,N,Q,R). The positive signals were due to different abundance of the protein, the degradation of the protein, different PCV2 strains, conformational epitopes, or linear epitopes, which interact with the antibody.

Locations of Binding Residues of MAbs (3H11, 4F6, and 6H11) on the 3-D Model of the VLP of PCV2
The crystal structure of capsid proteins (residues 42-231) of PCV2a [Protein Data Bank (PDB) with accession code 3JCI [54] was reported and Chimera software was used to spatially visualize the locations of critical binding residues of mAbs (3H11, 4F6, and 6H11). Both of the binding residues (residues 228-231) of mAbs (3H11 and 6H11) and the binding residues (residues 220-227) of mAb 4F6 were mapped onto the 3-D structure of VLP of PCV2a and a single capsid protein subunit. The residues ( 228 DPPL 231 ) were highlighted in red, residues ( 220 QFREFNLK 227 ) were labeled in yellow, and other residues ( 205 SKYDQDYNIRVTMYV 219 ) were labeled in blue ( Figure 3A-F). These aforementioned color residues represented the peptide of immunogen (P64) that was depicted on the ribbon and surface maps of a single capsid protein subunit. The critical binding residues of mAbs (3H11 and 6H11) extend toward the VLP exterior ( Figure 3E,F). However, the binding residues of mAb 4F6 were not accessible on the surface of the VLP ( Figure 3E,F), excluding two residues ( 225 N and 227 K). Although these two residues ( 225 N and 227 K) were obviously on the surface of a single capsid protein subunit ( Figure 3C), they were almost buried after the self-assembling of capsid proteins into a VLP ( Figure 3E,F).

Locations of Binding Residues of MAbs (3H11, 4F6, and 6H11) on the 3-D Model of the VLP of PCV2
The crystal structure of capsid proteins (residues 42-231) of PCV2a [Protein Data Bank (PDB) with accession code 3JCI [54] was reported and Chimera software was used to spatially visualize the locations of critical binding residues of mAbs (3H11, 4F6, and 6H11). Both of the binding residues (residues 228-231) of mAbs (3H11 and 6H11) and the binding residues (residues 220-227) of mAb 4F6 were mapped onto the 3-D structure of VLP of PCV2a and a single capsid protein subunit. The residues (228DPPL231) were highlighted in red, residues (220QFREFNLK227) were labeled in yellow, and other residues (205SKYDQDYNIRVTMYV219) were labeled in blue ( Figure 3A-F). These aforementioned color residues represented the peptide of immunogen (P64) that was depicted on the ribbon and surface maps of a single capsid protein subunit. The critical binding residues of mAbs (3H11 and 6H11) extend toward the VLP exterior ( Figure 3E-F). However, the binding residues of mAb 4F6 were not accessible on the surface of the VLP ( Figure 3E-F), excluding two residues (225N and 227K). Although these two residues (225N and 227K) were obviously on the surface of a single capsid protein subunit ( Figure 3C), they were almost buried after the self-assembling of capsid proteins into a VLP ( Figure 3E-F).   [54], and images were generated using UCSF Chimera version 1.14 from the Resource for Biocomputing, Visualization, and Informatics at the University of California, San Francisco, USA [55].

Immunoreactivities of Peptides with Swine Sera Which Were from Vaccinated Piglets or Control Ones
For primary screening, there was no significant difference between the vaccine group and the control group in non-capsid or non-C-terminus of the capsid peptides ELISA tests ( Figure 4A). However, the striking result to emerge from the data was that significantly different interactive affinity between these two groups in the same C-terminus of the capsid peptides ELISA ( Figure 4B) and the VLP-A ELISA tests( Figure 4A). structure of the single capsid protein is represented as a ribbon diagram in one VLP. (E) 3-D model of the single capsid protein subunit rendered as a solid surface. (F) The surface map of all capsid proteins subunits in one VLP. The structural model of the capsid protein and the VLP of PCV2a coordinates were retrieved from the Protein Data Bank (PDB) entries for the capsid protein of PCV2a (PDB code: 3JCI) [54], and images were generated using UCSF Chimera version 1.14 from the Resource for Biocomputing, Visualization, and Informatics at the University of California, San Francisco, USA [55].

Immunoreactivities of Peptides with Swine Sera Which Were from Vaccinated Piglets or Control Ones
For primary screening, there was no significant difference between the vaccine group and the control group in non-capsid or non-C-terminus of the capsid peptides ELISA tests ( Figure 4A). However, the striking result to emerge from the data was that significantly different interactive affinity between these two groups in the same C-terminus of the capsid peptides ELISA ( Figure 4B) and the VLP-A ELISA tests( Figure 4A). C-terminal capsid peptides (P30, P47, P65, P71, P73, P84, P97, P98, P106, and P108) had a significantly higher binding interaction with pig sera from the control group than with sera from the vaccine group ( Figure 4B). It is worth noting that most of the aforementioned peptides belong to PCV2a. In another way, the VLP-A had a significantly higher binding interaction with pig sera from the vaccine group than with sera from the control group. Although Trible et al. indicated the peptide (169STIDYFQPNNKR180, P93) was an immunological decoy that provides the basis for diagnostic approaches that can differentiate infected from vaccinated animals [36,47,48], it was unlikely in this result C-terminal capsid peptides (P30, P47, P65, P71, P73, P84, P97, P98, P106, and P108) had a significantly higher binding interaction with pig sera from the control group than with sera from the vaccine group ( Figure 4B). It is worth noting that most of the aforementioned peptides belong to PCV2a. In another way, the VLP-A had a significantly higher binding interaction with pig sera from the vaccine group than with sera from the control group. Although Trible et al. indicated the peptide ( 169 STIDYFQPNNKR 180 , P93) was an immunological decoy that provides the basis for diagnostic approaches that can differentiate infected from vaccinated animals [36,47,48], it was unlikely in this result ( Figure 4A). Both control and vaccinated animals had suffered PCV2 infection ( Figure S1). Remarkably, the capsid peptides (P101, 132 KATALTYDPYVNYSS 146 , and P103, 117 GVGSSAVILDDNFVT 131 ) had higher mean OD values of binding interaction with pig sera from the control group than with sera from the vaccine group, but their data did not show a significant difference in this experiment since several potential shortcomings needed to be considered in this primary screening test.
Previous studies demonstrated that each anti-viral protein IgG had a different curve of the profile in the same herd after vaccination [56]. To verify C-terminus of the capsid peptides can differentiate vaccinated from control animals, anti-viral peptide IgG in archival sera from pigs at different age were used in this study. Some C-terminus of the capsid peptides (C3, P47, P64, P97, P106, and P107) had a significantly higher binding interaction with pig sera from the control group than with sera from the vaccine group aged 15 weeks to 24 weeks (Figures S2 and 5B-F). The peptide P30 had a similar effect on pig sera aged 18 weeks to 24 weeks ( Figure 5A). Significantly, using PCV2a peptides (P47 or P106) reacted with pig sera, which had a larger difference in OD value between the control group and vaccine group than that using the same length of PCV2b peptides (P30 or P107), particularly swine aged 15 weeks to 24 weeks. Furthermore, two 220 QFREFNLKDPPLKP 233containing peptides (P97 or P106) had better results of this phenomenon than did other 225 NLKDPPLKP 233 (or 225 NLKDPPLNP 233 )-containing peptides (C3, P30, P47, P64, and P107). These peptides can be used as a tool to differentiate vaccinated from non-vaccinated animals which had suffered PCV2 infection. The cut-off value of anti-P97 IgG and anti-P106 IgG ELISA was 1.9 and 1.4, respectively.
( Figure 4A). Both control and vaccinated animals had suffered PCV2 infection ( Figure S1). Remarkably, the capsid peptides (P101, 132KATALTYDPYVNYSS146, and P103, 117GVGSSAVILDDNFVT131) had higher mean OD values of binding interaction with pig sera from the control group than with sera from the vaccine group, but their data did not show a significant difference in this experiment since several potential shortcomings needed to be considered in this primary screening test.
Previous studies demonstrated that each anti-viral protein IgG had a different curve of the profile in the same herd after vaccination [56]. To verify C-terminus of the capsid peptides can differentiate vaccinated from control animals, anti-viral peptide IgG in archival sera from pigs at different age were used in this study. Some C-terminus of the capsid peptides (C3, P47, P64, P97, P106, and P107) had a significantly higher binding interaction with pig sera from the control group than with sera from the vaccine group aged 15 weeks to 24 weeks ( Figure S2, 5B-F). The peptide P30 had a similar effect on pig sera aged 18 weeks to 24 weeks ( Figure 5A). Significantly, using PCV2a peptides (P47 or P106) reacted with pig sera, which had a larger difference in OD value between the control group and vaccine group than that using the same length of PCV2b peptides (P30 or P107), particularly swine aged 15 weeks to 24 weeks. Furthermore, two 220QFREFNLKDPPLKP233-containing peptides (P97 or P106) had better results of this phenomenon than did other 225NLKDPPLKP233 (or 225NLKDPPLNP233)-containing peptides (C3, P30, P47, P64, and P107). These peptides can be used as a tool to differentiate vaccinated from non-vaccinated animals which had suffered PCV2 infection. The cut-off value of anti-P97 IgG and anti-P106 IgG ELISA was 1.9 and 1.4, respectively. Figure 5. Comparison of anti-C-terminus of the capsid peptide-specific IgGs between two groups of pigs at different ages. Among vaccinated (n = 10) and control (n = 10) animals, the anti-PCV2 peptide-specific IgGs were detected in the serum samples from pigs at different ages (3,6,9,12,15,18,21, and 24 weeks of age). The C-terminus of the capsid peptides (P30, P47, P64, P97, P106, and P107) were used as coating antigens in the iELISA test. The peptide P30 had a significantly higher binding interaction with pig sera from the control group than with sera from the vaccine group aged 18 weeks to 24 weeks (A). Further, the others (P47, P64, P97, P106, and P107) also had a similar effect on pig sera aged 15 weeks to 24 weeks (B-F). The data were analyzed by using ANOVA and Tukey's Studentized Range multiple comparisons test using the SAS Enterprise Guide 7.1 ® software. The data represent the mean ± standard error. Significant p values are indicated as * p < 0.05. Figure 5. Comparison of anti-C-terminus of the capsid peptide-specific IgGs between two groups of pigs at different ages. Among vaccinated (n = 10) and control (n = 10) animals, the anti-PCV2 peptide-specific IgGs were detected in the serum samples from pigs at different ages (3,6,9,12,15,18,21, and 24 weeks of age). The C-terminus of the capsid peptides (P30, P47, P64, P97, P106, and P107) were used as coating antigens in the iELISA test. The peptide P30 had a significantly higher binding interaction with pig sera from the control group than with sera from the vaccine group aged 18 weeks to 24 weeks (A). Further, the others (P47, P64, P97, P106, and P107) also had a similar effect on pig sera aged 15 weeks to 24 weeks (B-F). The data were analyzed by using ANOVA and Tukey's Studentized Range multiple comparisons test using the SAS Enterprise Guide 7.1 ® software. The data represent the mean ± standard error. Significant p values are indicated as * p < 0.05.
As anticipated, this experiment showed the immunogen VLP-B had a significantly higher binding interaction with pig sera from the vaccine group than with sera from the control group whose age was 9 weeks to 15 weeks ( Figure S2G). The middle region of the capsid peptide (C2) and ORF3 peptide (N1) had a significantly higher binding interaction with pig sera from the control group than with sera from the vaccine group at 24 weeks and 15 weeks of age, respectively ( Figure S2B,D). The N-terminus of the capsid peptide (C1), ORF6 peptide (N2), ORF9 peptide (N3), the binding residues (P82) of mAb 4F6, ORF10 peptide (P114), and ORF11 peptide (P115) could not be used as a tool to differentiate vaccinated from non-vaccinated animals that had suffered PCV2 infection ( Figure S2A,E,F,H-J).

Discussion
According to previous work, the chemical synthesis of peptide C3 (residues 195-233), can mimic the native form of the C-terminus of the capsid protein of PCV2b [42] and its epitopes [33]. The peptide P64 (residues 205-233) of the capsid protein of PCV2a was synthesized and used in this study. These experiments corroborate previous results. It is not only the anti-P64 mouse sera that recognized the native form of viral protein in PCV2-infected PK cells by IFA but also the peptide P64 showed good immunoreactivity with sera from PCV2-infected pigs. Subsequently, anti-P64 mAbs were generated and characterized in this study. This study showed that anti-P64 mAb (3H11 or 6H11) bound the two minimal linear epitopes ( 229 PPLKP 233 and 228 DPPLNP 233 (or 229 PPLNP 233 )), which were located at the C-terminus of the capsid proteins of PCV2a and PCV2b, respectively. This phenomenon lends support to previous findings about two types of C-terminus of the capsid protein of PCV2 (PCV2b and PCV2d) that were recognized by anti-C3 mAbs with pluripotency of binding [33]. The single most conspicuous observation to emerge from these phenomenon comparisons was C-terminus of the capsid protein of PCV2a could elicit the specific antibody against both PCV2a and PCV2b, then PCV2b one could elicit the specific antibody against both PCV2b and PCV2d. It is very similar to the previous study reports that widely used PCV2a-based vaccines lead to a decrease in the prevalence and viremia level of PCV2a and PCV2b in field pig serum collected from 12 states in the United States in 2012 compared to data from the pre-vaccination period of 2006 [57]. The shift in circulating genotype prevalence to PCV2d may reflect the lack of available commercial PCV2b-based vaccines, which warrants further study for confirmation. Although some studies demonstrated that the PCV2a-based vaccine was effective in reducing PCV2d viremia [58], their data showed that serum viremia and shedding of the virus were still observed at 3 weeks or 6 weeks after challenge in vaccinated pigs [58,59].
The data showed the binding interaction of anti-P64 mAb 6H11 and P85 (OD value: 0.22 ± 0.02) was weaker than that of mAb 6H11 and P84 (OD value: 0.92 ± 0.05). The P85 ( 230 PLKP 233 ) should be the minimal linear epitope, although the mAb 3H11 did not react with P85. However, the mAb 3H11 reacted with P84 ( 229 PPLKP 233 ) and recognized the PCV2 virus (strain: T657) in PCV2-infected PK cells. This is in good agreement with previously reported conformational epitopes within residues 230-233 of PCV2a [31]. According to the crystal structure of capsid proteins (residues 42-231) of PCV2a, this study showed that the residues 228 DPPL 231 extend toward the VLP exterior. The residues 232 KP 233 were not shown in this crystal structure (PDB code: 3JCI) [54], but 232 KP 233 was supposed to extend toward the VLP exterior ( Figure 3E-F). The data also showed that the binding interaction of mAb 3H11 (or 6H11) and P65 ( 228 DPPLKP 233 ) was stronger than that of the mAb 3H11 (or 6H11) and P84 (Figure 1). It is expected that the antigenicity of P65 ( 228 DPPLKP 233 ) was better than P84 ( 229 PPLKP 233 ) in swine immune response, but the truth was: not yet ( Figure 4B). This study also revealed that the mAb 4F6 bound strongly to the minimal linear epitope (P82 and 220 QFREFNLK 227 ), but it neither bound the VLP of PCV2a nor recognized native viral protein in PCV2-infected PK cells by IFA. Although residues in this epitope (P82 and 220 QFREFNLK 227 ) are located on the interior surface of a single capsid protein subunit ( Figure S3D), major residues (P89 and 220 QFREFN 225 ) of this epitope are completely buried in the VLP ( Figure 3E-F). This epitope (P82 and 220 QFREFNLK 227 ) showed mild immunoreactivities (OD value: 0.59 ± 0.04) with sera from PCV2-infected pigs at 24 weeks of age ( Figure S2H).
It has been suggested [35,36,60] that an immune response against the interior (or decoy) epitope on the disassembled monomer or fragments of capsid protein is produced during virus replication. They also demonstrated that swine immunity following VLP vaccination favored PCV2 neutralizing activity; whereas PCV2 infection and disease produced high levels of non-neutralizing antibody, primarily directed against that decoy epitope (residues 169-180)-containing polypeptide (residues 160-233 of capsid protein) [36] or ubiquitin protein (residues 43-233 of capsid protein) [35]. Therefore, they indicated that the level of anti-decoy epitope (residues 169-180) antibody can differentiate infected from vaccinated animals (DIVA) [35,36,47,61]. According to the aforementioned hypothesis and results, the interior epitope (P82 and 220 QFREFNLK 227 ) might be an unrevealing decoy epitope. However, the data showed that interior epitope (P82 and 220 QFREFNLK 227 ) could not be used as the tool to differentiate vaccinated from non-vaccinated animals that had suffered PCV2 infection. The most striking observation to emerge from the data comparison was that 229 PPLKP 233 -containing peptides had a significantly higher binding interaction with pig sera from non-vaccinated animals than with sera from VLP-vaccinated animals. Particularly, 227 KDPPLKP 233 -containing peptides or 220 QFREFNLKDPPLKP 233 -containing peptides (P97 or P106) did that test which had a larger difference in OD value between the control group and vaccine group whose ages 15 weeks to 24 weeks. Those residues are just local at the C-terminal epitope ( 229 PPLKP 233 ) and the interior epitope (P82 and 220 QFREFNLK 227 ). A previous study indicated that the C-terminus of the capsid protein of PCV2 is critical to VLP assembly, cell entry, and propagation [53]. Therefore, C-terminal epitopes should play an important role in PCV2-infected swine herds.
In general, these results confirm the conformational epitope ( 229 PPLKP 233 ) and interior epitope ( 220 QFREFNLK 227 ) of the C-terminus of the capsid proteins of PCV2a. This study revealed that the 229 PPLKP 233 -containing peptides had a significantly higher binding interaction with pig sera from non-vaccinated animals than those sera from VLP-vaccinated animals. Particularly, 220 QFREFNLKDPPLKP 233 -containing peptides (P97 or P106) did that test. It may be assumed that PCV2a avoids swine immune response against this important conformational epitope ( 229 PPLKP 233 or 230 PLKP 233 ) by eliciting anti-decoy epitope-specific antibodies which do not recognize this conformational epitope ( 229 PPLKP 233 or 230 PLKP 233 ) of the C-terminus. For this reason, those decoy peptides should be connected to this conformational epitope ( 229 PPLKP 233 ) of the C-terminus to elicit specific antibodies to replace anti-229 PPLKP 233 antibody production. It is recommended that further research should be undertaken on this interaction of host and virus. At last, this approach has the potential to be used as a tool to differentiate VLP-vaccinated from non-vaccinated animals that had suffered PCV2 infection in swine herds, but the major spreading PCV2 strains in the field must be taken into account.

Conclusions
To sum up, this work generated anti-P64 mAbs against the C-terminus of the capsid peptide (residues 205-233) of PCV2a. This study confirmed the mAb (3H11 or 6H11) bound the two minimal linear epitopes ( 229 PPLKP 233 and 228 DPPLNP 233 (or 229 PPLNP 233 )), which were located at the C-terminus of the capsid proteins of PCV2a and PCV2b, respectively. Both mAb 3H11 and 6H11 produced positive staining in PCV2-infected PK cells by IFA. The mAb 4F6 bound strongly to the minimal linear epitope ( 220 QFREFNLK 227 ), but it neither bound the VLP of PCV2a nor produced positive staining in PCV2-infected PK cells. These data showed that the residues of 229-233-containing capsid peptides of PCV2a had a significantly higher binding interaction with pig sera from the control group than with sera from the VLP vaccine group in primary screening. Furthermore, two 220 QFREFNLKDPPLKP 233containing peptides (P97 or P106) had better results of this phenomenon than did other residues of 225-233-containing peptides, particularly swine aged 15 weeks to 24 weeks. These peptides can be used as a tool to differentiate VLP-vaccinated from non-vaccinated animals which had suffered PCV2 infection. In the structural view, residues 228-231 extend toward the VLP exterior, otherwise, residues 220-227 were not accessible on the surface of the VLP on the 3-D Model. This protruding epitope cooperates with the buried epitope to elicit specific IgG in PCV2-infected pigs, particularly which had not been inoculated with the VLP-B vaccine, but the sample size needs to be considered.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/v14112373/s1, Figure S1: Both vaccinated (n = 10) and control (n = 10) swine sera had been examined with the immunofluorescence assay and polymerase chain reaction; Figure S2: Comparison of anti-PCV2 peptide-specific IgGs between two groups of pigs at different ages; Figure S3: Location of amino acid residues of designed peptide (P64) and critical binding residues on the single capsid protein of PCV2a.
Author Contributions: Conceptualization, methodology, validation, formal analysis, investigation, resources, data curation, writing-original draft preparation, writing-review and editing, visualization, supervision, project administration, funding acquisition, L.-C.H. All authors have read and agreed to the published version of the manuscript.