Neutralization Mechanism of a Monoclonal Antibody Targeting a Porcine Circovirus Type 2 Cap Protein Conformational Epitope

PCV2 is associated with several clinical manifestations collectively known as PCV2-associated diseases (PCVADs). Neutralizing antibodies play a crucial role in the prevention of PCVADs. We demonstrated previously that a MAb, MAb 3A5, neutralizes the PCV2a, PCV2b, and PCV2d genotypes with different degrees of efficiency, but the underlying mechanism remains elusive. Here, we report the neutralization mechanism of this MAb and the structure of the PCV2 virion in complex with MAb 3A5 Fabs, showing a binding mode in which one Fab interacted with more than two loops from two adjacent capsid proteins. This binding mode has not been observed previously for PCV2-neutralizing antibodies. Our work provides new and important information for vaccine design and therapeutic antibody development against PCV2 infections.

ABSTRACT

Porcine circovirus type 2 (PCV2) is an important pathogen in swine herds, and its infection of pigs has caused severe economic losses to the pig industry worldwide. The capsid protein of PCV2 is the only structural protein that is associated with PCV2 infection and immunity. Here, we report a neutralizing monoclonal antibody (MAb), MAb 3A5, that binds to intact PCV2 virions of the PCV2a, PCV2b, and PCV2d genotypes. MAb 3A5 neutralized PCV2 by blocking viral attachment to PK15 cells. To further explore the neutralization mechanism, we resolved the structure of the PCV2 virion in complex with MAb 3A5 Fab fragments by using cryo-electron microscopy single-particle analysis. The binding sites were located at the topmost edges around 5-fold icosahedral symmetry axes, with each footprint covering amino acids from two adjacent capsid proteins. Most of the epitope residues (15/18 residues) were conserved among 2,273 PCV2 strains. Mutations of some amino acids within the epitope had significant effects on the neutralizing activity of MAb 3A5. This study reveals the molecular and structural bases of this PCV2-neutralizing antibody and provides new and important information for vaccine design and therapeutic antibody development against PCV2 infections.

IMPORTANCE PCV2 is associated with several clinical manifestations collectively known as PCV2-associated diseases (PCVADs). Neutralizing antibodies play a crucial role in the prevention of PCVADs. We demonstrated previously that a MAb, MAb 3A5, neutralizes the PCV2a, PCV2b, and PCV2d genotypes with different degrees of efficiency, but the underlying mechanism remains elusive. Here, we report the neutralization mechanism of this MAb and the structure of the PCV2 virion in complex with MAb 3A5 Fabs, showing a binding mode in which one Fab interacted with more than two loops from two adjacent capsid proteins. This binding mode has not been observed previously for PCV2-neutralizing antibodies. Our work provides new and important information for vaccine design and therapeutic antibody development against PCV2 infections.

INTRODUCTION

Porcine circovirus type 2 (PCV2) is a member of the genus Circovirus in the family Circoviridae. PCV2 is nonenveloped and is one of the smallest animal viruses, with a diameter of approximately 17 nm. This virus contains a covalently closed circular single-stranded DNA (ssDNA) genome that contains five major open reading frames (ORFs), namely, ORF1, ORF2, ORF3, ORF4, and ORF5 (1–8). As with other ssDNA viruses, PCV2 is characterized by a high evolutionary rate, leading to the emergence of variants with different biological and epidemiological features. ORF2-based classification criteria have been collectively adopted to define PCV2 genotypes, and eight PCV2 genotypes exist, namely, PCV2a, PCV2b, PCV2c, PCV2d, PCV2e, PCV2f, PCV2g, and PCV2h (9, 10). PCV2a, PCV2b, and PCV2d are the major genotypes, and PCV2d strains are currently dominating worldwide (9, 10). Thus far, only one serotype has been found among PCV2 strains. Although small and simply structured, PCV2 is associated with various syndromes, including postweaning multisystemic wasting syndrome, respiratory disease complex, reproductive disorders, and enteric diseases (11). These syndromes are collectively designated PCV2-associated diseases (PCVADs), and they have a severe impact on the worldwide swine industry (11).

ORF2 encodes the virus structural capsid protein (Cap protein). The circular genome is packaged by 60 Cap protein subunits arranged in 12 pentamer-clustered units, resulting in a homopolymer icosahedral virion (12). Cap protein can bind host cell receptors and induces specific immunity (13–18). Although PCV2-infected pigs produce high levels of Cap-specific antibody, the onset and severity of PCVADs are correlated with the absence or decreased levels of PCV2-neutralizing antibodies (19, 20), suggesting a crucial role for neutralizing antibodies in the prevention of PCVADs. Therefore, the neutralizing epitopes and neutralization mechanisms need to be elucidated. Several antigenic domains on the PCV2 Cap protein have been identified in experiments using porcine polyclonal antibodies or mouse monoclonal antibodies (MAbs) (21–24). Genotype-specific domains (six amino acid residues at positions 86 to 91 and four at positions 190, 191, 206, and 210) on the PCV2 Cap protein have also been identified using multiple sequence alignment (25, 26). Lekcharoensuk et al. identified at least five overlapping conformational epitopes within residues 47 to 85, 165 to 200, and 230 to 233 of the PCV2 Cap protein, using chimeric PCV1/PCV2 infectious clones (21). Conformational epitopes recognized by MAbs with neutralizing activity against PCV2 have been determined in transfected PK-15 cells, and residues 145 to 162, 175 to 192, and 231 to 233 participate in the formation of conformational epitopes (24). MAbs with different immunoperoxidase monolayer assay (IPMA) reactivity or neutralization phenotypes with different strains have been used to identify critical amino acids of conformational epitopes, and mutations at positions 59, 60, 190/151, and 131/191 were shown to result in a switch in IPMA reactivity and neutralization phenotype (27–29).

Although many methods have been used to study the conformational epitopes of PCV2, the overall appearance of the neutralizing conformational epitope remains unclear. Cryo-electron microscopy (cryo-EM) is an electron microscopy (EM) technique used for macromolecular structure determination that does not require crystallization, and many structures of virus-Fab complexes have been determined to map conformational epitopes (30–34).

Here, we report that MAb 3A5 neutralizes PCV2 strains from PCV2a, PCV2b, and PCV2d genotypes. To further explore the structural basis of neutralization, we resolved the cryo-EM structure of a PCV2 virion in complex with the Fab fragments of MAb 3A5. Based on the footprints of the antibodies revealed by the structure, we tested the neutralization abilities of MAb 3A5 for manually rescued virus strains, each of which had a substituted amino acid in the epitope. These studies are the first structural investigations of a broad-spectrum, conformational, and dominant epitope of the PCV2 Cap protein.

RESULTS

MAb 3A5 blocks PCV2 attachment.

We previously produced and identified a murine MAb, MAb 3A5, with potent neutralization effects on the PCV2a/LG strain (35), but its potential mechanism is still unclear. To determine whether this MAb blocks virus attachment, we first measured the 50% inhibitory concentration (IC₅₀) values of MAb 3A5 IgG and its Fab using virus neutralization assays, and we found that the IC₅₀s were 0.16 ng/μl and 4 ng/μl, respectively (Fig. 1A). We then performed an indirect immunofluorescence assay (IFA) to assess the influence of MAb 3A5 IgG on PCV2 cell attachment. Purified PCV2a/LG was incubated for 1 h at 37°C with MAb 3A5 IgG or MAb 10B10 against foot-and-mouth disease virus (FMDV), and then the mixture was used to infect PK15 cells for 1 h at 4°C. MAb 8G12 against PCV2 was used to stain PCV2 on the surface of PK15 cells after three washes. As shown in Fig. 1B, the PCV2 signal (green) was not detected on the PK15 cell surface in the PCV2 and MAb 3A5 incubation group but was present in the PCV2 and MAb 10B10 incubation group. There was no PCV2 signal in the mock-infected PK15 cells (Fig. 1B). These findings suggest that MAb 3A5 neutralizes the virus by blocking virus attachment on the cell surface.

FIG 1.

Neutralization of PCV2 by MAb 3A5. (A) Neutralizing activity of MAb 3A5 IgG and Fab against PCV2a/LG. A mean neutralizing activity of >50% was considered to represent neutralization. The minimum antibody concentration that reduced the number of PCV2-infected cells by 50% was defined as the IC₅₀ for the antibody. The dotted line indicates the cutoff value. Error bars indicate the standard deviations. IC₅₀s were calculated and are shown on the right. (B) Detection of PCV2 by IFA. Purified PCV2a/LG was incubated with MAb 3A5 IgG or MAb 10B10 for 1 h at 37°C, and then the mixture was used to infect PK15 cells for 1 h at 4°C. The cells were fixed with 4% (wt/vol) paraformaldehyde, stained with anti-PCV2 MAb 8G12 (PCV2) (green) and DAPI (nucleus) (blue), and observed by fluorescence confocal microscopy. Scale bars, 10 μm. Mock-infected PK15 cells were used as a negative control.

Cryo-EM structure of a PCV2 virion complexed with MAb 3A5 Fab fragments.

To elucidate the structural basis for the neutralizing activity of MAb 3A5, we resolved the cryo-EM structures of a PCV2 and a complex of the virion and MAb 3A5 Fab fragments (virion-Fab complexes). Electron micrographs of PCV2 virions or virion-Fab complexes were collected using a Talos F200C microscope with a Ceta charge-coupled device (CCD). In contrast to the native virions, the particles of the complexes became spiky in appearance in both negative-staining images (Fig. 2A and C) and cryo-EM images (Fig. 2B and D), indicating that the Fabs had attached to the surface of virions. A total of 20,111 particles were picked from 433 micrographs, using RELION auto-picking, for reconstruction of the PCV2 virion. For virion-Fab complexes, 8,618 particles were selected from 2,774 images and used for two-dimensional (2D) classification (Fig. 2E); approximately 3,500 particles were used for three-dimensional (3D) reconstruction. According to the 0.143 Fourier shell correlation (FSC) criterion, the final cryo-EM structures of the PCV2 virion and virion-Fab complex were determined with overall resolutions of 6.7 Å and 7.2 Å, respectively (Fig. 3A and B and Fig. 4A and B). Although the resolution of the cryo-EM structures (Fig. 4B) was moderate, the densities of the virion and the Fab fragments were well defined. In contrast to those of the virion surface binding regions, the densities of the distal regions of the Fab were not well resolved, suggesting that these regions were flexible. Fab fragments were bound on the surface of the virion, showing an apical feature (Fig. 4B and C). Given that the capsid of the virion consists of 60 copies of the Cap protein, theoretically a total of 60 Fab fragments would be needed to obtain saturation. We then used a block-based reconstruction method to evaluate the saturation status based on the reconstructions, and the results showed that about 50% of the epitopes were bound with Fabs.

FIG 2.

EM micrographs of PCV2 virions and PCV2-Fab complexes. (A) EM image of negatively stained PCV2 virions. Scale bar, 200 nm. (B) Image of virions embedded in vitreous ice, recorded with a Ceta CCD on an FEI Talos cryo-EM microscope operated at 200 kV at liquid-nitrogen temperature. Scale bar, 50 nm. (C) EM image of negatively stained PCV2-Fab complexes. Scale bar, 100 nm. (D) Image of PCV2-Fab complexes embedded in vitreous ice, recorded with a Ceta CCD on an FEI Talos cryo-EM microscope operated at 200 kV at liquid-nitrogen temperature. Scale bar, 50 nm. (E) Reference-free 2D class averages of PCV2-Fab complexes.

FIG 3.

Gold-standard FSC curve for the final map. Resolution assessments of the PCV2 virion 3D reconstruction (A) and the PCV2-Fab complex 3D reconstruction (B), based on the 0.143 criterion for the reference-based FSC coefficient.

FIG 4.

Cryo-EM reconstructions of the PCV2 virion and PCV2-Fab complex. (A and B) Radially colored cryo-EM 3D reconstructions of the PCV2 virion (A) and the PCV2-Fab complex (B), viewed along a 5-fold axis. (C) Cutaway view showing DNA (gray), capsid protein (green), and Fab (yellow). An enlarged density map of a Cap protein in complex with a Fab is also shown, with ribbons representing the homology models of a Fab (yellow) and Cap protein (green) fitted in.

To obtain the atomic coordinates of the complex, we fitted homology models of the Cap and Fab fragment into the virion-Fab complex density map. The resulting structure was further refined against the density map in real space to obtain a final structure of the complex. Based on the structure of the virion-Fab fragment complex, the complementarity-determining regions (CDRs) of the Fab variable domains and the Cap protein interfaces were defined. The CDRs of each Fab interacted directly with two adjacent Cap proteins. The binding was mediated mainly by the CDRs of the heavy chain of the Fab and the flexible loops (B-C, C-D, E-F, G-H, and H-I loops) and one C terminus of the two adjacent capsid proteins (Fig. 5). Compared with the other loops, the B-C and E-F loops were more prominent on the surface of the virion. It is speculated that the interactions of amino acid residues on these two loops with Fab may be more critical.

FIG 5.

Binding mode of MAb 3A5 to two Cap proteins. The structures are ribbons represented in different colors (medium blue, Fab light chain; steel blue, Fab heavy chain; forest green, one Cap protein; cyan, the adjacent Cap protein). The names of the interacting loops from both the Fab and Cap proteins are indicated in the enlarged interface structure.

Footprint of MAb 3A5.

To obtain information on the epitope of MAb 3A5, the Fab interactions were analyzed using Chimera and the footprints of MAb 3A5 on the virion surface were defined by the atoms in the virus that were closer than 4 Å to any atom in the bound Fab fragments using RIVEM. The epitope was approximately 671 Ų, as shown by the PISA program (www.ebi.ac.uk/pdbe/pisa), and consisted of 23 amino acid residues, of which 55Y, 56T, 57V, 58K, 59A, 60T, 61T, 62V, 68A, 69V, 70D, 128N, 131T, 132K, 133A, 134T, 135A, 136L, 137T, and 204N were located in one Cap protein; the remaining 3 residues (88K, 189T, and 231L) were located in the adjacent Cap protein (Fig. 6A). To further analyze whether these predicted amino acids on the surface of the capsid were exposed to facilitate Fab binding, we analyzed the surface accession ability of the amino acids. Among these footprint-defined residues, most were located in the flexible loops on the surface of the virion, including the B-C loop (58K, 59A, 60T, 61T, and 62V), the E-F loop (128N, 131T, 132K, 133A, 134T, 135A, 136L, and 137T), the C-D loop (88K), and the G-H loop (189T). The flexibility of the loops facilitates the binding between the virion and Fabs. In addition, two amino acids (55Y and 56T) were located in β-strand B and one residue (231L) in the C-terminal region. However, 57V, 68A, 69V, 70D, and 204N were shown to be located in the bottom of a small groove and could not be directly contacted by the residues in the Fab (Fig. 6B). Therefore, we defined the epitope as being composed of 18 amino acid residues, including the residues 55Y, 56T, 58K, 59A, 60T, 61T, 62V, 128N, 131T, 132K, 133A, 134T, 135A, 136L, 137T, 88K, 189T, and 231L.

FIG 6.

View of the Fab binding surface of the PCV2 capsid. (A) Fab footprint on the surface of the virion. The surface of the virion is shown as a stereographic projection, in which the polar angles θ and ϕ represent latitude and longitude, respectively. The MAb 3A5 Fab footprints are shown in orange red. The locations of the 2-fold, 3-fold, and 5-fold icosahedral symmetry axes are indicated as black ovals, triangles, and pentagons, respectively. In the expanded view, amino acid residues of the virion are indicated. (B) Interface shown on the surface of two adjacent capsid proteins using Chimera 1.13.1 (University of California, San Francisco). Residues 57V, 68A, 69V, 70D, and 204N are colored orange, yellow, green, magenta, and cyan, respectively, and the rest of the amino acids are shown in cornflower blue.

Conservation of the MAb 3A5 epitope.

Given that MAb 3A5 was capable of reacting with the PCV2 strains of genotypes PCV2a, PCV2b, and PCV2d (35), we then analyzed the conservation of the amino acids in the epitope. A total of 2,273 Cap protein sequences of PCV2, which were all available in GenBank, were used in the analysis. The results showed that most of the residues, including 55Y, 56T, 61T, 62V, 128N, 132K, 135A, 137T, 189T, and 231L, were almost completely conserved (>99%), while the rest showed variable levels of conservation, with 5 highly conserved (86.89 to 98.33%) positions (positions 58, 60, 131, 133, and 136), 1 moderately conserved (approximately 60%) position (position 134), and 2 highly variable (<25%) positions (positions 59 and 88) (Fig. 7A and B). The majority of the amino acids in the epitope were conserved among different PCV2 strains, which might explain the relatively broad-spectrum reactivity of MAb 3A5 with PCV2 strains.

FIG 7.

Conservation analysis of the MAb 3A5 epitope. (A) Location of amino acids within the MAb 3A5 epitope on the PCV2 Cap protein. A homology model of the PCV2a/LG capsid protein was generated with Discovery Studio Client (version 2.5), using the PCV2 VLP cryo-EM structure (PDB accession number 3JCI) (44) as a template. The figure was generated using Chimera 1.13.1. (B) Extent of conservation. Multiple alignments of amino acid sequences of the Cap protein from 2,273 PCV2 strains were performed using MEGA X. PCV2 Cap protein amino acids within the epitope were analyzed, and the conservation degree was calculated for every amino acid with the following formula: 100% × (number of conserved sequences/number of total sequences). The extent of conservation is depicted by color; red indicates conservation of 90 to 100%, orange indicates conservation of 70 to 89%, yellow indicates conservation of 50 to 69%, and green indicates conservation of 10 to 29%.

Effects of amino acid mutations within the MAb 3A5 epitope on the neutralization activity of MAb 3A5.

To evaluate the effects of amino acids within the epitope on the neutralization activity of MAb 3A5 with PCV2a/LG, we selected 18 amino acids in the epitope of MAb 3A5 and conducted alanine-scanning mutagenesis (except for alanine codons, which were mutated to serine), and the corresponding 18 mutants were rescued in PK15 cells, including M-Y55A, M-T56A, M-K58A, M-A59S, M-T60A, M-T61A, M-V62A, M-K88A, M-N128A, M-T131A, M-K132A, M-A133S, M-T134A, M-A135S, M-L136A, M-T137A, M-T189A, and M-L231A. The supernatants of mutant-, parental clone-, and mock-transfected PK15 cells were collected at 72 h posttransfection and titrated in PK15 cells. The titers of M-T61A, M-K132A, and M-T189A were lower than that of the parental clone (P < 0.05), while those of M-T56A and M-K88A were significantly lower than that of the parental clone (Fig. 8A). Furthermore, the titer of M-Y55A was below the detection limit, suggesting that the Y55A mutation was lethal (Fig. 8A). The titers of the other mutants were similar to those of the parental clone (Fig. 8A).

FIG 8.

Neutralization of PCV2 mutants by MAb 3A5. (A) Virus titers of the mutant viruses and the parental virus. The dotted line indicates the cutoff value. The P values were determined by an unpaired t test, with Welch’s correction, from virus titer values. *, P < 0.05; **, P < 0.01; ***, P < 0.001. (B) IC₅₀s of MAb 3A5 for the PCV2 mutants. (C) Neutralization activity of the mutants and parental viruses with MAb 3A5 at a concentration of 0.16 ng/μl. The P values were determined by an unpaired t test, with Welch’s correction, from the neutralizing activity values. *, P < 0.05; **, P < 0.01; ***, P < 0.001. (D) Levels of binding of mutant viruses to MAb 3A5. The normalized means and standard deviations (error bars) in panels A, B, C, and D were calculated from three replicates with GraphPad Prism 8.0.1.

With the exception of M-Y55A, the mutant viruses, together with the parental virus, were used to conduct neutralization assays with MAb 3A5 and MAb 10B10 (anti-FMDV). The single mutations T56A, T131A, T134A, A135S, L136A, and T189A within the epitope reduced the neutralization efficacy of MAb 3A5 by 5- to 125-fold, while the mutations A59S, T60A, and N128A reduced the neutralization efficacy of the MAb by <5-fold (Fig. 8B). At a concentration of 0.16 ng/μl, MAb 3A5 neutralization activities for M-K88A and M-K132A were lower than that for the parental virus (P < 0.05) (Fig. 8C). Thus, the mutations K88A and K132A also affected neutralization efficiency. To identify whether the reduced neutralization efficacy of MAb 3A5 with these mutants was related to the loss of MAb binding to the mutants, a capture enzyme-linked immunosorbent assay (ELISA) based on MAb 3A5 was used to detect the binding reactivity of the MAb with the mutants. The single mutations N128A, T131A, K132A, T134A, A135S, L136A, and T189A in the Cap protein resulted in >50% loss of binding reactivity of MAb 3A5, while the mutation A59S reduced the binding reactivity of the MAb by approximately 30% (Fig. 8D). The binding levels of M-T56A and M-T60A were similar to that of parental virus (Fig. 8D); however, their neutralization activities were significantly lower than that of parental virus. Therefore, antibody neutralization and binding are related but are not the same.

MAb 3A5 neutralizes the PCV2a, PCV2b, and PCV2d genotypes with different degrees of efficiency.

We demonstrated previously that MAb 3A5 reacted with PCV2 strains from the PCV2a, PCV2b, and PCV2d genotypes (35). We then tested the IC₅₀ of MAb 3A5 for two PCV2a strains, two PCV2b strains, and two PCV2d strains, and the results are shown in Fig. 9A. Consistent with our previous findings (35), MAb 3A5 exhibited potent activity against PCV2a strain PCV2a/CL, with an IC₅₀ of 0.16 ng/μl, which is equivalent to that for the PCV2a/LG strain. In contrast, this MAb did not show as powerful neutralization activities against the PCV2b and PCV2d strains as it did against the PCV2a strains, with IC₅₀ values of 20 ng/μl, 4 ng/μl, 4 ng/μl, and 20 ng/μl for the strains PCV2b/YJ, PCV2b/2206, PCV2d/2209, and PCV2d/2811, respectively. Taken together, these results strongly suggest that MAb 3A5 neutralizes PCV2 strains from all three genotypes with different degrees of efficiency. The Cap protein sequences from the aforementioned six PCV2 strains were analyzed with MegAlign (Fig. 9B). We found that 59A, 88K, and 134T were not conserved among the six PCV2 strains but they were conserved within specific genotypes (Fig. 9B). Furthermore, the conservation levels for 59A, 88K, and 134T were medium or low in the 2,273 PCV2 sequences (Fig. 7A and B). To further elucidate the functions of these three amino acids in the neutralization of MAb 3A5, we mutated amino acids at positions 59, 88, and 134 in PCV2a/LG, PCV2b/2206, and PCV2d/2209, rescued the mutant viruses (PCV2a/LG-A59K, PCV2a/LG-K88P, PCV2a/LG-T134N, PCV2b/2206-K59A, PCV2b/2206-P88K, PCV2d/2209-K59A, PCV2d/2209-P88K, and PCV2d/2209-N134T), and determined the neutralization efficacy of MAb 3A5 against these mutants. The single mutations A59K, K88P, and T134N reduced the neutralization efficacy of MAb 3A5 against PCV2a/LG (Fig. 9C). The K59A mutation increased the neutralizing effect of MAb 3A5 against PCV2b/2206, and the same effect was observed with the N134T mutation in PCV2d/2209 (Fig. 9D and E). These findings indicated that positions 59 and 134 were involved in the differential neutralization efficiency of MAb 3A5 against PCV2a/LG, PCV2b/2206, and PCV2d/2209.

FIG 9.

MAb 3A5 neutralization of the PCV2a, PCV2b, and PCV2d genotypes with different degrees of efficiency. (A) Neutralizing activities of MAb 3A5 against PCV2a, PCV2b, and PCV2d strains. The dotted line indicates the cutoff value. Error bars represent the standard deviations of the neutralizing activities. IC₅₀s were calculated and are shown on the right. (B) Comparison of Cap protein amino acids among the strains. Multiple alignments of amino acid sequences of the Cap protein of six PCV2 strains were performed using ClustalW (version 7.0; DNASTAR), and amino acids within the epitope are shown. (C to E) IC₅₀ values of MAb 3A5 for PCV2a/LG and its mutants (C), PCV2b/2206 and its mutants (D), and PCV2d/2209 and its mutants (E). The normalized means and standard deviations (error bars) were calculated from three replicates with GraphPad Prism 8.0.1.

DISCUSSION

Neutralizing antibodies are the predominant contributors to antiviral immunity, and they are also important indicators of successful vaccination against PCV2. Therefore, the antibody-mediated neutralization mechanism needs to be studied to provide useful information for vaccine design and therapeutic antibody development for PCV2 infections. MAb 3A5 has been shown to be highly effective in binding PCV2 strains from the PCV2a, PCV2b, and PCV2d genotypes in IPMAs and neutralizing PCV2 in PK15 cells (35). However, no positive reaction between MAb 3A5 and the PCV2 Cap protein was observed with Western blotting (35). It was speculated that MAb 3A5 might bind to a conformational epitope of PCV2 virions. Although MAb 3A5 is of murine origin, the epitope of this MAb is dominant in PCV2-infected and PCV2-immunized pigs. Specific antibodies produced by PCV2-infected and PCV2-immunized pigs could be detected using a competitive ELISA based on MAb 3A5 (35). Conformational epitopes are difficult to completely identify using traditional methods (21, 24, 27–29). Therefore, we resolved the structure of a PCV2-Fab complex by using cryo-EM single-particle analysis and mapped this neutralizing and conformational epitope on the surface of the virion.

The epitope of MAb 3A5 is located at the topmost edges around the 5-fold axes of the virion and is composed of amino acids from two adjacent Cap proteins. These findings thus suggested that the complete PCV2 capsid structure is necessary for the production of these neutralizing antibodies. The results are consistent with a previous study that demonstrated that the PCV2 Cap monomer elicited strong antibody responses, predominantly against a decoy epitope of Cap (residues 169 to 180, inside the viral capsid) that did not protect swine from PCV2 infection (36); in contrast, neutralizing antibodies were successfully induced using PCV2 virus-like particles (VLPs) as antigens in swine, indicating that these neutralizing antibodies recognized conformational epitopes of PCV2 VLPs assembled from Cap (36).

Most of the amino acid residues (15 of 18 residues) of the epitope were highly conserved among the 2,273 PCV2 sequences (Fig. 7B). This may be the reason why MAb 3A5 reacted to all PCV2a, PCV2b, and PCV2d isolates in our laboratory (35). However, there was a significant difference in the neutralization efficiency of MAb 3A5 against PCV2 strains of different genotypes in this study (Fig. 9A). This difference might be associated with genotypic amino acid differences in the MAb 3A5 epitope (residues 59 and 134) (Fig. 9B). The neutralizing effect of MAb 3A5 on the mutant strains (PCV2a/LG-A59K, PCV2a/LG-K88P, and PCV2a/LG-T134N) was significantly reduced, compared to that on the parental virus, while the neutralizing effect of MAb 3A5 on the mutant strains (PCV2b/2206-K59A and PCV2d/2209-N134T) was enhanced, compared to that on their parental viruses (Fig. 9C to E). Mutations at these genotypic difference sites (residues 59 and 134) reduced the binding capacity of MAb 3A5 (Fig. 8D), which in turn led to a decrease in the neutralizing effect of the MAb for PCV2 strains of other genotypes.

A single amino acid mutation in the MAb 3A5 epitope not only could affect the neutralizing efficiency of the MAb against the mutant PCV2 (Fig. 8B and C) but also could alter the reproductive capacity of the mutant virus (Fig. 8A). Some PCV2 mutant strains had no significant differences in reproductive characteristics, compared with the parent strains, while some mutants, such as M-Y55A, had very low or undetected titers (Fig. 8A). The amino acid at position 55 is located in β-strand B and is highly conserved. Therefore, this mutation may affect the structure of PCV2 and be a lethal mutation. Although some mutant strains could be passaged (data not shown), the titers of these strains were significantly lower than that of the parent virus. The point mutations may affect virus attachment, virus entry, or other steps of the replication cycle, thereby reducing the infectivity of the virus.

PCV2 could be neutralized by MAb 3A5 at the preattachment stage through two mechanisms. First, MAb 3A5 could neutralize PCV2 by preventing the absorption of the virus to the cell surface. Second, MAb 3A5 IgG might cause PCV2 aggregation. Both the IgG and Fab of MAb 3A5 could neutralize PCV2, but the IgG showed higher efficiency than the Fab (Fig. 1A). Previous studies showed that some antibodies inhibited infectivity by binding to viruses and causing them to aggregate (37). This could explain the additional efficiency displayed by the MAb 3A5 IgG in our study, for the Fab without the Fc fragment lost its virus-aggregating capability.

In summary, we found that MAb 3A5 binding to a PCV2 Cap conformational epitope neutralized PCV2a, PCV2b, and PCV2d strains with different neutralization efficiencies. Furthermore, MAb 3A5 neutralized PCV2 by blocking viral attachment to PK15 cells. To understand the neutralization mechanism, we resolved the structure of a PCV2 virion in complex with the Fab fragments of MAb 3A5 using cryo-EM. The epitopes of MAb 3A5 were found to be located at the topmost edges around the 5-fold axes, with each epitope covering 18 amino acids from two adjacent Cap proteins. Fifteen amino acid residues of the epitope were highly conserved. Finally, mutations of some amino acids within the epitope were shown to affect the neutralization activity of MAb 3A5 for PCV2. This study reveals the molecular and structural bases of PCV2-neutralizing antibody activity and suggests that the complete PCV2 capsid structure is the basis for ensuring good immunogenicity of PCV2.

MATERIALS AND METHODS

Cells, viruses, and antibodies.

PK15 cells (ATCC CCL-33) free of PCV1 and PCV2 were grown in Dulbecco’s modified Eagle’s medium (DMEM) (Invitrogen, Carlsbad, CA, USA) containing 10% heat-inactivated fetal bovine serum (FBS) (Gibco, Grand Island, NY, USA), 100 U/ml penicillin, and 0.1 mg/ml streptomycin. SP2/0 cells cultured in DMEM containing 10% FBS, 100 U/ml penicillin, and 0.1 mg/ml streptomycin were used for MAb preparation. Six different PCV2 strains were used in this study. Strain origins, genotypes, and accession numbers are shown in Table 1. MAbs 3A5 and 8G12 and PCV2-positive serum were prepared and maintained in our laboratory (17, 35). MAb 10B10 against FMDV (38) was a gift from Yu Li and was used as an isotype control.

TABLE 1.

Origins of the PCV2 strains used in this study

Isolate name	Isolation region	Isolation year	Genotype	Genome length (nucleotides)	GenBank accession no.
LG	Jilin	2008	PCV2a	1,768	HM038034
CL	Jilin	2007	PCV2a	1,768	HM038033
YJ	Heilongjiang	2008	PCV2b	1,767	HM038032
2206	Jilin	2018	PCV2b	1,767	MK347362
2209	Liaoning	2018	PCV2d	1,767	MK347401
2811	Shandong	2017	PCV2d	1,767	MK347389

PCV2 purification.

The PCV2a/LG strain was cultured in PK15 cells and purified by CsCl gradient centrifugation, as described previously (12, 17). Purified particles were stored at –80°C.

Fab/antibody preparation.

MAb 3A5 was purified from ascites using a protein G IgG purification kit (GE Healthcare, Uppsala, Sweden), and its Fab was then generated from the purified MAb using a Fab preparation kit (Thermo Fisher Scientific, Rockford, IL, USA).

Virus neutralization assays.

Serial dilutions of MAb 3A5 IgG and its Fab (20 ng/μl, 4 ng/μl, 0.8 ng/μl, 0.16 ng/μl, and 0.032 ng/μl) were incubated for 1 h at 37°C with 10^4.0 times the 50% tissue culture infective dose (TCID₅₀)/ml PCV2a/LG. MAb 3A5-virus complexes were added to PK15 cells (50% confluence) in 96-well plates. After 60 min, the cells were washed three times with DMEM and then cultured at 37°C in fresh DMEM supplemented with 2% FBS. Thirty-six hours later, the plates were fixed with 100% cold methanol at –20°C. An IPMA with PCV2-positive serum was performed as described previously (16, 17). The number of infected cells per well was determined by light microscopy (Eclipse TE2000-S; Nikon, Japan). The procedure was performed for the PCV2-positive serum group, the isotype control (MAb 10B10 against FMDV) group, and the mock-infected cell control group. The neutralizing activity of an antibody is expressed as the percent reduction in the number of infected cells, compared with that for the MAb 10B10 control. The neutralizing activities of different concentrations of antibody for viruses were calculated. The minimum antibody concentration that reduced the number of PCV2-infected cells by 50% was defined as the IC₅₀ of the antibody. Six different PCV2 strains (two PCV2a strains, two PCV2b strains, and two PCV2d strains) were used to test the neutralizing activity of MAb 3A5 IgG (100 ng/μl, 20 ng/μl, 4 ng/μl, 0.8 ng/μl, 0.16 ng/μl, and 0.032 ng/μl) for different PCV2 genotypes with preattachment neutralization assays. The neutralizing activities of different concentrations of antibody for viruses and their IC₅₀ values for different strains were calculated.

IFA.

For determination of whether MAb 3A5 could inhibit the adsorption of PCV2 to PK15 cells, the virus adsorbed onto the cells was tested by IFA. Purified PCV2a/LG (10^6.5 TCID₅₀/ml) was mixed with MAb 3A5 IgG (80 ng/μl) at 37°C for 1 h and then added to PK15 cells (50% confluence) at 4°C for 1 h. Cells were washed three times with cold DMEM. The cells for IFA were fixed with 4% (wt/vol) paraformaldehyde in phosphate-buffered saline (PBS) for 10 min at room temperature. After the samples were washed, PCV2 was stained by incubating the cells for 1 h at 37°C with anti-PCV2 MAb 8G12 (35), followed by 1 h of incubation at 37°C with fluorescein isothiocyanate (FITC)-conjugated goat anti-mouse IgG1 (1:400 in PBS; Molecular Probes). For visualization of the nuclei, the cells were stained with 4′,6-diamidino-2-phenylindole (DAPI) for 10 min at room temperature, after three washes with PBS. Stained cells were analyzed by fluorescence confocal microscopy (LSM 800 confocal laser scanning microscope with Airyscan; Zeiss, Germany). Mock-infected PK15 cells and PCV2-infected PK15 cells with MAb 10B10 as an isotype control MAb were used as negative and positive controls, respectively.

Fab/antibody preparation.

MAb 3A5 was purified from ascites using a protein G IgG purification kit (GE Healthcare, Uppsala, Sweden), and its Fab was then generated from the purified MAb using a Fab preparation kit (Thermo Fisher Scientific).

Sequence determination of the MAb 3A5 variable domain.

Hybridoma cells that secreted MAb 3A5 were pelleted by centrifugation at 1,000 rpm for 5 min, and RNA was extracted using an RNeasy minikit (Qiagen, Hilden, Germany). MAb 3A5 cDNAs were synthesized from the treated RNA with a RevertAid first-strand cDNA synthesis kit (Thermo Fisher Scientific). The cDNAs were used as the template for PCR and amplified using KOD FX Neo DNA polymerase (Toyobo, Osaka, Japan). PCR amplification was performed with the amplification primers described previously by Wang et al. (39). PCR products were cloned into the pMD18-T vector (TaKaRa, Dalian, China) and then sequenced for the variable domains of the heavy and light chains.

Negative-staining EM.

Purified MAb 3A5 Fab fragments were incubated for 1 h at 4°C with purified PCV2 particles, at a ratio of 60:1. PCV2 virions and virion-Fab complexes were dipped into a carbon-coated grid for 30 s, and excess solution was removed with filter paper. The grids were immediately negatively stained for 20 s with 2% phosphotungstic acid (pH 6.5). The grids were examined with a Hitachi H7650 transmission electron microscope at an accelerating voltage of 80 kV and were photographed at a magnification of 30,000×.

Cryo-EM imaging.

For cryo-EM grid preparation, a 3.5-μl sample aliquot was applied to a fresh glow-discharged Quantifoil Cu grid (R1.2/1.3, 200 mesh; Quantifoil Micro Tools) using an FEI Vitrobot instrument with a force setting of −4, a blotting time of 5 s, and a wait time of 1 s, and the grid was dipped into a liquid ethane bath. The grids were stored and handled in liquid nitrogen for data collection. Data were collected at a cryogenic temperature on an FEI Talos F200C electron microscope operating at 200 kV, using a dose rate of <35 e/Ų s. Images were captured with a Ceta CCD (4,096 × 4,096) at a nominal magnification of 92,000×, with defocus values from −1.5 to −3.0 μm. The CCD pixel size was 1.14 Å/pixel. The cryo-EM data collection parameters are shown in Table 2.

TABLE 2.

Cryo-EM data collection and atomic model refinement statistics

Parameter	PCV2-Fab complex	PCV2
Data collection and reconstruction
EM equipment	FEI Talos F200C	FEI Talos F200C
Voltage (kV)	200	200
Detector	Ceta CCD (4,096 × 4,096)	Ceta CCD (4,096 × 4,096)
Pixel size (Å)	1.14	1.14
Electron dose (e−/Å2)	35	35
Defocus range (μm)	−1.5 to −3.0	−1.0 to −3.0
Reconstruction software	RELION 2.0	RELION 3.0
No. of collection micrographs	2,774	433
No. of particles used	8,618	20,111
Symmetry imposed	I	I
Final resolution (Å)	7.2	6.7
Model building software	Coot	Coot
Refinement software	Phenix	Phenix
Model refinement and validation
PDB accession no.	6L62	6LM3
EMDB accession no.	EMD-0838	EMD-0916
Root-mean-square deviation
Bonds (Å)	0.01	0.01
Angles (°)	1	1
Ramachandran plot (%)
Favored	93.83	95.21
Allowed	6.17	4.79
Outliers	0	0
Rotamer outliers	0.55	0.58

Model refinement, building, and analysis.

Contrast transfer function (CTF) estimation and correction were performed using GCTF (40). For PCV2 virions, 20,111 particles were picked from 433 micrographs using RELION autopicking. The structure of the PCV2 virion was obtained after one round of reference-free 2D classification and reference-free 3D classification and two rounds of 3D autorefinement with lower initial offset range and initial offset step. For PCV2-Fab complexes, 8,618 particles were manually selected from 2,774 selected micrographs. After one round of reference-free 2D classification using 10 classes, the selected particles were then subjected to 3D classification, which generated two good 3D maps with the particles. One map of the 3D classification was then low-pass filtered to 60 Å and used as a reference for the final 3D density map reconstruction with I symmetry, postprocessing, and gold-standard FSC estimations, which were all performed with RELION (41). Taking advantage of the known cryo-EM structure of the VLP of circovirus, the homology model of PCV2 and Fab was built using the “build homology models” tool in Discovery Studio Client (version 2.5). Amino acid sequences of the target proteins were provided as input, PCV2 (Protein Data Bank [PDB] accession number 3JCI) (42) and Fab (PDB accession number 15C8), based on a BLAST search (DS server), were then identified as suitable templates, and the resulting model with the top-ranked global model quality estimate (GMQE) was chosen as the homology model. Atomic models were fitted into the segmented volume of the final cryo-EM coulombic potential map using the “fit in map” tool in Chimera (www.cgl.ucsf.edu/chimera). Atomic positions of the residues were iteratively modified using Coot (43) and refined using Phenix (44) five times in a real-space refinement program, to maximize the correlation coefficient between the EM density map and a calculated map based on the coordinates of all atoms. The collection parameters for cryo-EM data can be found in Table 2. A block-based reconstruction method (45) was used to determine the occupancy of the Fabs. Briefly, this method was used to clip the data set of PCV2-Fab complexes. The asymmetric unit of the icosahedral virus was divided into one block (interface of the epitope), resulting in 547,440 particles that were 3D classified using C1 symmetry with no image alignment and a mask generated from the complex with a single MAb 3A5 Fab fragment bound.

The Fab interactions were analyzed using Chimera and the PISA server (www.ebi.ac.uk/pdbe/pisa) (46), with zone distances of <4 Å for interactions. Fab density in the different maps was projected on a stereographic sphere, and the amino acid residues of the MAb 3A5 epitope were colored using RIVEM (47).

Conservation analysis of the MAb 3A5 epitope on PCV2 capsids.

Multiple alignments of amino acid sequences of the Cap proteins from 2,273 PCV2 strains were performed using MEGA X. PCV2 Cap amino acids within the epitope were analyzed, and the conservation degree of every amino acid was calculated with the following formula: 100% × (number of conserved sequences/number of total sequences).

Amino acid mutagenesis and virus rescue.

The interface of PCV2a/LG Cap with MAb 3A5 Fab was subjected to amino acid mutagenesis to generate 26 mutant viruses, as described previously (27). Eighteen of them underwent alanine-scanning mutagenesis (except for alanine codons, which were mutated to serine), and others were subjected to amino acid (positions 59, 88, or 134) exchange among PCV2a/LG-Cap, PCV2b/2206-Cap, and PCV2d/2209-Cap. Mutant names and primers are listed in Table 3. Mutant and infectious clone constructs were generated as described previously (27). Mutant viruses were rescued as described below. Mutated plasmids were digested with HindIII, and the subsequently purified genomic DNA fragments were self-ligated for 20 min at room temperature using T4 DNA ligase (Invitrogen). The DNA (1.2 μg) was then transfected into PCV-negative PK15 cells with Lipofectamine 3000 (Invitrogen), according to the manufacturer’s instructions. Empty plasmid-transfected PK15 cells were included as a mock-transfected control. In total, 26 mutants were rescued and named M-Y55A, M-T56A, M-K58A, M-A59S, M-T60A, M-T61A, M-V62A, M-K88A, M-N128A, M-T131A, M-K132A, M-A133S, M-T134A, M-A135S, M-L136A, M-T137A, M-T189A, M-L231A, PCV2a/LG-A59K, PCV2a/LG-K88P, PCV2a/LG-T134N, PCV2b/2206-K59A, PCV2b/2206-P88K, PCV2d/2209-K59A, PCV2d/2209-P88K, and PCV2d/2209-N134T. Virus titration was performed as described previously (20).

TABLE 3.

Primers used for construction of the mutant clones

Clone	Primer	Primer sequence
M-Y55A	Y55A-F	CCTGCACCTTCGGAGCGACTGTCAAGGCTAC
	Y55A-R	GTAGCCTTGACAGTCGCTCCGAAGGTGCAGG
M-T56A	T56A-F	GCACCTTCGGATATGCGGTCAAGGCTACCAC
	T56A-R	GTGGTAGCCTTGACCGCATATCCGAAGGTGC
M-K58A	K58A-F	CTTCGGATATACTGTCGCCGCTACCACAGTCAGA
	K58A-R	TCTGACTGTGGTAGCGGCGACAGTATATCCGAAG
M-A59S	A59S-F	GGATATACTGTCAAGTCCACCACAGTCAGAACG
	A59S-R	CGTTCTGACTGTGGTGGACTTGACAGTATATCC
M-T60A	T60A-F	GATATACTGTCAAGGCTGCGACAGTCAGAACGCC
	T60A-R	GGCGTTCTGACTGTCGCAGCCTTGACAGTATATC
M-T61A	T61A-F	CTGTCAAGGCTACCGCGGTCAGAACGCCCTC
	T61A-R	GAGGGCGTTCTGACCGCGGTAGCCTTGACAG
M-V62A	V62A-F	GTCAAGGCTACCACAGCCAGAACGCCCTCCTGG
	V62A-R	CCAGGAGGGCGTTCTGGCTGTGGTAGCCTTGAC
M-K88A	K88A-F	GGAGGGGGGACCAACGCCATCTCTATACCCTTTG
	K88A-R	CAAAGGGTATAGAGATGGCGTTGGTCCCCCCTCC
M-N128A	N128A-F	GCTGTTATTCTAGATGATGCGTTTGTAACTAAGGCCAC
	N128A-R	GTGGCCTTAGTTACAAACGCATCATCTAGAATAACAGC
M-T131A	T131A-F	CTAGATGATAACTTTGTAGCTAAGGCCACAGCCCTAACC
	T131A-R	GGTTAGGGCTGTGGCCTTAGCTACAAAGTTATCATCTAG
M-K132A	K132A-F	GATAACTTTGTAACTGCGGCCACAGCCCTAACC
	K132A-R	GGTTAGGGCTGTGGCCGCAGTTACAAAGTTATC
M-A133S	A133S-F	GATAACTTTGTAACTAAGTCCACAGCCCTAACCTATGAC
	A133S-R	GTCATAGGTTAGGGCTGTGGACTTAGTTACAAAGTTATC
M-T134A	T134A-F	CTTTGTAACTAAGGCCGCGGCCCTAACCTATGAC
	T134A-R	GTCATAGGTTAGGGCCGCGGCCTTAGTTACAAAG
M-A135S	A135S-F	GTAACTAAGGCCACATCCCTAACCTATGACCC
	A135S-R	GGGTCATAGGTTAGGGATGTGGCCTTAGTTAC
M-L136A	L136A-F	GTAACTAAGGCCACAGCCGCCACCTATGACCCCTATG
	L136A-R	CATAGGGGTCATAGGTGGCGGCTGTGGCCTTAGTTAC
M-T137A	T137A-F	CTAAGGCCACAGCCCTAGCCTATGACCCCTATGT
	T137A-R	ACATAGGGGTCATAGGCTAGGGCTGTGGCCTTAG
M-T189A	T189A-F	CTTTGGCTGAGACTACAAGCCTCGGCAAATGTGGAC
	T189A-R	GTCCACATTTGCCGAGGCTTGTAGTCTCAGCCAAAG
M-L231A	L231A-F	TAATCTTAAAGACCCCCCAGCCAAAACCCTAAATGAAT
	L231A-R	ATTCATTTAGGGTTTTGGCTGGGGGGTCTTTAAGATTA
PCV2a/LG-A59K	Forward	CGGATATACTGTCAAGAAAACCACAGTCAGAACGC
	Reverse	GCGTTCTGACTGTGGTTTTCTTGACAGTATATCCG
PCV2a/LG-K88P	Forward	AGGGGGGACCAACCCCATCTCTATACCCTTTG
	Reverse	CAAAGGGTATAGAGATGGGGTTGGTCCCCCCT
PCV2a/LG-T134N	Forward	TGTAACTAAGGCCAACGCCCTAACCTATG
	Reverse	CATAGGTTAGGGCGTTGGCCTTAGTTACA
PCV2b/2206-R59A	Forward	GATATACTATCAAGGCTACCACAGTCAGAAC
	Reverse	GTTCTGACTGTGGTAGCCTTGATAGTATATC
PCV2b/2206-P88K	Forward	GGGGCTCAAACAAACGCTCTGTGCC
	Reverse	GGCACAGAGCGTTTGTTTTGAGCCCC
PCV2d/2209-K59A	Forward	GTTATACTGTCAAGGCTACCACAGTCAGAAC
	Reverse	GTTCTGACTGTGGTAGCCTTGACAGTATAAC
PCV2d/2209-P88K	Forward	GGGGCTCAAACAAACTCACTGTGC
	Reverse	GCACAGTGAGTTTGTTTGAGCCCC
PCV2d/2209-N134T	Forward	CTTTGTAACAAAGGCCACAGCCCTAACCTATGAC
	Reverse	GTCATAGGTTAGGGCTGTGGCCTTTGTTACAAAG

Neutralization assay with mutants.

For evaluation of the function of amino acids within the epitope, the neutralizing activities of MAb 3A5 (20 ng/μl, 4 ng/μl, 0.8 ng/μl, 0.16 ng/μl, and 0.032 ng/μl) for different PCV2 mutants were analyzed using the neutralization assay, as described above. The number of infected cells per well was determined by light microscopy (Nikon Eclipse TE2000-S). The neutralizing activities of antibodies (different concentrations) for mutant viruses were calculated. The IC₅₀ values of MAb 3A5 IgG for different mutants were calculated and analyzed using GraphPad Prism software (GraphPad Software, CA, USA).

Antibody binding assay.

A capture ELISA based on MAb 3A5 (35) was used to test the binding reactivity of MAb 3A5 to different PCV2 mutants. In brief, MaxiSorp ELISA plates (Nunc, Glostrup, Denmark) were coated for 16 h at 4°C with purified MAb 3A5 diluted in 0.05 M carbonate buffer (10 μg/ml) and then blocked for 2 h at 37°C with 1% bovine serum albumin (BSA) diluted in PBS (1% BSA-PBS). Titers of M-T56A, M-A59S, M-T60A, M-N128A, M-T131A, M-K132A, M-T134A, M-A135S, M-L136A, M-T189A, and the parental virus were adjusted to 10^3.67 TCID₅₀/ml with 1% BSA-PBS; virus was then distributed in each well and incubated for 1 h at 37°C. After three washes with PBS containing Tween 20 (PBST), MAb 3A5 conjugated with horseradish peroxidase was diluted in dilution buffer (1:4000), added to the plates, and incubated for 1 h at 37°C. The colorimetric reaction was developed for 20 min by adding 0.21 mg/ml of 2,2-azino-di-(3-ethylbenzthiazoline sulfonic acid) in 0.1 M citrate (pH 4.2) containing 0.003% hydrogen peroxide, after three washes with PBST. The reaction was stopped by adding 50 μl of 1% NaF. The optical density at 405 nm (OD₄₀₅) was measured using a microplate reader (BioTek, Shoreline, WA, USA). Experiments were repeated in triplicate. The binding level of the parental virus was set to 100%. The binding level of each mutant was calculated according to the following formula: 100% × (OD₄₀₅ of mutant/OD₄₀₅ of parental virus).

Data availability.

Cryo-EM density maps of PCV2 virions and virion-Fab complexes were deposited in the Electron Microscopy Data Bank (EMDB) under accession numbers EMD-0838 and EMD-0916, respectively. PCV2 and PCV2-Fab complex coordinates were deposited in the Protein Data Bank under accession numbers 6L62 and 6LM3, respectively.