The Carboxyl Terminus of the Porcine Circovirus Type 2 Capsid Protein Is Critical to Virus-Like Particle Assembly, Cell Entry, and Propagation

The carboxyl terminus (CT) of porcine circovirus type 2 (PCV2) capsid protein (Cap) was previously reported to be associated with immunorecognition, alterations of viral titer in swine sera, and pathogenicity. However, the molecular mechanisms underlying these effects remain unknown. In this study, roles of the critical residues and motifs of the CT are investigated with respect to virus-like particle (VLP) assembly, cell entry, and viral proliferation. The results revealed that the positively charged ²²⁷K of the CT is essential for both cell entry of PCV2 VLPs and virus proliferation. Our findings, therefore, suggest that the CT should be considered one of the key epitopes, recognized by neutralizing antibodies, for vaccine design and a target for drug development to prevent PCV2-associated diseases (PCVADs). Furthermore, it is important to respect the function of ²²⁷K for its role in cell entry if using either PCV2 VLPs for nanoscale DNA/drug cell delivery or using PCV2 VLPs to display a variety of foreign epitopes for immunization.

ABSTRACT

The capsid protein (Cap) is the sole structural protein and the main antigen of porcine circovirus type 2 (PCV2). Structural loops of the Cap play crucial roles in viral genome packaging, capsid assembly, and virus-host interactions. Although the molecular mechanisms are yet unknown, the carboxyl terminus (CT) of the PCV2 Cap is known to play critical roles in the evolution, pathogenesis, and proliferation of this virus. In this study, we investigated functions of CT. Removal of this loop leads to abrogation of the in vitro Cap self-assembly into virus-like particles (VLPs). Likewise, the mutated virus resists rescue from PK15 cell culture. A conserved PXXP motif in the CT is dispensable for VLP assembly and subsequent cell entry. However, its removal leads to the subsequent failure of virus rescued from PK15 cells. Furthermore, substituting either the PCV1 counterpart or an AXXA for the PXXP motif still supports virus rescue from cell culture but results in a dramatic decrease in viral titers compared with wild type. In particular, a strictly conserved residue (²²⁷K) in the CT is essential for VLP entry into PK15 cells, and its mutation to alanine greatly attenuates cell entry of the VLPs, supporting a mechanism for the failure to rescue a mutated PCV2 infectious DNA clone (K227A) from PK15 cell culture. These results suggest the CT of the PCV2 Cap plays critical roles in virus assembly, viral-host cell interaction(s), and virus propagation in vitro.

IMPORTANCE The carboxyl terminus (CT) of porcine circovirus type 2 (PCV2) capsid protein (Cap) was previously reported to be associated with immunorecognition, alterations of viral titer in swine sera, and pathogenicity. However, the molecular mechanisms underlying these effects remain unknown. In this study, roles of the critical residues and motifs of the CT are investigated with respect to virus-like particle (VLP) assembly, cell entry, and viral proliferation. The results revealed that the positively charged ²²⁷K of the CT is essential for both cell entry of PCV2 VLPs and virus proliferation. Our findings, therefore, suggest that the CT should be considered one of the key epitopes, recognized by neutralizing antibodies, for vaccine design and a target for drug development to prevent PCV2-associated diseases (PCVADs). Furthermore, it is important to respect the function of ²²⁷K for its role in cell entry if using either PCV2 VLPs for nanoscale DNA/drug cell delivery or using PCV2 VLPs to display a variety of foreign epitopes for immunization.

INTRODUCTION

Porcine circoviruses (PCVs), a virus in the Circoviridae family, contain two main genotypes, namely, porcine circovirus type 1 (PCV1) and type 2 (PCV2) (1). Both types have the same genomic organization and share ∼70% similarity of nucleotide acid sequence. PCV1 was first isolated from the porcine kidney cell line (PK15) as a nonpathogenic contaminant agent by Tischer in 1974 (2). PCV2 has been identified as the causative agent of the postweaning multisystemic wasting syndrome (PMWS), which causes huge economic losses in swine-producing areas and countries (3). PCV2 is a small and nonenveloped animal virus with respect to both the genome size (1,766 to 1,768 nucleic acids) and the virus particle size (∼17 nm). The virion contains a circular, single-stranded DNA genome, which is packaged tightly in an icosahedral capsid assembled from 60 subunits of the PCV2 capsid protein (Cap) (4, 5). Two main open reading frames (ORF1 and ORF2) in the genome are separated by a replication origin (ori) with a stem-loop structure. ORF1 and its alternative splicing isoform encode Rep and Rep’ proteins, respectively, both of which are indispensable for viral DNA replication (6). ORF2 encodes the sole structural protein of the PCV2 Cap (molecular weight [MW], 27.8 kDa). Its N terminus is predicted to be a nuclear localization signal (NLS) and responsible for the viral genomic DNA packaging; however, it is not required for PCV2 virus-like particle (VLP) assembly (4). Based on genome or cap gene analysis of PCV2 isolates, PCV2 was divided into 5 main genotypes (2a, 2b, 2c, 2d, and 2e) (7, 8), among which PCV2b and 2d are currently the predominant strains circulating in commercial farms globally (9, 10). Recently, PCV2 has been updated into 8 genotypes from 2a to 2h, according to the cap gene of PCV2 genome sequences deposited in GenBank (11). Meanwhile, PCV2 is evolving at a high rate of 1.2 × 10⁻³ substitution/site/year through both a series of point mutations on the cap gene and genome recombination (12–14) at a rate comparable to RNA viruses (15).

The PCV2 Cap is the major immunogenic polypeptide, and the epitopes of the Cap have been mapped by different research groups (16–19). The Cap was successfully expressed in Escherichia coli (20), yeast (21), and insect cells (baculovirus expression vector system [BEVS]) (22, 23), and 60 PCV2 Cap monomers can self-assemble into a VLP in vitro. Importantly, the PCV2 VLP has been extensively used as a core antigen capable of inducing neutralizing antibodies (NAbs) to efficiently control PCV2-associated diseases (PCVADs) on farms and brings significant economic benefits to the swine industry worldwide (24). Although the Cap monomer is able to elicit strong humoral immune responses in swine, the majority of the antibodies recognize a linear decoy epitope (169-STIDYFQPNNKR-180) of the Cap, which lacks neutralization activity, and were often detected in diseased pigs with PMWS (25, 26). Therefore, appreciation of the assembly mechanism of PCV2 VLPs will greatly help PCV2 vaccine design and applications. In addition, PCV2 VLPs have also been extensively used to study the structure of the PCV2 capsid (4, 5), PCV2 binding receptors on cell surface (27), PCV2 entry and infection (28, 29), as well as the mechanism of the PCV2 capsid assembly (30). In addition, PCV2 VLPs have been used as a vehicle to display various foreign epitopes or functional peptides in vivo (31–34). Understanding the mechanism of PCV2 VLP assembly will also greatly advance applications of the VLPs as a novel multivalent vaccine and a vector for gene delivery.

The three-dimensional (3D) structure of the PCV2 Cap is comprised of a canonical viral jelly roll, which consists of eight β-strands and has been often found in other icosahedral viruses (4). The eight β-strands are connected by seven loops (loops BC, CD, DE, EF, FG, GH, and HI), which contribute to the PCV2 surface features (except loop FG) and stabilize the capsid via extensive interactions with neighboring loops (35). Loop CD is exposed on the capsid surface and frequently exploited as a site to display foreign epitopes on the surface for multivalent vaccine design (33, 34). In loop DE, a unique cysteine was reported to play a critical role in the assembly of PCV2 VLPs since the point mutant (C108S) failed to self-assemble into VLPs in vitro (30). A possible explanation is that the Cap forms homodimers through intermolecular disulfide bonds, and the homodimers function as a basic unit responsible for VLP assembly in vitro. The CT, an unstructured stretch, is located at the carboxyl end of the Cap and is also exposed on the surface of the PCV2 capsid (10, 35). This loop contains both a linear and a conformational epitope that are recognized by NAbs (18, 19, 36). The previous in silico study from our laboratory revealed that the structure alterations of the CT have potential effects on immunorecognition (10). Notably, an extra residue (²³⁴K) is exclusively present at the end of the CT in a newly emerging PCV2b mutant, which is now designated a novel genotype of PCV2d (37) and has been frequently isolated from diseased pigs on farms (9, 10, 38). The additional ²³⁴K residue has been reported to be responsible for the increased virulence (39) and higher viral genome copies in swine sera (40). However, the underlying molecular mechanism remains to be determined. In addition, the CT has also been chosen as a site to display various tags or foreign epitopes in genetically modified PCVs to develop new-generation vaccines (41–43). In summary, the CT is crucial for virus proliferation, virus-host cell interactions, immune recognitions, and pathological lesions.

In this study, we created a series of mutants to study the functions of the CT. Our data revealed that removal of the CT (²²⁵NLKDPPLNP²³³) completely abolished the capacity of the Cap assembly into VLPs and led to the failure of virus rescue from an infectious DNA clone in PK15 cell culture. A strictly conserved lysine residue (²²⁷K) in this loop is critical for VLP entry into PK15 cells and virus rescue. Also, accurate single-sequence prediction (called SPOT-Disorder-Single), together with previous studies of 3D structures and immunological recognitions, suggested that the last several residues of the CT featured an intrinsically disordered region (IDR) on the compact capsid, which may play critical roles in the PCV2 life cycle.

RESULTS

The CT was indispensable for assembly of PCV2 VLPs in vitro.

To investigate the role of the Cap CT in VLP assembly, we constructed and expressed a series of CT mutants (Fig. 1A) based on the codon-optimized PCV2 cap gene. The Cap (wild type) can be one-step purified from soluble supernatant by Ni²⁺-affinity chromatography (Fig. 1B, wild type [WT]). Consistent with the previous studies that PCV2 Cap is capable of self-assembling into VLPs in vitro (4, 44), our transmission electron microscopy (TEM) and gel filtration chromatography demonstrated that the wild-type PCV2 Cap we purified above was also able to self-assemble into VLPs efficiently (Fig. 2, WT). In contrast, the CT-truncated (CTT) Cap failed to self-assemble into VLPs (Fig. 2, CTT), despite that its concentration was compared with the wild type (Fig. 1B, CTT). These results suggested that the removal of the CT abolished the capacity of the Cap assembly into VLPs, indicating the CT is one of critical elements responsible for PCV2 VLP assembly in vitro.

FIG 1.

Preparation of wild-type and various CT-mutated PCV2 Caps. Schematic diagrams of wild type and mutants of the PCV2 Caps (A). Numbers indicate the positions of the NLS and various loops on the PCV2 Cap. Residues of the CT are shown for both wild type and various mutants. In the CTT mutant, all residues of the CT are deleted; in CT-D2, the conserved motif (PXXP) is removed; and in CT-M1, the PXXP motif is replaced with LNK (red) from the counterpart of PCV1 Cap. Point mutations of the CT in CT-M2 to M10 are also labeled in red. These PCV2 Caps were prepared via one-step purification of Ni²⁺-affinity chromatography, and protein purity was evaluated by SDS-PAGE analysis (B).

FIG 2.

Assembly of PCV2 VLPs from wild type and various CT-mutated PCV2 Caps. Purified PCV2 Caps (Fig. 1B) were dialyzed in assembly buffer, and subsequently, the proteins were subject to examination by gel filtration chromatography and TEM. As a control, chromatographic analyses of the PCV2 Cap subunit (left panel at the bottom). Chromatographic separation of standard proteins (thyroglobulin [T], 669 kDa; ferritin [F], 440 kDa; aldolase [A], 158 kDa; conalbumin [C], 75 kDa; ovalbumin [O], 44 kDa) are shown on the bottom right. Note: x axes indicates elution volume (ml) and y axes mean protein sample absorbance at UV 280 nm in the ÄKTA FPLC system. Each image represents the results of at least three independent experiments.

The CT contributed to the surface features around the edge of the 5-fold axes of PCV2 capsid.

To appreciate the CT-associated molecular mechanism, first, we aligned and compared the Cap sequences derived from 28 representative PCV strains (25 PCV2 and 3 PCV1) (see Fig. S1A in the supplemental material). The alignment suggested that the numbers of residues in the CT were variable, from 8 in PCV1 to 9 or 10 in PCV2. A PXXP motif at the end of the CT was strictly conserved in PCV2 but absent in PCV1. Two charged residues (²²⁷KD²²⁸) were also conserved among all the strains of PCV1 and PCV2. In general, strictly conserved residues in viral structure proteins, particularly when exposed on the surface of the virus, may play important roles in the life cycle of the virus, such as virus-host interactions, virion assembly, cell entry, and/or egress from host cells (45, 46). Second, the 3D structure from X-ray crystallography (4) of PCV2 VLPs demonstrated that the CT (red) is located adjacent to the edge of the 5-fold axes of the capsid, which was decorated by loop BC (blue) (Fig. 3). Also, the 5 residues of the CT (²²⁷KDPPL²³¹) were exposed on the surface of the PCV2 capsid (Fig. 3B). The last two residues of the CT (²³²NP²³³), however, were not resolved by X-ray crystallography and are also absent in the cryo-EM structure of the PCV2 capsid (5). Thus, the structure data suggested that both residues were intrinsically disordered on the compact capsid. Finally, we also predicted the intrinsic disorder of the Cap via an online server (https://sparks-lab.org/server/spot-disorder/); the prediction also indicated the presence of the potential intrinsic disorder in the CT (Fig. S1B). In addition, we noted that PCV evolution has been accompanied by an increase of positively charged amino acids around the five-axes (loop BC and the CT) of the PCV capsid. Both of them account for 30% (3/10) positively charged residues on the capsid surface of PCV2a, ∼45% (4/9) of PCV2b, and 50% (5/10) of PCV2d (Table 1). Thus, in the subsequent experiments, we designed a series of the CT mutants based on the analyses of residue conservation and 3D structure of the CT.

FIG 3.

3D structure of the CT in the PCV2 Cap and its orientation on the PCV2 capsid. (A) The cartoon of the PCV2 Cap subunit (left panel) is shown with 8 typical β-strands (arrow ribbons) and loops. The surface features of the Cap are demonstrated with 50% transparency. The orientation of this subunit in the capsid is shown in the right panel. (B) Residues (²²⁵NLKDPPL²³¹) of the CT on the surface of the PCV2 capsid are shown in different colors. Loop BC-decorated 5-fold axes are labeled in blue (left panel). One magnified area from this capsid is shown in the right panel. PyMOL was used to display all the structures. 3D structures of the Cap and capsid were retrieved from the PDB database (https://www.rcsb.org/) (accession number 3R0R).

TABLE 1.

Data for amino acids with positive charges exposed on the surface of PCV2 capsids

Genotype	Positions of positively charged residues (lysine and arginine)a	Reference GenBank accession no.
PCV2a	51R, 58K, 73R, 88K, 132K, 186R, 191R, 206K, 227K, 232K	BAB69436
PCV2b	51R, 58K, 59R, 63K, 73R, 89R, 132K, 186R, 227K	AAT79580
PCV2d	51R, 58K, 59K, 63R, 73R, 132K, 169R, 186R, 227K, 234K	KP112484

^aNumbers indicate residue positions of the PCV2 Caps (35). Residues in loops BC and CT are denoted by bold text.

A conserved PXXP motif within the CT did not affect PCV2 VLP assembly and entry into PK15 cells.

To test whether the conserved PXXP motif in the CT is required for PCV2 VLP assembly and entry into PK15 cells, we deleted this motif (Fig. 1A, CT-D2) or replaced it with the counterpart (LNK) of the PCV1 Cap (Fig. 1A, CT-M1) or with AXXA (Fig. 1A, CT-M2). TEM images clearly showed that intact VLPs, with a diameter of approximately 20 nm, were successfully assembled from the three Cap mutants (Fig. 2, CT-D2, CT-M1, and CT-M2). Furthermore, these mutations had no negative effect on the efficiencies of VLP assembly (chromatographic images in Fig. 2). Immunofluorescence assays (IFAs) demonstrated that these PCV2 VLPs were also able to enter PK15 cells, and statistical analysis indicated that their entry was as efficient as that of the wild type (Fig. 4A, C to E, and J). Therefore, the conserved PXXP motif in the PCV2 Cap was not required for PCV2 VLP assembly and entry into PK15 cells (Table 2).

FIG 4.

Immunofluorescence assays of PCV2 VLPs in PK15 cells. Cofocal microscopic images of PCV2 VLPs assembled from PCV2 Cap-WT (A) or different Cap mutants (B to H). Green indicates PCV2 VLP distributions in PK15 cells, and nuclei are stained by DAPI (blue). As a control, PK15 cells without PCV2 VLPs are shown in I. Bar indicates 10 μm. (J) Statistical analyses of percentages of PK15 cells with green fluorescence. Data were collected from 100 PK15 cells in different fields.

TABLE 2.

Summary of assembly and cell entry of PCV2 VLPs derived from wild type or Cap mutants

Name	Construct(s)	Assembly	% of cell entry
WT	N/Aa	Yes	100
CTT	Deletion of 225NLKDPPLNP233	No	0
CT-D2	Deletion of 230PLNP233	Yes	100
CT-M1	230PLNP233 to 230LNK232	Yes	100
CT-M2	230PLNP233 to 230ALNA233	Yes	100
CT-M3	K227A	Yes	∼10
CT-M4	D228A	Yes	100
CT-M5	227KD228 to 227AA228	Yes	100
CT-M6	P229A	Yes	100
CT-M7	P230A	Yes	100
CT-M8	L231A	Yes	100
CT-M9	N232A	Yes	100
CT-M10	P233A	Yes	100

^aN/A, not applicable.

The conserved ²²⁷K is essential for PCV2 VLP entry into PK15 cells.

Two charged residues (²²⁷KD²²⁸) of the CT were strictly conserved in both Caps of PCV1 and PCV2 (Fig. S1A). The 3D structure demonstrated both residues were located on the capsid surface (Fig. 3B), indicating that they may be involved in some critical biological processes, such as virus-host cell interactions via electrostatic interactions. To test the hypothesis, first, we prepared two single point mutants, in which ²²⁷K (Fig. 1A, CT-M3) or ²²⁸D (Fig. 1A, CT-M4) was replaced with alanine. TEM images suggested that both mutants were capable of self-assembling into VLPs in vitro (Fig. 2, CT-M3 and CT-M4). In addition, analyses of gel-filtration chromatography indicated both Cap mutants were efficiently assembled into VLPs, similar to the wild type (chromatographic images in Fig. 2). However, the VLPs assembled from CT-M3 (K227A) had less capability to enter PK15 cells (Fig. 4A, F, and J). IFAs showed intracellular fluorescence (green) was detected only in ∼10% of PK15 cells, suggesting that there were fewer VLPs entering PK15 cells than these VLPs assembled either from the Cap mutant of CT-M4 or from the wild type (Fig. 4J). These results strongly suggested that the conserved, positively charged residue (²²⁷K) plays a critical role in VLP entry into PK15 cells. To rescue the low entry efficiency of the Cap mutant (CT-M3), we prepared another mutant (Fig. 1A, CT-M5), in which the negatively charged ²²⁸D was also substituted by alanine. Thus, the CT of Cap-CT-M5 had no charged residues. TEM images and gel-filtration results suggested this Cap mutant was able to efficiently self-assemble into VLPs in vitro (Fig. 2, CT-M5). Surprisingly, the VLPs assembled from the Cap-CT-M5 were as highly efficient in entering PK15 cells as was the wild type (Fig. 4A, H, and J). Therefore, mutating the negatively charged residue (²²⁸D) into alanine completely rescued the low efficiency of cell entry of the VLPs (assembled from CT-M3). Besides the residues of ²²⁷KD²²⁸, another 5 residues (²²⁹PPLNP²²³) of the CT are exposed on the capsid surface (Fig. 3). Thereafter, we also evaluated the importance of these residues on VLP assembly and cell entry. Five Cap mutants, among which a residue at each position was replaced with alanine (Fig. 1A), were expressed and purified (Fig. 1B, M6-M10). TEM and chromatographic images showed that the five Cap mutants (M6-M10) were capable of self-assembling into VLPs in vitro (Fig. 5A). IFAs demonstrated that VLPs assembled from each of the Cap mutant (M6-M10) retained a similar capacity and efficiency to enter PK15 cells as its wild type (Fig. 5B). Thus, the statistical results indicated these exposed residues of the CT, except the positively charged lysine, may not engage VLP cell entry (Table 2).

FIG 5.

Characterization of PCV2 VLPs assembled from various Cap point mutants (CT-M6 to M10). (A) Assembly of PCV2 VLPs from the CT mutants and the wild type of the Cap. (B) Immunofluorescence assays of PCV2 VLPs in PK15 cells, and statistical analyses of percentages of PK15 cells with green fluorescence.

The CT mutations had adverse effects on virus rescue and propagation in PK15 cells.

As shown previously (47), PCV2 can be rescued from PK15 cells transfected with an infectious DNA clone containing two tandem copies of the viral genomes. To test the effects of the CT mutations on virus rescue and propagation, a series of infectious DNA clones with various CT mutations were constructed and transfected into PK15 cells. PCV2 was successfully rescued by transfection of the wild-type infectious DNA clone into PK15 cells. This also yielded a stable copy number of viral genomic DNA until passage 10 (Fig. 6, WT with two copies). As a control, PCV2 could not be rescued from the DNA clone harboring only one copy of the PCV2 genome (Fig. 6, WT with one copy). Of note, the viral genomes in PK15 cell culture were undetectable after passage 4 for the mutant of the CTT and passage 6 for both CT-D2 and CT-M3 (Fig. 6), whereas the genomes of PCV2 rescued from other CT mutants (CT-M1, CT-M2, CT-M4, and CT-M5) were detectable (Fig. 6). However, their DNA copy numbers gradually decreased with continuous passages in the cell culture; by passage 10, they were at least 100-fold (2 log₁₀) lower than that of the wild type. These results strongly suggested that the CT plays critical roles in virus rescue and propagation in PK15 cell culture.

FIG 6.

PCV2 virus was rescued from wild-type or various CT mutated infectious DNA clones. PK15 cells were transfected with different PCV2 infectious clones (wild type or the CT mutants), and then the cells were grown for more than 10 passages. PCV2 genome copies in each passage were detected by quantitative real-time PCR. (* indicates 0.01< P < 0.05, ** indicates P < 0.01 compared with WT in each group).

DISCUSSION

The PCV2 Cap, besides being the structural protein responsible for packaging the viral genome, also plays crucial roles in virus entry into host cells (27), controlling viral replication (48), and cell death (49), through unknown mechanisms. A previous study suggested that the majority of the PCV2 capsid surface is constituted by the loops (loops BC, CD, DE, EF, GH, and HI) and the CT, and these structures, located on the capsid surface, are hot spots of amino acid mutations in the course of PCV2 evolution (35). The 3D structure demonstrates the CT is located on the surface and determines the features of the capsid surface (Fig. 3B). Of note, the structure information of the last two residues of the CT is absent in both 3D structure models built up either from X-ray crystal chromatography or cryo-EM, but it can be deduced that both residues should extend farther away from the capsid surface. Therefore, these structural data indicate that the end of the CT is highly dynamic on the capsid surface and might coordinate with other surface loop(s) for some specific biological processes, such as VLP stability, virus-receptor interactions, and perhaps virus entry into host cells. In addition, it is also very difficult to investigate the precise functions of the CT on these biological processes due to the lack of their structure. Our results indicated that PCV2 VLP assembly failed if the CT was removed from the PCV2 Cap. The truncated Cap irregularly aggregates in vitro, which was frequently observed by TEM (Fig. 2, CTT). Thus, the CT is indispensable for PCV2 Cap stability and VLP assembly. Removal of the CT will abolish the capacity of the Cap assembling into VLPs in vitro, which may lead to the failure of virus rescue in cell culture (Fig. 6, CTT). Furthermore, point mutation experiments revealed the critical function of the conserved ²²⁷K residue of the CT on PCV2 VLP entry into PK15 cells (Fig. 4A, F, and J; Fig. 5; Table 2). K227A mutation (Fig. 1A, CT-M3) greatly decreased the capacity of the VLP entry into PK15 cells, which may, thereafter, result in failure of virus rescue of this mutant (Fig. 6, CT-M3). Interestingly, replacement of adjacent negatively charged ²²⁸D with alanine (Fig. 1A, Cap-M5) completely restores the capacity of the VLP entry into PK15 cells to the level of the wild type (Fig. 4A, H, and J) and leads to virus successfully rescued from the cell culture, even though the genome copy number of this mutant is lower than that of the wild type (Fig. 6, WT and CT-M5). Similar to the effect of CT-M5 (KD was replaced with AA), the single point mutation of D228A (CT-M4) also caused a low copy number of the PCV2 genome in the virus rescue experiment. These results strongly suggest that both strictly conserved residues (²²⁷KD²²⁸) of the CT play critical roles in the life cycle of the virus. The 3D structure suggested that both residues are located on the PCV2 capsid surface and may interact with the proximal residue(s) of the neighboring Cap in the course of VLP assembly and/or bind to the receptor(s) on the cell membrane for cell entry of the virus. Although no specific cell receptor(s) for PCV2 entry into host cells have been reported so far (50), both heparan sulfate and chondroitin sulfate B have been identified as general receptors for the virus-cell attachment. A potential PCV2 heparan sulfate binding site was predicted in the Cap (⁹⁸IRKVKV¹⁰³); however, it is located inside the PCV2 capsid and unlikely to be the binding site unless it externalizes (4). Very recently, Khayat’s lab reported that there are multiple asymmetric heparan sulfate binding sites on the surface of the PCV2 capsid and that the ²²⁷K residue of the CT is one of critical residues contributing to the interactions between PCV2 and negatively charged heparan sulfate (51). It is also possible that the extra lysine residue (²³⁴K) present at the end of the CT among the novel genotype PCV2d might further contribute to the binding of heparan sulfate due to the intrinsic disorder of this residue. Therefore, an increase of the binding of PCV2 capsid to heparan sulfate on the cell membranes may be positively correlated with higher viral genome copy of PCV2d in swine sera (52). However, the detailed molecular mechanism remains to be determined.

Molecular simulation suggests the last 4 residues (PXXP) of the CT also form an intrinsically disordered stretch (Fig. S1B). In addition, the last 2 residues are absent in both 3D models from X-ray crystallography (4) and cryo-EM (5) due to their high flexibility, which further confirms the intrinsically disordered nature of these several residues at the end of the CT. Over the past decade, a large number of intrinsically disordered proteins or IDR have been found in both mammalian and viral proteomes. Importantly, IDR in one protein may implement multiple functions in cells because of its unstructured feature. Therefore, in PCV2 Cap, the high flexibility of this intrinsically disordered stretch within the CT provides more chances for this IDR to function together with various elements/motifs on the capsid surface, such as interactions between the virus and receptor(s) of host cells. It is well known that interactions of PXXP motifs with SH3 domain-containing proteins activate a number of signal pathways in cells (53). Due to the size limitation of the viral genome, a virus usually hijacks the cellular processes of host cells to produce viral progeny and modulate the host immune system by using a simple and economic motif, such as the PXXP motif. For instance, the viral protein of Nef of HIV binding to the SH3 domain of Hck via a PXXP motif is associated with higher viral titers and the pathogenesis of AIDS (54). In Ebola virus, the viral matrix protein VP40 exploits its conserved consensus proline-rich motifs (i.e., PPXY, P(T/S) AP, and YPX_(n)L/I) in the l-domain to interact with a number of cellular proteins to regulate the virus/VLP assembly and budding in host cells (55, 56). Among the PCV2 genotypes, all the Caps contain a strictly conserved PXXP motif in the CT, which is absent in the PCV1 Caps (Fig. S1A) (35). Recently, Zhang et al. reported that the PXXP motif in the CT is a key factor capable of altering mRNA expression of SH3 domain-containing tyrosine kinases (both Src and Abl families) (57). Moreover, two specific inhibitors (STI-571 and PP2) for Abl and Lck/Fyn kinases reduced production of PCV2 progeny in PK15 cell culture (57). Our results show this motif plays a critical role in virus propagation in cell cultures since PCV2 cannot be successfully rescued and no viral genome was detected after passage 6 when this motif was deleted (Cap-D2) (Fig. 6). However, the loss-of-function of the mutant (Cap-D2) was partially rescued by adding the counterpart (LNK) of the PCV1 Cap to the end of the PCV2 Cap (Fig. 6, CT-M1). Furthermore, mutating both prolines of the PXXP motif to alanines dramatically decreased the genome copy numbers of the virus in PK15 cell culture (Fig. 6, CT-M2). In addition, Meng’s group reported that the viral genomic DNA of genetically engineered PCV2 (generated by insertion of various tags and foreign epitopes to the 3′ terminus of the cap gene) is 2-log₁₀ lower than that of the parent virus in swine sera (42). This result indicates that even adding additional residues to the end may have negative effects on the functions of the CT. Therefore, it suggests that the conserved motif (PXXP) of PCV2 interacting with potential SH3 domain-containing protein(s) in host cells might contribute to a higher propagation titer of this virus in PK15 cell culture. However, the effect of the PXXP motif on the pathogenesis of PCVAD has yet to be investigated, and further studies are warranted.

MATERIALS AND METHODS

Chemicals, genes, plasmids, cells, and PCV2 infectious clones.

All chemicals were purchased from Sigma-Aldrich (St. Louis, MO, USA), unless otherwise stated. The cap gene (GenBank accession number KF700357) was optimized and synthesized by GenScript (Nanjing, China) in order to elevate the Cap expression in E. coli. The fragments of the cap gene and various mutants were amplified via polymerase chain reactions (PCR) with distinct primer pair (Fig. 1A and Table 3) using this cap gene as a template. The PCR products were separated by 1% agarose DNA gel and purified via a gel purification kit (ProCell, Changsha, China). Finally, these fragments were cloned into a protein expression vector of pET100-TOPO, according to the manufacturer’s protocol (Invitrogen, Carlsbad, CA, USA) and then confirmed by DNA sequencing (GenScript). Competent cells of BL21 plus stain (TransGen, Beijing, China) were used for protein expression. PK15 cells free of PCV1 were stored in our laboratory and propagated in Dulbecco’s modified Eagle medium (DMEM) (Gibco BRL, Grand Island, NY, USA) with 5% fetal bovine serum (FBS) (Gibco BRL) and maintained at 37°C in 5% CO₂. To generate various PCV2 infectious DNA clones (Fig. 7), overlap PCR was applied with a distinct primer pair (Table 4) in the following reaction mixture which contained 25 μl of 2× FastPfu PCR SuperMix (TransGen), 0.4 μM of each primer, and 100 ng of PCV2 genomic DNA (GenBank accession number MK281580). PCR was performed on a thermocycler under the following conditions: 5 min at 94°C; followed by 25 cycles of 20 s at 94°C, 20 s at 60°C, and 1 min at 72°C; and a final step of 10 min at 72°C. The resulting PCR products containing 1.1 copy of the PCV2 genomic DNA with two stem loops (Fig. 7B) were harvested as described above, after which the products were inserted into a cloning vector of pSP72 (Promega, Madison, WI, USA) by using two restriction sites of KpnI and HindIII. Finally, positive clones analyzed by restriction were further confirmed by DNA sequencing.

TABLE 3.

Oligonucleotide primers used for PCR amplification of the PCV2 cap gene

Primer name	Sequence (5′–3′)	Primer size (bp)
pPCV2 cap-F-NdeI	CGCCATATGCGTGGCTCTCACCATCACC	28
pPCV2 cap-CTT-R-BamHI	CGCGGATCCTTAAAATTCACGGAACTGG	28
pPCV2 cap-CT-D2-R-BamHI	CGCGGATCCTTACGGGTCTTTCAGGTTA	28
pPCV2 cap-CT-M1-R-BamHI	CGCGGATCCTTATTTGTTCAGCGGGTCTTTCAGGTTA	37
pPCV2 cap-CT-M2-R-BamHI	CGCGGATCCTTACGCGTTCAGCGCCGGGTCTTTCAGGTTA	40
pPCV2 cap-CT-M3-R-BamHI	CGCGGATCCTTACGGATTCAGCGGCGGGTCCGCCAGGTTAAATTCACGG	49
pPCV2 cap-CT-M4-R-BamHI	CGCGGATCCTTACGGATTCAGCGGCGGCGCTTTCAGGTTAAATTCACGG	49
pPCV2 cap-CT-M5-R-BamHI	CGCGGATCCTTACGGATTCAGCGGCGGCGCCGCCAGGTTAAATTCACGG	49
pPCV2 cap-CT-M6-R-BamHI	CGCGGATCCTTACGGATTCAGCGGCGCGTCTTTCAGGTTAAATTCACGG	49
pPCV2 cap-CT-M7-R-BamHI	CGCGGATCCTTACGGATTCAGCGCCGGGTCTTTCAGGTTAAATTCACGG	49
pPCV2 cap-CT-M8-R-BamHI	CGCGGATCCTTACGGATTCGCCGGCGGGTCTTTCAGGTTAAATTCACGG	49
pPCV2 cap-CT-M9-R-BamHI	CGCGGATCCTTACGGCGCCAGCGGCGGGTCTTTCAGGTTAAATTCACGG	49
pPCV2 cap-CT-M10-R-BamHI	CGCGGATCCTTACGCATTCAGCGGCGGGTCTTTCAGGTTAAATTCACGG	49

FIG 7.

Construction of PCV2 wild-type infectious DNA clones and various CT mutants. One (A) or 2 (B) copies of stem loops within PCV2 genomic DNA were inserted into a cloning vector of pSP72. (C) A cartoon showing that two copies of PCV2 genomic DNA with two stem loops in tandem and numbers indicate positions of both cap and rep genes on PCV2 infectious DNA clone (wild type). (D) PCV2 CT-mutated Caps were encoded by the viruses rescued from various infectious DNA clones. Residues among various CT mutants were shown in a rectangle, and residues in red indicate point mutations in each mutant.

TABLE 4.

Oligonucleotide primers used for PCR amplification of PCV2 infectious DNA clones

Primer name	Sequence (5′–3′)	Primer size (bp)	Product length (bp)
pPCV2IF-A-F1-Kpn I	CGGGGTACCTCCTTGGATACGTCATATCTGAAAACG	36	1,882
pPCV2IF-B-R2-Hind III	CCCAAGCTTTCTTTTTGCTGGGCATGTTGCTGC	33
pPCV2IF-WT-A-R1	CCATTACGAAGTGATAAAAAAGACTCAGTAATTTATTTCATATGG	45
pPCV2IF-WT-B-F2	GAAATAAATTACTGAGTCTTTTTTATCACTTCGTAATGG	39
pPCV2IF-CTT-R1	GTACAATTCAGAGAATTTTGAATAATAAAAACCATTAC	38	1,852
pPCV2IF-CTT-F2	TCAAAATTCTCTGAATTGTAC	21
pPCV2IF-CT-D2-R1	GAATTTAGTCTTAAAGACCCCTGAATAATAAAAACCATTAC	41	1,867
pPCV2IF-CT-D2-F2	TCAGGGGTCTTTAAGACTAAATTC	24
pPCV2IF-CT-M1-R1	CCCTAAATAAATGAATAATAAAAACCATTACG	32	1,876
pPCV2IF-CT-M1-F2	GTTTTTATTATTCATTTATTTAGGGGGTCTTTAAGATTA	39
pPCV2IF-CT-M2-R1	CCGCCCTTAACGCCTGAATAATAAAAACC	29	1,882
pPCV2IF-CT-M2-F2	ATTCAGGCGTTAAGGGCGGGGTCTTTAAGATTAAAT	36
pPCV2IF-CT-M3-R1	TGCCGACCCCCCACTTAACC	20	1,882
pPCV2IF-CT-M3-F2	GTTAAGTGGGGGGTCGGCAAGATTAAATTCTCTG	34
pPCV2IF-CT-M4-R1	AGCCCCCCCACTTAACCCTTGA	22	1,882
pPCV2IF-CT-M4-F2	GGTTAAGTGGGGGGGCTTTAAGATTAAATTC	31
pPCV2IF-CT-M5-R1	TGCGGCGCCCCCACTTAACCCTTGA	25	1,882
pPCV2IF-CT-M5-F2	GTTAAGTGGGGGCGCCGCAAGATTAAATTCTCTG	34

Protein expression and purification.

For protein expression, after the plasmid was transformed into BL21 plus competent cells, a single colony was picked and grown in Luria-Bertani (LB) broth plus ampicillin (0.1 mg/ml) at 37°C with shaking (200 rpm) until cells reached mid-log growth phase, and then isopropyl-thiogalactopyranoside was added into the culture for a final concentration of 1 mM to induce protein expression at 37°C for 3 h. After induction, cells were centrifuged at 3,000 × g for 15 min at 4°C and the cell pellet was washed once with cool phosphate-buffered saline (PBS) and then resuspended in lysis buffer (100 mM NaH₂PO₄·2H₂O, 0.5 M NaCl, 100 mM KCl, 10 mM Tris-HCl, 0.5% Triton X-100, 5 mM β-mercaptoethanol, and 0.1 mM phenylmethylsulfonyl fluoride [PMSF] [pH 8.0]). After sonication, the homogenate was cleared by centrifugation (22,000 × g, 25 min at 4°C), and the supernatants containing PCV2 Caps were rapidly one-step purified through prepacked HisTrap FF crude columns (catalog number 11-0004-58; GE-Healthcare, Shanghai, China) connected to an automated fast protein liquid chromatography (FPLC) system (ÄKTA, GE-Healthcare). After samples were loaded, the column was washed with 5 column volumes (CVs) of buffer B (50 mM NaH₂PO₄·2H₂O, 0.5 M NaCl, 50 mM imidazole, and 10% glycerol) to remove nonspecific binding proteins from the column. Finally, proteins of interest were eluted with 10 CV of buffer C (50 mM NaH₂PO₄·2H₂O, 0.3 M NaCl, 300 mM imidazole, and 10% glycerol), and each fraction (1 ml) was collected. Protein expression and purification were evaluated by SDS-polyacrylamide gel electrophoresis (SDS-PAGE) and Western blots. For Western blots, proteins were transferred onto a polyvinylidene difluoride (PVDF) membrane (Merck Millipore, Billerica, MA, USA) after being separated by a 1-mm-thick 10% SDS-PAGE (Bio-Rad, Hercules, CA, USA), and then the membrane was soaked in 3% bovine serum albumin (BSA) blocking solution for 1 h before incubation with anti-6× His mouse monoclonal antibody (1:6,000 dilution) (Roche, Basel, Switzerland) for 1 h at room temperature, followed by incubation with anti-mouse IgG (1:5,000 dilution) (Promega). The membranes were washed three times with PBS buffer, and bound antibodies were detected with alkaline phosphatase (Promega). Protein concentrations were determined by a bicinchoninic acid (BCA) protein assay kit (Sangon, Shanghai, China).

Morphology of PCV2 VLPs was observed by TEM, and assembly efficiency of PCV2 VLPs was evaluated by gel filtration chromatography.

After purification, the full-length of the PCV2 Cap and various Cap mutants (Fig. 1) were dialyzed against the VLP assembly buffer (100 mM NaH₂PO₄·2H₂O, 0.5 M NaCl, 100 mM KCl, 10 mM Tris-HCl, 0.5% Triton X-100, and 0.1 mM PMSF [pH 7.4]) at 4°C for 48 h. In the meantime, the buffer was replaced three times. Following dialysis, protein samples were allowed to absorb onto carbon-coated copper grids for 7 min. Then, the grids were dried using filter paper and negatively stained with 3% phosphotungstic acid for 20 s. Finally, samples were examined using a transmission electron microscope (JEOL, Tokyo, Japan). To determine the assembly efficiency of PCV2 VLPs, samples, after dialysis, were injected into a Superloop (1 ml) for gel filtration performed by FPLC with a HiPrep 16/60 Sephacryl S-300 high-resolution column (GE-Healthcare) and ran at a flow rate of 0.5 ml/min. The elution profile was followed by a continuous assay of the optical density at 280 nm. Commercially supplied molecular mass standards (catalog number 28-4038-42; GE-Healthcare) were used to differentiate PCV2 VLPs and the Cap monomer.

PCV2 VLP entry into PK15 cells.

PK15 cells grown to 50% confluence were washed once with fresh medium and incubated with various PCV2 VLPs at 37°C for 30 min, followed by another wash to remove unbound VLPs. The cells were allowed to incubate for another 12 h in fresh medium at 37°C and then were fixed with 4% (wt/vol) paraformaldehyde in PBS for 20 min at room temperature, followed by wash briefly in PBS three times, after which cells were permeabilized with Triton X-100 (0.1% in PBS) for 10 min at room temperature. In order to stain PCV2 VLPs, cells were incubated with anti-PCV2 mouse polyclonal antibodies (prepared and stored in our lab) for 1 h at room temperature, were washed three times with PBS, and then were incubated with Alexa Fluor 488 donkey anti-mouse IgG (H+L) antibody (Invitrogen) for 1 h at room temperature. Finally, the cells were mounted in ProLong Gold antifade reagent with 4′,6-diamidino-2-phenylindole (DAPI; Invitrogen), and images were collected by an inverted Nikon TiE microscope connected to an UltraVIEW Vox system spinning disc (PerkinElmer, USA) with an EMCCD camera (catalog number C9100-13; Hamamatasu, Japan). Fluorescein isothiocyanate (FITC) and Hoechst were excited with 488- and 405-nm lasers, respectively. For image analysis, mean fluorescence intensity of each cell was measured and processed by Image J software (National Institutes of Health, MD, USA). To summarize the infection rate of each group, the average fluorescence intensity of 100 cells from the negative-control group was recognized as a threshold value, and cells with a higher mean fluorescence intensity than the threshold value were defined as positive cells. A total of 100 cells of each group were evaluated according to the mean fluorescence intensity, and these experiments were repeated 3 times.

Cell transfection and virus rescue.

PK15 cells were seeded in 6-well plates at a density of 2.5 × 10⁵ cells per well with approximately 60% to 80% confluence and transfected with 1.2 μg of various PCV2 infectious DNA clones using Lipofectamine 2000 (Invitrogen) according to the manufacturers’ protocol. After removal of the transfection reagent, cells were maintained in DMEM containing 10% FBS for 36 h. Viruses rescued from transfected cells were passaged for 10 generations. Genomic DNA copy numbers of PCV2 wild type and its mutants in infected cells of each passage were measured by using a quantitative PCR (qPCR) assay of SYBR green master mix (Vazyme, Nanjing, China) mixed with a primer pair (5′-GGGTTATGGTATGGCGGGAG-3′ and 5′-CCCTCACTGTGCCCTTTGAA-3′), and template DNA was extracted by a DNA purification kit (catalog number 69504; Qiagen, Dusseldorf, Germany). Briefly, the qPCR was performed using the ABI StepOne real-time PCR system (Applied Biosystems, Waltham, MA, USA) under the following program: 1 cycle of 95°C for 5 min; and 40 cycles of 95°C for 10 s and 60°C for 30 s.

Prediction of protein intrinsic disorder.

The single-sequence method (SPOT-Disorder-Single) was used to predict the protein intrinsic disorder of PCV2 Cap (https://sparks-lab.org/server/spot-disorder/) (58).

Data analysis.

All statistical analyses were performed using SPSS software (version 16; Chicago, USA). Differences between groups were considered significant at P values of <0.05 and are indicated by * (P < 0.05), ** (P < 0.01), and *** (P < 0.001).