Identification and Nearly Full-Length Genome Characterization of Novel Porcine Bocaviruses

The genus bocavirus includes bovine parvovirus (BPV), minute virus of canines (MVC), and a group of human bocaviruses (HBoV1-4). Using sequence-independent single primer amplification (SISPA), a novel bocavirus group was discovered with high prevalence (12.59%) in piglet stool samples. Two nearly full-length genome sequences were obtained, which were approximately 5,100 nucleotides in length. Multiple alignments revealed that they share 28.7–56.8% DNA sequence identity with other members of Parvovirinae. Phylogenetic analyses indicated their closest neighbors were members of the genus bocavirus. The new viruses had a putative non-structural NP1 protein, which was unique to bocaviruses. They were provisionally named porcine bocavirus 1 and 2 (PBoV1, PBoV2). PBoV1 and PBoV2 shared 94.2% nucleotide identity in NS1 gene sequence, suggesting that they represented two different bocavirus species. Two additional samples (6V, 7V) were amplified for 2,407 bp and 2,434 bp products, respectively, including a partial NP1 gene and the complete VP1 gene; Phylogenetic analysis indicated that 6Vand 7V grouped with PBoV1 and PBoV2 in the genus of bocavirus, but were in the separate clusters. Like other parvoviruses, PBoV1, PBoV2, 6Vand 7V also contained a putative secretory phospholipase A2 (sPLA2) motif in the VP1 unique region, with a conserved HDXXY motif in the catalytic center. The conserved motif YXGXF of the Ca2+-binding loop of sPLA2 identified in human bocavirus was also found in porcine bocavirus, which differs from the YXGXG motif carried by most other parvoviruses. The observation of PBoV and potentially other new bocavirus genus members may aid in molecular and functional characterization of the genus bocavirus.


Introduction
Animals play an important role in the spread of human diseases. Since the development of civilization, animals, especially livestock and fowls, have had a close association with human beings. Previous studies have indicated that 61% of pathogens known to infect humans can be transmitted between humans and animals, and that 75% of emerging pathogens are zoonotic, especially protozoa and viruses [1]. The transmission of bird flu to humans is a typical example. Likewise, members of the family Parvoviridae also represent significant pathogens in human and animal diseases [2][3][4]. The family Parvoviridae includes two subfamilies, Parvovirinae and Densovirinae; the former infect vertebrates and the latter infect invertebrates. The subfamily Parvovirinae can be divided into six genera: Parvovirus, Erythrovirus, Dependovirus, Amdovirus, Bocavirus, and a new proposed genus Hokovirus [4][5][6].
Members of the genus Bocavirus are virions consisting of an isometric, non-enveloped capsid, which is round with icosahedral symmetry. The bocavirus genome is not segmented and contains a single molecule of linear, positive-or negative-sense, single-stranded DNA of 4,000-6,000 nucleotides in length. Known members of bocavirus include bovine parvovirus (BPV), minute virus of canine (MVC) and Human bocaviruses 1-4. The MVC genome shares about 43% identity with that of BPV at the nucleotide level, with the NS1, VP1, and NP1 proteins being 33.6%, 41.4%, and 39% identical to those of BPV, respectively [7,8]. BPV was first identified in 1961 in samples from calves with diarrhea [9], while MVC was first isolated from canine fecal samples in 1970 [10]. Subsequent studies indicated that these two viruses had close relationships with respiratory and enteric infections in animals, especially in younger animals [4,7]. In 2005, Allander et al. reported the detection of a new human parvovirus in children with acute respiratory tract infections, showing 42-43% amino acid identity to the nearest neighbor MVC and BPV in both major ORFs that encode the NS1 and VP1 proteins. It was provisionally categorized into the genus bocavirus and named Human Bocavirus (HBoV) [11]. Recently, HBoV2, HBoV3, and HBoV4 have been discovered and all have been categorized into the genus bocavirus [12,13]. These human bocaviruses had close relationships with respiratory and enteric infections in humans, but their connection with these infections has not been proven by animal model [11][12][13][14][15][16].
In recent years, with the development of random amplification methodologies and high-throughput sequencing, more new viral pathogens have been identified in samples from humans and animals [17][18][19]. Sequence-independent single primer amplification (SISPA) is a primer-initiated technique by which nucleic acids of an unknown sequence can be amplified with sequenceindependent PCR methods using a single primer [20]. Since this method has been used, some new human and animal viruses, including new bovine parvovirus species, bungowannah virus, new torque teno mini virus species, and human bocavirus 1, have been discovered [11,18,21,22]. Using this method, we identified a novel group of viruses in the stools of piglets, the closest neighbors to which are MVC, BPV and HBoV. We provisionally named them porcine bocavirus (PBoV) and report their genomic characterization here.

Study participants and sample collection
From September to November of 2006, 397 fecal specimens were collected from healthy piglets (,15 days of age) from three different farms and several sporadically distributed families that raise pigs in Lulong County, China. All fecal specimens were frozen at 280uC until further processing.

Identification of novel viruses
We used SISPA to screen and identify viral agents, according to previous protocols [11,18,23]. Ten stool samples, diluted 1:5 (wt/vol) in phosphate-buffered saline, were mixed together. To remove bacteria and cells present in the feces, the mixed samples were centrifuged (10,0006g, 10 min), and the suspensions were then separated through 0.45 mm and 0.22 mm filters (Ultrafree-MC, Millipore). The filtered suspensions were ultracentrifuged (200,0006g, Beckman rotor, 180 min) to pellet the viral particles; the pellet was then resuspended in 200 mL of sterile saline solution. Next, 100 units of DNAse I (Promega) and 3 mL RNAse A (Qiagen) were added to the viral resuspension and it was incubated at 37uC for 90 min. Viral DNA was extracted from the stool resuspension using the QIAampH Viral DNA Mini kit (Qiagen), according to the manufacturer's protocol. The DNA was redissolved in 50 mL of RNase-and DNase-free water and stored at 280uC until further processing.
To detect viral DNA, 20 mL of extracted DNA was mixed with 2 mL of 10 mM primer FR26RV-N (GCCGGAGCTCTGCAGA-TATCNNNNNNNNNN), 1 mL of 10 mM dNTP, and 2.5 mL of 106Ecopol buffer (New England Biolabs). The samples were then incubated at 94uC for 3 min and chilled on ice for 2 min; then, 2.5 units (0.5 mL) of 39-59exo-Klenow DNA polymerase (New England Biolabs) was added, before incubating at 37uC for 1 h, followed by chilling on ice for 2 min. Then, 2.5 units (0.5 mL) of 39-59exo-Klenow DNA polymerase was added, and the reaction was incubated again at 37uC for 1 h, followed by enzyme inactivation at 75uC for 10 min and chilling on ice for 2 min. Next, 5 mL of the above reaction mix was used as a template in the following PCR, mixed with 40 pmol of primer FR20RV (GCCGGAGCTCTG-CAGATATC), 2.5 mM MgCl 2 , 1 mL of 10 mM dNTP, 5 mL of 16GeneAmp PCR buffer II and 0.5 mL of AmpliTaq Gold DNA polymerase (Applied Biosystems) in a 50 mL reaction. After 10 min at 94uC, 40 cycles of amplification (94uC for 1 min, 55uC for 1 min, and 72uC for 2 min) were performed, followed by a final extension at 72uC for 10 min. PCR products were separated on a 1.5% agarose gel and fragments between 300-1500 bp in length were excised and extracted using the QIAquick Gel Extraction kit (Qiagen). The extracted products were cloned, and the plasmid inserts were sequenced.
Contigs or singlet sequences that failed to assemble were aligned against the NCBI nr database using the BLASTn and BLASTx algorithms. An E-value cutoff of 1610 25 was used according to previous study [23].

Detection and genomic sequences of porcine bocavirus
According to the sequence obtained from SISPA, two primers (forward: 59-AAACTGGTTCCTGAGC-39 and reverse: 59-CAGTGAAACAGCGTCT-39) were designed to amplify a 208nt region within the ORF of NS1. After 5 min at 94uC, 35 cycles of amplification (94uC for 45 s, 50uC for 45 s, and 72uC for 1 min) were performed, followed by a final extension for 10 min at 72uC. PCR products were verified by sequencing and submitted to GenBank (Acc. No. HM053674-HM053692).
To obtain complete sequences of PBoV, specific primer PCR and genome walking (kit D316; TaKaRa) were used (all primer sequences are available on request). The positive PCR products were cloned, and the plasmid inserts were sequenced. The nearly full-length sequences, except the termini, were submitted to GenBank (Acc. No. HM053693-HM053694). Two additional sequences (6V and 7V) obtained from this process were also submitted to GenBank (Acc. No. HM053672-HM053673).

Sequence and phylogenetic analyses
The nucleotide and deduced amino acid sequences of PBoV were compared with entries in the GenBank database. The sequences were aligned and manually adjusted using ClustalW. Phylogenetic trees were determined by the neighbor-joining (NJ) method, implemented in the Molecular Evolutionary Genetics Analysis (MEGA) 4.1 software package. Various nucleotide substitution models were examined and phylogenetic trees of similar topology were generated. A bootstrap resampling (1,000 replicates) was used to assess the reliability of individual nodes for each phylogenetic tree. Recombination analysis between PBoV and other parvoviruses was conducted using SimPlot.

Molecular virus screening of porcine stool samples
Two pools of viral resuspension were extracted, amplified and, cloned. In total, 120 clones were sequenced, and the sequences were searched against those in the NCBI databases using the BLASTx and tBLASTx programs. The clones included gene fragments of human, phage, virus, bacteria, and unknown sequences. Of the 120 clones analyzed, 18 were significantly similar to viral sequences (E,1610 25 ); 16 of these matched porcine kobuvirus, porcine circovirus, and frog virus with high sequence identity, whereas the other two sequences, sequences A and B, were found to be Parvovirus-like sequences. These two sequences did not share similarity with database sequences at the nucleotide level by BLAST searching using default parameters. Nevertheless, the deduced amino acid sequence A matched the NS1 gene of MVC (40% identity, E = 8610 214 ) and HBoV2 (44% identity, E = 4610 26 ) using the tBLASTx algorithm. Likewise, the deduced amino acid sequence B matched the VP1 gene of MVC (57% identity, E = 6610 234 ) and HBoV (68% identity, E = 1610 228 ) using the tBLASTx algorithm. We provisionally indicated the virus as ''porcine bocavirus (PBoV).''

Detection of PBoV in healthy piglets
Two primers were designed based on sequence A to amplify a partial NS1 gene of PBoV in 397 stool samples from healthy piglets. Fifty (12.59%) samples were identified as positive for PBoV; most of them were collected in October and November 2006. These NS1 gene sequences displayed 94.3-100% identity with each other. The phylogenetic tree was constructed by the neighbor-joining method with 1,000 bootstrap replicates, using the MEGA 4.1 software, and revealed that all of these sequences were grouped together in one cluster, but separate from other bocaviruses ( Figure 1).

Genomic analysis of porcine bocavirus
According to sequences A and B, specific primer PCR and genome walking PCR were used to amplify the complete sequences of PBoV. Two nearly full-length genomic sequences except the termini were obtained: PBoV1 (5,173 nt; Acc. No. HM053693) and PBoV2 (5,186 nt; Acc. No. HM053694). The base composition of PBoV1 and PBoV2 was 27.4-28% A, 25.0-25.1% C, 29.4-29.9% G, and 17.5-17.7% T, with 54.4-54.9% GC. The two nearly full-length sequences showed the highest identity to MVC at the nucleotide level in the BLAST search, using default parameters.
Phylogenetic tree was constructed based on the alignments of PBoV1, PBoV2 and other viruses in the Parvoviridae family, the results indicated that PBoV was grouped with other bocaviruses (HBoV, MVC and BPV) but formed a distinct cluster ( Figure 2).
Putative open reading frames (ORFs) of PBoV were predicted using NCBI's ORF Finder, and the map of the genomic organization of PBoV and HBoV was constructed using the Geneious 4.8.5 software (Figure 3). The genomic organization of PBoV was closest to other known bocavirus. Similar to other bocaviruses, PBoV had three major ORFs encoding two nonstructural proteins (NS1 and NP1) and two capsid proteins (VP1 and VP2). VP2 is a truncated version of the VP1 protein with an Nterminal deletion of 138 amino acids (VP1-unique region). Similar to HBoV, the NS1 and NP1 gene of PBoV are separate, with a short length of overlapping sequence between the 39 terminus of NP1 and the 59 terminus of the VP1 gene. The genus bocavirus has a non-structural NP1 protein encoded by an ORF in the middle of Although they share a similar genomic organization, the two strains, PBoV1 and PBoV2, had different genomic sequences. At the nucleotide level, they shared 93.6% and 94.2% identity in the nearly full-length and NS1 gene sequences. At the amino acid sequence level, PBoV1 and PBoV2 showed 93.3%, 87.8%, and 92.8% identity in the NS1, NP1, and VP1 proteins. Additionally, when compared with PBoV1, PBoV2 has two and four extra amino acids in its NP1 and VP1 proteins, respectively.

Additional bocavirus found in piglets
In the process of amplifying the complete PBoV sequences in various samples, two additional samples (6V, 7V) were positive for 2,407 bp and 2,434 bp amplification products, respectively, which included a partial NP1 gene and complete VP1 gene. Interestingly, these two sequences were not grouped with PBoV1 and PBoV2, and they only shared 53.9-55.6% DNA and 41.9-43.3% amino acid sequence similarities with them. When compared with the VP1 amino acid sequences of other Bocaviruses, samples 6V and 7V showed 45.4-45.7% identity with MVC, 44.1-44.5% with BPV, and 43.1-43.3% with HBoV. Furthermore, 6V and 7V were

Comparison with Pbo-likeV
In 2009 in Sweden, Blomstrom et al. [24] reported the detection of a novel porcine boca-like virus (Pbo-likeV) with a 1,879-bp sequence, which included a complete NP1 gene and partial VP1/VP2 gene. Sequence analysis demonstrated that it was most closely related to HBoV, MVC, and BPV, sharing 33.1-35.6% identity with the NP1 gene of these bocaviruses Compared with PBoV1 and PBoV2, Pbo-likeV shared only 47.3-49.2% identity in NP1 protein sequence.
Because the data for 6V, 7V, and Pbo-likeV were not complete sequences, we aligned them with other parvoviruses. By comparing only the regions in common from the middle of the NP1 gene to the middle of the VP1 gene (about 1,300-1,500 bp in length), Pbo-likeV showed 54.5-59.9% identity with PBoV1, PBoV2, 6V, 7V, and other bocaviruses (MVC, BPV, HBoV) in this region. Phylogenetic tree of this common region among these parvoviruses also indicated that Pbo-likeV, PBoV1, PBoV2, 6V and 7V formed three distinct clusters, but grouped in the bocavirus genus ( Figure 5).

Putative phospholipase A 2 (PLA 2 ) motif in bocaviruses and other parvoviruses
VP1 protein sequences of bocaviruses and some other parvoviruses were aligned, and a conserved domain in the VP1 unique protein (VP1U) was found ( Figure 6). Similar to other parvoviruses, PBoV1, PBoV2, 6V, and 7V have a putative secretory phospholipase A 2 (sPLA 2 ) motif in the VP1U. In the alignment, a conserved HDXXY motif in the catalytic center of sPLA 2 was found in the VP1U sequence of bocaviruses and other parvoviruses. Furthermore, the conserved Ca 2+ binding loop of sPLA 2 is an 'YXGXF' motif in bocaviruses, rather than the 'YXGXG' motif found in most parvoviruses, except AAV1.

Recombination analysis between PBoV and other parvoviruses
Multiple alignments revealed that PBoV1 and PBoV2 share 28.7-55.6% and 30.2-56.8% DNA sequence identity with other members of the subfamily Parvovirinae, respectively. No recombination signal was found in either PBoV1 or PBoV2 with other bocaviruses by SimPlot analysis (data not shown).

Discussion
In the current study, we identified a group of novel candidate porcine bocaviruses (PBoV) in the stools of piglets, the closest neighbors of which were MVC, BPV, and HBoV. The nearly full-length genomic sequences, except the termini, of two strains (PBoV1, PBoV2) were determined. Additionally, about 2,400-bp sequences was determined in two other strains (6V, 7V). Phylogenetic analyses revealed that these PBoVs were not closely related to other known porcine parvoviruses. Furthermore, they  had a non-structural NP1 protein encoded by an ORF in the middle of the genome, a unique structure found in bocaviruses, but not in most other parvoviridae members. The NP1 protein is a highly phosphorylated protein, although its function has not yet been determined [4].
The PBoVs grouped with other bocaviruses (HBoV, MVC, BPV) in phylogenetic trees. However, they share only about 37.4-45.0% identity with other bocaviruses in the NS1 gene, confirming them as a new member in the bocavirus genus, parvovirinae subfamily, parvoviridae family. PBoV1 and PBoV2 only share 94.2% identity in their NS1 gene sequences at the nucleotide level. Species with nonstructural gene genetic homology of less than 95% are defined as a new species in the bocavirus genus, in accordance with the ICTVb criteria (http://www.ictvdb.org/Ictv/fs_parvo.htm). Thus, these two strains can be defined as two new species of PBoV.
Pbo-likeV was detected in 2009 by Blomstrom et al. in the background of porcine circovirus type 2-induced post-weaning multi-systemic wasting syndrome (PMWS). They proposed that it was a novel porcine parvovirus with closest relationship to bocaviruses. But they only obtained a 1,879-bp sequence with a complete NP1 gene and partial VP1/VP2 gene [24]. Subsequently in China, Zhai et al. [25] detected this virus based on the VP1/2 gene of the Swedish Pbo-likeV strain, and the expected product size was 496-bp. Alignment results showed that these fragments have 99-100% identity with the Swedish Pbo-likeV strain. It was reasonable that the virus they detected should be named Pbo-likeV strains in China instead of porcine bocavirus (PBoV).
Although sequence analyses suggested PBoV1-2, 6V-7V and Pbo-likeV all belong to the bocavirus genus, they showed marked variation in their NS1, NP1 and VP1/VP2 sequences, and phylogenetic analyses also indicated they were in three separate clusters, suggesting the likelihood of there being more than one new porcine bocavirus species in this genus. However, complete sequences for Pbo-likeV and 6V-7V were not obtained, and a new species is defined by non-structural (NS1) gene homology, so whether these are new bocavirus species is, as yet, unclear.
Previous studies have discovered conserved sequences of about 40 amino acids in the N terminal region of the VP1 protein in most parvoviruses, whereas the remaining sequences of VP1 vary greatly; this conserved domain has been demonstrated to have sPLA 2 -like activity [26][27][28][29]. This activity has been suggested to be a key for the efficient transfer of the viral genome from late endosomes/lysomes to the nucleus to initiate viral replication in parvoviruses, and amino acid substitution in the active site of the sPLA 2 motif would inactivate enzymatic activity, eliminating viral infectivity [26,28,[30][31][32]. A conserved HDXXY motif in the catalytic center of sPLA 2 , and the YXGXF motif, similar to the YXGXG motif in the conserved Ca 2+ -binding loop of sPLA 2 , were also found in the VP1U sequences of PBoV1-2 and 6V-7V. Although the conserved Ca 2+ -binding loop of sPLA 2 in all bocaviruses contains YXGXF, rather than the YXGXG motif found in most other parvoviruses, we have demonstrated previously that HBoV with an YXGXF motif also possesses sPLA 2 activity [31]. This suggests that the alternate motif found in PBoV may also have sPLA 2 activity, although further study is needed.
Previous studies have suggested the occurrence of recombination phenomena within parvoviral species [12,13,[33][34][35]. In the current study, no recombination signal was found for either PBoV1 or PBoV2 with other bocaviruses, indicating that PBoV is not a recombination of other parvoviruses, but rather a novel member of the genus bocavirus. In addition to obtaining more PBoV genome sequences in the future, further study is needed to address exact phylogenetic relationships between various porcine bocavirus species and strains.
The severity of disease caused by parvoviruses varies by virus [34,36]. Some parvoviruses cause mild disease or asymptomatic infections [37,38], whereas others can result in fetal death and abortion in animals [34,39,40]. The members of genus bocavirus have a close relationship with respiratory and enteric diseases, especially in younger children and animals [4,7,[11][12][13][14][15][16]. In the further study by Blomstrom et al in 2010, they detected Pbo-likeV in pigs with and without PMWS with infection rates of 88% and 46%, respectively [41], indicating that Pbo-likeV may have relationship with PMWS. Zhai's study [25] showed that the positive rates of Pbo-likeV (38.7%, 74/191) in pigs suffered from respiratory tract symptoms were significantly higher than those (7.3%, 3/41) in healthy pigs, so they proposed that this virus might be an emerging virus for swine respiratory tract diseases. The PBoVs found in the current study had a high prevalence (12.59%) in stool samples of piglets, suggesting that swine is a host for PBoV, and that they may multiply in the intestinal tract of piglets, as do other bocaviruses. With the inclusion of only healthy piglets and the absence of serologic and cytology data, it is not yet known whether PBoV can cause gastroenteritis or other diseases in swine. We are conducting a case control study to clarify the role of PBoV in gastroenteritis in piglets.
In the current study, to the best of our knowledge, several different bocaviruses with two nearly full-length genome sequences were identified for the first time in porcine samples and these are proposed to be new bocavirus members in pigs. Our findings indicate that many new potential bocavirus species exist in pigs. Further study is needed to explore their exact roles, including the clinical significance and epidemiology of their infection, and the precise evolutionary relationships among them. Although MVC, BPV, and human bocaviruses were discovered in recent decades, the biological characteristics of the genus of bocavirus, such as virus replication and assembly, and relatedness to diseases are still not fully understood. The observation of PBoV and other potential new members of the bocavirus genus can perhaps help in the molecular and functional characterization of bocaviruses in the near future.