Genetic Diversity Analysis of Genotype 2 Porcine Reproductive and Respiratory Syndrome Viruses Emerging in Recent Years in China

Porcine reproductive and respiratory syndrome virus (PRRSV) is characterized by its extensive genetic diversity. Here we analyzed 101 sequences of NSP2 hypervariable region, 123 ORF3 sequences, and 118 ORF5 sequences from 128 PRRSV-positive clinical samples collected in different areas of China during 2008–early 2012. The results indicated that the amino acid identities of the three genes among these sequences were 87.6%–100%, 92.5%–100%, and 77%–100%, respectively. Meanwhile, 4 novel patterns of deletion and insertion in NSP2 region or GP5 were first found. The phylogenetic analysis on these 3 genes revealed that the Chinese PRRSV strains could be divided into three subgroups; majority of genes analyzed here were clustered in subgroup 3 with multiple branches; the strains with 30-aa deletion in NSP2-coding region were still the dominant virus in the field. Further phylogenetic analysis on four obtained complete genomic sequences showed that they were clustered into different branches with the Chinese corresponding representative strains. Our analyses suggest that the genetic diversity of genotype 2 PRRSV in the field displays a tendency of increasing in recent years in China, and the 30-aa deletion in NSP2-coding region should be no longer defined as the molecular marker of the Chinese HP-PRRSV.


Introduction
Porcine reproductive and respiratory syndrome (PRRS) characterized as reproductive failure in sow and respiratory disorder in all-age pigs [1] is regarded as one of the major concerns for disease controlling in pig farms [2][3][4][5]. The first outbreak of PRRS in Western Europe and North America was almost concurrently documented during the late 1980s and early 1990s [6,7]. Within the succeeding years, PRRS was an endemic disease in North America, Europe, and Asia [7][8][9][10][11]. Since then, PRRS has become the most economically devastating disease for global pig industry [4,5].
The causal agent, porcine reproductive and respiratory syndrome virus (PRRSV), is classified into the order Nidovirales, family Arteriviridae, together with equine arteritis virus (EAV), lactate dehydrogenase-elevating virus (LDV), and simian hemorrhagic fever virus (SHFV) [12,13]. According to the genetic diversity, PRRSV can be divided into two genotypes: type 1 (European) PRRSV with prototype Lelystad and type 2 (North American) PRRSV with prototype VR-2332. Although the two types of PRRSV can cause similar syndrome to the infected pigs, they share only 55%-70% nucleotide and 50%-80% amino acid similarity in their various genes [14]. The single positive-strand RNA genome of PRRSV is approximately 15 kb in length, encoding at least 10 open reading frames (ORF) [15][16][17][18]. The ORF1a and ORF1b encode replication-related polymerase proteins, which can be autoproteolytically cleaved into at least 13 nonstructural proteins (NSP) [19][20][21][22]. And the rest of ORFs 2 to 7 encode viral structural proteins [15,17,23,24]. Among them, the largest nonstructural protein gene-NSP2, ORF3 encoding minor glycosylated structural protein-GP3, and ORF5 that encodes major envelope protein-GP5 are often selected for variation investigation and phylogenetic analyses for their 2 BioMed Research International genetic diversities [25,26]. The genetically extensive variation with genetic/antigenic diverse strains in the field is regarded as an important reason for vaccination failure and occasional outbreaks of more severe forms of PRRS [21,26].
Since the first outbreak of PRRS in China was documented at the end of 1995 [27], this disease has been accompanying the Chinese swine industry [28]. Considering China has the largest number of pig farms with diversity of size and different levels of biosecurity control and management, the economical cost caused by PRRS in China should be higher than that in the United States, which was estimated to be $664 million per year [4,29]. Especially in 2006, a large-scale outbreak caused by the highly pathogenic PRRSV (HP-PRRSV) was characterized by prolonged high fiver, rubefaction on the skin, and increased morbidity and mortality in all ages of pigs, resulting in unprecedented damage to the Chinese swine industry [25,30,31]. The phylogenetic analyses have indicated that the causative pathogen HP-PRRSV was evolved by a gradual variation and accumulation progress of genome changes from the early Chinese domestic strain [25,26]. In the following years, the HP-PRRSV has been becoming the dominant strains in the field [25]. In the year 2011, the Chinese HP-PRRSV-derived commercial vaccines, which were attenuated by serial passaging on the MARC-145 cells, were approved to put on the domestic market. In the same year, the European PRRSV isolates were first reported in China [32]. Considering the risk of potential reversion to virulence and recombination, the two events increased the complexity of PRRSV epidemic situation in China, which will attract more attention on the molecular epidemiology analysis.
In this study we phylogenetically analyzed the PRRSV NSP2 hypervariable (HV) region and ORF3 and ORF5 genes, which were directly amplified from the clinical samples collected from various pig farms around the pig-producing areas of China, during the period from 2008 to early 2012. Simultaneously, we described the complete genomic sequences of four new Chinese PRRSV isolates including one strain from Tibet mini-pig and three sharing novel characteristic genetic variations and compared their genetic characterization with previous strains. Finally a phylogenetic tree based on the fulllength genomic sequence is conducted in order to analyze the evolutionary relationship of these strains.

Sample Collection and Geographic Distribution.
During the period from 2008 to early 2012, 128 clinical samples, including lung, brain, spleen, lymph node, and sera, which were positive for PRRSV by conventional laboratory detection and diagnosis, were collected from pig farms distributed in 18 regions of China. These samples were further used for PRRSV isolation or NSP2 HV region and ORF3 and ORF5 genes amplification and sequencing.

RNA Extraction and RT-PCR Amplification and
Sequencing. Total RNA was extracted from 250 L of tissue homogenates or serum by using TRIzol LS reagent (Invitrogen Corporation, Auckland, NY, USA). Then reverse transcription was performed by using M-MLV reverse transcriptase (Promega, Madison, WI, USA) and specific antisense primers (Table 1). Resulting cDNA was amplified by using PrimeSTAR HS DNA polymerase (TaKaRa Biotechnology Co., Dalian, China) in the following process: 34 cycles of denaturation at 98 ∘ C for 12 s, annealing at 56 ∘ C for 10 s, and extension at 72 ∘ C for 1 min/kb. The PCR products were examined by gel electrophoresis and purified by using Agarose Gel DNA Extraction Kit (BioDev Co., Beijing, China) and then subjected to BGI (Beijing, China) for sequencing.

Cells and
Virus. MARC-145 cells were grown at 37 ∘ C in Dulbecco's minimum essential medium (DMEM) supplemented with 10% fetal bovine serum (FBS) and antibiotics. The pulmonary alveolar macrophage (PAM) cells were prepared as described previously [33] and maintained in 10% FBS 1640 medium. Serum or supernatant of tissue homogenates from PRRSV-positive samples were used to inoculate the MARC-145 cells or PAM for PRRSV isolation.

Full-Length Genomic Sequencing of PRRSV Isolates.
Fourteen pairs of primers for genotype 2 PRRSV (Table 1), covering the full-length genomes, were designed, based on JXwn06 (Accession number EF641008). Each fragment of the isolates was amplified and cloned into pEASY-Blunt vector (Transgen Tech Co, Beijing, China) as described previously [34]. The 5 and 3 ends region was amplified using 5 and 3 full RACE kit (TaKaRa, Dalian, China) according to the manufacturer's instructions. The PCR products or plasmid with cloned PRRSV fragments was subjected to BGI (Beijing, China) for sequencing.

Sequence Alignment and Phylogenetic Analysis.
The nucleotide and deduced amino acid sequences were aligned by ClustalW in software Lasergene (DNASTAR Inc., Madison, WI, USA) to determine sequence homology. And phylogenetic and molecular evolutionary analyses were conducted using MEGA version 5 (Tamura, Peterson, Peterson, Stecher, Nei, and Kumar 2011), along with multiple sequences of representative PRRSV available in GenBank from various countries and areas (Supplementary Table S1, see Table S1 in Supplementary Material available online at http://dx.doi.org/10.1155/2014/748068).

Number of NSP2 HV Fragment and ORF3 and ORF5
Genes Amplified from PRRSV-Positive Clinical Samples. The fragments of NSP2 HV region and ORF3 and ORF5 genes amplified from PRRSV-positive samples were sequenced. The results showed that totally 101 NSP2, 123 ORF3, and 118 ORF5 sequences were successfully obtained from 128 PRRSVpositive samples collected during the period from 2008 to early 2012 (Table 2).

Sequence Alignment and Phylogenetic Analysis of NSP2
HV Region. The amplified NSP2 HV region exhibited various sizes in length (Table 2). Nucleotide and deduced amino acid sequences analysis revealed that 86 out of 101 NSP2 HV       Table S1). The phylogenetic tree revealed that all 101 NSP2 sequences belonged to genotype 2 of PRRSV and all Chinese PRRSV strains could be classified into three main subgroups ( Figure 1). The JL1101 and GZ1101 were located in subgroup 1 with the representative strains VR-2332, BJ-4, and RespPRRS MLV, the other 99 were clustered into the subgroup 3 with multiple branches, together with the representative strains HB-1(sh)/2002, JXwn06, JXA1, and JXA1 P80. No strains in this study were clustered into subgroup 2 with representative strain CH-1a, the earliest Chinese strain. This means that the genetic diversity of NSP2 still existed and the strains with 30aa deletion in NSP2-coding region remain to be the dominant viruses in the field. Compared with the data from 2006 to 2007, the percentage of NSP2-deleted strains increased [25]. However, these subgroups did not appear to be associated with epidemiological features based on geography or date. Interestingly, a minor branch with JXA1 P80, the HP-PRRSV JXA1 derived vaccine strain, was observed in the NSP2 phylogenetic tree. Four strains HB1105, HB1201, SC1101, and BJ1101, collected later than the year 2011 when the JXA1-derived vaccine was launched commercially, were also clustered in this branch, whereas the parental strain JXA1 was out of this branch, suggesting that there is the possibility that the four strains directly derived from the vaccine strain JAX1 P80. However few earlier strains were also clustered into this minor branch. Even though the analysis from this study does not fully reflect that a great number of emergence of PRRSV were due to the use of HP-PRRSV-derived MLV, the potential risk of the reversion of MLV to virulent strains, and the recombination between the vaccine virus and field viruses are worthy to pay more attention to in the future [35].

Sequence Alignment and Phylogenetic Analysis of ORF3
Gene. All the obtained ORF3 genes in this study had the same size of 725 nt. The sequences alignments indicated that they shared 92.5%-100% amino acid similarity with each other and 89.4%-95.3% amino acid similarity with JXwn06, as well as 80.7%-85.0% with VR-2332. The regions residues 33-46, 120-133, and 162-198 were conserved among these strains; otherwise, majority of amino acid substitutions were located in two hypervariable regions, the residues 58-71 and 216-226. Especially, 63 out of 123 contained the I66-T66 mutation, comparing with those in JXwn06 and VR-2332.
The phylogenetic analysis of deduced amino acid sequences of ORF3 indicated that all Chinese genotype 2 strains were distributed into three subgroups (Figure 2). Three genes JL1101, HB1103, and GZ1101 were clustered into subgroup 1 with the representative strains VR-2332 and BJ-4, and no strains in this study were clustered into subgroup 2 with the representative strains CH-1a, HB-1(sh)/2002, and HB-2(sh)/2002. All the other strains were clustered into subgroup 3, which contained most Chinese strains collected later than 2004.

Sequence Alignment and Phylogenetic Analysis of ORF5
Gene. Except for the GZ1101 which had one amino acid deletion at the position aa 34 in ORF5-coding region, the other 117 genes had the same size of 603 nt as that of VR-2332. Sequences alignments showed that the amino acid similarity among the 117 ORF5 genes ranged from 77.0% to 100%, and they shared 78%-99% amino acid similarity with VR-2332, as well as 86.5%-99% with JXwn06. Similar as previous report, the residue 3-39, the putative signal sequence was the most variable region, whereas, the regions 40-57, 67-90, 107-120, 138-160, and 165-184 were relatively conserved [25]. However a novel substitution E170-G170, which was conserved in the Chinese strains collected during the period from 2006 to 2007, was observed in recent strains.
The phylogenetic tree conducted by using the deduced amino acid sequences of ORF5 showed that the Chinese strains of genotype 2 PRRSV could be divided into 3 different subgroups (Figure 3). Three strains JL1101, HB1103, and GZ1101 were in subgroup 1 with the representative strains VR-2332, BJ-4, and CH-1a, and the SD1003 was the only strain clustered in subgroup 2 with the representative strain MN184A; all other 115 strains were clustered into subgroup 3 with multiple branches, which were completely composed of Chinese strains with the representative JXwn06 and HB-1(sh)/2002. Similar to the NSP2 phylogenetic tree, a minor branch with the JXA1 P80 contained the strains collected both earlier and later than 2011.

Full-Length Genomic Analysis of 4 New PRRSV Isolates.
Three strains, SD0901, LN1101, and GZ1101, with characteristic deletion or insertion in NSP2 or ORF5 genes, and another strain BJ1102 were successfully isolated from the clinical samples using MARC-145 cells or PAMs. The four strains were subjected to full-length genomic sequencing after plaque purification of three rounds. The SD0901 (Gen-Bank Accession number NJ256115) and BJ1102 (GenBank Accession number KF751237) shared same size of complete genome with 15,320 nt in length, excluding the ploy (A) tails. The genome sizes of LN1101 (GenBank Accession number KF751238) and GZ1101 were 15,356 nt and 15,404 nt, respectively. The BJ1102 was isolated from clinical samples of Tibet mini-pig with acute PRRS symptom in a pig farm where HP-PRRSV-derived vaccine was used before importing Tibet mini-pig.
Sequence alignments indicated that the 5 UTR of the four strains shared nucleotide identities of 91.0%-100% with the representative genotype 2 PRRSV strains. A nucleotide "A" insertion at the position nt 75 of GZ1101 5 UTR was first observed in this study. It was shown that major variations were located in NSP2-coding region including 3-aa deletion at the position aa 593-595 in GZ1101, 18-aa deletion at the position aa 482-499 in LN1101, 30-aa deletion at the positions aa 482 and aa 533-561 in BJ1102, and 31-aa deletion at the positions aa 468, aa 482, and aa 533-561 and an amino acid "P" insertion between aa 585 and aa 586 in SD0901 (Figure 4). In addition, a new deletion at the position aa 34 of GP5 was found in GZ1101 ( Figure 5). The individual homology analysis of the other genes was also summarized in Supplementary  Tables S2-S5. To further classify the evolutionary relationship of these 4 isolates, the phylogenetic tree was conducted based on their 8      full-length genomic sequence, together with both genotype 1 and genotype 2 representative strains. It was shown that the SD0901 and BJ1102 were clustered in the subgroup of Chinese HP-PRRSV and HP-PRRSV-derived vaccine virus, sharing high identity 98.7% and 98.4% with JXwn06, respectively; in addition, the LN1101 was the neighbor of HB-1(sh)/2002 in the same minor branch, which share 98.8% identity with each other. The GZ1101 was close to the minor branch with prototype VR-2332 and BJ-4 ( Figure 6). The four strains exhibited 88.3%-97.8% nucleotide identity with each other.
The findings suggest that various PRRSV strains from different clusters simultaneously circulate and spread in pig farms in China.

Discussion
PRRSV is characterized of its extensive genetic/antigenic variation in the field [36]. Low replication fidelity of RNA polymerase, abundance of quasispecies, RNA recombination, and immune pressure selection are regarded as the    Figure 2: Phylogenetic tree based on the deduced amino acid sequence of ORF3. The bootstrap consensus tree is shown. The sequence downloaded from GenBank had a suffix "ORF3. " The representative strains were labeled with "black triangle" and the vaccine strains were labeled with "black diamond. " The bootstrap values were shown close to the branches. mechanisms of generating viral heterogeneity and diversity which promotes the evolution of PRRSV [37][38][39]. The emergence and reemergence of acute form PRRS is often influenced by the genetics of PRRSV [36]. Since the PRRS outbreak in China was first documented in 1995; this virus is always accompanied with the Chinese pig industry [27]. In 2006, an unparalleled, large-scale, atypical PRRS outbreak was reported in China [25,30,31]. In the following 1-2 years, the HP-PRRSV with 30-aa deletion in NSP2-coding region rapidly became the dominant in the field, meanwhile the classical and low-pathogenic strains could also be isolated from pig farms [25]. In 2011, the HP-PRRSV-derived MLV was licensed and widely used afterward in the field. This situation might greatly increase the immune selective pressure in pig herds to accelerate the variation and evolution of PRRSV [39]. Meanwhile, the European genotype 1 PRRSV strains also emerged in China [32], resulting in the complexity of PRRS. Therefore it is meaningful to continually survey the BioMed Research International       Figure 3: Phylogenetic tree based on the deduced amino acid sequence of ORF5.The bootstrap consensus tree is shown. The sequence downloaded from GenBank had a suffix "ORF5. " The representative strains were labeled with "black triangle" and the vaccine strains were labeled with "black diamond. " The bootstrap values were shown close to the branches. diversity of PRRSV and analyze the phylogenetic relationship and evolutionary process of field strains.
In this study we amplified and gained 101 NSP2 HV region sequences from 128 PRRSV-positive clinical samples. Out of them, 86 had the same 30-aa deletion in NSP2-coding region as that of JXwn06 and other early HP-PRRSV strains. The 86 new sequences shared 87.6%-100% amino acid similarity with each other, as well as 66.6%-69.5% with VR-2332, which were both lower than previous corresponding data, 93.4%-99.8% and 77.1%-77.8%, we obtained in 2006-2007 [25]. Meanwhile, 3 novel patterns of deletion or insertion in NSP2-coding region were first found in this study. These results suggest     branch. As more and more low pathogenic strains have been found to have 30-aa deletion in NSP2-coding region, this deletion will no longer be defined as the molecular marker of HP-PRRSV. The ORF3 sequences alignment showed that the residues within 3 regions including 33-46, 120-133, and 162-198 were relatively conserved among the strains in this study, while the nt 58-71 and 216-226 of ORF 3 gene were hypervariable regions. More than 50% strains contained the I66-T66 mutation, which was located at the identified epitope in GP3 [40][41][42][43]. A previous clue suggested that the residue substitution at this position may be related with inducing neutralizing antibody [40]. Whether this mutation is associated with the immune pressure selection or immune invasion still needs further investigation.

S L G G D V P N S W E D L A V S S P F D L P T P P E P A T P S S E L V I V S S P Q C I F R P A T P L S E P A P I P A P R G T V S R P V T P L S E P I P V P A P R R K F Q Q V K R L S S A A A I P
Even if the ORF5 is the highest variable region of PRRSV structural proteins, the deletion in this region is little recognized. In this study one amino acid deletion at the position aa 34 of GP5 was first found in GZ1101, and the sequence alignment showed that this strain had higher homology with VR-2332 and RespPRRS MLV, implying that the virus might be evolved from the vaccine virus. The lowest amino acid similarity of ORF5 among these strains was 77.0%, which was lower than the data (84.1%) in our previous research, supporting that the diversity of strains has increased since 2008. The phylogenetic tree based on deduced amino acid sequence of ORF5 showed that the Chinese PRRSV strains could be clustered into 3 different subgroups. Compared with Shi Mang's phylogenetic result based on more than 8,000 sequences, subgroup 1 was composed of representative strains located in lineage 8 (VR-2332 and CH-1a), lineage 5.1 (VR-2332 and BJ-4), and lineage 7 (SP and prime Pac), subgroup 2 contained representative strains MN184a from lineage 1, and the other HP-PRRSV in subgroup 3 was late clustered into lineage 8 in Yanyan Ni's modified phylogenetic tree, even if the information of Chinese HP-PRRSV had not been included in Shi's analysis [44,45].
Because of having novel genetic characterization or being isolated from special host Tibet mini-pigs, four strains, SD0901, LN1101, GZ1101, and BJ1102, in our study were subjected to full-length genomic sequencing in order to better understand their characterizations of whole genome. Comparative analysis showed that their complete genome sequence homology ranged from 88.3% to 97.8%, and they were clustered into different branches of genotype 2 PRRSV, further indicating that PRRSV strains with genetic diversity simultaneously exist in the field in China. In this study, the molecular sequence data of PRRSV was utilized to characterize the epidemiology and evolutionary process in phylogenetic analysis, expecting that it could provide an important clue for modification of diagnosis methods and design of novel vaccine. Hopefully, these analyses will be useful for PRRS control strategy. Considering that the modern transportation in pork supply chains can easily spread the virus nationwide or even internationally, and meanwhile the wide use of attenuated PRRSV live vaccine will raise the risk of reversion to virulence and increase the possibility of recombination between vaccine strains and field strains, the RPRSV diversity will be continually expanded and the epidemic situation in the field will be more and more complicated. So if we try to gain a deeper view of the PRRSV epidemiology, the long-term investigation, linked observation between genetic diversity and phenotypic difference, and effort of explaining the mechanism of how HP-PRRSV strains gain the dominance in field should be first concerned in future.

Conclusion
Our analysis results indicated that the genetic diversity of PRRSV in the field further increased in recent years in China, due to the dramatic variations of NSP2 and ORF5 genes of PRRSV, and the 30-aa deletion in NSP2-coding region should be no longer defined as the only molecular marker of the Chinese HP-PRRSV as the PRRSV strain with same deletion and low pathogenicity emerged in the field and the attenuated live vaccines derived from HP-PRRSV were widely used in pig farms.