Full Genomic Analysis of New Variants of Porcine Reproductive and Respiratory Syndrome Virus Revealed Multiple Recombination Events Between Different Lineages and Sublineages

Porcine reproductive and respiratory syndrome virus (PRRSV) has had a devastating impact on the pig industry in China, and monitoring its genetic diversity is important for epidemiological surveillance and understanding its evolution. Here, we determine the complete genome sequences of two PRRSV strains, GXYL1403 and GXNN1839. Comparative, phylogenetic, and recombination detection program analyses show that the two isolates are recombinant strains with large-fragment amino acid deletions in nsp2. GXYL1403 possesses a unique deletion region of 124 amino acids in nsp2, and GXNN1839 contains a deletion of 131 amino acids in nsp2 as compared with VR2332. Further analysis of the full-length sequence suggests that GXYL1403 is a natural recombinant between sublineages 8.1 (CH-1a like) and 8.3 (JXA1-like). The recombination site of GXYL1403 is located in nsp9–nsp12 (8961nt−11181nt). GXNN1839 is a natural recombinant between the lineage 5 (VR-2332-like) and lineage 1 (NADC30-like) strains. The recombination events occurred in nsp9 (7872nt-8162nt) and in ORF2 (12587nt−13282nt) in the genome of GXNN1839. These results provide new evidence that PRRSV strains circulating in the environment have undergone recombination among the different lineages or sublineages of field strains, and these add to our understanding of RNA combination events that occur in PRRSV.

The increasing diversity of the type 2 PRRSV is related to point mutations and genetic recombination among the field strains or in between the field and modified live vaccine (MLV) strains, which can lead to some new emerging antigenic variant strains (6,(18)(19)(20)(21)(22)(23)(24). In the present study, we identify two novel recombinant PRRSV strains circulating in swine farms in Guangxi, a province of southern China. We perform genome-wide sequence analysis of the two isolates and find new combination patterns in PRRSV evolution.

Cell Culture and Virus Isolation
Meat Animal Research Center-145 (MARC-145) cells and porcine alveolar macrophages (PAMs) were maintained in cell cultures as described in our previous study (25). Field samples (sera) were collected from clinically diseased pigs in Guangxi province, China. Samples were detected by RT-PCR using specific primers for the ORF5 gene of genotype 2 PRRSV (26). One positive sample was from 56 dying piglets exhibiting severe respiratory syndrome on a pig farm located in YuLing city, Guangxi Province, which has a total pig population of 823. One positive sample was from 18 pregnant sows suffering reproductive failure in a farm located in Nanning city, Guangxi Province in South China, which has a total sow population of 312. The positive samples were used for virus isolation using MARC-145 cells or PAMs. Cells were observed daily for the appearance of cytopathic effects (CPE). When 75% CPE was observed, the culture media were harvested and used for blinding passage in MARC-145 cells or PAMs. The PRRSV strain (GXYL1403) isolated from MARC-145 cells was then plaque-purified three times on MARC-145 cells for subsequent complete genome sequencing. The third passage of each virus strain was identified by RT-PCR and an indirect immunofluorescence assay (IFA).

RT-PCR Amplification and Complete Genome Determination
The viral genomic RNA was extracted from PRRSV-infected PAMs or MARC-145 cells and reverse transcribed into cDNA using M-MLV reverse transcriptase (Takara) according to the manufacturer's instructions. PCR was performed to amplify the complete genomic sequences of PRRSV isolates GXYL1403 and GXNN1839 by using PRRSV-specific primers as described previously (27,28). The PCR products were purified using an E.Z.N.A.TM Gel Extraction Kit (OMEGA, USA) and cloned into a pMD18-T vector (TaKaRa, Dalian, China) for nucleotide sequencing. The complete genomic sequences of PRRSV isolates were assembled using the SeqMan program of DNAstar software, version 7.0.

Immunofluorescence Assay (IFA)
IFA was used to detect PRRSV N protein expression in PAMs and MARC-145 cells as described previously (25). At 72 h infection, the isolate-infected PAMs or MARC-145 cells were washed with phosphate-buffered saline (PBS) twice and fixed using cold methanol for 15 min at room temperature and then blocked with 3% BSA (fraction V bovine serum albumin) (Roche, Mannheim, Germany) for 30 min. The cells were treated with a monoclonal antibody (mAb) against PRRSV N protein (SR30-A) (Rural Technologies, Inc., Brookings, SD, USA) as the primary antibody and incubated for 2 h at 37 • C. Then, the cells were washed five times with PBS and incubated with goat antirabbit IgG H&L (Alexa Fluor R 488, Abcam) for 1 h at 37 • C. Following five further washes with PBS, the cells were observed under a fluorescence microscope.

Sequence Comparison and Phylogenetic Analysis
Differences of sequences from the two isolates and reference strains from all over the world were analyzed and aligned using the MegAlign program in DNAstar 7.0 software (DNASTAR Inc., Madison, WI, USA). Phylogenetic trees, including the complete genomes and ORF5 from this study and the representative strains from China and other countries, were constructed by neighbor-joining in MEGA 6.0, using the maximum composite likelihood and bootstrap confidence values from 1,000 replicates. Detailed information on the PRRSV reference sequences is shown in Table 1.

Recombination Analyses
The complete genomic sequences of GXYL1403 and GXNN1839 were compared with PRRSV reference strains JXA1, CH1a, VR-2332, and NADC30 using the Recombination Detection Program (RDP) v4.66 with six different algorithms (RDP, GENECONV, MaxChi, BootScan, SiScan, and 3Seq) (29). The results were presented using the RDP method. Recombinant events were confirmed by Bootscan analysis in Simplot software (v3.5.1, JHK University, Baltimore, MD, USA) with the default parameters. A series of phylogenetic trees based on each of the sequence regions were constructed to further identify these putative recombination events.

Isolation and Identification of PRRSV Strains
Two PRRSV-positive sera were used to inoculate the MARC-145 cells and PAMs for PRRSV isolation. Typical CPE, characterized by cell fusion and shedding, was induced in both MARC-145 cells and PAMs 48 h post-inoculation (hpi) (data not shown). To confirm the isolation of the PRRSV strains, IFA was conducted using a mAb against PRRSV N (SR-30A). PAMs and MARC-145 cells infected with PRRSV isolates reacted with the specific antibody against the PRRSV N protein (data not shown). These results indicate successful isolation of the infectious PRRSV strains. These PRRSV isolates were designated as GXYL1403 and GXNN1839, respectively.

Genomic Characteristics and Phylogenetic Analysis of the Two PRRSV Isolates
The full-length genomes of GXYL1403 and GXNN1839 were determined and submitted to the GenBank database with accession numbers MN660069 and MN660070. The entire genomes of GXYL1403 and GXNN1839 are 15,018 and 15,019 nucleotides in length, respectively, excluding the poly (A) tails as  Nsp2 is one of the most heterogeneous proteins in PRRSV and contains different patterns of amino acid insertions and deletions. Compared with VR-2332, GXYL1403 contains a discontinuous 124-amino acid deletion in the hypervariable region (HVR) of nsp2. GXNN1839 has a discontinuous 131-amino acid deletion in the HVR of nsp2, which has the same amino acid deletion pattern as the NADC30-like strain (Figure 1). This data suggest that novel PRRSV strains with amino acid deletions in nsp2 are emerging.
The phylogenetic tree based on GP5 shows that the type 2 PRRSV strains can be classified into nine different lineages. GXYL1403 belongs to sublineage 8.3, which is represented by the reference strain JXA1. GXNN1839 belongs to lineage 3, which is represented by the reference strain NADC30 (Figure 2A). The phylogenetic tree based on the complete genome also shows that GXYL1403 belongs to sublineage 8.3, and GXNN1839 belongs to lineage 3 ( Figure 2B).

Recombination Analysis
To determine the recombinant events in the genomes of the GXYL1403 and GXNN1839 isolates, we performed recombinant detection using RDP4 and SimPlot by comparing GXYL1403 and GXNN1839 with the representative PRRSV strains from different lineages in the phylogenetic tree, including strains from lineage 1 (NADC30), lineage 3 (QYYZ), lineage 5 (VR2332), and lineage 8 (CH-1a and JX1A). The results show that putatively recombinant breakpoints located at positions 8961nt−11181nt in the GXYL1403 genome were detected with a high degree of reliability ( Table 4). The breakpoints in the GXYL1403 genome separated its genome into three regions (Figure 3A). The phylogenetic trees also paralleled and confirmed the recombinant events as shown in Figure 3B. The region 1nt−8961nt is FIGURE 2 | (A) Phylogenetic tree based on ORF5 sequences of isolates from this study and reference ORF5 sequences from PRRSV strains originating from China and other countries. The phylogenetic tree was constructed using the neighbor-joining algorithm based on the p-distance model. The PRRSV ORF5 sequences collected in this study are marked with black triangles. (B) Phylogenetic tree based on full-length genomic sequences of isolates. The GXYL1403 and GXNN1839 in this study and PRRSV reference strains are available in GenBank. The strains isolated in this study are labeled with black triangles.
closely genetically related to the JXA1-like strain. The region 8961nt−11181nt is closely genetically related to the CH-1alike strain, and the region 11182nt−5018nt is closely related to the JXA1-like strain ( Figure 3B). It was suggested that GXYL1403 is a natural recombinant between sublineage 8.1 (CH-1a-like) and sublineage 8.3 (JXA1-like). Two putatively major recombinant breakpoints in the GXNN1839 genome were detected with high reliability, and they are located at the nsp9 coding region (7872nt−8162nt) and the ORF2 coding region (12587nt−13282nt), respectively ( Table 4). The breakpoints in the GXNN1839 genome separated its genome into five regions ( Figure 4A). The phylogenetic trees also paralleled and confirmed the recombinant events as shown in Figure 4B. This suggests that GXNN1839 is a natural recombinant between linage 5 (VR-2332-like) and lineage 1 (NADC30-like).

DISCUSSION
PRRSV has had a large impact on the Chinese swine industry since its initial outbreak in China in 1996. In 2006, a large outbreak of PRRS caused by HP-PRRSV (10,11,26) emerged in most areas of China and caused major economic losses for swine farming. With the emergence of lineages 3 and 1 PRRSV in some areas in China in 2010 and 2015 (12,(14)(15)(16)(17), respectively, the genetic diversity of RPRSV has significantly expanded and has further complicated vaccine development. In the present study, we isolated two PRRSV strains (GXYL1403 and GXNN1839) and determined their complete genome sequences. Consistent with previous studies, phylogenetic trees based on GP5 show type 2 PRRSV strains isolated in China are divided into four lineages (1, 3, 5, and 8) (9). The GXYL1403 and GXNN1839 strains reported in this study are grouped into lineages 1 and 3, respectively. The nsp2 is the most heterogeneous protein in PRRSV (26,30,31). Compared to strain VR-2332, the nsp2 of the two isolates in this study contain different amino acid deletions, which suggest that novel PRRSV strains with amino acid deletions in nsp2 are emerging.
Recombination plays a critical role in the genetic evolution of PRRSV and presents great challenges for the prevention and control of the disease it causes. Recombination can occur during the virus-replication process. It has been shown that there are different patterns of recombination among the different lineages and sublineages of PRRSV strains (14,(32)(33)(34)(35)(36)(37)(38)(39). In the present study, whole genome sequence  comparison and recombination analysis reveals that the two isolates (GXYL1403 and GXNN1839) show novel recombination patterns. GXNN1839 is a recombinant virus that originated from recombination events that occurred in nsp9 and GP2 between strains of lineages 1 (NADC30-like) and 5 (VR2332-like). Many recombinant PRRSV strains that originated from recombination events between NADC30-like and VR2332-like strains have also been identified in previous studies (23,32,38,40). Indeed, the NADC30-like strains seem to be prone to recombine with other PRRSV strains. Numerous recombinant PRRSV strains originated from recombination events between NADC30-like strains and other lineages of PRRSV strains that were found in the field (14,23,24,34,(40)(41)(42). GXYL1403 was detected as a recombinant virus from strains of sublineages 8.1 (Ch-1a-like) and 8.3 (JX1A-like) that have circulated in China in recent years. Recombinant PRRSV strains that originated from recombination events that occurred in nsp9 between sublineages 8.1 (Ch-1a-like) and 8.3 (JX1A-like) are also reported in a previous study (42). The emergence of new PRRSV recombination variants poses a major obstacle to the effective control of the spread of virus infection and complicates vaccine development for adequate preventative measures (43). The exact mechanism of genetic recombination in PRRSV remains unknown. The breakpoints of recombination sites in the genome among different PRRS strains can be found at each region of the PRRSV genome and seem to appear randomly. A recently published study shows that nsp9 and GP2-GP3 are the hot spots for PRRSV RNA recombination (44). In this study, we find the two strains (GXYL1403 and GXNN1839) both display a breakpoint in nsp9, and GXNN1839 has an additional breakpoint in GP2, indicating that PRRSV gains genetic diversity through the frequency of recombination events at specific regions with PRRSV strains located in China.
In summary, we determined the complete genome sequences of two novel PRRSV isolates (GXYL1403 and GXNN1839). Both of these are natural recombinant strains that contain amino acid deletions in nsp2. Our findings suggest that PRRSV strains circulating in southern China have undergone recombination among the different lineages or sublineages of field strains, and RNA recombination contributes to their genetic diversity.

DATA AVAILABILITY STATEMENT
The datasets generated for this study can be found in the nucleotide sequences generated in present study were submitted to the NCBI GenBank with accession numbers, MN660069 and MN660070.