Evolutionary Analysis of Four Recombinant Viruses of the Porcine Reproductive and Respiratory Syndrome Virus From a Pig Farm in China

The porcine reproductive and respiratory syndrome virus (PRRSV) is one of the most important pathogens causing substantial economic losses to the Chinese swine industry. In this study, we analyzed the complete genome sequences of four PRRSV isolates (PRRSV2/CN/SS0/2020, PRRSV2/CN/SS1/2021, PRRSV2/CN/L3/2021, and PRRSV2/CN/L4/2020) isolated from a single pig farm from 2020 to 2021. The genomes of the four isolates were 14,962–15,023 nt long, excluding the poly (A) tails. Comparative analysis of the genome sequences showed that the four isolates shared 93.2–98.1% homology and they had no close PRRSV relatives registered in the GenBank (<92%). Furthermore, PRRSV2/CN/SS0/2020 and PRRSV2/CN/SS1/2021 had characteristic 150-aa deletions (aa481+aa537-566 +aa628–747) that were identical to the live attenuated virus vaccine strain TJM-F92 (derived from the HP-PRRSV TJ). Further analysis of the full-length sequences suggests that the four isolates were natural recombinant strains between lineages 1 (NADC30-like), 3 (QYYZ-like), and 8.7 (JXA1-like). Animal experiments revealed discrepancies in virulence between PRRSV2/CN/SS0/2020 and PRRSV2/CN/L3/2021. The strain with high homology to HP-PRRSV demonstrates higher pathogenicity for pigs than the other isolate with low homology to HP-PRRSV. Taken together, our findings suggest that PRRSVs have undergone genome evolution by recombination among field strains/MLV-like strains of different lineages.

All PRRSV strains isolated to date have been classified into two major genotypes, namely, the European (type 1) and North American (type 2) genotypes, with type 2 PRRSV being predominant in China since its initial report there in 1996 (9,10). PRRS is considered one of the most important infectious diseases in Chinese swineherds as it causes severe economic losses in the swine industry every year. Currently, the Chinese type 2 PRRSV strains can be classified into different lineages, including lineage 1 (NADC30-like), 3 (QYYZ-like), 5 (VR2332like), and 8 (JXA1-like and CH-1a-like) based on phylogenetic analysis of ORF5 sequences (11,12). The diversity of the Chinese type 2 PRRSV has been increasing due to recombination events among the different PRRSV lineages since the emergence of NADC30-like PRRSVs in China in 2012 (12)(13)(14)(15)(16)(17)(18)(19)(20)(21)(22)(23). In the present study, we identify four recombinant PRRSV strains circulating in a swine farm in Fujian province, China. Phylogenetic and molecular evolutionary analyses indicated that these strains evolved from natural recombination among the NADC30-, QYYZ-, and JXA1-like variants. To further understand the four strains, we genetically characterized the complete genomes of the viruses.

Sample Collection and Viral Isolation
From 2020 to 2021, a severe reproductive and respiratory disease was observed in pigs from four independent pens in a swine farm in Fujian province, China. This farm created a PRRSpositive stable herd by inoculating 0.05 dose/pig of MLV TJM-F92 vaccine (derived from the HP-PRRSV TJ) from 2018. The affected pigs from the two pens exhibited high fever (40.3-41.8 • C), severe respiratory syndrome, and high morbidity (30%) and mortality (20%), whereas those from the other two pens exhibited fever, severe respiratory syndrome and reproductive problems, and especially high abortion rate (10%) was observed in the sows. PRRSV was detected using an RT-PCR kit (Beijing Anheal Laboratories Co., Ltd., Beijing, China) according to the manufacturer's instructions. PRRSV was positive when the CT value was ≤30. MARC-145 CD163 cell lines (a stable cell line highly expressing porcine CD163) were used for PRRSV isolation from the positive samples (12). The tissues (lung, lymph nodes, and serum samples) were homogenized with DMEM containing antibiotics and antimycotics to obtain a 50% (w/v) suspension. After freeze-thawing thrice, the samples were centrifuged at 10,000 × g for 10 min and the supernatants were sieved through a 0.22-µm filter. MARC-145 CD163 cells were mixed with 100 µL of supernatant and incubated at 37 • C with 5% CO 2 for 3-5 days.
PRRSVs were passaged three times in MARC-145 CD163 cells for subsequent analysis after being plaque-purified three times.

RNA Isolation and RT-PCR
RNA was extracted from the positive tissues (lungs, serum, and lymph nodes) and the virus using a Viral RNA Mini Kit (TIANGEN, Beijing, China) according to the manufacturer's instructions. cDNA was generated using the HiScript R III 1st Strand cDNA Synthesis Kit (Vazyme Biotech Co., Ltd, Nanjing, China) following the supplier's instructions. The fulllength viral genomes were amplified using six viral-specific primers (Supplementary Table 1). Subsequently, we confirmed the complete genome using other PCR primers as described previously (24,25). The purified PCR products were cloned into the pEASY R -Blunt Simple Cloning Vector (TransGen Biotech, Beijing, China). To determine a consensus sequence of each fragment, at least three recombinant clones were sent to the Ruibo Life Technologies Corporation (Beijing, China) for sequencing by using the Sanger approach in both directions for each fragment.

Complete Genomic Sequence and Recombination Analysis
Forty-four representative PRRSV strains available in GenBank, including the type 1 and type 2 strains, were used for the comparative sequence analysis in this study ( Table 1). Multiple sequence alignments and genome analyses were performed using the MEGA 7.0 software and the DNAStar 7.0 package. Phylogenetic trees were constructed using the neighbor-joining method in MEGA 7.0 and bootstrap confidence values from 1,000 replicates.
Potential recombination within the whole genome sequences was determined using seven methods (RDP, BootScan, GENECONV, Chimera, Maxchi, SiScan, and 3Seq) via the recombination detection program 4.10 (RDP 4.10) (26). A recombination event was identified when at least five of the seven methods reported recombination signals in RDP 4.10 with the highest acceptable p of 0.05 (27). The possible recombination event was further confirmed using the SIMPLOT software (version 3.5.1) with a 200-bp window width and a 20-bp step size (28).

Challenge Experiment
To evaluate the pathogenicity of the recombinant strains, two strains PRRSV2/CN/SS0/2020 (98.3% homology with JXA1) and PRRSV2/CN/L3/2021 (86.3% homology with JXA1) were selected for animal experiments. Fifteen 4-week-old pigs confirmed to be free of PRRSV and PCV2 were randomly divided into 3 groups (5 pigs/group). The pigs in group 1 were intranasally administered with 2 mL of PRRSV2/CN/SS0/2020 containing 2×10 5 TCID50 and those in group 2 were intranasally administered with 2 mL of PRRSV2/CN/SS0/2020 containing 2×10 5 TCID50. The pigs in the negative control group were inoculated with 2 mL of Dulbecco's Modified Eagle Medium (DMEM). Rectal temperature was recorded daily from 0 to 14 days post-challenge (dpc). Serum was collected at 0, 4, 7, 11, and 14 dpc to detect PRRSV-specific antibodies using PRRSV using ELISA kit (IDEXX Laboratories Inc., Westbrook, ME, USA). Viral load in the sera of each group was detected by an IFA-microtitration infectivity assay as previously described (29). All animals were euthanized at 14 dpc for necropsy. In addition, lung tissues of pigs were collected at necropsy and fixed in 10% neutral-buffered formalin and routinely processed for histological examination. The procedures of animal handling and experimentation performed in this study were approved by the Longyan University Animal Ethics Committee (Permit number: permit no.LY20210010X).

Data Analysis
Statistical analysis in this study was performed using a one-or two-way ANOVA analysis of variance in GraphPad Prism software (version 6.0), and the results were considered statistically significant when p < 0.05.  Table 2).

Complete Genome Sequence Analysis of PRRSV
To evaluate the genomic characteristics of the four PRRSV isolates, each region of the genomes of the four strains was further compared with four viruses from different lineages, including NADC30-like strains, QYYZ-like strains, VR2332-like strains (VR-2332 and BJ-4), JXA1-like strains, and the LV (the prototype of type 1 PRRSV) strain (  ORF1a and ORF1b encode the non-structural proteins (Nsp) of PRRSV. A nucleotide sequence comparison of ORF1 showed that in the four strains, ORF1b was relatively conserved as compared to ORF1a. Of all the Nsp sequences, the most variable ones were from Nsp1β and Nsp2 within ORF1a, which had 72.4-94.9% and 63.4-86.9% amino acid identity with the reference strains, respectively ( Table 2).

Phylogenetic Analysis of PRRSV
Four respective phylogenetic trees were constructed based on the full-length genome sequences, Nsp2, ORF2-7, and ORF5 nucleotide sequences of the four PRRSV isolates and 27 representative strains of type 2 PRRSV.
Interestingly, the tree based on the full-length genome sequences showed that the four PRRSVs formed a separate branch that was in the middle of lineages 3 and 8.7, which are represented by QYYZ and JXA1, respectively ( Figure 1A). The tree based on the ORF2-7 or ORF5 sequences indicates that the four strains were grouped in lineage 3, together with the QYYZlike strains (Figures 1C,D). The tree constructed based on Nsp2 indicates that PRRSV2/CN/L3/2021 and PRRSV2/CN/L4/2020 were clustered into lineage 1, together with the NADC30like strains, and PRRSV2/CN/SS0/2020 was clustered into lineage 8.7, together with the JXA1-like strains; however, PRRSV2/CN/SS1/2021 was in a separate branch (Figure 1B).

Amino Acid Analysis of NSP2 and ORF5
The Nsp2-coding region is recognized as one of the most variable proteins with different deletions and insertions. Therefore, Nsp2 is often used to analyze the genetic variations and molecular epidemiology of PRRSV. As shown in Figure 2, two isolates, namely PRRSV2/CN/L3/2021 and PRRSV2/CN/L4/2020, had a unique discontinuous deletion of 131-aa in the Nsp2coding region, which was identical to that of the NADC30 and NADC30-like strains. Meanwhile, PRRSV2/CN/L4/2020 had an additional 1-aa deletion at the 15th aa. Interestingly, both PRRSV2/CN/L3/2021 and PRRSV2/CN/L4/2020 had two additional 2-aa insertions at amino acid positions 224-225 (Figure 2A). Notably, PRRSV2/CN/SS0/2020 and PRRSV2/CN/SS1/2021 had characteristic 150-aa deletions (aa481+aa537-566+aa628-747) that were identical to those of the live attenuated virus vaccine strain TJM-F92 (derived from the HP-PRRSV TJ) (30) and TJbd14-1 (13), which is an MLV-like strain that evolved from the HP-PRRSV vaccine virus TJM-F92 ( Figure 2B). Notably, comparative sequence analysis showed that PRRSV2/CN/SS0/2020 and PRRSV2/CN/SS1/2021 exhibited the highest nucleotide similarity (89.5-95.3%) and amino acid similarity (87.0-94.0%) with TJbd14-1 and TJM-F92.   Frontiers in Veterinary Science | www.frontiersin.org AR37-51, and AR192-200) at the N-terminus of the four strains were the most similar to the QYYZ-like strains but differed from the JXA1-like, VR-2332-like, and NADC30-like strains. Additionally, seven unique amino acids distributed in GP5 were only identified in QYYZ-like strains and the four strains (I 26 , Y 38 , C 66 , S 92 , F 117 , I 152 , and H 199 ) (Figure 3). Studies have shown that the 13th and 151st amino acids of GP5 are related to the virulence of the virus (33). In the present study, four strains had both R 13 Q and R 151 G mutation as compared with the VR2332 strain.

Recombination Analysis
The recombination analysis using the RDP version 4.10 software revealed that the four isolates are the result of recombination between JXA1-like, QYYZ-like, and NADC30-like strains circulating in China (Table 3, Figure 4, Supplementary Figure 2). Additionally, the putative recombinant events and statistically incongruent phylogenetic trees were further confirmed using SimPlot v3.5.1 and statistically incongruent phylogenetic trees (Figure 4,  Supplementary Figure 2). From the similarity plot, the breakpoints of recombination events of four strains are mainly located in nsp1, nsp2, nsp3, nsp11, and nsp12 (Figure 4,  Supplementary Figure 2).
The PRRSV was isolated from the serum of pigs challenged with the two strains at 4 dpc. The mean virus titers of PRRSV2/CN/SS0/2020 or PRRSV2/CN/L3/2021 reached peak levels at 7 dpc (10 6.2 TCID50/mL and 10 5.0 TCID50/mL, respectively) and high viremia persisted until 14 dpc in every PRRSV-infected group (Figure 5C). Furthermore, the mean virus loads for PRRSV2/CN/SS0/2020-infected group were significantly higher than those of the PRRSV2/CN/L3/2021infected group at 7 and 11 dpc (p < 0.05). Meanwhile, no PRRSV was detected in the control group throughout the experiment. The specific antibodies of PRRSV N protein in the sera of pigs were detected by a commercial ELISA kit. All pigs in the PRRSV2/CN/SS0/2020-infected group became     seroconverted at 7 dpc, whereas PRRSV2/CN/L3/2021-infected were seroconverted at 11 dpc ( Figure 5D). In contrast, PRRSVspecific antibodies were not detected in the control group throughout the experiment.

DISCUSSION
Porcine reproductive and respiratory syndrome virus is one of the most important pathogens causing substantial economic losses to the Chinese swine industry. According to the global PRRSV classification system, four different lineages of PRRSV (lineages 1, 3, 5, and 8) co-existed in Chinese swine herds, and two of them (lineages 1 and 8) are currently the predominant strains circulating in China. With the emergence of NADC30-like PRRSVs in China since 2012, recombinant strains have been frequently observed in the field. Notably, multiple studies have shown that PRRSV variants evolved from recombination events between lineage 1 and one or two other lineages (12)(13)(14)(15)(16)(17)(18)(19)(20)(21)(22)(23). In this study, four PRRSV strains (PRRSV2/CN/SS0/2020, PRRSV2/CN/SS1/2021, PRRSV2/CN/L3/2021, and PRRSV2/CN/L4/2020) were also identified as recombinant viruses from three lineages of type 2 PRRSV (NADC30-like, QYYZ-like, and JXA1-like viruses). Importantly, the four strains in this study had a low wholegenome nucleotide similarity with other PRRSV sequences in GenBank (< 92%), which raised the concern of the introduction of new viruses into other large-scale pig farms. Therefore, to further our knowledge about these recombinant PRRSV strains, we characterized them genetically and phylogenetically to obtain insights that link genotypic with phenotypic data.
Nsp2 and ORF5 are the most variable proteins in the viral genome and are usually used as target genes for the molecular epidemiological surveillance of PRRSV (27). Compared with VR2332, the Nsp2 gene of PRRSV contains different patterns of amino acid insertions and deletions. For example, an outbreak in China with highly pathogenic PRRSVs (HP-PRRSV, JXA1like) in 2006 had a unique 30-aa deletion in the Nsp2 region, whereas a unique 131-aa (111aa+1aa+19aa) deletion pattern in Nsp2 is the footprint of NADC30 and NADC30-like strains (11,41). In the present study, the Nsp2 gene of the two strains (PRRSV2/CN/L3/2021 and PRRSV2/CN/L4/2020) had the 131aa deletion, which is identical to NADC30-like strains, suggesting that these two strains may belong to lineage 1. Additionally, 2-aa insertions were also found in PRRSV2/CN/L3/2021 and PRRSV2/CN/L4/2020. Previous studies showed that MLV TJM-F92 has a unique 150-aa deletion (aa481+aa537-566+aa628-747) signature in its Nsp2-coding region (30). Interestingly, two strains identified in our study (PRRSV2/CN/SS0/2020 and PRRSV2/CN/SS1/2021) had characteristic 150-aa deletions that were identical to TJbd14-1 (an MLV-like strain that evolved from TJM-F92) and TJM-F92, suggesting that the PRRSV genome may have evolved more compactly by eliminating dispensable genomic regions (42). Furthermore, the phylogenetic analysis based on the ORF5 gene demonstrated that the four strains belonged to the QYYZ-like virus. In addition, the GP5 antigenic regions in the four strains were similar to those of the related lineage 3 representative strains but different from other lineage representative strains. The discordance between Nsp2 patterns and ORF5 lineages resulted from the recombination of PRRSVs. In other words, there is limited understanding of the recombinant PRRSV genomic data, especially if quantifying PRRSV diversity has only focused on ORF5 and Nsp2 analysis. In addition, the 137th aa of the GP5 protein is assumed to differentiate the attenuated vaccine strain (A 137 ) and the wild strain (S 137 ). Residue S 137 was present in all four strains, suggesting that the four strains may be wild viruses (43). Additionally, a retrospective survey found that a massive vaccination with HP-PRRSV live attenuated vaccine (TJM-F92) was effective in preventing and controlling PRRS in this farm since 2018. Thus, we propose that PRRSV2/CN/SS0/2020 and PRRSV2/CN/SS1/2021 may have evolved from recombination among MLV TJM-F92 and NADC30-like or other PRRSV strains. Whether this strain directly evolved from MLV TJM-F92 needs to be further investigated.

DATA AVAILABILITY STATEMENT
The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found in the article/Supplementary Material.

AUTHOR CONTRIBUTIONS
JLiu and CW performed the experiments. YX, CL, and YY conceived and designed the experiments. LL, CH, and JLi performed the data analyses. JLiu completed the writing of the manuscript. All authors contributed to the article and approved the submitted version.