Emergence of mosaic recombinant strains potentially associated with vaccine JXA1-R and predominant circulating strains of porcine reproductive and respiratory syndrome virus in different provinces of China

A total of 293 clinical lung samples were collected from pigs from different swine herds that experienced high fever and reproductive and respiratory syndrome in 16 provinces of China in 2015. A portion of each sample was homogenized for RNA extraction and stored at −70 °C. Meanwhile, the clinical symptoms, morbidity, mortality, and changes in autopsy tissues of pigs were examined.

Total RNA was extracted from tissue homogenates using TRIzol reagent (Life Technologies, New York, NY, USA). cDNA was constructed from 7 μL of RNA using M-MLV reverse transcriptase (Promega, Madison, WI, USA). Two microliters of cDNA was used as a template for subsequent PCR analysis. Specific PRRSV nucleotide sequences of 143 bp were amplified with the primers 5′-AGC TGT GCC ARA TGY TGG-3′ and 5′-GGT RAA GTG ATG YCT GAC-3′ targeting ORF7. The cycling conditions were 95 °C for 1 min, followed by 35 cycles at 95 °C for 30 s, 56 °C for 30 s, 72 °C for 30 s, and a final extension at 72 °C for 2 min.

The presence of PRRSV antigens in lung tissues was examined by IHC analysis. Briefly, the lung tissues were fixed in 10% neutral buffered formalin for 48 h, sectioned, and stained with monoclonal antibodies specific for the N protein of PRRSV for detection of antigen-positive cells, as previously described [25].

The 28 PRRSV infected samples (Table 1) were chosen for further whole genome sequencing. Viral RNA purification from lung tissues and cDNA synthesis were performed as described previously [26], and PCR amplification of 14 fragments covering the PRRSV genome was performed with primers described in a previous report with minor modifications [27] (Additional file 1: Table S1). The quality of the PCR products was confirmed by electrophoresis on a 1% agarose gel and then purified using the PureLink Quick Gel Extraction Kit (Life Technologies). Then, the PCR products were quantified, cloned into the pMD18-T vector (Takara Bio, Inc., Shiga, Japan), and sequenced with an automated genome sequencer (Genetic Analyzer 3730XL; Applied Biosystems, Foster City, CA, USA).

A total of 651 genomic sequences, which included 28 sequences from the present study and all available sequences of type 2 PRRSVs (n = 623) retrieved from the GenBank database in March 2017 (https://​www.​ncbi.​nlm.​nih.​gov/​genbank/​) were aligned using the MUSCLE multiple alignment program (v3.8.31) [28] under default settings with minor manual adjustments afterward. Phylogenetic trees were constructed using the neighbor-joining algorithm with the Kimura 2-parameter substitution model and bootstrap tests of 1000 replicates using the MEGA6 sequence alignment tool [29]. Interactive tree of life (http://​itol.​embl.​de) [30] was employed to display and annotate the phylogenetic trees. The genetic distances of intra- and inter- clusters were computed using MEGA 6 with Kimura 2-parameter model and 1000 bootstrap replicates. Sub-data sets with a diversity level of >10% were further divided into smaller monophyletic clusters and sublineages within lineages were identified with a cut off of 1.4%.

To accurately identify potential recombinants, sequences of the 28 strains identified in this study were combined with up to 623 isolate sequences for recombination screening. Recombination events were only considered significant when supported by at least three of the following five methods: RDP [31], GENECONV [32], MAXCHI [33], CHIMAERA [34], and 3SEQ [35] implemented in RDP v4.16 [36]. In order to check the recombination signals and estimate the breakpoint locations, sequences of potential recombinants isolated in this study or similar sequences of the parental lineage and those retrieved from the GenBank database were chosen for similarity plot analysis implemented in SIMPLOT v3.5.1 [37], with a window of 500 bp and step size of 10 bp. Furthermore, recombination events were confirmed with a neighbor-joining phylogenetic tree.

Only the strains that fit all of the following three criteria, as modified from a previous report [23], were recognized as JXA1-R (JXA1 P80)-like strains: (1) nested in the phylogenetic tree within the same cluster as the vaccine strain JXA1-R, (2) sharing the highest sequence identity with the JXA1-R strain (>99%), and (3) presenting amino acids (aa) unique to JXA1-R strain derivatives.

A total of 293 clinical samples were collected from different pig farms in 16 provinces in China in 2015. Of these samples, 117 (39.93%) were positive for PRRSV detection by RT-PCR, which was similar to that report in 2010 (45.2%) [17], suggesting that PRRSV was still prevalent in China in 2015. Representative sick pigs exhibited high and continuous fever, labored breathing, blue ears, blue nose, erythematous blanching rashes, pimples in the ears or on the back, and loss of appetite (Fig. 1a, b). Pathological examination results showed hyperplasia, edema, and hemorrhage in the lungs (Fig. 1c), and severe hemorrhagic spots in the lymph nodes (Fig. 1d). IHC results indicated obvious PRRSV antigen in the lungs of diseased piglets (Fig. 1e), as compared with healthy piglets (Fig. 1f).

The full length sequences of 28 PRRSV strains from 24 pig farms in nine provinces were amplified and sequenced for further analysis. Strain names, areas, accession numbers, and clinical backgrounds are summarized in Table 1.

As shown in Fig. 2, the phylogenetic tree showed that the 28 isolates and another 623 isolates of PRRSV were clustered into 10 lineages, which generally have intralineage diversities of less than 10% (Table 2). The pairwise homology of the 28 isolates ranged from 76.3 to 99.9%, and the 28 isolates were divided into two lineages: 9 and 10 (Additional file 2: Table S2; Fig. 2). Meanwhile, the vaccine strain (ATP) and its parental strain (JA142) belonged to lineage 1. MLV RespPRRS/Repro and its parental, the prototype north American strains VR-2332, belonged to lineage 2. Some local endemic strains in Hong Kong and Taiwan, like isolates HK2, were clustered into lineage 3, and isolate CA-2, a virulent Korean strain, was clustered into lineage 4. MN184 variants, virulent MN414, and NADC31 were nested within lineage 5, 6, and 7, respectively. QYYZ-like strains and it recombinants isolated from southern China were clustered into lineage 8.

Seven strains (15LN1, 15ZJ1, 15HEN4, 15JX1, 15HEN1, 15SC3, and 15LN3) in this study were nested in lineage 9 together with NADC30-like strains and its recombinants. A total of 350 isolates mainly from China were clustered into lineage 10, which was divided into eight sublineages (10.1–10.8).

CH-1a (isolated in 1996) was located in sublineage 10.1 and HB-1(sh)/2002 (isolated in 2002) was nested in sublineage 10.2. 15HEB1 isolated in this study and recombinant SH1211 were nested within sublineage 10.3. The early HP-PRRSV isolates [15, 26] were clustered in sublineage 10.5 and 10.8. Four strains (15SN1, 15SN2, 15SN3, and 15GD4) isolated in this study and BJsy06 (isolated in 2006) were located in sublineage 10.5. HP-PRRSV JXA1, isolated in the first outbreak in 2006 [15], was clustered in subgroup 10.8. The isolates identified in this study (15GD2, 15GD3, 15HUN1, and 15HUN2) were potential recombinants associated with 2009–2010 HP-PRRSV-like strains and were located in sublineage 10.6. Sublineage 10.7 consisted of JXA1 derivatives by serial passaging in MARC-145 cells [38, 39] and potential vaccine JXA1-R (JXA1 P80)-like strains. In sublineage 10.7, 10 isolates (15ZJ2, 15ZJ3, 15HUN3, 15LN2, 15JX2, 15JX3, 15JX4, 15HEN3, 15SC1, and 15SC2) from the present study were identified as potential vaccine JXA1-R-like strains with a 30-aa deletion in Nsp2.

These results indicated that multiple subgenotypes of PRRSV co-existed in China in 2015, which included classic HP-PRRSV (sublineages 10.5 and 10.8), 2009–2010 HP-PRRSV-like strains and its recombinants (sublineage 10.6), potential vaccine JXA1-R-like strains (sublineage 10.7), and NADC30-like strains and its recombinants strains (lineage 9).

In sublineage 10.7 with vaccine strain JXA1-R as shown in Fig. 2, there were 44 JXA1-R-like strains, which included the 10 potential JXA1-R-like strains in this study and seven reported strains in China (NT1, NT2, NT3, 11NZ-GD, 11XX-GD, 11SH1-GD, and 11SH-GD) [23, 27]. These 27 Chinese strains were deposited in the GenBank database because they shared the highest nucleotide identities with JXA1-R strains (>99%) (Additional file 3: Table S3) and contained 6–11 amino acids unique to JXA1-R (Additional file 4: Figure S1). The geographical distribution results revealed that those potential JXA1-R-like strains were detected in all 16 provinces or autonomous cities (regions) with a wide geographical range in China (Fig. 3).

All recombination events in RDP were detected and confirmed by SIMPLOT and phylogenetic trees (Additional file 5: Figure S2). As shown in Fig. 4, nine of 28 isolates and one isolates from another laboratory were the result of recombination events between the JXA1-R and predominant circulating strains. Three kinds of recombination events were involved: inner-lineage, inter-lineage, and both inner-lineage and inter-lineage (Table 3). First, 15GD2, 15GD3, 15HUN1, 15HUN2, and 10FS1-GD were the products of inner-lineage recombination events between 2009 and 2010 HP-PRRSV-like strains (sublineage 10.6) and JXA1-R vaccine strains (sublineage 10.7). The genomes of 15GD2 and 15GD3 shared the same recombination pattern, as did the genomes of 15HUN1 and 15HUN2. Second, 15JX1, 15HEN1, and 15SC3 were inter-lineage recombination events between strains NADC30 (lineage 9) and JXA1-R. Third, 15HEB1, and 15LN3 were mosaic recombination events between strains NADC30, 2009–2010 HP-PRRSV, and JXA1-R. These results indicate that the recombination events in China were relatively complicated. Among them, Nsp2 or Nsp9 exhibited the complicated recombination events, involving eight breakpoints from six recombination events (Fig. 5).