The Emergence and Pathogenesis of Recombinant Viruses Associated with NADC34-like Strains and the Predominant Circulating Strains of Porcine Reproductive and Respiratory Syndrome Virus in Southern China

Since its recent appearance in China, the NADC30-like strains of porcine reproductive and respiratory syndrome virus 2 (PRRSV-2) have caused an expanding epidemic, and this has further expanded the genetic diversity of PRRSV. In this study, three NADC30-like strains—GXFCG20210401, GXQZ20210403 and GXNN20210506—were isolated from pig serum samples obtained in Guangxi, and their genomes were sequenced. A comparative analysis of the whole genomes showed that the three strains were most similar to NADC30 (88.3–88.7%). In particular, the non-structural protein coding regions (nsp1, nsp4-5, nsp7-8 and nsp9) showed the highest similarities to JXA1, and the ORF2a-ORF5 regions showed the highest similarities to NADC34. The three strains had same discontinuous deletions of 111+1+19 amino acids in the nsp2 region, which were similar to the NADC30-like strains. Phylogenetic tree analysis based on the ORF5 gene showed that the three PRRSV isolates were divided into lineage 1.5 along with the representative NADC34-like strains, but they were classified as NADC30-like strains with respect to the whole genome and nsp2 evolutionary trees. Recombinant analysis revealed complex recombination patterns in the genomes of the three strains, which likely originated from multiple recombination events among JXA1-like, NADC30-like and NADC34-like strains. The results from animal experiments showed that the GXQZ20210403 strain was 20% lethal to piglets and caused more severe clinical reactions than GXFCG20210401, and both recombinant strains were similar in terms of pathogenicity to the previously reported NADC34 strains. This study demonstrates that NADC34-like strains of PRRSV have been circulating in the southern provinces of China and have exchanged genomes with several other indigenous strains. In addition, differences in recombination patterns may cause different clinical pathogenicity and indicate the importance of the surveillance and preventive control of recombinant strains.


Introduction
Porcine reproductive and respiratory syndrome virus (PRRSV) is an economically important swine viral pathogen which causes reproductive and respiratory disease in pigs throughout the world. PRRSV belongs to the genus Porartevirus and the family Arteriviridae [1]. PRRSVs are single-stranded, enveloped and positive-sense RNA viruses. The genome length of PRRSV is approximately 15.3 kb, which contains at least ten open reading frames (ORFs), a 5 cap structure and a 3 poly (A) tail [2]. The two large replicase

Immunofluorescence Assay
The MARC-145 cells in the 6-well plates were inoculated with the isolated PRRSV strains. After 1 h of incubation, the cells were washed three times with PBS and then fixed with cold acetone. The acetone was removed, and the cells were then blocked with 5% bovine serum albumin. A monoclonal antibody against PRRSV N protein (SDOW17, Rural Technologies, Inc, Brookings, SD, USA) was added to the cells and incubated for 1 h. The cells were washed three times with PBS and then incubated with goat anti-mouse IgG H&L, (Alexa Fluor®488, Invitrogen, San Jose, CA, USA) secondary antibody for 1 h. Finally, the cells were washed three times with PBS and observed under a fluorescence microscope while in the PBS.

Viral RNA Extraction, Reverse Transcription and Complete Genome Determination
The viral RNA of the PRRSV strains was isolated from the sera using a Viral RNA Mini Kit (Axygen Biosciences, Union City, CA, USA) according to the manufacturer's instructions. PRRSV RNA was measured by using a RT-PCR, as described previously [30,31]. For the whole genome amplification, the viral RNA was reverse transcribed into cDNA using a Prime Script Reverse Transcriptase (TaKaRa, Japan). Twelve segments covering the whole genome were amplified by PCR using the Prime STAR GXL DNA Polymerase kit (Takara, Kyoto, Japan). The primers used for the amplification of the whole genome are listed in Supplementary Table S1. The amplicons were gel-purified with an E.Z.N.A.TM Gel Extraction Kit (OMEGA, Norcross, GA, USA) and then cloned into the pMD18-T vector (TaKaRa, Dalian, China), following the manufacturer's instructions. The positive clones were sequenced by using primers T7 or T3 (HuaDa Gene Inc., Guangzhou, China). The whole genomic sequences of the three PPRSV strains were assembled using the SeqMan program of DNAstar software, version 7.0, and then deposited in the GenBank database under the respective accession numbers: OK486522(GXFCG20210401), OK486523(GXQZ20210403) and OK486524(GXNN20210506).

Sequence Comparison and Evolutionary Analysis
The MegAlign program in DNAstar 7.0 software (DNASTAR Inc., Madison, WI, USA) was used to analyze the differences between the nucleotide and amino acid sequences from the three PRRSV strains reported in this study and some representative strains. MEGA 6.0 software with the neighbor-joining method was then used to perform the evolutionary analysis of these strains. Bootstrap values were estimated for 1000 replicates. The detailed information of the selected PRRSV reference strains is shown in Supplementary Table S2.

Recombination Analysis
The Recombination Detection Program (RDP) v4.66 with seven different algorithms (RDP, BootScan, SiScan, Chimaera, GENECONV, MaxChi and 3Seq) was used for recombination analysis. The recombinant events in the PRRSV genomes were confirmed by a Bootscan analysis in Simplot software (v3.5.1, JHK University, Baltimore, MD, USA), with the default parameters. These putative recombination events were further confirmed by constructing phylogenetic trees for each of the sequence regions.

Pathogenicity of the Recombinant PRRSV Strain in Piglets
The animal experiments were carried out in accordance with the guidelines issued by the Animal Care & Welfare Committee of Guangxi University (GXU2019-043). Fourteen 28-day-old piglets of both sexes were diagnosed as negative for PRRSV by RT-PCR and a PRRSV antibody ELISA kit (JNT PRRSV-Ab ELISA kit, Beijing, China). The piglets were Viruses 2022, 14,1695 4 of 17 randomly allocated into a control group (n = 4), a GXFCG20210401-challenged group (n = 5) and a GXQZ20210403-challenged group (n = 5). PRRSV-challenged piglets were inoculated intramuscularly and intra-nasally with 2 mL (10 4 TCID50/mL) at each site. The piglets in the control group were inoculated intramuscularly and intra-nasally with 2 mL DMEM.
All of the piglets were monitored daily for clinical signs of disease, and their rectal temperatures were recorded daily. The body weights of each piglet were measured at 0, 7 and 14 days post-infection (dpi). Nasal and rectal swabs were collected for detection viral RNA shedding by RT-PCR [32]. Sera were collected at 0, 3, 5, 7, 10 and 14 dpi for viremia detection by using qPCR. The PRRSV antibodies in sera were measured using an ELISA kit (JNT PRRSV-Ab ELISA kit, Beijing, China). All the piglets were subjected to necropsy after euthanasia at 14 dpi. The tissue samples, including the thymus, lymph node, tonsil, brain sections, heart, spleen, liver, lung, kidney, small intestine and stomach, were collected from each piglet for the detection of viral loads by qPCR [32].

Hematoxylin-Eosin and Immunohistochemistry
The tissue samples of the thymus, lymph node, tonsil, brain sections, heart, spleen, liver, lung, kidney, small intestine and stomach were dehydrated and then fixed in 4% paraformaldehyde for hematoxylin-eosin staining. Samples of the lung tissue were used for immunohistochemical studies. A monoclonal antibody (SDOW17, Rural Technologies, Inc, Brookings, SD, USA) for the PRRSV N protein and HRP-conjugated goat anti-mouse IgG (H+L) was used for immunohistochemistry staining. The stained sections were observed under a microscope (Nikon E100). According to previous report [18], the microscopic pathological lesions and the intensity of the IHC signal after the staining of the lung tissues in each group were graded and scored.

Statistical Analysis
In this study, t-tests and multiple comparisons were performed to compare the differences in the means of changes in rectal temperatures, body weights, antibody levels and virus copy numbers in each group of piglets. All the data in this report are shown as the means ± SDs. Data analysis was conducted using GraphPad Prism 5 software (San Diego, CA, USA). Statistical analyses were performed with the two-tailed, unpaired Student's t-test. When multiple comparisons were performed, one-way ANOVA followed by Tukey's test or one-way ANOVA with Dunnett's test were performed. The data presented met the assumptions of the statistical test employed. Differences were regarded as statistically significant at p < 0.05 and as extremely significant at a value of p < 0.01.

Isolation and Identification of PRRSV Strains
A distinct CPE in MARC145 cells inoculated with sera from diseased piglets showing respiratory syndrome was observed. Three PRRSV strains were designated as GXFCG20210401, GXQZ20210403 and GXNN20210506, respectively. The CPEs were characterized by cell rounding and shrinkage. Some cells were detached from the 6-wells plate surface ( Figure 1). IFA showed specific PRRSV N protein expression in the MARC-145 cells inoculated with GXFCG20210401, GXQZ20210403 and GXNN20210506, indicating that the isolates propagated in the MARC-145 cells (Figure 1).
The UTR and the amino acid sequence coded by each ORF of the three PRRSV isolates were then compared with those of the reference strains. As shown in Supplementary Table  S3, the 5′UTR of the PRRSV isolates GXFCG20210401 and GXNN20210506 exhibited the highest identities (97.4 and 96.8%, respectively) with that of JXA1, and the 5′UTR of GXQZ20210403 showed the highest identity (96.8%) with that of NADC30-like strains. The 3′UTR of the PRRSV isolates (GXFCG20210401, GXQZ20210403 and GXNN20210506) The UTR and the amino acid sequence coded by each ORF of the three PRRSV isolates were then compared with those of the reference strains. As shown in Supplementary and CHSX1401 strains. The ORF2a-ORF5 of GXQZ20210403 displayed the highest nucleotide (94.5-97.4%) and amino acid (93.0-98.6%) similarities with the NADC34 strain. These results indicate that the three PRRSV isolates may be made up of mosaic isolates that originated from the predominant circulating PRRSV strains and the emerging NADC34-like strains ( Table 1). Nsp2 is the most variable protein in PRRSVs and contains different patterns of amino acid deletions and insertions. The sequence alignments of the PRRSV strains showed that the three strains had same discontinuous deletions of 111 + 1 + 19 amino acids in the nsp2 region when compared with VR2332 (Supplementary Table S1a). These deletions, which were considered to be a genetic marker, also occurred in the genome of the NADC30-like strains.
GP5 is the most enriched and variable envelope glycoprotein of PRRSVs. An analysis of the GP5 sequences of the three isolates showed that the N-terminal signal peptide and the C-terminal cellular epitope regions were more variable when compared to the VR2332 strain, while the amino acids at positions 26-27 in the decoy epitope were changed (Supplementary Table S1b). An L41S amino acid substitution was present in the PNE region of the GXFCG20210401 strain. In addition, the K59S and R151K substitutions occurred in the three isolates at positions 59 and 151, and these are thought to be associated with viral virulence. Like other known NADC34-like strains, the three isolates contain five predicted glycosylation sites in GP5.

Phylogenetic Tree
The phylogenetic tree based on the ORF5 showed that the PRRSV strains could be divided into nine lineages. The three isolates belonged to sub-lineage 1.5, which is represented by IA/2014/NADC34 and NADC34-like strains isolated in China, namely, CH/2018/NCVAnheal-1, HLJDZD32-1901 and HLJZD30-1902 ( Figure 2a). The phylogenetic tree based on the nsp2 and the complete genome showed that the PRRSV isolates belonged to sub-lineage 1.8, together with the reference strain and the NADC30 and NADC30-like strain, CHSX1401 (Figure 2b,c). These trees suggest that the three PRRSV isolates may have originated from older NADC30-like stains as well as the emerging NADC34-like strains.
The phylogenetic tree based on the ORF5 showed that the PRRSV strains could be divided into nine lineages. The three isolates belonged to sub-lineage 1.5, which is represented by IA/2014/NADC34 and NADC34-like strains isolated in China, namely, CH/2018/NCVAnheal-1, HLJDZD32-1901 and HLJZD30-1902 ( Figure 2a). The phylogenetic tree based on the nsp2 and the complete genome showed that the PRRSV isolates belonged to sub-lineage 1.8, together with the reference strain and the NADC30 and NADC30-like strain, CHSX1401 (Figure 2b and 2c). These trees suggest that the three PRRSV isolates may have originated from older NADC30-like stains as well as the emerging NADC34-like strains.  [10]. The phylogenetic tree was constructed using the Mega 6.0 distance-based neighbor joining method, with a total of 1000 replicates. The virus strains obtained in this study are marked with red circles and red boxes. The Chinese NADC34-like isolates are marked with green triangles. The American and Peruvian 1-7-4 PRRSV strains are labelled with blue squares. The NADC30-like isolates are marked with pink triangles. In the evolutionary tree branches, the NADC30-like strains are indicated by brown lines, the NADC34-like strains are indicated by blue lines and the other lineages and sub-lineages are indicated by black lines.

Recombination Analysis
The SimPlot and RDP4 software packages were used to identify possible recombination events in the whole genomes of the three novel PRRSV strains. The result of the RDP4 software showed that the whole genomes of GXFCG20210401, GXQZ20210403 and GXNN20210506 showed high degrees of certainty according to the results of at least five detection methods (Supplementary Table S4). A similarity plot analysis showed that the GXFCG20210401 genome had seven recombination breakpoints (positions of alignment) that were located in nsp1α (nt 504 and nt 696), nsp2 (nt 2044), nsp9 (nt 7870 and nt 8786), nsp12 (nt 12088) and ORF6 (nt 14508) (Figure 3a). The breakpoints divided its genome into  [10]. The phylogenetic tree was constructed using the Mega 6.0 distance-based neighbor joining method, with a total of 1000 replicates. The virus strains obtained in this study are marked with red circles and red boxes. The Chinese NADC34-like isolates are marked with green triangles. The American and Peruvian 1-7-4 PRRSV strains are labelled with blue squares. The NADC30-like isolates are marked with pink triangles. In the evolutionary tree branches, the NADC30-like strains are indicated by brown lines, the NADC34-like strains are indicated by blue lines and the other lineages and sub-lineages are indicated by black lines.

Observations of Clinical Signs
In order to explore the pathogenicity of recombinant PRRSV containing HP-PRRSV-, NADC30-and NADC34-like fragments, the PRRSV strains GXFCG20210401 and GXQZ20210403 were selected for pathogenicity tests in 4-week-old piglets. Compared with the piglets in the control group, which showed a normal physiological temperature and good health throughout the entire study period, the piglets infected with GXQZ20210403 showed body temperatures of over 40 °C on 1, 4, 8 and 14 dpi (Figure 5a). Clinical symptoms such as a cough, runny nose, anorexia and redness and swelling of the conjunctiva began to appear within 3-5 dpi, and severe symptoms characterized by the inflammatory adhesion of the conjunctiva, tachypnea, unstable standing, drowsiness and dyspnea were observed from 7 to 14 dpi. Two piglets infected with GXQZ20210403 were extremely emaciated at 13 dpi, one of which died and the other showed extreme asthenia. Piglets infected with GXFCG20210401 showed a febrile response (39.7 °C) at 6 dpi and a peak of 40.4 °C at 14 dpi. From 6 to 14 dpi, the piglets showed an abnormally high physiological body temperature (≥ 39.7 °C) for nine consecutive days. (Figure 5a). Piglets infected with GXFCG20210401 exhibited milder clinical symptoms. These included coughing, redness of the conjunctiva, anorexia, lethargy, tachypnea and slow growth, which were observed at 6-14 dpi, but all the piglets survived for the duration of the study. To reflect the pathogenicity of the two recombinant PRRSV strains, we conducted blind evaluation scores on all the piglets from different groups. From 1 to 14 dpi, the mean clinical scores of the GXQZ20210403-and GXFCG20210401-infected groups were significantly higher than those of the control group, and the difference between the GXQZ20210403and the GXFCG20210401-infected groups was highly significant (p < 0.01; Figure 5c).

Observations of Clinical Signs
In order to explore the pathogenicity of recombinant PRRSV containing HP-PRRSV-, NADC30-and NADC34-like fragments, the PRRSV strains GXFCG20210401 and GXQZ20210403 were selected for pathogenicity tests in 4-week-old piglets. Compared with the piglets in the control group, which showed a normal physiological temperature and good health throughout the entire study period, the piglets infected with GXQZ20210403 showed body temperatures of over 40 • C on 1, 4, 8 and 14 dpi (Figure 5a). Clinical symptoms such as a cough, runny nose, anorexia and redness and swelling of the conjunctiva began to appear within 3-5 dpi, and severe symptoms characterized by the inflammatory adhesion of the conjunctiva, tachypnea, unstable standing, drowsiness and dyspnea were observed from 7 to 14 dpi. Two piglets infected with GXQZ20210403 were extremely emaciated at 13 dpi, one of which died and the other showed extreme asthenia. Piglets infected with GXFCG20210401 showed a febrile response (39.7 • C) at 6 dpi and a peak of 40.4 • C at 14 dpi. From 6 to 14 dpi, the piglets showed an abnormally high physiological body temperature (≥ 39.7 • C) for nine consecutive days. (Figure 5a). Piglets infected with GXFCG20210401 exhibited milder clinical symptoms. These included coughing, redness of the conjunctiva, anorexia, lethargy, tachypnea and slow growth, which were observed at 6-14 dpi, but all the piglets survived for the duration of the study. To reflect the pathogenicity of the two recombinant PRRSV strains, we conducted blind evaluation scores on all the piglets from different groups. From 1 to 14 dpi, the mean clinical scores of the GXQZ20210403-and GXFCG20210401-infected groups were significantly higher than those of the control group, and the difference between the GXQZ20210403-and the GXFCG20210401-infected groups was highly significant (p < 0.01; Figure 5c).

Antibody Levels, Viremia and Virus Load in Tissues and Swabs
A commercially available JNT ELISA kit was employed to measure PRRSV N proteinspecific antibodies in pig sera at 0, 7, 10 and 14 dpi. As shown in Figure 6, all piglets in both challenged groups remained negative for PRRSV antibodies at 7 dpi (<0.4) and showed serum conversion at 10 dpi and 14 dpi. PRRSV-specific antibodies in the control group remained negative throughout the experimental period. Virus titers in serum samples collected at 0, 3, 5, 7, 10 and 14 dpi from the GXFCG20210401-and GXQZ20210403-infected groups reached peak levels at 5 and 7 dpi, respectively (Figure 7b). Viraemia in serum samples from the control group remained negative throughout the experimental period. After the euthanasia of each group of piglets, organ tissues were collected from each group and were examined for viral load (Figure 7a). The results of the viral load in all organs showed that the virus was detected in all the tissues of the two infection groups, with the highest viral load in the heart, lungs, tonsils and lymph nodes. Viral RNA could not be detected in the control piglet tissue samples.

Antibody Levels, Viremia and Virus Load in Tissues and Swabs
A commercially available JNT ELISA kit was employed to measure PRRSV N tein-specific antibodies in pig sera at 0, 7, 10 and 14 dpi. As shown in Figure 6, all pi in both challenged groups remained negative for PRRSV antibodies at 7 dpi (< 0.4) showed serum conversion at 10 dpi and 14 dpi. PRRSV-specific antibodies in the con group remained negative throughout the experimental period. Virus titers in serum s ples collected at 0, 3, 5, 7, 10 and 14 dpi from the GXFCG20210401-and GXQZ20210 infected groups reached peak levels at 5 and 7 dpi, respectively (Figure 7b). Viraem serum samples from the control group remained negative throughout the experime period. After the euthanasia of each group of piglets, organ tissues were collected f each group and were examined for viral load (Figure 7a). The results of the viral loa all organs showed that the virus was detected in all the tissues of the two infection gro with the highest viral load in the heart, lungs, tonsils and lymph nodes. Viral RNA c not be detected in the control piglet tissue samples.

Antibody Levels, Viremia and Virus Load in Tissues and Swabs
A commercially available JNT ELISA kit was employed to measure PRR tein-specific antibodies in pig sera at 0, 7, 10 and 14 dpi. As shown in Figure 6 in both challenged groups remained negative for PRRSV antibodies at 7 dpi showed serum conversion at 10 dpi and 14 dpi. PRRSV-specific antibodies in group remained negative throughout the experimental period. Virus titers in s ples collected at 0, 3, 5, 7, 10 and 14 dpi from the GXFCG20210401-and GXQZ infected groups reached peak levels at 5 and 7 dpi, respectively (Figure 7b). V serum samples from the control group remained negative throughout the ex period. After the euthanasia of each group of piglets, organ tissues were coll each group and were examined for viral load (Figure 7a). The results of the v all organs showed that the virus was detected in all the tissues of the two infecti with the highest viral load in the heart, lungs, tonsils and lymph nodes. Viral R not be detected in the control piglet tissue samples.  In addition, nasal swabs were collected at 0, 3, 5, 7, 9 and 14 dpi for the determination of virus shedding. Viral titers in the nasal swabs from piglets infected with GXQZ20210403 ranged from 1.22 × 10 5 to 9.81 × 10 5 copies/mL within 3-14 dpi and peaked at 5 dpi with a titer of 9.81 × 10 5 copies/mL. In the GXFCG20210401-inoculated group, high levels of shedding titers could be detected within 7-14 dpi and peaked at 14 dpi with a titer of 3.06 × 10 5 copies/mL (Figure 7c). No virus shedding was detected in the control group during the whole experiment. In addition, nasal swabs were collected at 0, 3, 5, 7, 9 and 14 dpi for the determination of virus shedding. Viral titers in the nasal swabs from piglets infected with GXQZ20210403 ranged from 1.22 × 10 5 to 9.81 × 10 5 copies/mL within 3-14 dpi and peaked at 5 dpi with a titer of 9.81 × 10 5 copies/mL. In the GXFCG20210401-inoculated group, high levels of shedding titers could be detected within 7-14 dpi and peaked at 14 dpi with a titer of 3.06 × 10 5 copies/mL (Figure 7c). No virus shedding was detected in the control group during the whole experiment.

Macroscopic and Histopathological Analysis
The lung lesions of piglets in both infected groups were characterized by symptoms of interstitial pneumonia, emphysema, multifocal pulmonary hemorrhage, petechiae and other typical lesions of PRRSVs (Figure 8a). Some areas of the parenchymal organs were earthy yellow with an uneven coloration, especially in the kidneys. In addition, the lymph nodes in the mediastinal and inguinal regions showed enlargement and hyperplasia, and punctate hemorrhaging had occurred in the lymph nodes of some piglets (Supplementary Figure S1). In contrast, the piglets in the control group showed normal morphology and staining of all tissues, and the organs were without significant lung pathology. The mean lesion scores for the infected pigs were 56 and 60.6% for GXQZ20210403 and GXFCG20210401, respectively (Figure 7d). The average score of lung injury in the challenged group was significantly higher than that in the control group.

Macroscopic and Histopathological Analysis
The lung lesions of piglets in both infected groups were characterized by symptoms of interstitial pneumonia, emphysema, multifocal pulmonary hemorrhage, petechiae and other typical lesions of PRRSVs (Figure 8a). Some areas of the parenchymal organs were earthy yellow with an uneven coloration, especially in the kidneys. In addition, the lymph nodes in the mediastinal and inguinal regions showed enlargement and hyperplasia, and punctate hemorrhaging had occurred in the lymph nodes of some piglets (Supplementary Figure S1). In contrast, the piglets in the control group showed normal morphology and staining of all tissues, and the organs were without significant lung pathology. The mean lesion scores for the infected pigs were 56 and 60.6% for GXQZ20210403 and GXFCG20210401, respectively (Figure 7d). The average score of lung injury in the challenged group was significantly higher than that in the control group.
Histopathologically, lung slices from both infected groups showed the activated proliferation of type II cells in the alveoli, diffuse and multifocal interstitial pneumonia, the thickening of connective tissue in the alveolar walls and alveolar septa and bronchial hyperplasia. This was accompanied by a large infiltration of lymphocytes and plasma cells into the interstitial layer of the lungs (Figure 8b). In the parenchymal organs, the piglets in both infected groups showed some degree of degenerative changes. For example, vacuolar degeneration was found in the livers of both infected groups, with lymphocytic infiltration within the lobules and the hepatic portal vein areas (Supplementary Figure S2). That was most evident in the renal corpuscles of the renal sections, which showed granular degeneration, mild damage to the basement membrane, a small amount of lymphocytic infiltration in the renal interstitium and multifocal hemorrhage in the renal medullary areas. Histopathologically, lung slices from both infected groups showed the activated proliferation of type II cells in the alveoli, diffuse and multifocal interstitial pneumonia, the thickening of connective tissue in the alveolar walls and alveolar septa and bronchial hyperplasia. This was accompanied by a large infiltration of lymphocytes and plasma cells into the interstitial layer of the lungs (Figure 8b). In the parenchymal organs, the piglets in both infected groups showed some degree of degenerative changes. For example, vacuolar degeneration was found in the livers of both infected groups, with lymphocytic infiltration within the lobules and the hepatic portal vein areas (Supplementary Figure S2). That was most evident in the renal corpuscles of the renal sections, which showed granular degeneration, mild damage to the basement membrane, a small amount of lymphocytic infiltration in the renal interstitium and multifocal hemorrhage in the renal medullary areas.
In the lymphoid organs (the mediastinal and inguinal lymph nodes and the spleen), the lesions were characterized by follicular and paracortical hyperplasia, especially the active proliferation of reticular cells, and in the spleen, the were characterized by the incomplete and severe disintegration of splenic vesicles, the laxity of splenic trabeculae and a marked decrease in lymphocytes (Supplementary Figure S2). The lung slices were subjected to immunohistochemical staining, and both GXQZ20210403 and GXFCG20210401 strains induced positive brownish-red staining of the epithelial cells and macrophages in the piglet lungs. Similar lung microscopic lesions and IHC staining scores were observed in all of the challenged groups (Figure 7e,f), with no statistical differences (p > 0.05). In contrast, no brownish-red cells were identified in the piglet lung tissues of the control group (Figures 7f and 8c).

Discussion
As one of the most mutation-orientated RNA viruses, there are a wide range of PRRSV subtypes created by its variability that are prevalent throughout the world. This In the lymphoid organs (the mediastinal and inguinal lymph nodes and the spleen), the lesions were characterized by follicular and paracortical hyperplasia, especially the active proliferation of reticular cells, and in the spleen, the were characterized by the incomplete and severe disintegration of splenic vesicles, the laxity of splenic trabeculae and a marked decrease in lymphocytes (Supplementary Figure S2). The lung slices were subjected to immunohistochemical staining, and both GXQZ20210403 and GXFCG20210401 strains induced positive brownish-red staining of the epithelial cells and macrophages in the piglet lungs. Similar lung microscopic lesions and IHC staining scores were observed in all of the challenged groups (Figure 7e,f), with no statistical differences (p > 0.05). In contrast, no brownish-red cells were identified in the piglet lung tissues of the control group (Figures 7f and 8c).

Discussion
As one of the most mutation-orientated RNA viruses, there are a wide range of PRRSV subtypes created by its variability that are prevalent throughout the world. This has resulted in huge economic losses for the swine industry in numerous countries [12,13,20,21]. In China, since the HP-PRRS outbreak in 2006, the PRRSV population has undergone a phase of alternating dominance between HP-PRRSV-like and NADC30-like strains and may now be undergoing a further change [12,13,16]. According to Xu et al., the combined prevalence rate of NADC34-like strains in 14 provinces in China has increased from 11.5 to 28.6% in 2020-2021, indicating a possible epidemic trend [29].
In this study, three novel NADC30-like strains were isolated-namely, GXFCG20210401, GXQZ20210403 and GXNN20210506-from serum samples that had recombinant fragments of NADC34-and HP-PRRSV-like strains from three different cities in Guangxi Province.
Comparative analysis showed that the isolates showed extensive genetic variability with representative strains of NADC30, VR-2332, QYYZ, CH-1a and JXA1. The nucleotide similarities of the whole genomes of the three isolates GXFCG20210401, GXQZ20210403 and GXNN20210506 with PRRSV representative strains were low, only sharing the highest identities with the NADC30 strains (88.7, 88.6 and 88.3%, respectively). A comparison of the UTR and each ORF region of the genomes of the three PRRSV isolates with those of the reference strains showed that most parts of the structural coding ORFs displayed the highest nucleotide and amino acid similarities with NADC34-like strains. The other region of the three isolates showed the highest nucleotide and amino acid similarities to the JXA or NADC30 strains. The nsp2 region is the largest non-structural protein in the genome, and it can harbor a variety of mutations such as deletions, recombination and insertions [4,33,34]. When compared to strain VR-2332, there were three major regions of deletions (111 + 1 + 19 aa) in the nsp2 of the three PRRSV strains. These deletions, which were considered as genetic markers of the virus, also occurred in the genome of the NADC30-like strains [25,35].
GP5 is the main capsule membrane structural protein, and it is also highly variable in PRRSVs. It has the important function of inducing the production of neutralizing antibodies. Usually, GP5 possesses two to five potential N-linked glycosylation sites in the ectodomain in the field strains of these viruses [36,37]. The three novel isolates reported in this study have five predicted potential N-linked glycosylation sites that are similar to the NADC34 and NADC34-like strains isolated from China [38]. The presence of 7-9 amino acid substitutions in the GP5 region was detected in the three isolates, and the most variable regions were the N-terminal signal peptide and the C-terminal cellular epitope regions. In addition, amino acid substitutions were present in both the PNE and decoy regions. The gained N-glycosylation sites and the amino acid substitutions in the specific antigenic sites may allow these recombinant viruses to escape the immunity afforded by commercially used vaccines [39].
The three PRRSV isolates were found in the phylogenetic tree to be classified together with NADC34 into the branch of lineage 1.5 and were located in the independent branch formed by the NADC34-like strains found in China. According to the evolution of the whole genome, the NADC34-like strains prevalent in China can be further divided into two subgroups: A and B [40]. The three isolated strains in this study were clustered in the A subgroup represented by HLJDZD32-1902 in China, suggesting that the ORF5 sequence of the parent strain was closely related to the representative strains of this subgroup. However, phylogenetic trees based on the nsp2 region and the whole genome showed that all three isolates belonged to lineage 1.8, as represented by NADC30. This indicates that there may have been recombinant events between the three isolated strains and the emerging NADC34-like strains. In addition, this reflects the limitations of epidemiological investigations based on the evolutionary tree of only the ORF5 region, which can lead to certain recombinant strains being ignored. The above comparative analysis and phylogenetic trees analysis imply that the three isolates were recombinants with complex recombination events in the genome that originated from the prevalence of NADC30-like strains, with the participation of HP-PRRSV-like and domestic NADC34-like strains.
According to previous reports [30,41], frequent recombination events occurred between the strains of lineage 1, represented by NADC30, as well as other lineages. A whole-genome-recombination analysis of 13 isolated NADC30-like strains revealed that 10 of their genomes were replaced with various gene fragments of different lengths from other PRRSV strains [30]. The early circulating strains, FJZ03 and FJWQ16, are from lineage 1, and these have recombined with the recombinant fragment from the attenuated vaccine of lineage 8 of HP-PRRSV [41]. The 15JX1, 15HEN1 and 15SC3 strains are all the results of recombinant events that occurred between the NADC30-like and JXA1-R strains [34]. The frequent recombination of lineage 1 PRRSVs can occur between different lineage strains and even between three different lineages (lineages 8, 5 and 1), such as SCcd17 and SDhz1512 [42,43]. The recombinant analysis of the three isolated strains in our study showed that all three strains were recombinant variants resulting from different degrees of mosaicism with HP-PRRSV-like strains such as the WUH4-, SH1704-25-, JL-04/12-, NADC34-and HLJZD30-1902-like strains. These are based on the domestic NADC30-like strains such as FJZ03 and FJWQ16, which acted as the main source.
Studies by Yu et al. and Zhao et al. [33,34] reported that the strains from lineage 1 and 8 strains were more prone to recombination events than others. They found that the high-frequency recombination regions were mainly distributed between nsp2, nsp9 and ORF2-ORF3. In the present study, we found recombination events in the nsp9 regions of all three strains, suggesting an important role for nsp9 in the recombination mechanism of PRRSVs. In addition, the Chinese HLJZD30-1902-like strain was involved in the genomic recombination of all three novel strains, suggesting that the NADC34-like strain has already been endemic in southern China and was co-circulating with the predominant local PRRSV strains.
Since Zhang et al. first identified and reported the presence of NADC34-like strains in China in 2017, more than twenty strains associated with NADC34-like strains have been discovered in at least 10 provinces [25,29,40]. It is worth noting that, since NADC34-like strains are in the same lineage as NADC30, the NADC34-like strains may also undergo frequent mutation and recombination events, as the NADC30-like strains do. In China, the NADC30 mosaic recombinant strains, which are associated with the vaccine and local predominant strains, have become the major prevalent strains of PRRSVs [34,41]. However, the increasing prevalence of the NADC34-like strains will further exacerbate the existing complex recombination situation, and these have a strong potential to replace the NADC30and HP-PRRSV-like strains.
The altered biological properties of PRRSVs are mainly motivated by recombination events, as evidenced by marked changes in antigenicity and cell tropism, thereby reducing the protective immunity induced by existing vaccines [17,18,41]. It has been shown that the two NADC30-like PRRSV strains, JL580 and FJ1402, are recombinants of an HP-PRRSVlike strain (09HEN1/GD), and these were shown to be highly pathogenic in vivo [15,27]. Based on field observations, a NADC34 PRRSV strain was first reported to cause dramatic "abortion storms" in sow herds and a high mortality in piglets in the USA [20]. The NADC34-like PRRSVs were also reported to be associated with high abortion rates in sows (10 to 30%) as well as with a high mortality in suckling pigs (10 to 80%) in affected Chinese pig farms. Consistent with field observations, pigs infected experimentally with the NADC34 strains, such as IA/2014/NADC34, IA/2013/ISU-1 and IN/2014/ISU-5, had a persistent fever and retarded growth, and some of the infected pigs died late in the experiments [26,28,40]. In contrast, the Chinese NADC34-like PRRSV, HLJDZD32-1901, was found to be a milder pathogenic strain in piglets [23].
A recent study has shown that PRRSV recombinants, which originated from recombination events between NADC34-and QYYZ-like strains, can have a high pathogenicity to piglets [28]. Novel PRRSV variants with evidence of recombination between viruses in lineages 8 (JXA1-like) and 1.8 (NADC30-like) and the emerging lineage 1.5 (NADC34-like) have been reported [29], but very little is known regarding the pathogenesis of the PRRSV recombinants originating from NADC34-like and circulating strains. In the present study, the recombinant viruses GXQZ20210403 and GXFCG20210506, which contain a fragment of the NADC34-like gene, were evaluated for their pathogenicity in piglets. Pigs inoculated experimentally with GXFCG20210506 displayed mild clinical signs including a mild cough, anorexia, a slight increase in body temperature and retarded growth, and they all survived until the end of the experiment. In contrast, pigs inoculated with GXQZ20210403 displayed more severe clinical symptoms such as tachypnea, lameness, drowsiness and dyspnea, a persistent fever and retarded growth. One of the pigs in this group died, and another was unable to stand due to weakness late in the experiment. Each infected group had gross lesions, and the mean lesion scores for the infected pigs were 56 (GXQZ20210403) and 60.6% (GXFCG20210506). This was similar to what was found for the previously reported SDSU73 and IA/2014/NADC34 strains. Our results demonstrated that GXQZ20210403 was a higher pathogenic strain in piglets than GXFCG20210506.

Conclusions
In conclusion, we identified and characterized three novel recombinant variants of NADC34-like strains which formed between the dominant PRRSV-2 strains (NADC30and HP-PRRSV-like) in Guangxi Province, southern China and confirmed that they were moderately pathogenic to piglets. It is worth noting that the recombinant strains came from three different cities in Guangxi Province, China, suggesting that the recombinant NADC34-like parental strains and the recombinant offspring strains may already be widely present and circulating around this area. This highlights the importance of constant monitoring and the ability to prevent the spread of NADC30-and NADC34-like strains in pig farms worldwide.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/ 10.3390/v14081695/s1, Figure S1. (A) Anatomical changes in the heart between the three groups, with the challenged groups showing slight steatosis. Figure S2. Heart, a small amount of inflammatory infiltration was present in the myocardial tissues of the challenged groups when compared to the control group. Table S1. Primers for amplification of the whole-length PRRSV genome. Table S2. The detailed information of selected PRRSV reference strains. Table S3. Comparison of ORFs and amino acids of three PRRSV isolates with representative strains of other lineages or sub-lineages. Table S4. Recombination events of PRRSV isolates GXQZ20210403, GXFCG20210401 and GXNN20210506 detected by RPD5 software.  Institutional Review Board Statement: The animal study protocol was approved by the Animal Experiment Committee of Guangxi University with the approval number GXU2019-043.

Informed Consent Statement: Not applicable.
Data Availability Statement: The complete genome sequences of the PRRSV strains GXFCG20210401, GXQZ20210403 and GXNN20210506 obtained in this study have been deposited in the GenBank under the acces-sion number OK486522, OK486523 and OK486524, respectively.