The prevalent status and genetic diversity of porcine reproductive and respiratory syndrome virus in China: a molecular epidemiological perspective

Porcine reproductive and respiratory syndrome virus (PRRSV) has been epidemic more than 30 years in America and 20 years in China. It is still one of the most important causative agents to the worldwide swine industry. Here, we systematically analyzed the prevalence status of PRRSV in China by a molecular epidemiological perspective. Now both PRRSV-1 and PRRSV-2 are circulating and approximately more than 80% of pig farms are seropositive for PRRSV. For PRRSV-2, there are four lineages (lineage 1, lineage 3, lineage 5, lineage 8) circulating in the fields. Lineage 8 (CH-1a-like) and lineage 5 (BJ-4-like) appeared almost at the same time during 1995-1996. Notably, BJ-4 shares 99.6% and 99.8% identity with VR2332 and RespPRRS MLV, respectively. It means that lineage 5 is likely to be imported from America. Now highly pathogenic PRRSV (HP-PRRSV) which was considered to be evolved from local diversity of lineage 8 strains is predominant with different variants. Lineage 3 appeared in 2010 which is mainly sporadic in south of China. Lineage 1, also known as NADC30-like strains in China, has been prevalent since 2013 and leads to PRRS pandemic again. For PRRSV-1, although sporadic at present, more than 9 provinces/regions have been reported. All the circulating strains belong to subtype I. It should be paid more attention since there are no vaccines available. Our analysis would help to deeply understand the prevalent status of PRRSV in China and provide useful information for prevention and control of porcine reproductive and respiratory syndrome (PRRS).


Background
Porcine reproductive and respiratory syndrome (PRRS) is one of the most important economically diseases to the swine industry worldwide. It is estimated that the total economic losses caused by PRRS is about $664 million annually in the America, an increase from the $560 million annual cost estimated in 2005 [1]. PRRS was firstly described in America in 1987 [2], followed in Europe and Asia (1990Asia ( -1992 [3][4][5][6]. The causative agent, porcine reproductive and respiratory syndrome virus (PRRSV), was isolated in 1991 in the Netherlands and 1992 in the United States with originally isolated strains named Lelystad Virus (the European prototypic strain) and Swine Infertility and Respiratory Syndrome (SIRS) virus (also known as ATCC VR-2332, the North American prototypic strain), respectively [7,8]. The retrospective studies showed that the conserved serum samples were PRRSV antibody positive as early as 1979 in Canada [9]. It suggested that PRRSV had already circulated in swine population before 1979.
High genetic diversity is a significant characteristic of PPRSV. Mang Shi et al. constructed the global classification system of PRRSV based on the comprehensive analysis of the complete ORF5 gene sequence. According to the classification system, the PRRSV-1 was divided into three subtypes (subtype 1-3) and the PRRSV-2 was classified as 9 lineages with several sublineages of each lineage [18,19]. The virulence and antigenicity are also different and variable due to the genetic diversity. There are several highly pathogenic PRRSV (HP-PRRSV) strains which cause severe reproductive and/or respiratory diseases and the emergence of new HP-PRRSV often leads to the widespread re-pandemic of PRRS. For example, although the vaccine was used in the late 1990s, the "abortion storm" swept through the America and an atypical PRRSV strain MN184 was isolated in 2001 [20,21]. In Eastern Europe, a highly pathogenic PRRSV, Lena strain, was reported by Belarus in 2010 [22]. In China, the emergence of HP-PRRSV variants in 2006 led to atypical PRRS pandemics and 20% mortality in pigs [23]. Since 2013, PRRS became prevalent again in China caused by new PRRSV variants, NADC30-like strains, which are considered to be imported from North American and adapted in China [24,25].
Here, we retrospectively and phylogenetically analyzed the prevalence and genetic diversity of PRRSV in China through a molecular epidemiological perspective. We hope to provide valuable information and new insights for the PRRSV epidemic trends and control strategies. Genome analyses were performed using the DNASTAR package. The unrooted phylogenetic tree was generated by the distance-based neighbor-joining method using MEGA 6.0 and the amino acid sequences were aligned by Clustal W method (MEGA 6.0) [26]. The recombination events were detected by RDP4 program using seven different algorithms (RDP, GeneConv, BootScan, MaxChi, Chimera, SiScanm and 3Seq) with Bonferroni correction and a highest acceptable p-value of 0.01 [27].

PRRSV history and prevalent status in China
PRRS was firstly described in 1995 in China and the etiological agent, PRRSV, was isolated by Guo et al. (1996) and Yang et al. (1997). The isolated strains were named CH-1a (Genbank ID: AY032626) and BJ-4 (Genbank ID: AF331831), respectively. Both of them belong to PRRSV-2 [28,29]. Serological studies showed that the percent of positive serum is more than 40% in the east and north of China [30], and it means that PRRSV circulated in China earlier than 1995. It is not clear how the PRRSV was introduced into pig population in China. One possibility is the import of breeding pigs. The evidences are: (i) Chen et al. (1996) successfully isolated PRRSV strain from the imported breeding pigs from Canada [31]; (ii) The genome homology of BJ-4 with VR2332 is about 99.6% and the genome homology of CH-1a with JA142 (Genbank ID: AY424271) is about 94.5%. Both of them (especially BJ-4) are closely related to North American isolates and PRRSV was endemic during 1995~2005.
During the summer of 2006, a new PRRSV variant, also known as highly pathogenic PRRSV (HP-PRRSV), caused atypical PRRS and led to a devastating destruction to swine industry with 20% motality to pigs. The new PRRSV variants show a unique discontinuous 30 amino acids (482aa, 534-562aa) deletion in Nsp2 gene [22]. The representative virus strains are JXA1 (Genbank ID: EF112445), TJ (Genbank ID: EU860248) and HuN4 (Genbank ID: EF635006) which are also the parent strains for attenuated vaccine strains JXA1-R, TJM-F92 and HuN4-F112, respectively [32][33][34]. The HP-PRRSVlike strains are predominantly epidemic strains in the pig farms after 2006. Since 2013, there was a markedly increased PRRSV infection during routinely clinical investigation. The reason for this re-pandemic of PRRS is considered as the emergence of NADC30-like strains which were probably imported from North America and went though extensive variation in China [24,25,35]. Now both HP-PRRSV-like and NADC30-like strains are mainly circulating in the fields and have a high clinical detection rate.
For the PRRSV-1, B13 (Genbank ID: AY633973) was the earliest strain reported in China. It was conserved in Dalian Animal and Plant Quarantine Bureau, whose original source is not clear [36]. The retrospective studies showed that the PRRSV-1 strains could be detected from clinical samples as early as 2006 in mainland of China [37]. And now they are sporadic in more than nine provinces/regions (Inner Mongolia, Heilongjiang, Liaoning, Beijing, Fujian, Guizhou, Guangdong, Shanghai, Hong Kong) [38]. We should pay more attention to the prevalence of European type in China since more and more clinical investigation reported their existence and there are no authorised vaccines yet.

The epidemic of PRRSV-2 in China
Based on the global PRRSV classification system and ORF5 sequence information in the Genbank, we chose 127 PRRSV-2 ORF5 sequences, including 9 reference strains and 118 prevalent strains ( Table 1). As shown in Fig. 1, the prevalent PRRSV-2 strains in China were clustered into four lineages: lineage 1, lineage 3, lineage 5 (sublineage 5.1) and lineage 8 (sublineage 8.7). Lineage 8 is predominant since the emergence of PRRSV in China which included classical PRRSV strains (CH-1a-like) prevalent before 2006 and HP-PRRSV-like strains prevalent after 2006. The epidemic of lineage 1, also known as NADC30-like strains, which spread rapidly around the country since 2013 and now the clinical detection rate is comparable with lineage 8 (HP-PRRSV-like). Lineage 3 was another newly emerged variants since 2010 which is mainly circulating in south of China (Jiangxi, Fujian, Guangdong, Guangxi) and the clinical detection rate is less than 10% [39,40]. Although the lineage 5 (BJ-4-like/ VR2332-like) appeared as early as in 1996, it is always non-pandemic in China and the clinical detection rate is low. It is still a puzzle that how the lineage 8 (HP-PRRSV-like) and lineage 1 (NADC30-like) become the major epidemic strains, while lineage 8 (CH-1a-like) and lineage 5 are always endemic in China. One reasonable explain is that HP-PRRSV-like and NADC30-like strains show high genetic variations and incidence of recombination, compared with lineage 8 (CH-1a-like) and lineage 5. These characteristics probably made current vaccines ineffective and confer them much easier to escape the immune surveillance. Thus, they adapted well during the pig populations.

Lineage 8/sublineage 8.7
According to the phylogenetic assay of ORF5 sequence, sublineage 8.7 could be divided into five subgroupssubgroup I~subgroup V (Fig. 1). Subgroup I is PRRSV strains that are closely related with CH-1a. CH-1a was the first isolated PRRSV strain in China and was recognized as an ancestor strain of classical PRRSV. The genome homology is 94.5% with JA142 (Genbank ID: AY424271) and SDSU73 (Genbank ID: JN654458) (sublineage 8.9) and 91.5% with VR2332 (sublineage 5.1). Subgroup II was considered as intermediate subgroup between classical and HP-PRRSV. BJ0706 (Genbank ID: GQ351601) and NB-04 (Genbank ID: FJ536165) were the representative strains of transition [41]. They have been found 1aa deletion at the 481th residue site in Nsp2 gene and they showed a genome identity of 95.1~95.8% with CH-1a and 97.1~98% with JXA1. Subgroup III~V belong to HP-PRRSV derivative strains or HP-PRRSV vaccine derivative strains. HP-PRRSV leaded to the outbreak of PRRS in China in 2006. The representative strains JXA1, TJ, HuN4 show 95.3% of genome homology with CH-1a and 89.6% with BJ-4. At present more consistent view is that HP-PRRSV is originated from lineage8 strains that have been circulating in China for about 10 years before the outbreak in 2006 [19]. The genetic diversity is still expanding and shows different evolutionary clusters following circulation of HP-PRRSV and extensive vaccination of modified live vaccines (JXA1-R, TJM-F92, HuN4-F112). Subgroup III is genetically closer with JXA1, also called JXA1-like group. Subgroup IV is evolutionarily close to TJ and HuN4 strains. Subgroup V is closely related with HP-PRRSV attenuated live vaccines. Considering modified live vaccines could circulate in the fields, it is expected to see many vaccines related sequences. And this also pose a challenge for a rapidly accurate diagnosis and control of PRRS.  [29]. It shows 99.6% and 99.8% of genome homology with VR2332 and RespPRRS MLV (Boehringer Ingelheim, Germany), respectively. It suggests that sublineage 5.1 is more likely to be imported into China from North American. It is still not known why the prevalence of sublineage 5.1 is limited compared with lineage 8. Lineage 3 was firstly reported in 2010 and considered as a novel lineage originally [42,43]. It is mainly endemic in southeast of China (Jiangxi, Fujian, Guangdong, Guangxi) [38]. The representative strains GM2 (Genbank ID: JN662424) and QYYZ (Genbank ID: JQ308798) show 86.2~88.7% of genome homology with VR2332, CH-1a, BJ-4 and JXA1. Lineage 3 is genetically distant from previous epidemic strains and its source remains to be further clarified. Clinical detection rate of both sublineage 5.1 and lineage 3 is less than 10%.

Lineage 1
Lineage 1, also known as NADC30-like strains, has been reported since 2013 in China [44]. Now it's epidemic in more than nine provinces/regions including northeast, southeast, central and east of China. The clinical detection rate is increasing every year and up to 50% in Henan Province [45]. The representative virus strain HENAN-XINX (Genbank ID: KF611905) and JL580 (Genbank ID: KR706343) share 86.7~87.8% genome homology with MN184A (Genbank ID: DQ176019) isolated in 2001 in America, 92.8~95.4% with NADC30 isolated in 2008 in America and 82.2~87.2% with CH-1a, BJ-4, JXA1 and QYYZ. Like MN184A and NADC30, the NADC30-like virus strain in China shows the identical deletion pattern of nsp2 including 131aa discontinuous deletion. NADC30 is likely to be introduced into China by import of breeding pigs [24,25,35].
Compared with the other PRRSV lineages in China, NADC30-like strains show more recombination possibilities and pathogenic diversity ( Table 2). There are several reports about the recombination of NADC30-like strains

The epidemic of PRRSV-1 in China
Although there are more recent reports about the prevalence of European type PRRSV in China, the clinical detection rate is still low [37,[56][57][58]. Now the research mainly focuses on the molecular epidemiology of PRSSV-1 in China. Here we did a genetic diversity assay based on the ORF5 sequence by choosing 20 strains isolated in China and 23 reference strains ( Table 3). The results showed all the Chinese isolated strains belong to subtype 1 (pan-European type) and could be further divided into four subgroups (Fig. 2). Our analysis may provide valuble information for the classification of future PRRSV-1 strains for other researchers.

N protein polymorphism
Normally the N protein is highly conserved. In genotype 1 PRRSV, N protein, as well as the high variable ORF5 and Nsp2 genes, is found to be very pleomorphic. The length of N protein ranges from 124 to 132aa [59][60][61].

Pathogenicity and recombination
To our knowledge, until now only two Chinese PRRSV-1 isolations, GZ11-G1 and HLJB1, to be evaluated the pathigenicity. These two strains are pathogenic to piglets and can lead to classic PRRSV-specific lesions. The infectious pigs showed mild clinical signs and no pigs died, which indicated they are less virulent [58,63]. Interestingly, both of them belong to subgroup I-the Amervac PRRSV vaccine-related strains. HLJB1 and GZ11-G1 show 91.64% and 96.3% genome homology with the Amervac PRRSV virus, respectively [58,63]. Additionally, recombination analyses showed that HLJB1 is a recombinant from the Amervac vaccine and the BJEU06-1 isolate [64]. It is necessary to routinely monitor the prevalence of PRRSV-1 in the fields and more research needs bo be done for better understanding the biological and pathogenic characteristics of PRRSV-1 in China.   [55] shortening duration of viral shedding. They can provide an efficient protection against a lethal challenge with their respective parental HP-PRRSV isolates [32][33][34][65][66][67]. However, they can not completely prevent infection and establish sterilizing immunity. Also, the vaccine efficiency significantly decreases against heterologous challenge, and they can only provide partial or limited protection [47,65,68,69]. Massive vaccination with live attenuated vaccines also leads to safety concern and more genetic divergence: (i) It has been demonstrated that vaccine viruses could spread from vaccinated pigs to non-vaccinated, suggesting their ability to circulate in the fields and form different vaccine-associated clusters [70,71]; (ii) The extensive immune pressure maybe serve as a major driving force which greatly promotes genetic diversity of genotype 2 in China [72]; (iii) Revertants from the vaccine derivatives should be paid more attention to. Indeed, there are several reports about the attenuated vaccine reverting to a virulent type [49,73]; (iv) The frequency of recombination is increasing, especially the emergence of NADC30-like virus strains (lineage1) which display a wide broad recombination ability with HP-PRRSV (JXA1), classical PRRSV (CH-1a,VR2332) and attenuated vaccines (JXA1-P80, TJM-F92-like) [46,49,52,74]. It has been verified that recombination is associated with re-prevalence of HP-PRRSV during 2009 to 2010 and closely related with the pandemic of NADC30-like virus since 2013 in China [24,53]. Accordingly, veterinarians and farmers become more cautious and re-evaluate the previous immune strategies. Some farmers are trying to reduce and stop the vaccination of PRRSV in China by taking other comprehensive PRRS prevention and control measures.

Production system and biosecurity
It is confirmed that multiple production sites and high quality biosecurity system are important in PRRSV prevention and control [75,76]. Due to historical reasons, most of Chinese pig farms are one production site and the layout of gestation, farrowing, nursery and growing are also not reasonable. Many farms are still continuous flow sites and lack good biosecurity system. Following the development of swine production in China, more and more pig producers have recognized the significance of management, production system, environment control and good biosecurity. They have changed their views that only or over-dependent vaccination could prevent porcine diseases and would like to invest more in nutrition, management, environment, biosecurity and application of new technologies. For example, airfiltration ventilation systems were used by some breeding companines; gilts acclimation and strict batch production (also known as "All in All out", AIAO) were accepted and implemented by more commercial pig farms.
The strategy of "load, close, homogenise" (LCH) The load, close, homogenise (LCH) strategy (also known as load, close, expose) was extensively accepted in PRRV control and elimination for it's inexpensive and easy operability [75,77,78]. To perform this strategy, the producer should load enough gilts once for minimum of 200 days and then implement herd closure. The key points are gilts acclimation and good biosecurity system. Poul H. Rathkjen and Johannes Dall summarized ten golden rules for biosecurity management and showed a good model for successful implement of LCH [76]. Uniform PRRS status could be achieved either by simultaneous vaccination or by inoculation with serum containing resident virus. Notably, it should be very cautious when used the virus positive serum for acclimation because of the virulence of HP-PRRSV. It is a better choice to build regional elimination of PRRSV. This model was successfully implemented in the state of Minnesota of the United States and in Horne Peninsula of the Denmark [75,76]. They are good examples for HP-PRRSV regional elimination in China. This Fig. 2 The phylogenetic assay based on ORF5 gene sequences of PRRSV-1. All the Chinese isolated strains belong to subtype 1 (pan-European) and can be classified into 4 subgroups according to the phylogenetic relationships. The Chinese PRRSV strains are marked with black circles (•) project needs close collaboration among veterinarians, pig producers and regional governments.

Future perspective
Based on the ORF5 sequence analysis of PRRSV, we systematically retrospect the history and prevalence of PRRSV in China from a molecular perspective. PRRSV diverged earlier in the origin, and the early isolated strain CH-1a and BJ-4 only share 91.4% homology. CH-1a belongs to sublineage 8.7 and BJ-4 is clustered into sublineage 5.1. BJ-4 is likely to be imported from the North America or maybe as a vaccine strain since it shares 99.6% and 99.8% identity with VR2332 and RespPRRS MLV, respectively. It is still not clear why sublineage 5.1 is relatively stable and shows limited epidemicity. Nevertheless, sublineage8.7 is pandemic and goes through such genetic diversity, especially the emergence of HP-PRRSV which is devastating to swine production in China. Although the vaccines derived from HP-PRRSV play a role in reducing the severe clinical signs and slowing the prevalence of PRRS, it is still far away from expectation. And the massive usage of vaccines also lead to some side effects including virulence reversion of attenuated live vaccines [49,73], markedly increase in virus recombination and variation [50,53,72,74].
Recently, lineage 3 and lineage 1 are also circulating in Chinese swine herds. Lineage 3 is mainly limited in south of China (Jiangxi, Henan, Jiangsu, Guangdong, Guangxi) [38,43]. Lineage 1 was firstly reported in Henan province in 2013 [45,79], and then this lineage transmitted quickly in northeast, southeast, central and east of China. They lead to pandemic of PRRSV in China [38,40,80,81]. The commercial vaccines seem not efficient since the dissemination of lineage 1 is under the background of massive vaccination. Several studies also show the commercial vaccines only provide limited or partial protection [38,68,69]. These facts prompt people to re-evaluate the strategies of prevention and control of PRRSV in China. And people become more cautious on massive vaccination. More and more farms try to improve management, biosecurity and environment instead of only over-dependent on vaccines.
Several studies showed the prevalence of PRRSV-1 in China, and up to 9 provinces and regions have been reported until now [38,[56][57][58]. All the Chinese circulating strains share some special common amino acids compared with Lylestad virus. Although the PRRSV-1 PRRSV is sporadic in China, we should pay more attention since there are no vaccines to be provided.
It has been 30 years since the first PRRS report in 1987 in America. Originally, we call it Mystery Swine Fig. 3 The amino acid alignment of ORF5 for PRRSV-1. The amino acid sequences were aligned by Clustal W method in MEGA 6.0. The signal sequence(1-32aa) and hypervariable domains (33-67aa, 89-110aa) were indicated by solid boxes. Potential N-glycosylation sites were shown in the red boxes. The hypervariable mutation sites were marked with dark spot (•) and the highly conserved mutation sites were indicated by asterisk (*) Disease (MSD) [3,7]. Thirty years later, although we have a deep understanding on clinical onset characteristics of PRRS, molecular epidemiology, virus proliferation, pathogenic mechanism, immune response and immune escape mechanism, PRRSV is still mysterious to us and we have not found a successful prevention and eradication strategy. There are still more questions need to be elucidated: i) What is the molecular mechanism of high variation and recombination frequency of PRRSV; ii) The broad spectrum neutralizing antibodies need to be analyzed further; iii) It remains to be elucidated that determination factors of virulence difference among different PRRSV strains; iv) How to look for a new vaccine development strategy which is safe, high efficient and have a broad spectrum protection to different PRRSV lineages. The clarification of these problems would provide great help and support for ultimate PRRSV control.

Conclusion
PRRSV is still one of the most important causative agents to the swine production worldwide and causes huge economic losses every year. In China, although several commercial attenuated live vaccines have been widely used, PRRS is still severe in the pig industry. Moreover, the genetic diversity and complexity of PRRSV were further increasing. In this review, we systematically analyzed the PRRSV prevalent history and genetic evolution in China using a molecular epidemiological perspective. Additionally, we summarized the effective strategies and discussed the current problems in the prevention and control of PRRS. Our analysis would provide valuable information and new insights for the PRRSV epidemic trends and control strategies in China.  Fig. 4 The alignment assay based on amino acid sequence of N protein for PRRSV-1. The sequence alignment was performed using Clustal W method. A2-12, B25-30, C40-46, D51-67 and D80-90 represent known epitopes in N protein; E50-58, E64-72, E105-113 and E113-121 indicate four porcine T -cell epitopes. The insertion site of amino acid was indicated by triangle (△). The highly conserved mutation sites in Chinese PRRSV-1 strains were marked with asterisk (*)