Continuous evolution of influenza A viruses of swine from 2013 to 2015 in Guangdong, China

Southern China is considered an important source of influenza virus pandemics because of the large, diverse viral reservoirs in poultry and swine. To examine the trend in influenza A virus of swine (IAV-S), an active surveillance program has been conducted from 2013 to 2015 in Guangdong, China. The phylogenetic analyses showed that the external genes of the isolates were assigned to the Eurasian avian-like swine (EA) H1N1 and/or human-like H3N2 lineages with multiple substitutions, indicating a notable genetic shift. Moreover, the internal genes derived from different origins (PB2, PB1, PA, NP: pdm/09 (pandemic influenza virus 2009)-origin, M: pdm/09- or EA-origin, NS: North American Triple Reassortant (TR)-origin have become the dominant backbone of IAV-S in southern China. According to the origins of the eight gene segments, the isolates can be categorized into five genotypes. The results of mice experiment showed that the YJ4 (genotype 1) and DG2 (genotype 4) are the most pathogenic to mice, and the viruses are observed in kidneys and brains, indicating the systemic infection. The alterations of the IAV-S gene composition supported the continued implementation of the intensive surveillance of IAV-S and the greater attention focused on potential shifts toward transmission to humans.


Introduction
Influenza A virus (IAV) belongs to the Orthomyxoviridae family and contains a genome composed of eight single-stranded RNA genomes. According to the antigenic properties of hemagglutinin (HA) and neuraminidase (NA), IAV can be subtyped into 16 HA and 9 NA types in aquatic birds, and 2 HA and 2 NA subtypes have been detected in bats [1,2]. Pigs play a pivotal role in the circulation and evolution of IAVs and are regarded as "mixing vessels" for the generation of novel reassortant viruses [3]. Three main subtypes (H1H1, H1N2 and H3N2) have been detected in pigs [4]. The simple infection of pigs with influenza A virus presents mild clinical signs in the field and under experimental conditions [5]. However, influenza A virus of swine (IAV-S) can impair the host immune system in a variety of ways, subsequently suppressing the immune response to other pathogens [6]. a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 Multiple lineages of IAV-S have emerged and became established in pigs. Classical swine influenza virus (CS) and Eurasian avian-like swine influenza virus (EA) have been isolated in pigs in the 1930s and 1970s, respectively [7,8]. Additionally, cases of pig infection by H1N1 and H3N2 human seasonal (HS) influenza viruses and avian influenza viruses (AIVs) have been sporadically reported worldwide [9][10][11], increasing the IAV-S gene pool and giving rise to numerous reassortant viruses such as the North American triple reassortant (TR) influenza virus [12,13]. In April 2009, a reassortant variant, pandemic influenza virus 2009 (pdm/09), sustained human-to-human transmission and caused the first pandemic in the 21 st century. This variant possessed TR-origin PB2, PB1, PA, NP, HA (H1), and NS genes and EA-origin NA (N1) and M genes [14].
After the pandemic, the 2009 pandemic virus was repeatedly transmitted to pigs in many countries [15][16][17] and reassorted with endemic viruses [18,19], which increased the genetic diversity of IAV-S circulating in pig herds. Phylogenetic and genetic analyses revealed that pdm/09-origin internal genes became established and evolved variants [20]. To research the trend in IAV-S in pig herds, an active surveillance program has been conducted in pigs from 2013 to 2015 in Guangdong, China. The origin, gene diversity and genetic markers of the isolates in this program were estimated through phylogenetic and molecular analysis. Mouse is one of the most mature mammal models for influenza virus infection, which has been used for evaluating the virulence of avian-and mammal-origin influenza viruses [21]. The pathogenicity of the strains to mice has been assessed in this study.

Sample collection and virus isolation
An active surveillance program of IAV-S in pig herds has been performed from 2013 to 2015. The nasal swabs utilized in this program were collected from commercial farms and abattoirs in Guangzhou, Foshan, Dongguan, Yangjiang, and Jiangmen, among other locations in China. The samples were transported to the laboratory at 4˚C with ice bags and preserved at −80˚C. The swabs were inoculated into 9-11 days old specific-pathogen-free (SPF) embryonated eggs and/or Madin-Darby Canine Kidney(MDCK) cell cultures in Eagle's minimum essential medium with trypsin (2 μg/ml). The embryonated eggs incubated at 37˚C for 72h and the cell culture tubes were incubated for up to 7 days. Virus isolates were passaged and identified using hemagglutination test with chicken red blood cells. The positive allantoic fluid and/or supernatant were collected were used to extract the viral RNA for whole genome sequencing.

RT-PCR and sequencing
Reverse transcription-polymerase chain reaction (RT-PCR) was performed to amplify the viral RNA for sequencing and phylogenetic analysis. Viral RNA was extracted using TRIzol TM Reagent (Thermo Fisher Scientific, USA) and amplified in one reaction with reverse transcriptase M-MLV (RNase H-), recombinant RNase inhibitor (TaKaRa, Japan) and the oligonucleotide universal primer 5'-AGCAAAAGCAGG-3'. The genome was amplified by PCR using the Platinum TM Pfx DNA polymerase kit (Thermo Fisher Scientific, USA) with a series of primers [22]. The products were then purified using an agarose gel DNA purification kit (TaKaRa, Japan). To sequence the genes, the products were cloned into the pJET 1.2 blunt-end cloning vector (Thermo Fisher Scientific, USA) and transformed into competent DH5α (Tiangen Biotech Beijing Co., Ltd.). At least three clones per gene were sequenced by Sanger sequencing (Thermo Fisher Scientific, USA). Molecular analysis. The sequences were compiled and edited using SeqMan and Editseq (DNASTAR, Lasergene. v7.1). BLAST analysis was carried out using NCBI. The sequences of the isolates were uploaded to the GISAID Epiflu database. The consensus sequences referenced in this study were downloaded from GenBank and the GISAID Epiflu database, purged for redundant sequences (>99.9% similarity), and aligned using MAFFT [23] and further edited using MEGA version 5 [24]. The number of amino acid residues of HA (H1 and H3 numbering after removal of the signal peptide) and other proteins (open reading frame, ORF) was based on the downloaded sequences. Potential asparagine-linked glycosylation sites (PNGs) in the HA proteins were predicted by examining the N-X-S/T motifs using the NetNGlyc 1.0 Server, and X could be any amino acid except proline (http://www.cbs.dtu.dk/services/NetNGlyc/). Phylogenetic analysis. A total of 872 sequences, 70-100 sequences per gene segment, were processed for phylogenetic analysis. These sequences included several endemic strains, mainly in mainland China and Hong Kong. The analysis were based on the following sequences: H1 HA, nt 33 The best-fit nucleotide substitution model was predicted using ModelTest [25]. The time of the most recent common ancestor (tMRCA) was determined using the Bayesian Markov Chain Monte Carlo (MCMC) method in BEAST v1.8.3 [26,27], under the HKY substitution model with gamma distributed rates amongst sites (HKY+γ), using a Bayesian skyline coalescent tree [28] prior to selecting a relaxed uncorrelated lognormal model [29]. MCMC chains were run for 5~10×10 7 generations for each dataset. The tMRCA was estimated using the collection date adjusted to year format in this study. Tracer v1.6 was used for evaluating the MCMC output results by caculating the the ESS (effective sample size) [30]. The trees were generated using TreeAnnotator v1.8.3 (http://beast.bio.ed.ac.uk/TreeAnnotator) with 10% burn-in and 95% highest probability density (HPD) values. The trees were visualized and redacted using FigTree v1.4 (http://beast.bio.ed.ac.uk/figtree).

Mice experiment
6-week-old female BALB/c mice (Guangdong Medical Laboratory Animal Center, Guangdong, China) were used in this study. Thirteen mice per group were anaesthetized with dry CO 2 and inoculated intranasally with 10 6.0 EID 50 viruses in a volume of 50μL. Three mice were euthanized on day 3 post-inoculation with dry CO 2 , and the virus titrations of the lungs, kidneys, brains, and turbinates were tested in chicken embryos eggs. The samples were homogenized in 1 mL per 1g tissue of PBS supplemented with penicillin (1,000 U/mL) and streptomycin (1,000 U/mL) and were centrifuged at 4,000 ×g to isolate supernatant fluids. The supernatants were serially diluted and inoculated into the allantoic cavity of 9-10 days old chicken embryonated eggs (100 μL per egg). The eggs were incubated at 37˚C for 48 h. The virus titers were detected by the hemagglutina tion test and calculated using the method of Reed and Muench method. The weight loss and survival of the remaining ten mice were recorded daily for a total of 14 days.
The animal experiments were carried out in ABSL-3 facilities in compliance with the biosafety committee of South China Agriculture University (SCAU) protocols. All animal experiments were approved by the Institutional Animal Care and Use Committee at SCAU and carried out in accordance with the approved guidelines. Animals determined to have lost > 25% of body weight or to be moribund (as indicated by increased respiratory rate and inability to ambulate) were euthanized.

Nasal swab collection and virus isolation
From June 2013 to December 2015, a total of 4,056 swine nasal swabs were collected randomly from 35 commercial farms and 5 abattoirs in Guangdong, China. The swabs were collected from Guangzhou (n = 1276), Foshan (n = 1180), Dongguan (n = 225), Yangjiang (n = 458), Jiangmen (n = 278) and Shenzhen (n = 639). Fourteen viruses were isolated using chicken embryos and/or MDCK cells and were identified using the hemagglutination test and were confirmed by RT-PCR and genomic sequencing. The sequences of the isolates have been uploaded to the GISAID Epiflu database, and the available IDs were listed in Table 1. According to the results of BLAST, the isolates can be categorized into three subtypes, including eight H1N1 stains, five H1N2 strains and one H3N2 strain.

Analysis of H1 HA
Among the 14 isolates identified in this survey, 13 HA genes belonged to the H1 subtype (Table 1), shared 94.7-99.9% nucleotide sequence identity and were assigned to the EA lineage ( Fig 1A,  . Among these sublineages, multiple substitutions were observed in the antigenic sites, including E155G (H1 numbering) in Sa, N184T, A190T and R193Q in Sb, and K142N in Ca. Furthermore, many substitutions were identified in the HA1 protein, such as R102K, E172K, Q208K (Fig 1A), and no studies have shown whether these substitutions affect the virus characteristics.
The H1HA genes in this study were all assigned to the Sw/HK/72-like sublineage and showed several unique substitutions in the antigenic sites (Fig 2), such as G202E in EP and YJ28, N194H in BRT15, and T72A in FSC17. At the receptor binding pocket, amino acid residues 225G were conserved in all H1 subtype isolates in this program and predicted to have no affinity for mammalian cell-surface receptors [31]. The N-linked glycosylation is essential for protein folding and maturation. Changes in the number or location of NLG sites in the HA protein can affect the biological activity of IAV. Five potential Potential asparagine-linked

Analysis of the FS4 HA
The remaining FS4 HA gene belonged to the H3 subtype. To assess its origin, 13 major human-origin H3 HA genetic clusters that had been circulating since 1968 were estimated [32] and represented in the H3 HA phylogenetic tree (Fig 1B,  Hong Kong/1071/2012)-like. Several substitutions at antigenic sites were observed between Hu/Perth/16-like and Sw/HK/1071-like, including T135K (H3 numbering), S137A, I140K and R143G in site A, D188Y, K189S, F193S in site B, E50G in Site C and R201I in site D ( Fig 1B). As shown in the phylogenetic tree, the FS4 HA gene was assigned to the Sw/HK/1071-like sublineage (Fig 1B, red), and substitutions such as G275D in site C and G218V in site D were found in the HA1 protein (Fig 3). With respect to receptor binding affinity, 226I and 228S in H3 HA at the receptor binding pocket were fairly conserved in FS4 and are usually detected in human viruses. Seven PNGs (22NGT24, 38NAT40, 63NCT65, 122NES124, 126NWT128, 246NST248 and 285NGS287) were present in the Hu/Perth/16-like and Sw/HK/1071-like viruses. The T135K and N165K substitutions, which eliminated two potential glycosylation sites at position 133NGT135 and 165NVT167, were observed in Sw/HK/1071-like viruses. In this study, a new PNG was observed at position 45 in FS4 for the S45N substitution (Fig 3).

Phylogenetic and molecular analysis of NA genes
Eight NA genes shared 96.7-99.9% sequence identity and belonged to the EA lineage (Fig 4A,  red). Similar to the EA-origin HA genes, the EA-origin NA genes differentiated from the AIV lineage and are prevalent in pig herds (Fig 4A). The EA-origin NA genes isolated in China since 2000 exhibited I40T (N1 Numbering), T48I, N221K, N386K substitutions and differentiated into two main sublineages, termed Sw/HK/2481-like and Sw/ZJ/1 (A/swine/Zhejiang/1/ 2007)-like. As shown in the phylogenetic tree, all N1 NA genes in this study were assigned to the Sw/ZJ/1-like sublineage. The DL2 and CH8 NA genes were distinct from the rest of the isolates and showed several unique substitutions, including S46P, I255V, E287K, I289T and P340S. Furthermore, multiple substitutions were observed in the N1 NA genes of the isolates, such as the N42D and R130K substitutions in CH8, S95N, and K217R in YJ10, I17T, S70N, T332I and D316G in BRT15, and N307D, T332V, R382G and R430Q in FSC17 (Table 2).
The phylogeny of the NS genes differed from that of the rest of the internal genes. In this study, the NS genes shared 96.9-100.0% nucleotide sequence identity, and all belonged to the TR lineage (Fig 5F, red). As shown in the phylogenetic tree, the TR-origin NS genes were divided into two distinct sublineages, Hu/Korea/CY02 (A/Korea/CY02/2002)-like and Sw/ HK/1562 (A/swine/Hong-Kong/1562/2005)-like, which were established in Korea and southern China, respectively. In this study, all NS genes were assigned to the Sw/HK/1562-like sublineage and displayed T18V, N48S, Q109H and P216S in NS1 and M31I and T48S in NEP (Table 2). Furthermore, several unique substitutions were observed in the NS genes of the isolates, such as L15I, V60I, H169I in NS1 and V32I, E67K, K88R in NEP of the novel isolates ( Table 2).

Drug resistance analysis
Oseltamivir is an important antiviral drug used in the treatment of infectious influenza viruses [34]. E119G/V, D199G, I223K/R/V, S247N, H275Y and N295S (N1 numbering) in N1NA, E119D/V, Q136K, I222L, R292K, N294S, and deletion of 245-248 (N2 numbering) in N2 NA have been detected in viruses associated with oseltamivir resistance (data reported in WHO). None of the above substitutions were detected in the N1 and N2 proteins of the isolates in this study, indicating their susceptibility to oseltamivir. Adamantanes, a group of antiviral drugs that inhibit the function of the viral M2 proton channel, are used for the treatment of human influenza. In our study, K, A and G were fairly maintained at positions 26, 30 and 34, respectively. However, N was found at position 31 of M2 protein in the isolates, and the V27A substitution was observed in DG1 and DG2; these substitutions confer resistance to amantadine [35].

Classification of genotypes of the novel viruses
To describe the genetic diversity of the novel isolates, genotypes were defined on the basis of the original lineage of the eight gene segments. Phylogenetic analysis revealed that the genomes of these novel isolates were reassortant from three or four lineage origins, and divided into five distinct genotypes. All the isolates possessed the pdm/09-origin PB2, PB1, PA and NP gene segments, and TR-origin NS gene segment. Eight out of 14 isolates in this study possess the H1N1 EA-origin surface genes, and pdm/09-origin M genes (n 2) or the EA-origin M genes (n 6). Five H1N2 viruses isolated in this study possess the EA-origin HA (H1) gene and N2 human-origin NA gene (N2), with the pdm/09-origin M genes (n 3) or the EA-origin M genes (n 2). Only one H3N2 subtype influenza virus was isolated in this surveillance, possessing H3N2 human-originsurface gene segments, pdm/09-origin PB2, PB1, PA, NP and M gene segments, and TR-origin NS gene (Table 3).

Pathogenicity in mice
Based on the results of phylogenetic analysis, we selected YJ4, CH8, YJ28, DG2 and FS4 from five genotypes to investigate the virulence, respectively. The results showed that mice infected The evolution of IAV-S with Genotype 3 (CH8) lost 21% of their body weight at most. YJ4 (Genotype 1) and DG2 (Genotype 4) could cause more than 25% weight loss and killed 100% and 50% of the mice, respectively (Fig 6). The results showed that the YJ4 (Genotype 1) and DG2 (Genotype 4) were considered to be the most pathogenic to mice. Organs from each mouse were collected to evaluate virus replication titration on day 3 post-inoculation. Virus was detected in all lungs and turbinates of the mice inoculated with 5 genotypic viruses. The titrations of the lungs inoculated with genotype 4 (DG2), 3 (YJ28), 1 (YJ4) and 2 (CH8) were significantly higher than genotype 5 (p<0.01), and the titrations of turbinates inoculated with genotype 4 and genotype 1 were signicantly higher than the remaining three genotypic groups. Furthermore, viruses were detected in the kidneys of three mice inoculated with YJ4, and in two mice inoculated with CH8 and DG2. The virus were also deteced in the brains of two mice inoculated with YJ4 and DG2, and one mice inoculated with YJ28. The mice inoculated with genotype 5 (FS4) present slight decrease of body weight, and the viruses can be detected only in the lungs and turbinates, not in the kidneys and brains.

Discussion
Due to unique geographical and environmental factors, southern China is considered an important reservoir of influenza virus. In the first decade of the 21st century, Multiple lineages of IAVs-S have emerged and become established in pigs in southern China: classical swine H1N1 (CS), European avian-like H1N1 (EA) and triple-reassortant H1N2 viruses (TRIG). In 2001, the first case of infection with the EA-origin virus in pigs in Asia was reported in Hong Kong, and EA-origin viruses have since formed a stable phyletic clade in China [36]. In addition, TR-origin viruses have been regularly isolated from pigs in China since 2002 [37]. Since the pdm/09-origin virus outbreaks in humans, this virus has been repeatedly transmitted in pig herds [38][39][40]. Reassortant variants with pdm/09-origin gene segments and endemic genes were subsequently found in Asia [41][42][43]. The swine-origin H1N1 viruses were found reassorting with the H3N2 canine influenza viruses circulate endemically in Asian dogs [44]. Furthermore, the novel triple EA H1N1 and Human Like H3N2 reassortants, containing the CS H1N1 NS genes and the remaining five or four genes originating from H1N1/2009 pandemic, may have become established in pig herds in Southern China [40,45]. Notably, the reassortant EA H1N1 viruses with EA-origin M gene, pdm/09-origin internal genes and CS-origin NS gene have been reported in human infections in Hunan, China [46].
In this study, phylogenetic analyses assigned the external genes of the novel isolates to the EA-origin H1N1 and/or human-origin H3N2, and the isolates were categorized into H1N1, H1N2 and H3N2. The regular isolation of H1N1 and H1N2 viruses demonstrates their continuing presence in pig herds, which means that the EA-origin variants reassorted with pdm/ The evolution of IAV-S 09-origin internal genes have become the major swine influenza lineage prevailing in southern China [41]. In addition, one H3N2 reassortant virus containing the pdm/09-origin internal genes was isolated in this study, indicating that the genotype of internal genes has reassorted with other endemic influenza viruses, indicating the pdm/09-origin internal genes affected the prevalence of H3N2 viruses in pig herds. It has been reported that the pdm/09-origin internal genes has become predominant in pig herds in Guangdong [41]. All novel isolates possessed the pdm/09-origin PB2, PB1, PA and The evolution of IAV-S NP gene segments, indicating that the pdm/09-origin PB2, PB1, PA and NP genes had replaced the EA-or TR-origin gene segments. By contrast, the M genes were assigned to the pdm/09 lineage and EA lineage, partly due to the critical role of the pdm/09-origin M gene segment for high transmission efficiency in mammalian hosts [47]. The pdm/09-origin internal genes has already altered the IAV-S gene pool and influenced the prevalence of viruses in southern China. Moreover, all NS genes of the isolates belonged to the TR lineage and were assigned to the polyphyletic lineage formed in pigs in southern China, distinct from the TRorigin viruses found in the USA and Korea.
According to the phylogenetic analysis, the external genes of the isolates were assigned to the Eurasian avian-like swine (EA) H1N1 and/or human-like H3N2 lineages. On the other hand, the internal genes were assigned to pdm/09-like lineage(PB2, PB1, PA, NP, M), EA lineage(M) and/or TR lineage(NS). Five isolates in each genotype with the most representative functional mutations were selected for pathogenicity test in mice. The pathogenicity of genotype 1 and 2 were much higner than that of the other genotypes and could replicate systemically in mice. Especially, the genotype 1 virus could kill mice with 10 6.0 EID 50 viruses. Therefore, recombination of specific lineage of HA and NA genes could enhance the virulence in mice.
The molecular characteristics of amino acid residues at antigenic sites, host adaptation, virulence markers, and drug resistance were summarized. Amino acid residues at the receptor binding pocket of HA1 223Q (H3 sequence 226I) and 225G (H3 sequence 228S) were conserved in all isolates in this program and were predicted to have affinity for mammalian cellsurface receptors. Gly at position 222 in the HA protein has been reported in some patients with severe or fatal H1N1/2009 infection, and the D222G substitution in pdm/09-origin HA could alter receptor specificity and increase virulence in mice [48]. The E222G substitution was present in the HA genes of YJ4, YJ28 and DG2, but its significance in the EA-origin HA gene is unclear.
The functional sites in the pdm/09-origin internal genes were relatively conserved after 2009. Similar to pdm/09 in humans, 627E and 701D were present in the swine-origin pdm/ 09-origin PB2 gene segment, which is regarded as a determinant of host range and has been detected in the majority of avian viruses [49,50]. These substitutions have been reported to have no major impact on virus replication or pathogenesis in mammalian models [51,52]. By contrast, the T271A, T588I, G589R and Q590S substitutions were discovered in TR-and pdm/ 09-origin PB2, likely compensating for the lack of 627K and 701N in the pdm/09-origin PB2 protein [53]. Interestingly, N was found at position 701 in the FNPA PB2 gene and could contribute to the adaptation and pathogenesis of the virus to the mammalian hosts combined with other substitutions in this novel isolate.
The latest surveillance data showed that the pdm/09-origin variants emerged in central Mexico as a result of inter-hemispheric swine movements, which is closely related to longdistance pig transportation [54]. The mice experiment demonstrated that the novel reassortant viruses showed different pathogenicity, indicating a risk for the new potential pandemic. The alterations of IAV-S gene composition combined with the complex epidemic situation underlines the importance of continued swine surveillance in China to maintain public health. This is the active surveillance to understand the evolution trend of IAV-S. However, there are two potential limitations that should be considered. First, the lack of detailed background information about the isolates could introduce interpretation bias into the results analysis. Second, the drug resistance, the receptor binding capacity, and antigenic characteristics of isolates have been predicted in the amino acid analysis, but the further experiments need be performed for verification.