Genotype Diversity of H9N2 Viruses Isolated from Wild Birds and Chickens in Hunan Province, China

Three H9N2 avian influenza viruses were isolated from the Dongting Lake wetland, among which one was from fresh egret feces, the other two were from chicken cloacal swabs in poultry markets. Phylogenetic analyses suggested that eight genes of the egret-derived H9N2 virus might come from Korean-like or American-like lineages. The two poultry-derived H9N2 viruses were reassortants between the CK/BJ/94-like and G1-like viruses. Except the PB1 genes (90.6%), the nucleotide sequence of other internal genes of the two viruses exhibited high homology (>95%). In addition, they also exhibited high homology (96–98.3%) with some genes of the H7N9 virus that caused an epidemic in China in 2013. Nucleotide sequence of the poultry-derived and egret-derived H9N2 viruses shared low homology. Infection studies showed that the egret-derived H9N2 virus was non-pathogenic to both mice and chickens, and the virus was unable to infect chickens even through 8 passages continuously in the lung. On the other hand, the chickens infected by poultry-derived viruses showed obvious clinical symptoms and even died; the infected mice showed no noticeable clinical symptoms and weight loss, but viruses could be detected in their lungs. In conclusion, for the egret-derived H9N2 virus, it would take a long adaptation process to achieve cross-species transmission in poultry and mammals. H9N2 viruses isolated at different times from the same host species in the same geographical region presented different evolutionary status, and virus isolated from different hosts in the same geographical region exhibited genetic diversity. Therefore, it is important to continue the H9N2 virus surveillance for understanding their evolutionary trends so as to provide guidance for disease control and prevention.


Introduction
The H9N2 avian influenza virus (AIV) was first isolated from North American waterfowls in 1966 and it has disseminated worldwide since then [1]. Many studies have shown the presence of the H9N2 viruses in both waterfowls and poultry and two-way transmission between them [2,3]. Guan et al. found that the internal genes of the H5N1 AIV that caused human infections in Hongkong in 1997 originated from H9N2 subtype AIV [4]. The first human case infected by H9N2 AIV occurred in 1999 [5]. Studies have demonstrated that the characteristics of the H9N2 virus might change after genetic reassortment through multiple hosts. Therefore, it is important to strengthen the monitoring and tracking studies of H9N2 AIV.
Epidemiology and phylogenetic studies found that the viruses that caused H9N2 AIV outbreak in Hong Kong live poultry market could be divided into three distinct lineages: A/Chicken/ Beijing/1/94 (Ck/BJ/94-like), Quail/Hong Kong/G1/97 (G1like) and Duck/Hong Kong/Y439/97 (Y439-like or Korean-like) [6,7]. In mainland China, H9N2 AIV was first isolated from chickens in Guangdong province in 1994 [8]. The Ck/BJ/94-like and G1-like lineages have been prevalent in domestic poultry, such as quails and chicken, in southern China since the early 1990s [4,7,9]. The Y439-like lineage was isolated from ducks in Hong Kong in 1997 and was also isolated from sick chickens in South Korea in 1996. In mainland China, monitoring data in recent years indicated that novel genotypes of H9N2 AIV had kept emerging [10,24]. More importantly, six gene segments of H7N9 AIV that caused human infections in 2013 came from H9N2 viruses, which again demonstrated that the H9N2 AIV was a potential threat to humans [11].
A total of three H9N2 subtype AIV strains were isolated during 2010-2012. Of which one was isolated from fresh egret feces collected in Dongting Lake Egret Nature Reserve, the other two were isolated from chicken cloacal swabs collected in poultry markets. We carried out whole-genome sequencing for the three viruses, constructed phylogenetic trees and investigated the sources of the virus genomes. Pathogenicity tests of these H9N2 viruses were conducted in mice and SPF white Leghorn chickens. The study revealed the evolutionary status and likelihood of crossspecies transmission of H9N2 AIVs in wide birds and poultry in Dongting Lake wetland, and contributed data for molecular epidemiology of AIV.

Ethics Statement
All animal experiments were approved by the Institutional Animal Care and Use Committee (IACUC) of Laboratory Animal, Wuhan Institute of Virology, Chinese Academy of Sciences (Approval numbers: WIVA04201202 for mice and WIVA05201001 for chickens). This study was carried out in accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals. All procedures were performed under anesthesia. Animals were closely monitored and observed for development of disease once daily, and all efforts were made to minimize suffering.
Fresh migratory bird feces were collected at Dongting Lake Egret Nature Reserve (longitude: 112.819496, latitude: 29.514841) and no specific permissions were required for these locations. The cloacal swabs collected in poultry markets (longitude: 112.996925, latitude: 29.441994) and no specific permissions were required for these activities. All field studies did not involve endangered or protected species.

Sample collection
Sterile cotton swabs were used to collect fresh migratory bird feces and poultry cloacal samples. The swabs were placed into viral transport media. All the samples were placed in a hand-held portable 4uC refrigerator, transported to the laboratory within 24 hours, and frozen in 280uC immediately for future use.

Virus isolation
The swabs frozen in the 280uC freezer were transferred to a 4uC refrigerator for thawing. Then the sample was spun at 6000 rpm for 10 min at 4uC, and the supernatant was inoculated into 10-day-old SPF embryonated chicken eggs. The embryonated chicken eggs were cultured at 37uC for 48-72 hours. The embryos that died within the first 24 hours were discarded, for the rest live or dead embryos, allantoic fluid was collected and spun at 3000 rpm for 10 min at 4uC and then tested for hemagglutination. The allantoic fluid positive for hemagglutination was divided into aliquots and frozen at 280uC for further tests.

RNA extraction and nucleotide sequencing
RNA extraction and reverse transcription of virus RNA were performed according to the instructions of the RNA extraction kit (Takara) and mouse reverse transcription kit (MLV, Promega), respectively. Influenza virus gene fragments were amplified by PCR using a Phusion High-Fidelity PCR Kit (New England Biolabs). The primers used to amplify the eight genes were universal primers designed by Hoffman et al. [12]. The recovered gene fragments were cloned into pGEM-T vector and sequenced by the dideoxy method with an ABI 3730 DNA sequencer (Applied Biosystems). The complete genome sequences of influenza virus were edited and aligned with BioEdit version 7.0.5.2.

Phylogenetic analysis
Phylogenetic analysis was based on nucleotides 94-1,556 (1,463

Chicken study
Two routes of virus infection, intravenous (i.v) and intranasal (i.n) injection were used. For each route, 6-week-old SPF white Leghorn chickens were divided into two groups, eight each group. In i.v groups, virus was diluted 10 fold and then injected intravenously to the chickens at a dose of 0.2 ml each. In i.n groups, each chicken was anesthetized via intramuscular injection of ketamine (20 mg/kg) and xylazine (1 mg/kg) and given a dose of 10 6 EID 50 virus in 0.1 ml. Oropharyngeal and cloacal swabs were collected on day 1, 3, 5, 7 p.i.. Viral titration was conducted in 10-day-old SPF embryonated chicken eggs. The chickens were observed for 21 days post-inoculation (p.i) and their symptoms and survival were recorded. The symptoms, virus titration of lungs and death were as results of virus infection.
For lung passage experiment in SPF white Leghorn chickens, eight chickens were infected intranasally with each virus in a volume of 0.1 ml of 10 6 EID 50 virus after being anaesthetized. Three days after infection, the chickens were euthanized with ketamine and xylazine, and the lungs were taken out and homogenized with a total of 2 ml of PBS containing 0.1% BSA. The tissue homogenate was used for virus titration after removing cellular debris by centrifugation. Virus titration was performed as above. The remaining five chickens were observed for symptoms and survival for 21 days and post-recovery sera were collected for hemagglutination inhibition (HI) test. The same procedure was performed from Passage 0 (P0) to 8 (P8).

Mouse study
Specific-pathogen-free female BALB/c mice, aged 6 to 8 weeks old, were purchased from the Center for Disease Control and Prevention in Hubei Province, China. Mice were bred in the Animal Resource Center at the Wuhan Institute of Virology, Chinese Academy of Sciences. All mice were maintained in specific-pathogen-free conditions. The BALB/c mice were divided into three groups, including two experimental groups and one control group, twenty-one each group. For the experimental groups, each mouse was anaesthetized with pentobarbital sodium and infected intranasally with 10 6 EID 50 virus in 50 ml. The mice (n = 12) were observed for 21 days p.i. and their symptoms, mortality and weight loss were recorded. On day 3, 5 and 7 p.i., three mice per group were euthanized, and the lungs, brains, spleens and kidneys were taken out and homogenized with a total of 2 ml of PBS containing 0.1% BSA.

Histopathological analysis
For histopathological analysis, lungs of SPF chicken were removed immediately following euthanasia, and fixed with 10% formalin at 4uC. Subsequently, lungs were stained with Hematoxylin and Eosin (H&E), and examined for pathological changes corresponding to infection. Images were obtained on an Nikon80i light microscope at 400-fold original magnification.

Antigenic analysis
Sera were harvested from the four groups of inoculated chickens on day 21 p.i. for confirmation of seroconversion and antigenic analysis by HI assays with 1.0% chicken erythrocytes according to relevant recommendation of OIE.

Nucleotide sequence accession numbers
The nucleotide sequences obtained in this study are available from GenBank under accession numbers JX437684 to JX437691, KF714772-KF714787.

Statistics
The results of the test groups were evaluated by Student's t-test; if p-value is less than 0.05, the difference was considered significant. The survival rates of the mice in the test and control groups were compared by using Fisher's exact test.

Virus isolation and homology analysis
From October 2010 to October 2012, three H9N2 subtype AIVs were isolated from Dongting Lake wetland. One virus was isolated from fresh egret fececs collected in Dongting Lake Egret Nature Reserve and named A/Egret/Hunan/1/2012 (H9N2). The other two viruses were isolated from cloacal swabs collected in poultry markets and named A/Chicken/Hunan/12/2011(H9N2) and A/Chicken/Hunan/1/2012 (H9N2), respectively.
The HA and NA genes of the A/Egret/Hunan/1/2012 virus had about 83% homology with those of A/Chicken/Hunan/12/ 2011 and A/Chicken/Hunan/1/2012 viruses and homology of the NS gene was even lower (70%). The genes of two poultryderived viruses shared high homology except the PB1 gene. Nucleotide homology of the PB1 gene between the two poultry viruses was only 90.6%, lower than the PB1 gene homology between A/Chicken/Hunan/1/2012 and A/Egret/Hunan/1/ 2012 (96.8%). Some genes of the two poultry viruses also had high homology with those of the H7N9 virus isolated from pigeons in Shanghai in 2013 (e.g., M gene: 98.3%) ( Table 2).

Phylogenetic analysis
Phylogenetic analysis indicated that the HA and NA genes of the two poultry-derived viruses belonged to the CK/Bei/94-like lineage, which had two subgroups represented by A/Duck/ HongKong/Y280/97 and A/Chicken/Shanghai/F/98, respectively. The HA genes of the two poultry-derived viruses belonged to the subgroup represented by A/Duck/HongKong/Y280/97 (Figure1a, b). Except PB1 and PB2 genes, all the other gene segments were in either the CK/Bei/94 lineage or the G1 lineage (Figure1c-h, Table 3). The PB2 genes of these two H9N2 viruses shared high homology but they were in the DK1-like lineage. The PB1 genes of the two chicken H9N2 viruses were in different evolutional lineages; A/Chicken/Hunan/12/2011 virus belonged to the CK/Bei/94-like lineage, while A/Chicken/Hunan/1/2012 virus belonged to the Korean-like lineage. For A/Egret/Hunan/ 1/2012 virus, its NS and PA genes belonged to the American-like lineage and the other genes originated from the Korean-like influenza viruses (Table 3).

Molecular characterization
Amino acid sequence analysis revealed the connecting peptide of HA was ASDR/G in A/Egret/Hunan/1/2012 virus and RSSR/G in the two chicken isolates ( Table 4). The RSSR/G motif was consistent with the sequence of poultry H9N2 viruses prevalent in mainland China. In addition, each of the three sequences had one or two basic amino acids, which showed a feature of low pathogenic viruses [9,13]. The three isolates shared the same amino acid sequences at their HA receptor-binding site, having leucine (L) at position 226 (H3 numbering) and glycine (G) at position 228, which indicated that their HA proteins favor binding with a-2,6-linked sialic acid receptors and the possibility of the three isolates to infect humans [14]. The analysis of surface glycoproteins potential glycosylation sites revealed that the HA protein of A/Egret/Hunan/1/2012 and A/Chicken/Hunan/1/ 2012 viruses had eight potential glycosylation sites (positions: 29, 82, 141, 218, 298, 305, 492, 551), which were the same as most H9N2 viruses [15], whereas the HA protein of A/Chicken/ Hunan/12/2011 virus had an additional glycosylation site at position 331. NA protein of A/Egret/Hunan/1/2012 virus did not have the 3-amino acid deletion at positions 63-65 in the stem region while the NA proteins of the other two isolates had the 3amino acid deletion at positions 63-65, which leaded the deletion of one potential glycosylation site. It has been reported that this amino acid deletion might be necessary for virus adaptation from wild birds to poultry (chicken) [16]. The NA proteins of the three viruses also differed in glycosylation sites. The NA protein of A/ Egret/Hunan/1/2012 virus had seven potential glycosylation sites  [17]. In contrast, the two chicken isolates both had the S31N amino acid substitution, suggesting that they might have gained resistance to M2 ion channel inhibitors. The D92E substitution of NS1 protein could enhance virus resistance, but it was not present in any of the three isolates [18].
For the polymerase complex, no amino acid substitution was found in sites related to host specificity and viral replication (e.g., E627K and D701N in PB2, L356R in PA), indicating that these three isolates were low pathogenic to mammal and avian hosts [18][19][20][21]. The PB1-F2 protein of A/Chicken/Hunan/12/2011 virus had the N66S substitution, which indicated this isolate might have enhanced virulence, while the mutation was not found in the other two isolates.   Table 5). Chicken lung tissue pathology revealed obvious pulmonary inflammatory responses on day 3 and 7 p.i. in chickens infected with the poultry-derived H9N2 virus. Partial necrosis appeared in lung tissues on day 3 and 7 p.i., more severe on day 3. The results indicated that poultry-derived H9N2 virus could not only replicate effectively in chickens, but also cause cellular inflammatory response and lung tissue injury. In contrast, the egret-derived H9N2 virus was less pathogenic to chickens than chicken-derived virus ( Figure S1). The histopathology findings were consistent with the results of lung virus titer.

Replication of H9N2 viruses in chickens
In addition, A/Egret/Hunan/1/2012 virus was 8 passaged continuously in the lungs of SPF white Leghorn chickens. No chicken showed overt clinical symptoms (such as hoarse voice and listlessness) and they kept normal food and water intake. The P0 to P8 chicken lung homogenate were inoculated into SPF eggs and cultured for 48-72 hours in a 37uC incubator and the allantoic fluid of chicken embryos were performed hemagglutination test. The virus was detected only in the P0 sample, no virus were detected in lungs of egret-derived H9N2 virus-infected at 3d p.i. and no antibody was detected at 21d p.i. for the rest 8 passages.
All chickens infected with A/Chicken/Hunan/12/2011 virus had seroconversion on day 21 p.i with very high HI titer. In contrast, none of the chickens inoculated with A/Egret/Hunan/ 1/2012 virus, either via i.v or i.n, showed seroconversion (positive if HI.16).

Replication of H9N2 viruses in mice
A/Egret/Hunan/1/2012 and A/Chicken/Hunan/12/2011 viruses were also used for mice infection experiments. None of the mice showed overt symptoms and all mice survived. The egretderived isolate group showed slight weight loss, whereas the poultry-derived isolate group showed symptoms such as slight piloerection and trembling and more weight loss as compared to the egret-derived isolate group. Their weight started to rebound from day 7 p.i. (Figure 2).
Virus detection was performed on lung, brain, spleen and kidney tissues taken on day 3, 5 and 7 p.i.. The results indicated only the poultry-derived virus could replicate in the lungs of mice, but the viral load was low and the result was positive only on day 3 and 5 p.i.. No virus was detected in the other organs. The egretderived virus did not replicate in the lungs of mice and virus detection was negative for the other organs (Table 6).

Antigenic analysis
Antisera were collected on day 21 post-infection. The crossreactivity of the three viruses was investigated by HI assay. The antisera to A/Chicken/Hunan/12/2011 virus could cross-react well with A/Chicken/Hunan/1/2012 and A/Chicken/Hunan/ 12/2011 viruses ( Table 7). The HI titer for antisera to A/Egret/ Hunan/1/2012 virus reacting with the antigen virus was 16, suggesting no seroconversion following A/Egret/Hunan/1/2012 virus infection. Therefore, its antisera did not cross-react with the two chicken isolates. High HI titer was found when antisera were Table 2. Homology (%) of nucleotide sequences of eight genes of virus from chickens and relevant sequences of birds or H7N9.   Table 3. Gene constellation of H9N2 viruses isolated in this study.

Discussion
Last several years, human infections of H9N2 subtype avian influenza viruses have occurred, which have increasingly gained more attention [15,22,23]. Dongting Lake Wetland is a very important migratory habitat and wintering site for East Asian migratory birds, whereas poultry markets are hotbeds for AIVs and are closely related to human activities. In the present study, three H9N2 subtype AIVs were isolated. Phylogenetic tree of HA showed two chicken isolates clustered in the CK/BJ/94-like lineage, while the egret isolate belonged to Korean-like lineage.
In recent years, CK/BJ/94-like AIVs have been prevalent in poultry in central and southern China [10,24]. The two chicken isolates of this study had high homology with H9 subtype viruses isolated from poultry in southern China and they were isolated from the same area at different times, showed different evolution-ary status. Nucleotide sequence comparison showed some internal gene segments of the two H9N2 virus shared high homology with those of the H7N9 virus, and they also showed large differences among them. These results indicated that H9N2 AIVs from China's poultry markets presented complicated evolutionary characteristics.
Reports on H9N2 AIV isolated from wild birds have been rare recently. In the present study, one H9N2 virus was obtained from egrets in Dongting Lake, which, although clearly differed from the two poultry-derived H9N2 viruses, but shared 96.8% homology in PB1 nucleotide sequence with A/Chicken/Hunan/1/2012 virus, suggesting it exchanged gene segments with poultry-derived viruses.
Virus challenge experiments indicated that virus shedding occurred in most chickens infected with the poultry-derived viruses whereas the egret-derived virus could not replicate effectively in chickens. These findings indicated that the poultryderived H9N2 viruses could not only adapt well in chickens, but also could shed virus into the environment, which increased the possibility of AIV transmission and reassortment among susceptible animals. In contrast, the egret-derived H9N2 virus was not yet capable of infecting chicken. The M gene plays a decisive role in cross-species transmission and pathogenicity of AIV. The M gene in the two poultry-derived H9N2 viruses belonged to the G1like lineage, the same lineage as the M gene isolated from the Hongkong human H9N2 infection cases. Studies have shown that M gene is related to host range of influenza viruses and their adaptability to new hosts. Whether the M gene of the H9N2 poultry isolates in this study is directly related to cross-species transmission needs to be confirmed by further studies [25,26].
All chickens infected with the chicken-derived isolates showed seroconversion and HI test results showed little antigenic variation between the two H9N2 viruses. No chickens infected with the egret-derived isolate showed seroconversion. Cross-reactivity was observed between antisera to the chicken-derived isolate and the egret-derived virus, although the titers were low, indicating the HA of A/Chicken/Hunan/12/2011 and A/Egret/Hunan/1/ 2012 viruses shared antigenic sites. Results of HI test and the HA gene phylogenetic analysis were consistent in that the two viral genomes had big differences.
Results from lung passage of the egret H9N2 virus in SPF chicken indicated that this virus could not cross-species transmit effectively. Chickens could clear the virus by their own immune system. The chicken-derived H9N2 viruses could replicate in the  lungs of mice without adaptation (though the titers were low) whereas the egret-derived virus could not replicate in the lungs of mice. This finding indicated that it would take an ongoing adaptation process for un-adapted poultry H9N2 viruses to replicate stably and become pathogenic to mammalian hosts, and it would take even longer evolution for a wild bird derived H9N2 virus to adapt to mammals. This study performed preliminary analysis of the biological characteristics of the H9N2 viruses isolated from different sources and found big differences between them, and it seemed that some more mutations and further evolution would be required for the wild-bird derived virus to adapt to poultry. These findings laid a foundation for further studies on virulence enhancement and molecular mechanism of pathogenesis of influenza viruses. The ongoing molecular epidemiology surveillance for H9 viruses in poultry-markets, lakes and wetlands is necessary for understanding their roles in the evolution of highly pathogenic influenza viruses and to prevent their potential pandemic in poultry and humans. Figure S1 Histopathological changes of lungs are shown by Hematoxylin and Eosin (H&E) staining after virus infection. A, infected by E1 virus 3d p.i., there was pulmonary congestion, proliferation of bronchial epithelial cells in tertiary bronchus, cellular inflammatory infiltration in the lamina propria, and thickening of pulmonary capillary with red blood cell. B, infected by E1 virus 7d p.i., there was pulmonary congestion, thickening of small blood vessels and pulmonary capillary filled with red blood cell. C, infected by C12 virus 3d p.i., there was lung interlobular necrosis and inflammatory cells aggregation. D, infected by C12 virus 7d p.i., there was pulmonary slight gore, a small amount of pulmonary capillary serous fluid exudation. (TIF)