All sequences of H5Nx viruses (H5N1, H5N2, H5N6, and H5N8) from 1996 to 2016 in China that were available from the GISAID and GenBank databases were downloaded. And the sequences of H5N6 virus aligned using the MegAlign 6.06 program (DNASTAR) with the ClustalW method.

The clade 2.3.4.4 H5 subtype avian influenza viruses, including A/goose/Guangdong/SH7/2013 (H5N1) (SH7, EPI_ISL_259927), A/duck/Guangdong/673/2014 (H5N6) (673, EPI_ISL_259923), A/duck/Guangdong/674/2014 (H5N6) (674, EPI_ISL_259924), and A/goose/Jiangsu/1306/2014 (H5N8) (Js1306, EPI_ISL_259926) were isolated from sick ducks or geese. The H5N2 avian influenza virus, A/chicken/Hebei/LZF/2014 (H5N2) (LZF, EPI_ISL_259925) was isolated from a sick chicken. All viruses were propagated in 10-day-old SPF embryonated eggs at 37°C and frozen at -80°C. Human embryonic kidney cells (HEK293T), Madin-Darby canine kidney (MDCK), and Chicken Embryo Fibroblast (CEF) cells were grown in Dulbecco’s modified essential medium (DMEM; Gibco) with 10% fetal bovine serum (FBS; BI) and 1% penicillin/streptomycin (Gibco).

N1 gene segments were replaced with NAs from LZF, 673, 674, and Js1306, respectively, and were cloned into a pHW2000 plasmid system according to a previous study (Hoffmann et al., 2000; Xiao et al., 2016). HEK293T cells monolayers in 6-well plates were transfected at 80–90% confluency with 4 μg of the eight plasmids (500 ng of each plasmid) by using Lipofectamine 3000 (Invitrogen) according to the manufacturer’s instructions. Six hours later, the mixture was replaced with Opti-MEM (Gibco) containing 0.2% bovine serum albumin and 1 μg/ml TPCK trypsin (Sigma). After 48 h, the supernatant was harvested and injected into SPF embryonated eggs for virus propagation. Viruses were titrated in embryonated eggs using hemagglutination assays and sequenced. The five reassortant H5Nx viruses were named rH5N1, rH5N2, rH5N6, ΔrH5N6, and rH5N8.

The plaque assay was adapted from previously described procedure (Tian et al., 2012). MDCK cells were grown in DMEM with 1 mM L-glutamine and 10% FBS and seeded onto 6-well plates. Confluent monolayers were washed twice with phosphate-buffered saline (PBS) and infected with serial 10-fold dilutions of the virus at 37°C. After 2 h of incubation, the cells were washed twice with PBS and then overlaid with MEM containing 1% agarose. After 48–72 h of incubation at 37°C, the agarose was removed and cells were stained with 0.5% crystal violet in 10% formaldehyde solution. The plaques were visualized and manually counted.

The balance between HA binding and NA cleavage was measured by performing hemagglutination assays as previously described (Jean-Sébastien Casalegno, 2014). Viruses were prepared using serial twofold dilutions in 50 μl of PBS in 96-well plates and incubated with 50 μl of 0.8% guinea pig erythrocytes at 4°C for 75 min and then at 37°C for 2 h. HA titers were recorded at both temperatures, respectively. Each sample was assayed in triplicate.

A fluorescent substrate 2′ (4-methylumbelliferyl)-α-D-N-acetylneuraminic acid (MUNANA; Sigma) was used to measure NA enzymatic activity, according to the manufacturer’s recommendations. Viruses were prepared using serial twofold dilutions in 50 μl of calcium saline buffer in 96-well plates and 50 μl of 200 μmol MUN was added to each well, and then incubated for 60 min at 37°C in darkness. One hour later, stop solution was added and NA activity was quantified using a Synergy HT Multi-Detection microplate reader (BioTek, Winooski, VT, United States) with excitation and emission wavelength of 360 and 440 nm.

Confluent MDCK cells were infected at a multiplicity of infection (MOI) of 0.001 TCID₅₀/cell, CEF cells were infected at MOI of 0.0001 TCID₅₀/cell for 1 h at 37°C on 6-well plates, as described previously (Xiao et al., 2016). One hour later, plates were washed twice with PBS, and then incubated with DMEM containing 0.2% BSA at 37°C or 39°C with 5% CO₂. Culture supernatants were collected at 12, 24, 36, and 48 hour post-inoculation (h.p.i.). The virus titers were determined by performing 50% tissue culture infective dose (TCID₅₀) assays in MDCK cells.

Groups of eight female 4- to 6-week-old BALB/c mice (Vitalriver Company, Beijing, China) were anesthetized with CO₂ and inoculated intranasally (i.n.) with 10⁵ EID₅₀/50 μl of the five reassortant H5Nx viruses (rH5N1, rH5N2, rH5N6, ΔrH5N6, and rH5N8). Body weight and clinical symptoms were monitored daily for 14 days after infection. The mice were euthanized if they lost more than 25% of their initial body weight. Three mice from each group were euthanized on 4 days post-inoculation (d.p.i.), and the lung, brain and turbinate tissues were collected to determine the virus titers by EID₅₀ assay. The 50% mouse lethal dose (MLD₅₀) were determined by inoculating groups of five mice with serial 10-fold dilutions of virus ranging from 10¹ to 10⁵ EID₅₀/50 μl. The MLD₅₀ values were calculated using the Reed and Muench method.

Groups of seven 5-week-old SPF chickens (Merial Vital Company, Beijing, China) were inoculated intranasally with 10⁵ EID₅₀/200 μl of five reassortant H5Nx viruses, respectively. Seven uninfected chickens served as negative controls. All chickens were observed for clinical symptoms for 14 days. On 1 d.p.i., seven naive chickens (contact) were mixed with treatment group chickens. Cloacal and throat swabs were collected on 3, 5, and 7 d.p.i. from infected and naive chickens. Three inoculated chickens in each group were euthanized on 3 d.p.i. to test the virus replication in organs, including heart, lung, kidney, brain, spleen, and liver. Similar actions were performed on chickens if they died within 3 days. For the chickens in the naive contact group, the virus replication in heart, lung, kidney, brain, spleen, and liver were detected after the chickens died.

All 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 reviewed and approved by the Institutional Animal Care and Use Committee at SCAU and were carried out in accordance with the approved guidelines.

Until 2010, H5N1 subtype virus was the major genotype in China (Figure 1A). The H5N2 and H5N8 viruses were sporadically reported in China after 2011. The H5N6 virus was isolated in 2013 and became a major genotype in China. The clade 2.3.4.4 viruses were detected in most regions of China from 2013 to 2016 (Figures 1B,C). In addition, H5N2 virus was mainly isolated in northern China. The NA6 stalk region can be divided into two groups according to the presence of an amino acid deletion in the positions 58–68 of NA stalk region. The percentage of 11 amino acids deletion in NA stalk region of H5N6 viruses were 63% (24 of 38 viruses), 37% (69 of 186 viruses), 87% (359 of 408 viruses), and 100% (23 of 23 viruses) respectively, from 2013 to 2016 (Figure 1C). Seventeen H5N6 strains of clade 2.3.4.4 have caused human infections². Sixteen viruses had the 11 amino acid deletion at the 58th to 68th in the NA stalk region. The H5N8 virus has become the pandemic genotype in 2016 (Figure 1C).

To directly measure the balance between HA and NA, the loss of HA titer at 37°C was compared to the erythrocyte agglutination rate at 4°C. The binding and release of five reassortant viruses were quantitatively measured via guinea pig erythrocytes. At 4°C, the NA is inactive while HA can bind to its receptors and agglutinate the erythrocytes. At 37°C, the NA becomes active; thus, the loss of HA titer evaluates the equilibrium between HA affinity and NA activity. HA titers were reduced by five- and sixfold for the rH5N6 and rH5N8 viruses, respectively. The rH5N1, rH5N2, and ΔrH5N6 viruses did not show any change of HA titer at either temperature (Figure 2A). 2′ (4-methylumbelliferyl)-α-D-N-acetylneuraminic acid was used to evaluate the NA activities of five H5Nx reassortant virus (Potier et al., 1979). The rH5N2 and rH5N8 viruses had RFU values, and the rH5N6 and ΔrH5N6 viruses had lower activities, as compared to rH5N1 (Figure 2B).

We then characterized the cleavage properties of the HA protein combined with different NAs by performing plaque assays. All reassortant H5Nx viruses formed plaques on MDCK cells without TPCK trypsin. MDCK cells infected with rH5N8 virus grew the largest plaques (1.80 mm) when compared to rH5N6 and ΔrH5N6 viruses. The rH5N1 virus and rH5N2 virus formed pin-point plaques in MDCK cells (from 0.48 to 0.73 mm) (Figure 3).

The growth kinetics of reassortant H5Nx viruses on mammalian and avian cells (MDCK and CEF) were determined. In MDCK cells, the rH5N1 and ΔrH5N6 viruses replicated more efficiently and had the highest virus titers at 12, 24, and 36 h. By contrast, rH5N2 had the lowest virus titer at each time point. The rH5N6 and rH5N8 reassortant viruses exhibited an intermediate level of virus titers among H5Nx reassortant viruses. Growth kinetic of rH5N8 was performed using a significantly higher titer when compared to other reassortant H5Nx viruses at 48 h. And the H5N2 virus also appeared to surpass the rH5N6 virus at 48 h (Figure 2C). In avian CEF cells, ΔrH5N6 replicated more efficiently than the other reassortant H5Nx viruses. rH5N2 had the lowest virus titers at each investigated time point, except at 12 h. rH5N8 growth kinetics were similar to rH5N1 but higher than rH5N6 (Figure 2D). These results showed that the ΔrH5N6 grew efficiently in MDCK and CEF cells. In contract, rH5N2 virus had lower titer in these cells.

To investigate whether these five H5Nx reassortant virus also replicate efficiently in mammalian hosts, BALB/c mice were infected with 10⁵ EID₅₀/50 μl of these H5Nx reassortant viruses. All viruses were lethal to mice with 2 d.p.i., except for rH5N2 (at 5 d.p.i.). Mice infected with rH5N1, rH5N6, ΔrH5N6, and rH5N8 died within 6 d.p.i., whereas mice infected with rH5N2 showed delayed fatality on 8 d.p.i. (Figures 4A,B). All viruses replicated efficiently in lung and nasal tissues. As H5N1 IAV can cause systemic infection, neurotropism, and long-term effects on the central nervous system (CNS) (Shinya et al., 2011), the replication of H5Nx reassortant viruses in mice brain was tested. The rH5N1, rH5N6, and rH5N8 virus were found in all tested mice brains, while rH5N2 and ΔrH5N6 were only detected in 2/3 or 1/3 mice, respectively (Figures 4C–E). Subsequently, to study the virulence of H5Nx reassortant viruses, BALB/c mice were inoculated with serial 10-fold dilutions of the virus to determine viral MLD₅₀. The MLD₅₀ (lgEID₅₀) of rH5N1, rH5N6, ΔrH5N6, and rH5N8 were 2.5, 3.3, 2.68, 3.16, and 2.5, respectively.

Groups of seven 5-week-old SPF chickens were inoculated intranasally with 10⁵ EID₅₀/50 μl of the five H5Nx reassortants. Chickens infected with rH5N2 and ΔrH5N6 died within 4 d.p.i. ΔrH5N6, rH5N6, and rH5N8 viruses caused sudden death without symptoms of illness on 2 d.p.i. (Figure 5A). Reassortant viruses were isolated from lung, spleen, kidney, and brain samples at high virus titers. All H5Nx reassortant viruses replicated in lungs, with mean titers from 5.8 to 6.97 lgEID₅₀. The ΔrH5N6 virus replicated highly in brains, with a mean titer of 6.58 lgEID₅₀ (Table 1).

To evaluate the transmission of five H5Nx reassortants in chickens, after 1 d.p.i., seven SPF chickens were co-housed with infected chickens. The naive contact chickens co-housed with chickens infected with rH5N1, rH5N6, ΔrH5N6, or rH5N8 viruses infection died at 4 d.p.i. Naïve chickens co-housed with chickens infected with the rH5N2 virus died on 5 d.p.i. (Figure 5B). No seroconversion was found in surviving contact animals. Oropharyngeal and cloacal swabs from contact chickens were tested at 3 d.p.i. or 5 d.p.i. No virus was detected in oropharyngeal among contact chickens at 7 d.p.i. (Table 2). The virus was detected in the lung, spleen, kidney, and brain of contact chickens (Table 1). Overall, the H5Nx reassortant viruses were highly pathogenic to chickens, and rH5N6, ΔrH5N6, and rH5N8 exhibited enhanced pathogenicity and greater transmissibility in chickens.