Evolution and Antigenic Drift of Influenza A (H7N9) Viruses, China, 2017–2019

After a sharp decrease of influenza A(H7N9) virus in China in 2018, highly pathogenic H7N9 viruses re-emerged in 2019. These H7N9 variants exhibited a new predominant subclade and had been cocirculating at a low level in eastern and northeastern China. Several immune escape mutations and antigenic drift were observed in H7N9 variants.

. Evolutionary history of influenza A(H7N9) viruses, China, 2017China, -2019. A) Phylogenic tree of the hemagglutinin gene of H7N9 viruses. Colors indicate reference H7N9 viruses (n = 1,038) from each wave together with the H7N9 isolates from this study (panel B).
Red on the right of the tree indicates isolates from humans. All branch lengths are scaled according to the numbers of substitutions per site. The tree was rooted by using A/Shanghai/1/2013(H7N9), which was collected in February 2013. B) Hemagglutinin gene tree revealing a single cluster of highly pathogenic H7N9 viruses circulating during 2019. Red indicates the H7N9 isolates from this study. Scale bar represents number of nucleotide substitutions per site. C) Distribution of highly pathogenic influenza A(H7N9) viruses during 2019. The backgrounds indicate the sampling spaces of highly pathogenic influenza A(H7N9) viruses during 2019 in humans (red), environment (gray), and chickens (blue). The map was designed by using ArcGIS Desktop 10.4 software (ESRI, http://www.esri.com).
pathogenicity. Phylogenic analysis demonstrated that the HA and NA genes of all of these HPAI H7N9 viruses belonged to the Yangtze River Delta lineage and formed a new subclade (Figure 1, panel A), which exhibited a long genetic distance to the HPAI H7N9 viruses that persisted during 2017-2018. In particular, the HA and NA genes of A/chicken/ northeast China/19376-E5/2019(H7N9), A/chicken/northeast China/19254/2019(H7N9), and A/ chicken/northeast China/LN190408A/2019(H7N9) were genetically closely related to the human-infecting influenza A(H7N9) viruses from Gansu ( Figure  1, panel B; Appendix , implying the potential risk for the reemerged HPAI H7N9 viruses to infect humans. A root-to-tip regression analysis of temporal structure revealed aspects of the clock-like structure of 189 H7N9 viruses (correlation coefficient 0.89; R 2 0.95) during 2013-2019 ( Figure 2, panel A). The epidemic HPAI H7N9 viruses had circulated in China since 2017 and can be classified into 2 sublineages, A and B. The HA and NA genes of the HPAI H7N9 viruses in 2019 belonged to a new sublineage B, whereas the HPAI H7N9 viruses circulating in 2017-2018 grouped into sublineage A (Figure 2,panel B;Appendix Figures 4,5). Using the evolutionary rates of HA and NA, we estimated the times of origin (95% highest population density) of HPAI H7N9 viruses in sublineage B, which were September 2017-June 2018 for HA and April 2017-May 2018 for NA. Our HPAI H7N9 isolates exhibited traits of sublineages B-1 and B-2. We observed that the HPAI H7N9 viruses in eastern and northeastern China belonged to sublineage B-2 ( Figure  2, panel B). However, in mid-2019, the HPAI H7N9 viruses continued to evolve and formed sublineage B-1, which suggested that the estimated times to the most recent common ancestors were May 2019 for HA genes and February 2019 for NA genes. Also, the human-and chicken-origin HPAI H7N9 viruses from Liaoning, Gansu, and Inner Mongolia clustered together in sublineage B-1. These results indicate that the poultry-origin H7N9 virus in sublineage B-1 emerged before the human spillover event in March 2019.
Although no substantial difference surfaced in the substitution rate of HA genes between H7N9 viruses during 2017-2018 and the viruses during 2019, the increased substitution rate occurred in the first and second codons of reemerged HPAI H7N9 viruses (Appendix Table 4). In a maximum clade credibility tree of the HA gene, 9 independently occurring mutations gave rise to the new sublin- eage-B circulating in 2019, including A9S, R22K,  E71K, I78V, T116K, V125T, A151T, K301R, D439N (H7 numbering, https://www.fludb.org/brc/ha-Numbering.spg) (Figure 2,panel B), and only the V125T and A151T substitutions of the HA protein were reported as immune escape mutations (13). In addition, sublineage B-1 appeared to have acquired 3 parallel K184R, I499V, I520T (H7 numbering) mutations. The prevailing K184R substitutions of HPAI H7N9 viruses occurred during 2019. The K184R mutation was located in the antigenic site B and receptor binding region (Appendix Figure 6), suggesting that K184R was a potential mediator of viral antigenicity.
Next, we evaluated the protective efficacy of the new candidate H7N9 inactivated vaccine (H71903)-that is, reverse genetic recombinant carrying HA and NA of A/chicken/east China/ H7SD12/2019(H7N9) with internal genes of A/ duck/Guangdong/D7/2007(H5N2)-in chickens against the challenge of 4 HPAI H7N9 viruses prevailing in sublineage B in 2019. All of the control chickens challenged with the H7N9 viruses died within 6 days of challenge (Appendix Figure 8). However, virus shedding was not detected from any of the vaccinated chickens challenged with H7N9 viruses (Appendix Table 3), indicating that the new candidate H7N9 vaccine could provide sound protection for chickens against challenge with these reemerged H7N9 variants.

Conclusions
Our findings highlight that the HPAI H7N9 viruses that reemerged during 2019 had been cocirculating at a low level in eastern and northeastern China after the vaccination strategy was implemented. These HPAI H7N9 viruses continued to evolve and showed antigenic drift, posing a public health concern. Although vaccination can largely control the occurrence of H7N9 virus outbreaks, it can also accelerate the generation of novel variants. Therefore, comprehensive surveillance and enhancement of biosecurity precautions should be undertaken immediately to prevent the influenza virus epidemic from becoming a pandemic.

Acknowledgments
We acknowledge all contributors who submitted the sequence data on which this research is based to the GISAID EpiFlu Database. All submitters of data may be contacted directly through the GISAID website (http://www.gisaid.org).
This work was supported by the Key Research and Development Program of Guangdong Province (2019B020218004), the National Natural Science Foundation of China (31672586, 31830097, and 319410014)

All experiments with all available influenza A(H7N9) viruses were conducted in an animal biosafety level 3 laboratory and animal facility under South China Agricultural
University (SCAU) (CNAS BL0011) protocols. All animals involved in experiments were reviewed and approved by the Institution Animal Care and Use Committee at SCAU and treated in accordance with the guidelines (2017A002).

Sample collection and virus isolation
The cloacal and tracheal swab samples of chickens, ducks, and geese from live poultry markets were collected in 15 provinces of China, including Guangdong, Guangxi, Hebei, Shandong, Liaoning, Shaanxi, Hunan, Hubei, Sichuan, Jiangxi, Yunnan, Fujian, Henan, Chongqing, and Jilin provinces, during January-December 2019. Each sample was placed in 2 ml of the PBS supplemented with penicillin (5000 U/ml) and streptomycin (5000 U/ml). All the samples were inoculated in the allantoic cavities of 10-day-old embryonated chicken egg at 37°C. The allantoic fluid was collected and tested for hemagglutinin (HA) assay with 1% chicken red blood cells and then used in this study.

RNA was extracted from the suspension of virus isolates with the RNeasy Mini Kit
(Qiagen) as directed by the manufacturer. Two-step RT-PCR was conducted with universal primers as previously described (1), and HA and neuraminidase (NA) gene segments were amplified under standard conditions (1). PCR products were purified with a QIAamp Gel extraction kit (Qiagen) and sequenced with an ABI 3730 DNA Analyzer (Applied Biosystems).

Phylogenic analysis
All the available HA and NA genomic sequences with the complete coding regions of

Bayesian maximum clade credibility (MCC) phylogeny of influenza A(H7N9) viruses
We estimated rates of evolutionary change (nucleotide substitution) in the HA and NA distribution of trees obtained from BEAST analysis (with 10% of runs removed as burn-in) was also used to obtain the MCC tree for the HA and NA gene segments.

Structure-based mapping analysis
We predicted the HA monomer structure using the SWISS-Model website (https://swissmodel.expasy.org/), employing the A/Victoria/361/2011 (Protein Data Bank no. 4WE8) as a template. The amino acid corresponding to a three-dimensional (3D) amino acid structure of the HA protein was mapped using MacPymol (http://www.pymol.org/).

Hemagglutinin inhibition (HI) and cross-hemagglutinin inhibition assays
HI assay was used to determine the HI titers (1 The HI test assay was a standard beta test, whereby 4 HA units of H7N9 viruses in 96-well plates and the two-fold serially diluted serum prepared previously were added. The highest serum dilution that produced complete inhibition of HA activity was regarded as HI endpoint titers. Also, HI titers were used to calculate the antigenic relatedness (r values), calculated by the Archetti and Horsfall (6)  antigenic difference between the strains, r = 1 indicates the same antigenicity, whereas r < 0.5 indicates a significant antigenic difference between the strains.

Plasmid construction and reverse genetics
The internal gene segments from the A/duck/Guangdong/D7/2007(H5N2) (D7) strain were cloned into the Hoffmann's bidirectional transcription vector pHW2000 plasmid system

Vaccine test in chickens
The candidate H7N9 vaccine strain (H71903) contains the HA and NA genes from the H7SD12 and six internal genes from the D7. It is a formalin-inactivated oil-emulsion vaccine, with three parts inactivated allantonic fluid emulsified in two parts paraffin oil (volume/volume).
To evaluate the protective efficiency of the candidate H71903 vaccine in chickens, groups of three-week-old SPF chickens (n=10) were inoculated intramuscularly with 0.3 ml of the vaccine or with an equal volume of PBS as a control. The concentration of vaccine strains should be ≥ Tracheal and cloacal swabs in 5 day post-infection were collected from all the surviving chickens and titrated in 10-day-old embryonated eggs. The chickens were observed for signs of diseases and death for two weeks.

Statistical analyses
Data are presented as mean ± standard deviation and were analyzed with GraphPad Prism 5.0. An independent samples t test was used for analysis. Appendix