Evolutionary dynamics of avian influenza A H7N9 virus across five waves in mainland China, 2013–2017

Journal of Infection 77 (2018) 205–211

Contents lists available at ScienceDirect

Journal of Infection journal homepage: www.elsevier.com/locate/jinf

Evolutionary dynamics of avian influenza A H7N9 virus across five waves in mainland China, 2013–2017 Dan Xiang a,b, Zhiqing Pu b, Tingting Luo b, Fucheng Guo b, Xiaobing Li b, Xuejuan Shen b,c, David M. Irwin d,e, Robert W. Murphy f, Ming Liao b,c, Yongyi Shen b,c,∗ a

Shantou University Medical College, Shantou 515041, China College of Veterinary Medicine, South China Agricultural University, Guangzhou 510642, China Key Laboratory of Zoonosis Prevention and Control of Guangdong Province, Guangzhou 510642, China d Department of Laboratory Medicine and Pathobiology, University of Toronto, Toronto M5S 1A8, Canada e Banting and Best Diabetes Centre, University of Toronto, Toronto M5S 1A8, Canada f Centre for Biodiversity and Conservation Biology, Royal Ontario Museum, Toronto M5S 2C6, Canada b c

a r t i c l e

i n f o

Article history: Accepted 16 May 2018 Available online 25 May 2018 Keywords: H7N9 Evolutionary dynamics Reassortment Highly pathogenic

s u m m a r y Since its emergence in March 2013, novel avian influenza A H7N9 virus has triggered five epidemics of human infections in China. This raises concerns about the pandemic threat of this quickly evolving H7N9 subtype for humans. In this study, we evaluated all available genomes for H7N9 and H9N2 influenza A viruses. Our assessment discovered that H7N9 of the 1st wave had the lowest nucleotide diversity, which then experienced substantial and rapid population expansion from a small founder population. From the 2nd wave, their nucleotide diversity increased quickly, indicating that H7N9 viruses had acquired larger populations and mutations after their initial emergence in 2013. After the phylogeographic divergence in the 2nd wave, although the HA and NA genes from different regions differed, compared to previous epidemics, the evolving H7N9 viruses in the 5th wave lost most of their previous clades. The highly pathogenic avian influenza (HPAI) H7N9 viruses in the 5th wave clustered together, and clustered close to the low pathogenic avian influenza (LPAI) virus isolated from the Pearl River Delta in the 3rd and 4th waves. This result supports the origin of HPAI H7N9 viruses was in the Pearl River Delta. In the 5th wave, although both HPAI and LPAI H7N9 viruses were isolated from the Pearl River Delta, their HA and NA genes were phylogenetically distinct. © 2018 The British Infection Association. Published by Elsevier Ltd. All rights reserved.

Introduction Since the novel avian influenza A H7N9 virus emerged in March 2013,1 it has triggered five epidemics of human infections in China.2-4 This virus causes severe and fatal respiratory diseases in humans, and many people infected with the H7N9 virus died due to severe pneumonia and its complications.5 A total of 1564 laboratory-confirmed cases of human infection have been reported to the World Health Organization (WHO).6 There were 135, 316, 226 and 121 cases reported in waves 1–4, respectively, with a declining trend. The 5th wave started earlier than usual, and had a rapid growth of human infection cases (766 cases) from September 2016 to October 2017.2 H7N9 viruses have now surpassed H5N1 viruses in laboratory confirmed human infections.7 This raises concerns about the pandemic threat of H7N9 viruses to humans and ∗ Corresponding author at: College of Veterinary Medicine, South China Agricultural University, Guangzhou 510642, China. E-mail address: [email protected] (Y. Shen).

if this virus can potentially become more invasive to humans. It is worth mentioning that some H7N9 viruses transitioned from low to high pathogenicity for poultry in the 5th wave.8-10 In addition, HPAI H7N9 viruses can transmit via respiratory droplets among ferrets, and are more pathogenic than LPAI H7N9 viruses in mammals.11 This consistent with their potential pandemic threat to humans and poultry.12 H7N9 is a triple reassortant virus, with the hemagglutinin (HA) and neuraminidase (NA) genes originating from migratory birds along the East Asian flyway, and its six internal genes from chicken H9N2 avian influenza viruses.13 , 14 After its emergence, the HA and NA genes have evolved into different lineages, and have frequently reassorted, acquiring internal genes from other H9N2 viruses, expanding the genetic diversity of H7N9 viruses.15 , 16 In addition to epidemiological studies of human infections, monitoring the evolution of these quickly evolving viruses is also critically needed. The prevalence and molecular evolution of the HA and NA genes in H7N9, and the six internal genes shared by the H7N9 and H9N2 viruses are not fully understood. Herein, we analyzed the evolution

https://doi.org/10.1016/j.jinf.2018.05.006 0163-4453/© 2018 The British Infection Association. Published by Elsevier Ltd. All rights reserved.

206

D. Xiang et al. / Journal of Infection 77 (2018) 205–211

of HA and NA genes of H7N9, explored the role of H9N2 viruses in the evolution of H7N9 viruses, and the evolution of HPAI H7N9 viruses. Materials and methods Source of sequences and preliminary treatment All previously published H7N9 and H9N2 influenza A virus (1994–2017) genomes and related sequences of H7 and N9 were obtained from the Influenza Virus Resource at the National Center for Biotechnology Information (NCBI) (www.ncbi.nlm.nih.gov/ genomes/FLU), Global Initiative on Sharing Avian Influenza Data (GISAID) database (www.gisaid.org) and the Influenza Research Database (FluDB) (www.FluDB.org). Redundant sequences were removed, and laboratory strains were excluded. GenBank/GISAID accession numbers are listed in Supplementary Table 1. All eight gene segments were aligned separately using MAFFT v7.245.17 Poorly aligned positions were identified and removed using Gblocks version 0.91b,18 and adjusted manually to correct frame shift errors. The genetic identities of the eight gene segments of the H7N9 influenza viruses were calculated using Mega v6.06,19 and plotted as a heatmap by R v3.4.0 software.20 Genetic diversity and haplotypes analysis of HA and NA Sequences from 1442 H7 and 1397 N9 viruses were analyzed. Nucleotide diversity, identification of haplotypes, haplotype diversity, Tajima’s D test and Fu and Li’ D∗ test were performed using DnaSP 5.10.0.21 HA and NA sequences were classed into 575 and 542 haplotypes, respectively. Nucleotide substitution models for the HA haplotypes and NA haplotypes were estimated using JMODELTEST 2.1.4.22 Maximum Likelihood (ML) haplotype trees were reconstructed for HA and NA by using RaxML v8.2.423 with 10 0 0 bootstraps employing the GTRGAMMA model. To investigate relationships between sequences from the different waves, medianjoint networks of the HA and NA sequences were separately constructed with Network 5.0 (http://www.fluxus-engineering.com/ sharenet.htm). Genotypic analysis of six internal genes ML trees of the six internal genes (matrix protein (MP), nucleoprotein (NP), nonstructural protein (NS), polymerase acidic protein (PA), polymerase basic 1 (PB1) and polymerase basic 2 (PB2)) from all H7N9 and H9N2 were constructed using RAxML v8.2.4.23 Clades with median root-to-tip distances equal or less than 20% of the distances in the tree were selected as clusters, and then merged manually. To detect the reassortments of H7N9 and H9N2, based on Pu et al.’s genotype classification method,24 we classified their six internal genes into more detailed genotypes (D1-212, Supplementary Table 2). D1-158 belonged to G57, while for genotypes (D159-211), one of their six internal genes didn’t belong to G57, and for D212, all six internal genes did not belong to G57. Each combination of clade ID assignment with the six internal segments was given a unique genotype ID. Results Increase in genetic diversity of HA and NA from wave 1 to 5 The percent identities of the HA, NA and six internal genes of the H7N9 and related sequences were plotted as a heatmap (Supplementary Fig. 1, Supplementary Table 3). HA and NA genes

display 0.0–9.2% sequence difference, the six internal genes displayed 0.0–10% sequence difference within H7N9, while those isolated from Jiangxi province in 2009 (11 strains) were more divergent (> 10% difference) compared with the other H7N9 sequences. A total of 1442 H7 and 1397 N9 sequences were analyzed. Genetic diversity of H7 and N9 sequences in the five H7N9 waves was estimated (Table 1, Supplementary Fig. 2). The nucleotide diversity of H7 (0.00315) and N9 (0.00335) in the 1st wave was the lowest across the 5 waves, with their diversity increasing sharply to reach values of 0.02151 and 0.02281, respectively, in the 5th wave. For the six internal genes, nucleotide diversity of NP, NS, PA, and PB2 was the lowest in the 1st wave, while for M and PB1, their nucleotide diversity was similar in the 1st and 2nd waves (M: 0.0101 vs. 0.00861; PB1: 0.01585 vs. 0.01204). Subsequently, the nucleotide diversity of the six internal genes increased quickly (Table 1). Tajima’s D and Fu and Li D∗ tests are the most widely used neutral tests. Significant negative Tajima’s D and Fu and Li’s D∗ values could be an indication of deviation from neutrality that might have resulted from a selective sweep and/or population expansion. Both Tajima’s D and Fu and Li’s D∗ tests for eight genes were significantly negative in the first two waves (Table 1). In the 5th wave, there were 69 HA genes showed four types of HPAI motifs: -PKGKRIAR/GLF-(1/69), -PKGKRTAR/GLF-(19/69), PKRKRAAR/GLF-(1/69), -PKRKRTAR/GLF-(48/69). HA sequences of these HPAI H7N9 viruses (98.4−100%) had higher genetic identity at the nucleotide level than that of LPAI viruses (95.7−100%). While, NA and the six internal genes of the HPAI and LPAI H7N9 viruses showed similar diversity at the nucleotide level (Fig. 1, Supplementary Table 3). Phylogenetic and geographic analyses of HA and NA Among the 1442 H7 sequences, 575 haplotypes were identified, while 542 haplotypes were found in the 1397 N9 sequences (Supplementary Tables 4 and 5). Phylogenetic analysis classified H7 into haplogroups A, B, C, D, E1 and E2, and N9 into haplogroups F, G, I, J, K1 and K2 (Supplementary Fig. 3). The median joining network depicted the five waves of H7N9 (Fig. 2). In the 1st wave, most of H7N9 isolated from the Yangtze River Delta region, where haplotype 3 from haplogroup A placed centrally in the H7 diagram. Haplotype 212 from haplogroup F located centrally in the N9 network diagram, which mainly contained sequences isolated from the Yangtze River Delta region during the 1st wave. In the 2nd wave, H7N9 spread from the Yangtze River Delta to the Pearl River Delta, and both H7 and N9 had significant divergence in these two regions. In the 2nd wave, H7 of the Yangtze River Delta mainly distributed in haplogroups B and C, while haplogroup D dominated in the Pearl River Delta region (Supplementary Fig. 4). For N9, viruses of the Yangtze River Delta mainly distributed in haplogroup G and K1, with those from the Pearl River Delta region in haplogroup J (Supplementary Fig. 4). In the 3rd–5th waves, viruses from the Yangtze River Delta and the Pearl River Delta have some differences, most of H7 sequences of Yangtze River Delta assigned to haplogroup E1, while most sequences of Pearl River Delta belonged to haplogroup E2. In the 5th wave, H7 sequences of all of the HPAI H7N9 viruses belonged to haplogroup E2, and had closed relationship with LPAI H7N9 viruses that isolated from the Pearl River Delta in the 3rd and 4th waves. While for the 20 LPAI H7N9 viruses isolated in the Pearl River Delta, H7 sequences of 12 LPAI H7N9 viruses clustered with those LPAI H7N9 viruses isolated in the Yangtze River Delta in the 5th wave (haplogroup E1), the other 8 viruses were in haplogroup D. In the Pearl River Delta, except 5 viruses, all N9 sequences of HPAI H7N9 viruses had a single clade (involved in haplogroup K2) in the 5th wave, while for LPAI H7N9 viruses, 8 viruses were in haplogroup J, and 12 viruses clustered with those LPAI H7N9 viruses

D. Xiang et al. / Journal of Infection 77 (2018) 205–211

207

Fig. 1. Matrices showing the genetic diversity of the eight segments of the HP and LPAI H7N9 genomes in the 5th wave. (a) 230 HA sequences, (b) 203 NA sequences, (c) 189 M sequences, (d) 189 NP sequences, (e) 189 NS sequences, (f) 189 PA sequences, (g) 189 PB1 sequences, (h) 189 PB2 sequences. Both HA and NA sequences contains 69 HPAI H7N9, while other six genes contains 61 HPAIVs.

208

D. Xiang et al. / Journal of Infection 77 (2018) 205–211 Table 1 Genetic diversity of the eight gene segments. Population

Na

nHTb

HTdivc

πd

Tajima’s D

Fu and Li’s D∗

H7

Wave1 Wave2 Wave3 Wave4 Wave5

275 694 221 11 230

115 308 157 11 143

0.966 0.988 0.994 1 0.988

0.00315 0.00801 0.01467 0.01323 0.02151

−2.69340∗∗∗ −2.39365∗∗ −2.43088∗∗ −0.80810 −1.60313

−7.94853∗∗ −9.32510∗∗ −6.12702∗∗ −1.08992 −3.70 0 05∗∗

N9

Wave1 Wave2 Wave3 Wave4 Wave5

265 696 222 11 203

115 239 128 11 134

0.934 0.972 0.982 1 0.991

0.00335 0.00833 0.01119 0.01424 0.02281

−2.67025∗∗∗ −2.37395∗∗ −2.48477∗∗∗ −1.21427 −1.64581

−8.64944∗∗ −7.99440∗∗ −6.83136∗∗ −1.50630P −5.38220∗∗

PA

Wave1 Wave2 Wave3 Wave4 Wave5

220 617 177 11 189

61 277 141 11 136

0.679 0.976 0.997 1 0.994

0.00545 0.00796 0.02547 0.02818 0.02853

−2.52546∗∗∗ −2.50626∗∗∗ −1.59166 −0.25489 −1.26248

−3.98142∗∗ −8.84567∗∗ −1.78754 −0.63841 −3.58964∗∗

PB1

Wave1 Wave2 Wave3 Wave4 Wave5

220 617 177 11 189

62 301 139 10 150

0.817 0.988 0.996 0.982 0.996

0.01585 0.01204 0.02025 0.03289 0.03016

−1.24584 −2.24635∗∗ −1.84510∗ 0.36872 −1.44803

−4.69849∗∗ −7.91678∗∗ −3.36838∗∗ −0.01674 −2.54430∗

PB2

Wave1 Wave2 Wave3 Wave4 Wave5

220 617 177 11 189

72 193 142 10 138

0.682 0.969 0.997 0.982 0.995

0.00825 0.0196 0.03123 0.03289 0.03643

−2.46122∗∗ −1.88096∗ −1.39550 0.36872 −0.96008

−6.15121∗∗ −7.12170∗∗ −2.43793∗ −0.01674 −3.56467∗∗

M

Wave1 Wave2 Wave3 Wave4 Wave5

220 617 177 11 189

63 156 103 9 93

0.825 0.897 0.987 0.964 0.978

0.0101 0.00861 0.01504 0.01451 0.01829

−1.69192 −2.30168∗∗ −1.76898∗ −0.83549 −1.52883

−6.13795∗∗ −6.93877∗∗ −2.87196∗ −0.78391 −3.49065∗∗

NP

Wave1 Wave2 Wave3 Wave4 Wave5

220 617 177 11 189

98 175 121 11 121

0.879 0.944 0.993 1 0.988

0.00992 0.02157 0.02729 0.03177 0.02999

−2.38133∗∗ −1.76369∗ −1.20474 0.89182 −1.33438

−5.02615∗∗ −7.57399∗∗ −1.87862 −0.37667 −1.87861

NS

Wave1 Wave2 Wave3 Wave4 Wave5

220 617 177 11 189

73 168 119 10 100

0.711 0.949 0.991 0.982 0.982

0.00509 0.01128 0.01737 0.01333 0.01436

−2.57983∗∗∗ −2.26119∗∗ −2.14298∗∗ −1.30688 −2.10368∗

−4.87259∗∗ −8.42077∗∗ −3.66188∗∗ −1.42911 −4.10914∗∗

a

Total number of samples in each wave. Number of haplotypes. Haplotype diversity. d Nucleotide diversity. ∗∗∗ P < 0.001. ∗∗ P < 0.01. ∗ P < 0.05. b c

isolated from the Yangtze River Delta in the 5th wave (haplogroup K1 and K2). Genotypic analyses of six internal genes reveal that H7N9 in the 5th wave tend to have backbones from previous waves ML trees of the six internal genes were constructed separately (Supplementary Fig. 5). Clades were manually merged, based on distances in the trees and bootstrap values of the branching. H7N9 and H9N2 sequences for all six internal genes isolated from viruses before 2010 fell into one clade in each phylogeny. After 2010, within each prevalent clade for the five internal genes (PB2, PB1, PA, M and NP) the relationships could be further divided into three clades that associate with a few viruses that were isolated before 2010, while for the genes encoding NS only two clades for the 2010–2017 sequences, with a few early viruses, could be identified (Supplementary Fig. 5). On the basis of clade classification, 212 genotypes (D1-212) were found in H9N2 and H7N9 (Supplementary Table 2). 83.43% of the 2013–2017 Chinese H7N9 viruses could be classified into eleven genotypes: genotype 1 (D1), genotype 2 (D2), genotype 7

(D7), genotype 8 (D8), genotype 9 (D9), genotype 43 (D43), genotype 83 (D83), genotype 86 (D86), genotype 124 (D124), genotype 127 (D127) and genotype 130 (D130). H7N9 and H9N2 showed a strong correlation in the epidemiological process. That is, the most common genotype in H7N9 can always be found in H9N2 (Fig. 3, Supplementary Table 6). However, comparison of the distribution of genotypes in H7N9 and H9N2 revealed that the most popular genotypes of H9N2 did not always become major genotypes in H7N9 (Supplementary Fig. 6). In the 1st wave, 85.9% of the H7N9 sequences belonged to genotypes D2, D83 and D86, while only 3.6% of the H9N2 sequences were classified into these genotypes. Meanwhile, 42.3% of the H9N2 sequences belonged to four genotypes D46, D70, D81 and D212, yet these genotypes had low frequency in H7N9 (D70 and D81 were both 0.9%, D46 and D212 were not found). In the 2nd wave, 53.8% of H9N2 belonged to 9 genotypes D1, D2, D7, D29, D30, D70, D72, D149 and D212, and 89.5% of the H7N9 sequences belonged to 8 genotypes D1, D7, D9, D43, D83, D124, D127 and D130. D1 and D7 were the most popular genotypes in both H7N9 and H9N2. In the 3rd wave, genotypes D1, D2, D7, D20 and D43 dominated in H7N9 (65.0%), while D1 and D212 were most common in H9N2 (37.2%).

D. Xiang et al. / Journal of Infection 77 (2018) 205–211

209

Fig. 2. Phylogenetic networks display the relationships of HA and NA of H7N9 viruses in 5 waves. Hap_3 from haplogroup A was central in the H7 network diagram; Hap_212 in haplogroup F was central to the N9 network diagram. (a) and (b) The phylogenetic networks of HA and NA display the epidemic periods. (c) and (d) The phylogenetic networks of HA and NA display host (human or other host) and geographical distribution where viruses isolated (Yangtze River Delta region, Pearl River Delta region, others Chinese region and Foreign). (e) The phylogenetic networks of HP H7N9 viruses which display four types of HPAIV motifs in the 5th wave.

Fig. 3. Diversity of the genotypes of H7N9 and H9N2 viruses in the 5 waves. There were 212 genotypes in the H7N9 and H9N2 virus isolated from the Yangtze River Delta region (Y), the Pearl River Delta region (P), other Chinese regions (O) and Foreign regions (F).

D20 and D43 were not found while D2 and D7 were low (12.8%) in H9N2, D212 was not discovered in H7N9. In the 4th wave, the most common genotype in H7N9 was genotype D1 (36.4%), while in H9N2, it was D28 and D212 (20.5% and 43.6%). In the 5th wave, D1 (19.5%), D8 (51.6%) and D10 (7.8%) had high percentage in LPAI H7N9 viruses, D1 (65.6%) and D8 (11.5%) dominated in HPAI H7N9 viruses, while in H9N2, genotype D8 had the highest frequency (66.7%).

In all five waves, the genotypes for H7N9 were not only those identified from previous H7N9 and H9N2 viruses, but also some were new reassortments. For the 1st wave, 12 of the 23 H7N9 genotypes were new reassortments. For the 2nd wave, 11 of the 40 H7N9 genotypes were from previous H7N9 viruses, 14 of the 40 were from H9N2 genotypes, while 15 H7N9 genotypes were new reassortments. 16 of the 33 H7N9 genotypes were from previous H7N9 viruses, eight of the 33 were H9N2 genotypes and nine of

210

D. Xiang et al. / Journal of Infection 77 (2018) 205–211

the 33 were new reassortments in the 3rd wave. In the 4th wave, three of the six genotypes derived from previous H7N9, one from H9N2, and two were new reassortments. In the 5th wave, all genotypes of LPAI H7N9 came from previous H7N9 and H9N2 viruses, with nine of the 10 genotypes from H7N9 and one from H9N2, 10 of the 11 HPAI H7N9 genotypes were from previous H7N9 viruses and one were new reassortments (Fig. 3, Supplementary Table 6). Discussion Since 2013, five waves of the H7N9 epidemic have spread from eastern to southern and northern China. H7N9 viruses have now surpassed H5N1 in laboratory confirmed human infections.7 In the 1st wave, H7 and N9 genes of these viruses were derived from Eurasian wild birds,14 and did not have enough time to accumulate mutations, therefore, possessed low nucleotide diversity (Table 1, Supplementary Fig. 2). Substantial population expansion appears to have occurred after their emergence. From the 2nd wave, nucleotide diversity of all H7N9 genes increased quickly (Supplementary Fig. 2), and both Tajima’s D test and Fu and Li’s D∗ test showed significantly negative in the first two waves. This was typical of a rapid demographic expansion from a small founder population.25 Thus after the first outbreak of these new reassortment viruses, they experienced a rapid population expansion. They were distributed from the Yangtze River Delta to the Pearl River Delta region, and both H7 and N9 had significant divergence in these two regions (Fig. 2), that is, H7N9 viruses established multiple regionally distinct lineages.15 There were four haplogroups for both H7 and N9 (Haplogroups A, B, C, D in HA and haplogroups F, I, J, G in NA) in the 2nd wave, but three of these (haplogroups A, B, C in HA and haplogroups F, I, G in NA) are undetectable after the 3rd wave. It means that after the phylogeographic divergence in the 2nd wave, HA and NA genes of H7N9 isolated from the Yangtze River Delta and the Pearl River Delta tended to lost most of their previous lineages, and remained limited clades.26 H9N2 viruses are endemic to China and continue to contribute their six internal genes to H7N9,15 , 24 which greatly increases the genetic diversity of H7N9. The internal genes of H7N9 have greater diversity than the surface genes (Supplementary Fig. 1) raising questions about this curious evolutionary trait. This phenomenon is most likely due to frequent reassortments. Acquiring internal gene cassettes from H9N2, increases the diversity of the six internal genes of H7N9. Genotype G57 of H9N2 has become predominant in chickens since 2010.24 Our phylogenetic trees of the six internal genes display that all H7N9 cluster within G57 of the H9N2 lineages (Supplementary Fig. 5). In order to further survey the contribution of H9N2 to the diversity of H7N9, we further classified the six internal genes of H7N9 and H9N2 into 212 genotypes. 34 of the 212 genotypes occurred exclusively in H7N9, 137 only found in H9N2, while 41 of the genotypes were shared by H7N9 and H9N2. The genotypic diversity may be caused by sequential reassortments after the H7N9 viruses were introduced into the different regions. Although the presence of a H7N9 genotype can always be found in H9N2 (Fig. 3, Supplementary Table 6), comparisons of the distributions of the genotypes of H7N9 and H9N2 revealed that the most popular genotypes of H9N2 did not always become major genotypes in H7N9, thus reassortments of H7N9 and H9N2 were not random. Therefore, H7N9 should prefer some genotypes over others. The genotypes of H9N2 in different region are different, thus H7N9 cannot freely reassort with all genotypes of H9N2. Therefore, founder effect would be the alternative explaination for the preference of some internal genotypes of H7N9. In waves 1–4, genotypes of H7N9 were not only from previous H7N9 and H9N2, but also some new reassortments. In the 5th wave, all genotypes of LPAI H7N9 came from previous H7N9 and H9N2 viruses, while, 10 of the 11 HPAI H7N9 genotypes were from previous H7N9 viruses and

one were new reassortments (Fig. 3, Supplementary Table 6). This result suggested that H7N9 in the 5th wave tend to have internal backbones from previous waves rather than new reassorments. H7N9 was regarded as low pathogenic avian influenza virus in the previous four waves.27 In the 5th wave, some H7N9 viruses have an insertion of several basic amino acids at the HA cleavage site, which is associated with transition from low to high pathogenicity for chicken.8-10 Although the HPAI H7N9 viruses have four types of cleavage motifs, all HA of HPAI H7N9 viruses in the 5th wave cluster together (haplogroup E2, Fig. 2a). Then they clustered with other LPAI H7N9 viruses isolated from the Pearl River Delta (Fig. 2, Supplementary Fig. 3). While HA of LPAI H7N9 viruses isolated from the Yangtze River Delta in the 5th wave were mainly in haplogroup E1. Most of the NA of HPAI H7N9 viruses (35 of 39) were clustered together, while four sequences were clustered with LPAI H7N9 viruses isolated from the Yangtze River Delta. This result supported the HPAI H7N9 viruses derived from LPAI H7N9 of the Pearl River Delta.28 , 29 In addition, NA gene pool of the Yangtze River Delta also has some contribution for the origin of HPAI H7N9 viruses. In the 5th wave, although both HPAI and LPAI H7N9 viruses were isolated from the Pearl River Delta, their HA and NA genes located in different phylogentic positions (Fig. 2, Supplementary Figs. 3 and 4). The dominated genotype of internal backbones in LPAI H7N9 viruses was D8, while it was D1 in HPAI H7N9 viruses (Fig. 3, Supplementary Table 6). These results suggested that HPAI and LPAI H7N9 viruses had independent evolution in the 5th wave. In addition, they may prefer different internal backbones, however we cannot exclude the possibility of founder effect. HA sequences had high genetic identity within HPAI H7N9 viruses, while HPAI and LPAI H7N9 viruses had similar genetic diversity in NA and six internal genes (Fig. 1). This result further supported the HA genes of HPAI H7N9 viruses came from limited ancestors, in contrary, both the HPAI and LPAI H7N9 recruited the NA and internal genes from the same influenza gene pool. Avian influenza A H7N9 viruses have evolved substantially since their emergence in 2013. Chickens are highly susceptible to H7N9 viruses and while they are able to efficiently shed viruses, low pathogenicity was observed in waves 1–4.30 These characteristics allowed H7N9 viruses to spread throughout China. The predominant group of viruses currently circulating among poultry was genetically and antigenically different from previous viruses.31 The clinical severity of individuals in the 5th wave was similar to that of previous epidemics,2 and importantly, the number of cases of human infections increased sharply and some avian influenza A H7N9 viruses have evolved from LPAI to HPAI. HPAI viruses have a devastating economic impact for poultry industry. Given the pandemic potential of H7N9, it is vitally important to monitor their evolution. Our results reveal that only two haplogroups for both H7 and N9 ((haplogroups D and E in HA and haplogroups J and K in NA) survived in the 5th wave, while other haplogroups were undetected. In addition, their internal backbones (genotypes of the six internal genes) tend to come from previous waves rather than new reassorments. Although few human-to-human transmission cases are confirmed, circulating H7N9 has surpassed H5N1 as the major avian influenza virus that causes human infections. Extensive efforts are needed to prevent and control the spread of this subtype. Conflict of interest The authors declare not conflict of interest. Acknowledgments This work was supported by Guangdong Natural Science Funds for Distinguished Young Scholar (2014A030306046).

D. Xiang et al. / Journal of Infection 77 (2018) 205–211

Supplementary materials Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.jinf.2018.05.006. References 1. Gao R, Cao B, Hu Y, Feng Z, Wang D, Hu W, et al. Human infection with a novel avian-origin influenza A (H7N9) virus. N Engl J Med 2013;368(20):1888–97. 2. Wang X, Jiang H, Wu P, Uyeki TM, Feng L, Lai S, et al. Epidemiology of avian influenza A H7N9 virus in human beings across five epidemics in mainland China, 2013–17: an epidemiological study of laboratory-confirmed case series. Lancet Infect Dis 2017;17(8):822–32. 3. Iuliano AD, Jang Y, Jones J, Davis CT, Wentworth DE, Uyeki TM, et al. Increase in human infections with avian influenza A(H7N9) virus during the fifth epidemic – China, October 2016–February 2017. Morb Mortal Wkly Rep 2017;66(9):254–5. 4. Su S, Gu M, Liu D, Cui J, Gao GF, Zhou J, et al. Epidemiology, evolution, and pathogenesis of H7N9 influenza viruses in five epidemic waves since 2013 in China. Trends Microbiol 2017;25(9):713–28. 5. Chen Y, Liang W, Yang S, Wu N, Gao H, Sheng J, et al. Human infections with the emerging avian influenza A H7N9 virus from wet market poultry: clinical analysis and characterisation of viral genome. Lancet 2013;381(9881):1916–25. 6. WHO. Human infection with avian influenza A(H7N9) virus-China(Disease outbreak news, 26 October 2017). 2017. Available from: http://www.who.int/csr/ don/26- october- 2017- ah7n9- china/en/. 7. Webby RJ, Yang Z. The changing landscape of A H7N9 influenza virus infections in China. Lancet Infect Dis 2017;17(8):783–4. 8. Zhang F, Bi Y, Wang J, Wong G, Shi W, Hu F, et al. Human infections with recently-emerging highly pathogenic H7N9 avian influenza virus in China. J Infect 2017;75(1):71–5. 9. Ke C, Mok CKP, Zhu W, Zhou H, He J, Guan W, et al. Human infection with highly pathogenic avian influenza A(H7N9) virus, China. Emerg Infect Dis 2017;23(8):1332–40. 10. Shi J, Deng G, Kong H, Gu C, Ma S, Yin X, et al. H7N9 virulent mutants detected in chickens in China pose an increased threat to humans. Cell Res 2017;27(12):1409–21. 11. Imai M, Watanabe T, Kiso M, Nakajima N, Yamayoshi S, Iwatsuki-Horimoto K, et al. A highly pathogenic avian H7N9 influenza virus isolated from a human is lethal in some ferrets infected via respiratory droplets. Cell Host Microbe 2017;22(5):615–26. 12. Subbarao K.. Avian influenza H7N9 viruses: a rare second warning. Cell Res 2018;28(1):1–2. 13. Liu D, Shi W, Shi Y, Wang D, Xiao H, Li W, et al. Origin and diversity of novel avian influenza A H7N9 viruses causing human infection: phylogenetic, structural, and coalescent analyses. Lancet 2013;381(9881):1926–32.

211

14. Lam TT-Y, Wang J, Shen Y, Zhou B, Duan L, Cheung C-L, et al. The genesis and source of the H7N9 influenza viruses causing human infections in China. Nature 2013;502(7470):241–4. 15. Lam TT, Zhou B, Wang J, Chai Y, Shen Y, Chen X, et al. Dissemination, divergence and establishment of H7N9 influenza viruses in China. Nature 2015;522(7554):102–5. 16. Zhu H, Lam TT, Smith DK, Guan Y. Emergence and development of H7N9 influenza viruses in China. Curr Opin Virol 2016;16:106–13. 17. Katoh K, Toh H. Parallelization of the MAFFT multiple sequence alignment program. Bioinformatics 2010;26(15):1899–900. 18. Castresana J. Selection of conserved blocks from multiple alignments for their use in phylogenetic analysis. Mol Biol Evol 20 0 0;17(4):540–52. 19. Tamura K, Stecher G, Peterson D, Filipski A, Kumar S. MEGA6: molecular evolutionary genetics analysis version 6.0. Mol Biol Evol 2013;30(12):2725–9. 20. Kolde R.. Pheatmap: pretty heatmaps. Available from:, https://cran.r-project.org/ web/packages/pheatmap/index.html; 2015. 21. Librado P, Rozas J. DnaSP v5: a software for comprehensive analysis of DNA polymorphism data. Bioinformatics 2009;25(11):1451–2. 22. Darriba D, Taboada GL, Doallo R, Posada D. jModelTest 2: more models, new heuristics and parallel computing. Nat Methods 2012;9(8):772. likelihood-based phyloge23. Stamatakis A.. RAxML-VI-HPC: maximum netic analyses with thousands of taxa and mixed models. Bioinformatics 2006;22(21):2688–90. 24. Pu J, Wang S, Yin Y, Zhang G, Carter RA, Wang J, et al. Evolution of the H9N2 influenza genotype that facilitated the genesis of the novel H7N9 virus. Proc Natl Acad Sci USA 2015;112(2):548–53. 25. Avise JC. Phylogeography: the history and formation of species. Harvard University Press; 20 0 0. 26. Xiang D, Irwin DM, Shen Y. The H7N9 viruses have lost most of their previous lineages in the fifth wave. J Infect 2018;76(4):417–18. 27. To KK, Chan JF, Chen H, Li L, Yuen KY. The emergence of influenza A H7N9 in human beings 16 years after influenza A H5N1: a tale of two cities. Lancet Infect Dis 2013;13(9):809–21. 28. Qi W, Jia W, Liu D, Li J, Bi Y, Xie S, et al. Emergence and adaptation of a novel highly pathogenic H7N9 influenza virus in birds and humans from a 2013-human-infecting low pathogenic ancestor. J Virol 2018;92(2) e00921-17. 29. Yang L, Zhu W, Li X, Chen M, Wu J, Yu P, et al. Genesis and spread of newly emerged highly pathogenic h7n9 avian viruses in mainland China. J Virol 2017;91(23) e01277-17. 30. Zhang Q, Shi J, Deng G, Guo J, Zeng X, He X, et al. H7N9 influenza viruses are transmissible in ferrets by respiratory droplet. Science 2013;341(6144):410–14. 31. Uyeki TM, Katz JM, Jernigan DB. Novel influenza A viruses and pandemic threats. Lancet 2017;389(10085):2172–4.