Poultry Infection with Influenza Viruses of Wild Bird Origin, China, 2016

A influenza virus with 10 hemagglutinin (HA) subtypes has emerged in poultry (1), and the potential role of migratory birds in transmission of avian influenza virus has caused concern (2). We report infection with low pathogenicity avian influenza (LPAI) virus of HA subtype 13 (H13) among migratory birds (black-tailed gulls [Larus crassirostris]) and domestic poultry (chickens) in Weihai, Shandong Province, eastern China (online Technical Appendix Figure 1, https://wwwnc.cdc.gov/EID/ article/24/7/17-1220-Techapp1.pdf). Weihai is a breeding center for black-tailed gulls that congregate from northern Asia, eastern Asia, Southeast Asia, and North America. These gulls reside along the coastlines of the East China Sea and Japan and have been found as vagrants in Alaska, North America, and the Philippines (3). In China, black-tailed gulls perch at the Longxudao wharf (37°23′24.05′′N, 122°41′26.16′E), located in the northeastern corner of Weihai. In December 2016, we collected 149 fecal samples from black-tailed gulls at Longxudao wharf and screened them for evidence of influenza virus by reverse transcription PCR, DNA sequencing, and BLAST (http://www.ncbi.nlm.nih.gov/blast/Blast.cgi) analysis in the GenBank database. After independently inoculating positive fecal samples into the allantoic cavities of specific pathogenfree embryonated chicken eggs, we obtained 6 influenza H13N2 and 60 influenza H13N8 virus isolates. To assess the epidemiologic characteristics of these H13 isolates, we completely sequenced an H13N2 isolate (A/black-tailed gull/Weihai/115/2016) and an H13N8 isolate (A/black-tailed gull/Weihai/17/2016) (GenBank accession nos. MF461177–92). Phylogenetic analysis indicated that their HA and neuraminidase (NA) segments were derived from the Eurasian lineage, in accordance with their geographic distribution (online Technical Appendix Figures 2, 3). Moreover, the H13N2 and H13N8 isolates possessed high nucleotide sequence identity to the avian influenza virus subtypes previously isolated from Europe, Asia, and North America (online Technical Appendix Table 1). We speculate that avian influenza virus subtypes H13N2 and H13N8 are reassortants between the Eurasian and North American lineages (online Technical Appendix Figures 4, 5). We next analyzed the timing of the reassortment events that led to the emergence of subtype H13N2 (Figure, panel A). During July 2009, June 2012, July 2009, and June 2015, the following genes, respectively, were transferred from seagulls in Europe: HA, nucleocapsid protein (NP), matrix (M), and nonstructural (NS). During 2004, November 2011, and October 2014, the following genes, respectively, originated from waterfowl in Asia: polymerase basic (PB) 1, polymerase acidic (PA), and NA. In November 2007, the PB2 gene was transferred from avian influenza viruses circulating among wild waterfowl in North America. We also estimated the timing of the reassortment events that led to the emergence of subtype H13N8 (Figure, panel B). During June 2012, July 2013, July 2013, and June 2015, the following genes, respectively, were transferred from seagulls in Europe: NP, NA, M, and NS. During 2004, November 2011, and September 2012, the following genes, respectively, originated from waterfowl in Asia: PB1, PA, and HA. In November 2007, the PB2 gene was transferred from avian influenza viruses circulating among wild waterfowl in North America. According to these data, the generation of influenza virus subtypes H13N2 and H13N8 in seagulls seems to have been a complex process and was probably completed in the middle of 2015 (Figure). We also found that subtypes H13N2 and H13N8 possessed some molecular markers associated with increased virulence and transmission among mammals (online Technical Appendix Table 2). In April 2017, we analyzed serum samples collected from 48 chickens at a chicken farm at Songcun town (37°04′39.96′′N, 122°00′38.83′′E) in Weihai for serologic evidence of exposure to H13 viruses. We found detectable hemagglutinin inhibition (HI) antibody titers against H13N2 virus in 4 (8.3%) samples and detectable HI antibody titers against H13N8 virus in 14 (29.2%) samples (online Technical Appendix Table 3). When we evaluated reference serum samples known to contain HI antibodies against each of the virus subtypes for potential cross-reactivity, we observed no apparent cross-reactivity of H13