A mild case of influenza A(H7N9) virus infection was identified in a 4-year-old boy in a rural area of Jinshan District, Shanghai, reported on March 31. The patient had been exposed to poultry. His signs and symptoms included acute fever (maximum 39° axillary), cough, nasal drainage, and tonsillitis. A diagnosis of upper respiratory tract infection was made, and the child recovered after 5 days of antiviral drug therapy (9,10). Nasal swab specimens were positive for influenza A(H7N9) virus by using real-time RT-PCR (11), as recommended by the World Health Organization. Although his family members, unrelated workers, and chickens he may have had contact with were tested, none tested positive for influenza virus.

The whole genome sequence of the SH/JS01strain was amplified from the nasal swab specimen by using RT-PCR (primer sequences available upon request). Strict controls were used during PCR amplification; results were confirmed by another laboratory to exclude contamination with laboratory strains. We constructed maximum likelihood trees of each gene segment sequence using the general time-reversible model implemented in MEGA 5.1 (12), and estimated divergence time using the Bayesian Markov chain Monte Carlo method implemented in BEAST (v1.6.1) (13). We compared all known strains of the 2013 H7N9 subtype and other referenced influenza virus sequences deposited in GISAID (http://platform.gisaid.org/epi3/frontend#57f951) and GenBank (Table 1 and Technical Appendix Figures 1–8).

The critical mutations in the SH/JS01 strain associated with virulence and mammalian adaption were compared to 3 prevalent H7N9 subtype reference strains: A/Shanghai/1/2013 (SH/1), A/Shanghai/2/2013 (SH/2), and A/Anhui/1/2013 (AH/1). In the HA gene of SH/JS01, the only mammalian adapting substitution observed was 138A (H3 numbering); amino acid residues involved in receptor-binding specificity showed avian-like signatures, including 186G and 226Q, which were similar to SH/1 but distinctive from SH/2 and AH/1. In the internal genes of SH/JS01, we observed some human-like and mammalian-adapting signatures, including 89V in polymerase basic subunit (PB)2, 368V in PB1, 356R in the RNA polymerase acidic subunit, 42S in nonstructural gene 1, and 30D and 215A in matrix gene 1; however, some hallmark changes involved in mammalian adaptation still showed avian signatures, including 627E and 701D in PB2 and 100V and 409S in the RNA polymerase acidic subunit (Table 2). Most strikingly, SH/JS01 retained aa 69–73 (N9 numbering) in the stalk region of the NA gene. In contrast, deletion of aa 69–73, which is considered to occur when viruses adapt to terrestrial birds, prevailed in all the known 2013 H7N9 subtype isolates (14) (Table 2, and Figure). These findings indicate that SH/JS01 is genetically distinct from all the known human influenza A(H7N9) strains and carries more avian influenza-like signatures.

Phylogenetic analysis and divergence time estimation showed that the SH/JS01 HA gene diverged in October 2005 and was closely related to SH/1; nucleotide similarity was 99.7% (online Technical Appendix Figure 1). However, the NA gene, which is most closely related to A/northern shoveler/Hong Kong/MPL133/2010(H2N9) and A/duck/Jiangxi/21714/2011(H11N9) with nucleotide similarity of 99% and 99.3%, respectively, are estimated to have diverged in September 2010, earlier than that of known strains of the 2013 H7N9 subtype (estimated to have occurred in January 2011) (Table 1; Technical Appendix Figure 2).

On the basis of internal genes, 2013 H7N9 viruses have been divided into minor (m) and major (M) clades in the phylogenetic trees, and first 9, then 27 genotypes (5,6,8). Our data showed that SH/JS01 belongs to the m-PB2|PA|M or G3 genotype (Table 1; Technical Appendix Figures 3–8); its NS gene and that of A/Chicken/Dawang/1/2011(H9N2) (DW1), shared the highest nt identity (99.4%). The SH/JS01 NS gene had an estimated divergence time as early as October 1996 (Table 1; Technical Appendix Figure 3). The closest relatives of the SH/JS01 M gene were poultry H9N2 strains A/chicken/Jiangsu/CZ1/2012 and A/chicken/Jiangsu/NTTZ/2013 (nt identity 99.7%); BJ16 (98.6%) was not closely related. The divergence time of SH/JS01 M gene is estimated as May 2011, which is earlier than most 2013 H7N9 isolates (Table 1; Technical Appendix Figure 4). The PB2 and PA genes of SH/JS01 were positioned in minor clades and showed higher identities to A/chicken/Jiangsu/MYJMF/2012(H9N2) (99.2% and 99.3%, respectively) than to BJ16 (97.1% and 98%) (Table 1; Technical Appendix Figures 5, 6). In contrast, the PB1 and NP genes of SH/JS01 belong to the major lineage (Technical Appendix Figures 7, 8). Collectively, these data indicate that, except for PB1 and NP, the internal genes of SH/JS01 are more closely related to those in poultry H9N2 viruses identified before 2013 than to BJ16, which was an H9N2 virus considered to be the donor of most of internal genes of the 2013 H7N9 virus.

SH/JS01, a distinct H7N9 virus strain identified in 2013 during the early stage of the influenza A(H7N9) epidemic in China, provided information to define the evolution of the H7N9 subtype. Although identified in an infected human, SH/JS01 has more avian-prone properties and fewer mammalian-adapting mutations than other known human 2013 H7N9 subtypes. SH/JS01 has a waterfowl-like NA gene characterized by the absence of a deletion in the NA stalk and most of its internal genes are more closely related to avian H9N2 subtype strains isolated during the 2011–2012 influenza season than to other recently emerged strains of the H7N9 subtype. Molecular clock analysis further predicted an earlier divergence time in most genes of SH/JS01. These findings indicate that SH/JS01 might be a precursor strain of the H7N9 virus that co-circulated with more evolved viruses, although we cannot exclude that SH/JS01 may have been generated independently from the other H7N9 strains by reassortment of waterfowl strains with avian H9N2 strains and then transmitted directly to a human.

The sequences of SH/JS01 contained more avian-like signatures than those of other H7N9 isolates from humans; this underscores the potential of these viruses to infect humans. The phenotypic characteristics of SH/JS01, which might describe its zoonotic potential, remain to be investigated.

It is unclear whether other SH/JS01–like viruses are still circulating in poultry in China and if so, what the potential is for their evolution and ability to infect humans. Intensive influenza surveillance and additional influenza A virus genome sequences isolated from poultry and from humans with severe and mild manifestations of infection are needed to clarify the population dynamics of H7N9 viruses.

Phylogenetic analysis of avian influenza viruses, 1988–2013