Discussion
Avian influenza viruses have posed a significant threat to the global poultry industry, in addition to public health. Due to the wide host range, some H9N2 AIVs not only circulate in poultry and wild bird population, but have also been detected in mammals, and which elevates the risk of transmission to humans. Another concern is the extensive genetic reassortment with other influenza serotypes, with H9N2 AIV considered to be the most common and destructive LPAIV subtype. While vaccination is an effective method to control the spread of influenza viruses, the efficacy of the H9N2 AIV vaccine has been challenged by the perpetual antigenic drift. Our previous research demonstrated that H9N2 field isolates have undergone antigenic drift to evolve into distinct antigenic groups, which also resulted in significant antigenic differences with the commercial vaccines. Unfortunately, the key amino acids associated with those antigenic drift events remain elusive. Two discrete antigenic sites “H9-A” and “H9-B” (or group I and II) which include at least 46 amino acid sites have been identified (T. Peacock et al., 2016). We also analyzed the evolution of these 46 aa in isolates collected over the most recent five years and found that most of these sites were completely conserved among circulating H9N2 viruses. To identify which single mutations of these viruses are driving the antigenic drift, 6 high-frequency mutations including those at aa sites 164, 168, 171, 198, 200 and 201 were screened by comparing amino acid alignments of the H9N2 AIVs isolated from China in 2014-2019. It should be noted that 5 of these (164,168,198,200 and 201) are located both in the antigenic sites and RBS region (Kaverin et al., 2004). Three of them (164, 168 and 201) were under positive selection (data not shown) pressure (Su et al., 2020). Therefore, these sites were likely responsible for the significant antigenic drift observed in recent years.
A variety of mechanisms, namely changes to epitope structure, acquisition of additional glycosylation sites and modulation of receptor-binding avidity, can contribute to both actual and apparent antigenic change (Abe et al., 2004; Das et al., 2011; Hensley et al., 2009). In the present study, 2 sites containing a high-frequency of mutations were identified that contribute to the viral antigenic drift. The A168N and D201G substitutions resulted in significant antigenic changes.
The HA protein of influenza is a highly glycosylated. N-linked glycosylation of HA has been reported to contribute to immune escape and virulence of influenza (Gao et al., 2021), and obtaining new glycosylation sites is also an important mechanism of viral antigenic drift. None of the mutations characterized here introduced novel glycosylation sites into the virus. In response to escape under neutralizing antibody pressure of the virus, influenza A virus could evolve by regulating HA receptor avidity via amino acid substitutions in the HA1 globular head domain (Das et al., 2011), many of which simultaneously alter the antigenicity (Hensley et al., 2009; Wu & Wilson, 2020). Some studies have shown that receptor avidity may be a more important factor than antigenicity in avoiding neutralizing antibodies (Crowe, 2012). Here, the effects of two mutations at antigenic site II of residue 201 (D201G and D201A) on antigenicity were significantly altered, in which the glycine substitution had a significant effect both on antigenicity and immune escape, while alanine had no effect. In addition, a serine substitution at residue 201 can also lead to antibody escape (Wan et al., 2014). It was observed that the change of surface structure was a result of the D201G substitution (G: hydrophobic, D: hydrophilic). This in turn may be due to the change of hydrophilicity, thus leading to the change of receptor avidity and antigenicity of the virus. It is important to note that changes in receptor avidity were not validated at residue 201 in this study. Of note, 201G mainly exists in predominant subclade C (Figure 1), which may result in immune escape in chicken flocks in the future. The other sites that may cause a similar phenomenon are at residues 201 and 168, which are both in the antigenic epitopes and RBS region. It was observed that asparagine (N, 37.1%) and alanine (A, 35.7%) accounted for the largest proportion at residue 168 in 2019-2020, which was significantly different from the major site (aspartic acid, D, 52.9%, and A, 33.0%) in 2014 by amino acid comparisons (Table S1). The substitution of those three amino acids can affect the antigenicity of H9N2 viruses (T. Peacock et al., 2016).
Comprised of a weak genetic basis of preferences for alternative avian receptors and human-like receptors, residue 198 may be a multifunctional site which could function to modulate polyclonal antisera binding and receptor-binding avidity at the same time (T. P. Peacock et al., 2020). In the present study, T198A was capable of immune escape in a reciprocal HI assay using polyclonal antibodies, while escape was not observed in a microneutralization assay. One possible explanation for this result is that the HI receptor-binding avidity and neutralization antibody epitopes are different. Thus, the molecular basis requires an in depth analysis, which is beyond the scope of this report.
In conclusion, key amino acid substitutions that may drive antigenic drift of H9N2-AIV in a recent five year periods were identified. Two predominant substitutions, including A168N and D201G were demonstrated to significantly affect antigenicity, but did not change growth characteristics and cell tropism. It is worth noting that the D201G substitution not only changed the antigenicity, but also produced an immune escape variant of the parental virus. The data presented here provide a reference for the prediction of the evolutionary direction of H9N2-AIV, and the development of effective vaccines.