Full Genome Characterization of Human Influenza A/H3N2 Isolates from Asian Countries Reveals a Rare Amantadine Resistance-Conferring Mutation and Novel PB1-F2 Polymorphisms

Abstract

Influenza A viruses evolve at a high rate requiring continuous monitoring to maintain the efficacy of vaccines and antiviral drugs. We performed next generation sequencing analysis of 100 influenza A/H3N2 isolates collected in four Asian countries (Japan, Lebanon, Myanmar, and Vietnam) during 2012-2015. Phylogenetic analysis revealed several reassortment events leading to the circulation of multiple clades within the same season. This was particularly evident during the 2013 and 2013/2014 seasons. Importantly, our data showed that certain lineages appeared to be fitter and were able to persist into the following season. The majority of A/H3N2 viruses continued to harbor the M2-S31N mutation conferring amantadine-resistance. In addition, an S31D mutation in the M2-protein, conferring a similar level of resistance as the S31N mutation, was detected in three isolates obtained in Japan during the 2014/2015 season. None of the isolates possessed the NA-H274Y mutation conferring oseltamivir-resistance, though a few isolates were found to contain mutations at the catalytic residue 151 (D151A/G/N or V) of the NA protein. These variations did not alter the susceptibility to neuraminidase inhibitors and were not detected in the original clinical specimens, suggesting that they had been acquired during their passage in MDCK cells. Novel polymorphisms were detected in the PB1-F2 open-reading frame resulting in truncations in the protein of 24-34 aminoacids in length. Thus, this study has demonstrated the utility of monitoring the full genome of influenza viruses to allow the detection of the potentially fittest lineages. This enhances our ability to predict the strain(s) most likely to persist into the following seasons and predict the potential degree of vaccine match or mismatch with the seasonal influenza season for that year. This will enable the public health and clinical teams to prepare for any related healthcare burden, depending on whether the vaccine match is predicted to be good or poor for that season.

Keywords: PB1-F2; antiviral; evolution; full-genome; influenza A/H3N2; phylogenetic analysis; reassortment; vaccine.

Figure 1

Evolutionary relationship among human influenza A/H3N2 isolates from Asian countries. Full genome sequences of 100 H3N2 isolates were aligned and the phylogenetic tree for each genome segment was inferred using maximum likelihood analysis based on the best-fit nucleotide substitution model for each gene. Bootstrap support values ≥70%, which corresponds to a ≥95% probability that a given clade is real, are shown (Hillis and Bull, 1993). Full genome sequences of WHO-recommended vaccine strains for the seasons covered by the study were obtained from the Influenza Resource Database and included in the analysis for comparison. The vaccine strains are indicated in boldface italics. The season(s) covered by the vaccine strains are indicated next to the vaccine strain name in the HA tree: e.g., V11/12 for season 2011/2012, V12/13 for 2012/2013, V13/14 for 2013/2014, V14/15 for 2014/2015, and V15/16 for 2015/2016.

Figure 2

Frequency of PB1-F2 polymorphisms leading to a truncated protein. The PB1-F2 coding region of 5365 H3N2 isolates reported between 1968 and 2014 at the Influenza Virus Resource were analyzed to determine the polymorphisms leading to an early stop codon in the PB1-F2 open reading frame. The predicted protein/peptide length in amino acid is indicated in the figure.