Asparagine (N) at residue 701 improves the binding of PB2 to mammalian importin-a isoforms [36
], and for H5N1 and H7N7 avian influenza viruses it is associated with increased viral replication in mammalian cell lines, enhanced virulence in the mouse model, and more efficient transmission in the guinea pig model [12
]. The data presented in this study demonstrate that the PB2-D701N substitution also increased the H1N1pdm virus replication in primary human cells, pathogenicity in the mouse model, and transmission in the ferret model. Notably, this is the first demonstration that PB2-D701N substitution enhances the transmissibility of influenza A viruses in the ferret transmission model.
Two other studies that examined the PB2-D701N and other PB2 mutations in the context of rA/Netherlands/602/2009 and rA/California/04/2009 H1N1pdm viruses concluded that both the PB2-E627K and PB2-D701N mutations had little impact on pathogenicity and transmissibility [6
]. However, the supplemental data published by Yamada et al., showed the PB2-D701N substitution in rA/California/04/2009 H1N1pdm enhanced morbidity in BALB/cJ mice, which is consistent with our findings. However, the tissue titers of the rA/California/04/2009 PB2-701N virus were comparable with those of the PB2-701D virus at 3 dpi and 6 dpi, and the authors did not remark that PB2-D701N actually showed enhanced virulence as their supplemental data suggested [6
]. One limitation with the study by Yamada et al. [6
], is that tissue titers were not determined at early time points (e.g., 12, 24, and 48 hpi.), when higher levels of replication in the lungs are more pronounced (, and thus differences from the wild type virus early in the course of infection would not have been evident. The data from Herfst et al.,
suggested that the PB2-D701N substitution didn’t increase the transmission efficiency of the rA/Netherlands/602/2009 virus in their ferret transmission model [7
]. The inconstant results between that study and the data presented in this study may result from differences in the virus strains used (four residues are different between the NY/1682 and Netherlands/602 strains), or the differences in the experimental design for the ferret transmission experiments. It appears that the transmission model for the wild type H1N1pdm (A/Netherlands/602/2009) used by Herfst et al.[7
], is less stringent (4 of 4 contact ferrets are infected) than the transmission model used for this study, which consistently demonstrates that only two out of three contact ferrets show transmission with several other wild type H1N1pdm strains [33
]. This more stringent transmission model allows one to detect the increased transmission rate upon introduction of the PB2-D701N substitution into the H1N1pdm virus and the data also clearly shows increased transmission kinetics (. The faster transmission kinetics of the PB2-701N substitution may be related to the fact that this substitution enhances RdRp activity at cooler temperatures such as those found in the upper respiratory tract (.
Collectively, using more detailed experimentation including using various cell lines at various temperatures, primary human lung cells, a well-designed mouse model, and a stringent ferret transmission model, we showed that the PB2-701N substitution enhanced the replication, pathogenesis, and transmission of an H1N1pdm virus in mice, ferrets, and primary cultured human alveolar epithelial cells. Therefore, we speculate that H1N1pdm viruses containing a PB2-D701N substitution may be more virulent and/or transmissible in humans than the viruses that circulated in the early years of the pandemic. Surprisingly, of the approximate 4,000 H1N1pdm PB2 sequences available in the major influenza virus databases (Influenza Research Database [40
], NIH/NCBI Influenza Virus Resource [41
], EpiFlu Database [42
]), there is only one H1N1pdm virus (A/Wisconsin/51/2009) containing the PB2-D701N mutation (GISAID EpiFlu, accession#EPI273622, accessed 08-15-2012). The rarity of PB2 sequences with D701N mutation indicates that so far the mutation has not conferred better fitness to the H1N1pdm virus for transmission in humans, which is plausible since fitness and evolution in the human population is a polygenic trait. However, considering the estimated large number of infections worldwide [43
], additional viruses with the PB2-D701N mutation may exist in the population. The higher polymerase activity and virus replication observed in human cells, and lower IFN-λ induction may give the virus an advantage by increasing mRNA transcription and genome replication. PB2 has been shown to inhibit interferon-β expression through its N-terminal region [36
], and the mechanism underlying this IFN-λ suppression effect caused by a mutation in the C-terminal region is intriguing and warrants further investigation. With increased rounds of replication, additional mutations are more likely to occur, which may lead to emergence of other known or unknown virulence determinants [45
]. Those virulence factors alone, or in combination with the PB2-D701N mutation, may increase the replication and pathogenicity of H1N1pdm viruses.
In summary, our findings support a role for the asparagine residue at PB2-701 in contributing to the virulence of influenza A viruses and show that PB2-D701N in the H1N1pdm virus increases its viral RNA polymerase activity in human and other mammalian cells, confers more efficient replication in cell lines and in primary human alveolar epithelial cells, and results in increased pathogenicity in mice and increased transmissibility in ferrets. The true threat of the PB2-D701N substitution to humans is unknown; however, if H1N1pdm viruses containing PB2-701N do emerge as this unique virus continues to evolve in human populations, it may pave the way to a more severe disease burden [39