Enhanced Human-Type Receptor Binding by Ferret-Transmissible H5N1 with a K193T Mutation

Abstract

All human influenza pandemics have originated from avian influenza viruses. Although multiple changes are needed for an avian virus to be able to transmit between humans, binding to human-type receptors is essential. Several research groups have reported mutations in H5N1 viruses that exhibit specificity for human-type receptors and promote respiratory droplet transmission between ferrets. Upon detailed analysis, we have found that these mutants exhibit significant differences in fine receptor specificity compared to human H1N1 and H3N2 and retain avian-type receptor binding. We have recently shown that human influenza viruses preferentially bind to α2-6-sialylated branched N-linked glycans, where the sialic acids on each branch can bind to receptor sites on two protomers of the same hemagglutinin (HA) trimer. In this binding mode, the glycan projects over the 190 helix at the top of the receptor-binding pocket, which in H5N1 would create a stearic clash with lysine at position 193. Thus, we hypothesized that a K193T mutation would improve binding to branched N-linked receptors. Indeed, the addition of the K193T mutation to the H5 HA of a respiratory-droplet-transmissible virus dramatically improves both binding to human trachea epithelial cells and specificity for extended α2-6-sialylated N-linked glycans recognized by human influenza viruses.IMPORTANCE Infections by avian H5N1 viruses are associated with a high mortality rate in several species, including humans. Fortunately, H5N1 viruses do not transmit between humans because they do not bind to human-type receptors. In 2012, three seminal papers have shown how these viruses can be engineered to transmit between ferrets, the human model for influenza virus infection. Receptor binding, among others, was changed, and the viruses now bind to human-type receptors. Receptor specificity was still markedly different compared to that of human influenza viruses. Here we report an additional mutation in ferret-transmissible H5N1 that increases human-type receptor binding. K193T seems to be a common receptor specificity determinant, as it increases human-type receptor binding in multiple subtypes. The K193T mutation can now be used as a marker during surveillance of emerging viruses to assess potential pandemic risk.

Keywords: H5N1; N-linked glycan; glycan array; influenza; receptor binding; sialic acid.

FIG 1

Receptor specificities of WT H5 and the transmissible mutants. Glycan microarray analysis was used to determine the receptor specificities of VN1203 and INDO05 WT and mutant HAs, including WT VN1203 (A), WT INDO05 (B), VN1203 N224K Q226L (C), INDO05 Q226L G228S (D), VN1203 T160A N224K Q226L T318I (E), INDO05 H107Y T160A Q226L G228S (F), VN1203 T160A T318I (G), and INDO05 H107Y T160A (H), and human HA proteins of a recent H3N2 A/HK/6398/10 strain (I) and the 2009 pandemic H1N1 A/CA/04/09 strain (J). The mean signals and standard errors were calculated from six independent replicates. The data shown are representative of results from three independent assays. α2-3-linked sialosides (glycans 11 to 79 on the x axis) and α2-6-linked sialosides (glycans 80 to 135) are shown. Glycans 1 to 10 are nonsialylated controls (see also Table S1 in the supplemental material). RFU, relative fluorescence units.

FIG 2

Receptor specificities of two sets of VN1203 H5 mutants. Glycan microarray analysis was used to determine the receptor specificities of HAs of the VN1203 T160A N224K Q226L (A), VN1203 T160A Q226L G228S (B), VN1203 K193T N224K Q226L (C), VN1203 K193T Q226L G228S (D), VN1203 T160A K193T N224K Q226L (E), and VN1203 T160A K193T Q226L G228S (F) mutants. The mean signals and standard errors were calculated from six independent replicates. The data shown are representative of results from three independent assays. α2-3-linked sialosides (glycans 11 to 79 on the x axis) and α2-6-linked sialosides (glycans 80 to 135) are shown. Glycans 1 to 10 are nonsialylated controls (see also Table S1 in the supplemental material).

FIG 3

Determination of receptor-binding avidities of WT H5 and all the mutants. An ELISA-like assay was used to determine the binding avidities of WT VN1203 (A), WT INDO05 (B), VN1203 T160A N224K Q226L T318I (C), INDO05 H107Y T160A Q226L G228S (D), VN1203 T160A T318I (E), INDO05 H107Y T160A (F), VN1203 T160A N224K Q226L (G), VN1203 T160A Q226L G228S (H), VN1203 K193T N224K Q226L (I), VN1203 K193T Q226L G228S (J), VN1203 T160A K193T N224K Q226L (K), and VN1203 T160A K193T Q226L G228S (L). The mean signals and standard errors were calculated from six independent replicates in the ELISA-like assay. The data shown are representative of results from three independent assays. In this array, α2-3-linked sialylated di-LacNAc (3SLNLN), α2-6-linked sialylated di-LacNAc (6SLNLN), and nonsialylated di-LacNAc (LNLN) are shown.

FIG 4

Staining of chicken and human trachea tissues with VN1203 H5 mutants. Tissue staining of VN1203 T160A N224K Q226L, VN1203 K193T N224K Q226L, and VN1203 T160A K193T N224K Q226L at two different concentrations is shown. Tissue binding of HAs to either chicken or human tracheal sections was visualized by AEC staining.

FIG 5

Proposed three-dimensional model for how the N-glycan at position 158 would inhibit binding to a human-type N-glycan with relatively short LacNAc repeats. (A and B) Biantennary N-glycan with 3 LacNAc repeats (A) and 5 LacNAc repeats (B). (C and D) HA surface, in gray, with N-glycans modeled with 3 LacNAc repeats (C) and with 5 LacNAc repeats (D). The N-glycan on HA is shown with a green surface and clashes with the N-glycan with 3 LacNAc repeats in multiple adopted shapes.