Receptor specificity and transmission of H2N2 subtype viruses isolated from the pandemic of 1957

Abstract

Influenza viruses of the H2N2 subtype have not circulated among humans in over 40 years. The occasional isolation of avian H2 strains from swine and avian species coupled with waning population immunity to H2 hemagglutinin (HA) warrants investigation of this subtype due to its pandemic potential. In this study we examined the transmissibility of representative human H2N2 viruses, A/Albany/6/58 (Alb/58) and A/El Salvador/2/57 (ElSalv/57), isolated during the 1957/58 pandemic, in the ferret model. The receptor binding properties of these H2N2 viruses was analyzed using dose-dependent direct glycan array-binding assays. Alb/58 virus, which contains the 226L/228S amino acid combination in the HA and displayed dual binding to both alpha 2,6 and alpha 2,3 glycan receptors, transmitted efficiently to naïve ferrets by respiratory droplets. Inefficient transmission was observed with ElSalv/57 virus, which contains the 226Q/228G amino acid combination and preferentially binds alpha 2,3 over alpha 2,6 glycan receptors. However, a unique transmission event with the ElSalv/57 virus occurred which produced a 226L/228G H2N2 natural variant virus that displayed an increase in binding specificity to alpha 2,6 glycan receptors and enhanced respiratory droplet transmissibility. Our studies provide a correlation between binding affinity to glycan receptors with terminal alpha 2,6-linked sialic acid and the efficiency of respiratory droplet transmission for pandemic H2N2 influenza viruses.

Figure 1. Respiratory droplet and direct contact transmission experiments using avian H2N2 virus.

For each experiment, three ferrets were inoculated intranasally (i.n.) with 10⁶ EID₅₀ of Mallard/78 virus and placed in individual cages. A: Twenty four hours later, one naïve ferret was housed in a cage adjacent to each of the cages containing an inoculated ferret (respiratory droplet transmission), separated by perforated walls to allow the exchange of respiratory droplets. B: For direct transmission experiments, one naïve ferret was housed in the same cage with the inoculated ferret. Nasal wash samples were collected from the inoculated and contact ferrets on alternating days and titered in eggs. Bars with the same color represent an inoculated/contact ferret pair. The limit of virus detection was 10^1.5 EID₅₀/ml.

Figure 2. Respiratory droplet and direct contact transmission experiments using human H2N2 viruses.

For each experiment, three ferrets were inoculated i.n. with 10⁶ EID₅₀ of Albany/58, El Salv/57 or El Salv/57-Q226L mutant. Each ferret was placed in individual cages. A, C and E: Twenty four hours later, one naïve ferret was housed in a cage adjacent to each of the cages containing an inoculated ferret (respiratory droplet transmission), separated by perforated walls to allow the exchange of respiratory droplets. B and D: For direct transmission experiments, one naïve ferret was housed in the same cage with the inoculated ferret. Nasal wash samples were collected from the inoculated and contact ferrets on alternating days and titered by EID₅₀. Bars with the same color represent an inoculated/contact ferret pair. The limit of virus detection was 10^1.5 EID₅₀/ml. †, ferret found dead on day 8 p.i.

Figure 3. Structural model of human H2N2 viruses with α2,6 receptor.

A, Stereo view of homology based structural model of Alb/58 HA complexed with α2,6 oligosaccharide. This model was constructed using a representative human H2N2 HA – α2,6 receptor co-crystal structure (PDB ID: 2WR7) as template. The optimal contact between the critical Leu226 with the C6 atom of SA α2,6-linked to Gal motif is shown with dashed line. B, Stereo view of homology based structural model of ElSalv/57 HA complexed with α2,6 oligosaccharide. This model was constructed using an avian-like H2N2 HA (PDB ID: 3KU3) as template. The main differences in contacts between A and B are in positions 226 and 228 that make contacts with SA α2,6-linked to the penultimate Gal and in 156 position that makes contacts with additional sugars at the reducing end of penultimate Gal. The sugar is shown as a stick representation colored by atom (carbon in orange, oxygen in red and nitrogen in blue). The homology based structural models were constructed as described previously .

Figure 4. Glycan binding of Alb/58, ElSalv/57, and A/El Salv/57-Q226L viruses.

The binding signals are expressed as percentage of maximum binding of each virus to the glycan arrays. At saturating HA titers, both Alb/58 and A/El Salv/57-Q226L showed binding signals to long and short α2-6 glycans, but Alb/58 also showed high binding signals to long and short α2-3 glycans. ElSalv/57 virus showed binding affinity to all three types of α2-3 glycans but also displayed some binding to long α2-6 glycans at higher titers. The types of glycans used on the arrays are described (Table S2).