Determinants of glycan receptor specificity of H2N2 influenza A virus hemagglutinin

Abstract

The H2N2 subtype of influenza A virus was responsible for the Asian pandemic of 1957-58. However, unlike other subtypes that have caused pandemics such as H1N1 and H3N2, which continue to circulate among humans, H2N2 stopped circulating in the human population in 1968. Strains of H2 subtype still continue to circulate in birds and occasionally pigs and could be reintroduced into the human population through antigenic drift or shift. Such an event is a potential global health concern because of the waning population immunity to H2 hemagglutinin (HA). The first step in such a cross-species transmission and human adaptation of influenza A virus is the ability for its surface glycoprotein HA to bind to glycan receptors expressed in the human upper respiratory epithelia. Recent structural and biochemical studies have focused on understanding the glycan receptor binding specificity of the 1957-58 pandemic H2N2 HA. However, there has been considerable HA sequence divergence in the recent avian-adapted H2 strains from the pandemic H2N2 strain. Using a combination of structural modeling, quantitative glycan binding and human respiratory tissue binding methods, we systematically identify mutations in the HA from a recent avian-adapted H2N2 strain (A/Chicken/PA/2004) that make its quantitative glycan receptor binding affinity (defined using an apparent binding constant) comparable to that of a prototypic pandemic H2N2 (A/Albany/6/58) HA.

Figure 1. Glycan receptor-binding specificity of Alb58 HA.

A, shows dose-dependent direct glycan array binding of Alb58 HA which shows high affinity binding to human receptors in comparison with avian receptor binding. B, shows extensive staining of apical surface of human tracheal epithelia and observable staining of alveolar tissue section by Alb58 HA (in green) shown against propidium idodide staining (in red).

Figure 2. Glycan receptor-binding specificity of mutant forms of Alb58 HA.

Shown in A-C is the dose-dependent glycan array binding of Alb58-LG, Alb58-QG and Alb58-QS mutants respectively. A single amino acid change from Ser228→Gly (Alb58-LG mutant) leads to a loss of avian receptor binding observed in Alb58 HA. An additional Leu226→Gln mutation (on Alb58-LG) completely transforms the binding specificity by making the Alb58-QG mutant bind predominantly to avian receptors. Alb58-QS mutant shows loss of both avian and human receptor binding. D shows homology based structural model of Alb58-QS mutant (RBS part is shown as a cartoon in beige) with the human receptor. Both the Leu226 and Gln226 side chains are marked. The Gln226 in the mutant is positioned to interact with Ser228 hence making the 226 position less favorable for contacts with both human and avian receptors.

Figure 3. Glycan receptor-binding specificity of CkPA04 HA.

A, shows dose-dependent direct glycan array binding of CkPA04 HA which shows high affinity binding to avian receptors in comparison with human receptors. B, shows extensive alveolar staining and minimal staining of apical surface of the human tracheal epithelia by CkPA04 HA (in green) shown against propidium idodide staining (in red).

Figure 4. Homology-based structural model of HA-glycan receptor complexes.

A, stereo view of the RBS (shown as cartoon in cyan) of CkPA04 HA – avian receptor structural complex constructed using co-crystal structure of A/Chicken/NY/91-avian receptor (PRB ID: 2WR2) as a template. The resolved coordinates of the avian receptor (Neu5Acα2→3Galβ1→3GlcNAc) are shown using a stick representation (in green). B, stereo view of RBS (shown as cartoon in gray) of Alb58 HA – human receptor complex constructed using co-crystal structure of A/Singapore/1/57– human receptor (PDB ID: 2WR7) as the template. The resolved coordinates of the human receptor (Neu5Acα2→6Galβ1→4GlcNAcβ1→3Gal) are shown using a stick representation (in orange). The side chains of the key residues involved in interaction with glycan receptor are shown and labeled. The residues in the RBS that differ between CkPA04 and Alb58 HA are labeled in red.

Figure 5. Glycan receptor-binding specificity of mutant forms of CkPA04 HA.

Shown in the figure is the dose-dependent glycan receptor binding (A, C, E) and human tissue binding (B, D, F) of CkPA04-LS, CkPA04-TLS and CkPA04-RTLS mutants respectively. All the mutants show substantial improvement in the human receptor binding and reduction in avian receptor binding in comparison to the WT CkPA04 HA as observed in both the glycan array tissue-binding experiments.

Figure 6. Glycan receptor-binding affinities of the mutant forms of CkPA04 HA.

A, shows the theoretical binding curves (with the apparent binding constant K_d') that depict the differences in the binding affinity of the WT and mutant H2N2 HAs to the representative avian receptor (3′SLN-LN). B, shows the theoretical binding curves that depict the differences in the binding affinity of the WT and mutant H2N2 HAs to the representative human receptor (6′SLN-LN). The range of K_d' values (3–8 pM) is shown for CkPA04-TLS, Alb58 and CkPA04-RTLS that is contrasted with the K_d' value of CkPA04-LS. The binding curves were generated by fitting to the Hill equation (see Methods) and plotting the theoretically calculated fractional saturation (y-axis) against HA concentration (x-axis). The n value for all the binding events is around 1.3.