Antigenic Pressure on H3N2 Influenza Virus Drift Strains Imposes Constraints on Binding to Sialylated Receptors but Not Phosphorylated Glycans

Abstract

H3N2 strains of influenza A virus emerged in humans in 1968 and have continued to circulate, evolving in response to human immune pressure. During this process of "antigenic drift," viruses have progressively lost the ability to agglutinate erythrocytes of various species and to replicate efficiently under the established conditions for amplifying clinical isolates and generating vaccine candidates. We have determined the glycome profiles of chicken and guinea pig erythrocytes to gain insights into reduced agglutination properties displayed by drifted strains and show that both chicken and guinea pig erythrocytes contain complex sialylated N-glycans but that they differ with respect to the extent of branching, core fucosylation, and the abundance of poly-N-acetyllactosamine (PL) [-3Galβ1-4GlcNAcβ1-]_n structures. We also examined binding of the H3N2 viruses using three different glycan microarrays: the synthetic Consortium for Functional Glycomics array; the defined N-glycan array designed to reveal contributions to binding based on sialic acid linkage type, branched structures, and core modifications; and the human lung shotgun glycan microarray. The results demonstrate that H3N2 viruses have progressively lost their capacity to bind nearly all canonical sialylated receptors other than a selection of biantennary structures and PL structures with or without sialic acid. Significantly, all viruses displayed robust binding to nonsialylated high-mannose phosphorylated glycans, even as the recognition of sialylated structures is decreased through antigenic drift.IMPORTANCE Influenza subtype H3N2 viruses have circulated in humans for over 50 years, continuing to cause annual epidemics. Such viruses have undergone antigenic drift in response to immune pressure, reducing the protective effects of preexisting immunity to previously circulating H3N2 strains. The changes in hemagglutinin (HA) affiliated with drift have implications for the receptor binding properties of these viruses, affecting virus replication in the culture systems commonly used to generate and amplify vaccine strains. Therefore, the antigenic properties of the vaccines may not directly reflect those of the circulating strains from which they were derived, compromising vaccine efficacy. In order to reproducibly provide effective vaccines, it will be critical to understand the interrelationships between binding, antigenicity, and replication properties in different growth substrates.

Keywords: antigenic drift; influenza; phosphorylated glycans; receptor binding; sialylated glycans.

FIG 1

N-glycan profiles of chicken and guinea pig erythrocytes. The N-glycans on chicken or guinea pig erythrocytes were released by the use of sodium hypochlorite and analyzed by MALDI-MS. The glycan compositions were deduced based on the m/z value and are indicated at the top of the spectrum. Blue box, N-acetylglucosamine; green circle, mannose; yellow circle, galactose; purple diamond, N-acetylneuraminic acid; light blue diamond, N-glycolylneuraminic acid; red triangle, fucose.

FIG 2

Receptor binding profile of H3N2 drift strains isolated from 2001 to 2015 on defined N-glycan array (27). The profiles for the drift strains are shown with 2,3-Sia structures highlighted in blue and 2,6-Sia structures highlighted in green. Glycans corresponding to chart IDs 17 to 32 (i.e., the Sia-terminating structures) are shown below the graphs. Error bars represent standard deviations of data corresponding to binding to replicate microarray spots. See Fig. 1 legend for cartoon key.

FIG 3

Drift viruses binding to CFG arrays. (A) Each virus was fluorescently labeled and interrogated on CFG array version 5.3. The binding profiles represent the relative intensity of the fluorescence for each glycan structure, indicating virus recognition and binding. The chart IDs have been ordered such that the sialylated glycans are grouped in chart IDs 1 to 153. (B) The top 10 structures bound by each individual virus are plotted by frequency (the number of viruses that bind each structure) and average rank (the percentage of binding for that structure for each applicable virus) and identified by chart ID. The glycan structures corresponding to the highest rank and frequency IDs are shown in the chart. See Fig. 1 legend for cartoon key.

FIG 4

Glycan microarray comparison of drifted H3N2 strains and H3N2 NY04 and H1N1 Penn08 and Bris07 strains on CFG array. The left panel shows a correlation map encompassing binding to all structures present on the array. The right panel contains a force graph, where the color nodes indicate the virus strains and the gray nodes represent individual glycans. Binding is visualized by connections via the circles and nodes, with higher relative fluorescence units equating to shorter distance. The glycans with Neu5Ac (sialic acid) structures are highlighted in lime green. Example cartoons are included for structures bound only by the drift strains, bound by all strains, and not bound. The strain colors are consistent between the two panels.

FIG 5

Glycan recognition of H3N2 drift strains on the HL-SGM. (A) Numbers of relative fluorescence units per glycan fraction are represented on a sliding color scale with lighter colors corresponding to greater binding intensity. (B) The receptor binding profile of each drift strain is present along with the profiles of sialic acid-recognizing lectins SNA and MAL-I. The binding of the lectins is highlighted in a blue box, which is transposed on the profiles of the viruses to indicate Sia-terminating structures. Error bars represent standard deviations of data corresponding to binding to replicate microarray spots.

FIG 6

Glycans present in high-binding sialylated fractions and phosphorylated fractions on HL-SGM. For both the sialylated and phosphorylated glycan groups, the top three binding fractions shared by all drift viruses on the HL-SGM were analyzed for glycan content. Glycan structures were deduced from the corresponding m/z value. Some glycans appear in two or all of the top three fractions. See Fig. 1 legend for cartoon key.

FIG 7

Binding of clinical isolate viruses (H3N2) collected in 2018 to glycan microarrays. (Left panel) Three strains (denoted L, M, and N) (Table 1) were tested on the N-glycan array (top row), the CFG array (middle row), and the HL-SGM (bottom row). Sialylated glycans are present in IDs 17 to 32 of the N-glycan array (See Fig. 2 for cartoons). The CFG glycans have been ordered such that sialylated glycans are listed first in chart IDs 1 to 153. Phosphorylated glycans populate chart IDs 1 to 52, and sialylated glycans correspond to chart IDs 53 to 120 on the HL-SGM. In all charts, sialylated glycans are highlighted in blue boxes. Error bars represent standard deviations of data corresponding to binding to replicate microarray spots.