Avian Influenza A Viruses Reassort and Diversify Differently in Mallards and Mammals

Abstract

Reassortment among co-infecting influenza A viruses (IAVs) is an important source of viral diversity and can facilitate expansion into novel host species. Indeed, reassortment played a key role in the evolution of the last three pandemic IAVs. Observed patterns of reassortment within a coinfected host are likely to be shaped by several factors, including viral load, the extent of viral mixing within the host and the stringency of selection. These factors in turn are expected to vary among the diverse host species that IAV infects. To investigate host differences in IAV reassortment, here we examined reassortment of two distinct avian IAVs within their natural host (mallards) and a mammalian model system (guinea pigs). Animals were co-inoculated with A/wildbird/California/187718-36/2008 (H3N8) and A/mallard/Colorado/P66F1-5/2008 (H4N6) viruses. Longitudinal samples were collected from the cloaca of mallards or the nasal tract of guinea pigs and viral genetic exchange was monitored by genotyping clonal isolates from these samples. Relative to those in guinea pigs, viral populations in mallards showed higher frequencies of reassortant genotypes and were characterized by higher genotype richness and diversity. In line with these observations, analysis of pairwise segment combinations revealed lower linkage disequilibrium in mallards as compared to guinea pigs. No clear longitudinal patterns in richness, diversity or linkage disequilibrium were present in either host. Our results reveal mallards to be a highly permissive host for IAV reassortment and suggest that reduced viral mixing limits avian IAV reassortment in a mammalian host.

Keywords: avian influenza; mallards; mammals; reassortment; viral diversity.

Figure 1

Efficient viral replication was observed in mallards and guinea pigs. Infectious viral titers in (A) cloacal swab samples collected from mallards and (B) nasal lavage samples collected from guinea pigs are plotted against day post-inoculation. Each colored line represents an individual animal. Horizontal dashed line shows the limit of detection.

Figure 2

Viral populations in mallards showed lower frequencies of parental genotypes, higher genotype richness and higher diversity than those in guinea pigs. Results from mallards are shown in panels (A–D) and from guinea pigs in panels (E–H) Stacked plot showing frequencies of unique genotypes detected in one representative animal over time. The lowermost two sections, in blue and orange, represent the H3N8 and H4N6 parental genotypes, respectively. (B,F) Frequency of parental genotypes over time. The total frequency of H3N8 and H4N6 parental genotypes is plotted. (C,G) Genotype richness over time. (D,H) Diversity over time, as measured by the Shannon-Weiner index.

Figure 3

Lower levels of viral diversity in guinea pigs are only partially accounted for by high frequency of parental genotypes. (A) Scatter plot of Shannon-Weiner diversity against parental genotype frequency. (B) Histograms showing evenness of virus samples in mallards. (C) Histograms showing evenness of virus samples in mallards, following removal of parental genotypes. (D) Histograms showing evenness of virus samples in guinea pigs. (E) Histograms showing evenness of virus samples in guinea pigs, following removal of parental genotypes.

Figure 4

Pairwise linkage disequilibrium was typically lower in mallards than in guinea pigs. Linkage disequilibrium for each pairwise combination of influenza A virus (IAV) gene segments is shown. Each data point represents a mallard or guinea pig, with samples from all three time points aggregated for a given individual. p values indicate results of a Mann Whitney U test.

Figure 5

Genotype patterns detected in mallards and guinea pigs are consistent with random sampling of gene segments. (A) Number of theoretically possible unique genotypes in each category (brown) and the number detected in mallards (olive) and guinea pigs (grey). Data from all time points and all individuals of a given host species are combined. Number of unique genotypes is displayed above each bar. (B) Proportion of unique genotypes in each category that were detected. Observed results in mallards (solid, olive line) and guinea pigs (solid, grey line) are compared to simulated data sets generated by randomly sampling segments from reassortant viruses detected in each species (dashed lines). Error bars on the simulated data show the 95% range of calculated proportions from the 1000 simulated data sets.