Exploring the molecular epidemiology and evolutionary dynamics of influenza A virus in Taiwan

Abstract

The evolution and population dynamics of human influenza in Taiwan is a microcosm of the viruses circulating worldwide, which has not yet been studied in detail. We collected 343 representative full genome sequences of human influenza A viruses isolated in Taiwan between 1979 and 2009. Phylogenetic and antigenic data analysis revealed that H1N1 and H3N2 viruses consistently co-circulated in Taiwan, although they were characterized by different temporal dynamics and degrees of genetic diversity. Moreover, influenza A viruses of both subtypes underwent internal gene reassortment involving all eight segments of the viral genome, some of which also occurred during non-epidemic periods. The patterns of gene reassortment were different in the two subtypes. The internal genes of H1N1 viruses moved as a unit, separately from the co-evolving HA and NA genes. On the other hand, the HA and NA genes of H3N2 viruses tended to segregate consistently with different sets of internal gene segments. In particular, as reassortment occurred, H3HA always segregated as a group with the PB1, PA and M genes, while N2NA consistently segregated with PB2 and NP. Finally, the analysis showed that new phylogenetic lineages and antigenic variants emerging in summer were likely to be the progenitors of the epidemic strains in the following season. The synchronized seasonal patterns and high genetic diversity of influenza A viruses observed in Taiwan make possible to capture the evolutionary dynamic and epidemiological rules governing antigenic drift and reassortment and may serve as a "warning" system that recapitulates the global epidemic.

Figure 1. Phylogenetic relationships of the HA and NA gene segments of influenza A viruses in Taiwan.

Phylogenetic relationships of the (A) HA gene segment of H1N1 (B) NA gene segment of H1N1 (C) HA gene segment of H3N2 (D) NA gene segment of H3N2 influenza A viruses sampled from Taiwan during 1979–2009, estimated using ML method. The trees were rooted on the oldest isolate (A/Brevig Mission/1/1918 for H1N1 and A/Hong Kong 1/1/1968 for H3N2). Colored branches represent different lineages of related influenza seasons that are present in the inferred genealogies of all eight viral genome segments (Figure S2A–2F). Bootstrap values (1000 replicates) >70% are shown for key nodes. For H1N1 viruses (panel A and B), 1980–1990 isolates are colored red, 1991–1998 isolates are chartreuse, 1999–2000 isolates are green, 2001–2002 isolates are indigo, 2003–2004 isolates are light blue, 2005–2006 I isolates are orange, 2005–2006 II isolates are sky blue, 2007–2008 I isolates are pink, 2007–2008 II isolates are purple, and 2008–2009 isolates are light green. For H3N2 viruses (panel D and C): 1979–1989 isolates are cyan, 1990–1997 isolates are chartreuse, 1998–2002 isolates are pink, 2002–2003 isolates are green, 2003–2004 isolates are light blue, 2004–2005 I isolates are orange, 2004–2005 II isolates are sky blue, 2006–2007 I isolates are violet, 2006–2007 II isolates are purple, and 2008–2009 isolates are light green. Pink solid rectangle indicates 1980 H1N1 reassortant viruses, and grey solid rectangle indicates 2004 H3N2 reassortant viruses. Vaccine strains are highlighted in Italic boldface.

Figure 2. History of genetic diversity events in the evolution of influenza A Virus in Taiwan.

The eight segments shown for each virus code for the following corresponding proteins of influenza A virus (top to bottom): polymerase PB2, polymerase PB1, polymerase PA, hemagglutinin, nuclear protein, neuraminidase, matrix proteins, and nonstructural proteins. The segments of the human influenza A H1N1 and H3N2 recombined in 1980 but did not continue to circulate in the human population. H1N1 and H3N2 have co-circulated in the population of Taiwan for more than 10 years. Histograms represent the incidence of flu in Taiwan during the past epidemics, and different subtypes were colored as follow: H1N1 in blue, H3N2 in red. The patterns due to reassortment are different in H1N1 and H3N2 but the mechanism remains unknown.

Figure 3. Dynamics of genetic and antigenic properties of influenza A (H3N2) virus during 1999–2009 in Taiwan.

The influenza A viruses undergo the multiple reassortment involving all eight segments of the viral genome. Colored rectangles represent the genetic relationship and antigenicity for viruses isolated in summer, which is a non-epidemic period. This representation shows that the majority of the strains emerging in summer were closely related to (both genetically and antigenically), and were more likely to be the progenitor of the subsequent winter epidemic strains of that subtype. Histograms represent the incidence of flu in Taiwan during the past epidemics; color schemes are same as in Figure 1C.

Figure 4. Phylodynamics of influenza viruses in Taiwan.

H3N2 and H1N1 phylodynamic patterns are shown on left and right end panels, respectively. (A) Bayesian skyline plots based on HA and (B) NA sequence data. The Y-axis represents a measure of relative genetic diversity and reflects the number of effective infections established by the virus. The black line is the median posterior value; the gray lines represent the 95% HPD intervals. (C) Surveillance of influenza activity in Taiwan. The different variants detected in the epidemics are represented by the same colors used in the MCC trees below. (D) MCC trees of the HA segment. Different seasons are represented in different colors, according to the legend in the figure. Trees are scaled in units of time with tips constrained to strain sampling dates.

Figure 5. Co-evolving sites identified by Bayesian Graphical Models.

(A) Amino acid sites of the H1N1 HA and NA glycoproteins were depicted as groups showing different antigenic sites in boldface and in different colors, antigenic sites Sa, Sb, Ca2, Cb are in red, orange, purple, and green, respectively. Each connection is associated with posterior probabilities for two dependencies. The rectangle contains the H1 numbering and oval contains the N1 numbering. (B) Amino acid sites of the H3N2 HA and NA proteins were depicted as groups showing different antigenic sites in boldface and in different colors, Antigenic sites A, B, D, E are in red, orange, purple and green, respectively. Each connection is associated with posterior probabilities for two dependencies. The rectangle contains the H3 numbering and oval contains the N2 numbering.