Codon usage bias and the evolution of influenza A viruses. Codon Usage Biases of Influenza Virus

Abstract

Background: The influenza A virus is an important infectious cause of morbidity and mortality in humans and was responsible for 3 pandemics in the 20th century. As the replication of the influenza virus is based on its host's machinery, codon usage of its viral genes might be subject to host selection pressures, especially after interspecies transmission. A better understanding of viral evolution and host adaptive responses might help control this disease.

Results: Relative Synonymous Codon Usage (RSCU) values of the genes from segment 1 to segment 6 of avian and human influenza viruses, including pandemic H1N1, were studied via Correspondence Analysis (CA). The codon usage patterns of seasonal human influenza viruses were distinct among their subtypes and different from those of avian viruses. Newly isolated viruses could be added to the CA results, creating a tool to investigate the host origin and evolution of viral genes. It was found that the 1918 pandemic H1N1 virus contained genes with mammalian-like viral codon usage patterns, indicating that the introduction of this virus to humans was not through in toto transfer of an avian influenza virus.Many human viral genes had directional changes in codon usage over time of viral isolation, indicating the effect of host selection pressures. These changes reduced the overall GC content and the usage of G at the third codon position in the viral genome. Limited evidence of translational selection pressure was found in a few viral genes.

Conclusions: Codon usage patterns from CA allowed identification of host origin and evolutionary trends in influenza viruses, providing an alternative method and a tool to understand the evolution of influenza viruses. Human influenza viruses are subject to selection pressure on codon usage which might assist in understanding the characteristics of newly emerging viruses.

Figure 1

CA of human (seasonal H1-H3 and H5) and avian influenza viruses. Each viral gene is displayed in a 3-dimensional representation. The X, Y and Z axes are in arbitrary scales generated by the CA and the weight of each codon in these axes varies in different segments. (A) Human influenza colored by subtype (B) seasonal human H1 subtype colored by year of isolation and (C) seasonal human H3 subtype colored by year of isolation. The codon usage trends with time of viral isolation are indicated by arrows. The orientations of NA of H1 and HA of H3 in the graphs were altered for better presentation.

Figure 2

CA of human (seasonal H1-H3, H5 and pandemic H1/09), avian and swine influenza viruses. The viral hosts are differentiated by color. Viral genes derived from A/Brevig Mission/1/1918 (1918) and pandemic H1N1/2009 (pdm H1N1/2009) are indicated by arrows. The X, Y and Z axes are codon usage and are in arbitrary scales generated by the CA for each segment.

Figure 3

Condon usage of influenza virus in the context of human coding sequences. RSCU values of all six human and avian viral segments and human RefSeq mRNA sequences were subjected to a joint CA. (A) CA of all influenza viral and human mRNA. (B) and (C) are subsets of the data shown in (A). (B) Human and avian influenza viral datasets by gene extracted from panel A. (C) Seasonal human H1 (left) and seasonal human H3 (right) datasets extracted from panel A. The unidirectional trend on the X coordinates found in the H1 (PB2, PB1, PA, NP and NA) and H3 (PB2, HA, NP) genes is indicated by an arrow. Examples of outliers (e.g. H1N2) are marked by broken blue circles.

Figure 4

Trends in correlation between viral and human codon usage by year of viral isolation. (A) Correlation trend for human H1N1 and H3N2 subtypes and avian influenza. (B and C) Correlation of codon usage between H1N1 (B) or H3N2 (C) viral genes and genes expressed in human bronchial epithelial cells by year of virus isolation. The linear regression line and the correlation coefficient of each dataset are shown.

Figure 5

tRNA adaptation index of influenza viruses. The tAI of each viral gene is shown by year of virus isolation. Viral host and subtypes are indicated by color.