Origin and evolution of the 1918 "Spanish" influenza virus hemagglutinin gene

Abstract

The "Spanish" influenza pandemic killed over 20 million people in 1918 and 1919, making it the worst infectious pandemic in history. Here, we report the complete sequence of the hemagglutinin (HA) gene of the 1918 virus. Influenza RNA for the analysis was isolated from a formalin-fixed, paraffin-embedded lung tissue sample prepared during the autopsy of a victim of the influenza pandemic in 1918. Influenza RNA was also isolated from lung tissue samples from two additional victims of the lethal 1918 influenza: one formalin-fixed, paraffin-embedded sample and one frozen sample obtained by in situ biopsy of the lung of a victim buried in permafrost since 1918. The complete coding sequence of the A/South Carolina/1/18 HA gene was obtained. The HA1 domain sequence was confirmed by using the two additional isolates (A/New York/1/18 and A/Brevig Mission/1/18). The sequences show little variation. Phylogenetic analyses suggest that the 1918 virus HA gene, although more closely related to avian strains than any other mammalian sequence, is mammalian and may have been adapting in humans before 1918.

Figure 1

Complete coding sequence of the HA gene of the 1918 influenza virus. The sequence for A/South Carolina/1/18 is shown with a theoretical translation of the HA1 and HA2 domains. The numbering of the nucleotide sequence is aligned to PR/8/34 and refers to the sequence of the gene in the sense (mRNA) orientation. The sequence coding for the signal peptide is underlined. The cleavage site (nucleotides 1,062–1,064) between the HA1 and HA2 domains is shown in bold. Sequence differences between this strain and the other 1918 strains are shown as double-underlined nucleotides (nucleotides 416 and 748). The sequences were confirmed by sequencing overlapping RT-PCR products and by replicate RT-PCRs for each case. The GenBank accession number is AF117241 for A/South Carolina/1/18, AF116576 for A/New York/1/18, and AF116575 for A/Brevig Mission/1/18. The theoretical translation of the gene is shown above the nucleotide sequence. Boxed amino acids indicate potential glycosylation sites as predicted by the sequence (26). Receptor-binding sites (open diamonds; ref. 23), C_b antigenic site (open circles), S_a antigenic site (closed squares), S_b antigenic site (closed diamonds), and C_a antigenic site (closed triangles; refs. and 37). Some of the receptor-binding residues are also in known antigenic sites. For these sites, symbols for both are shown.

Figure 2

Phylogenetic tree of the influenza virus HA gene segment, HA1. Sequences were aligned with lasergene software (DNAstar, Madison, WI) and analyzed for phylogenetic relationships by the NJ method with the proportion of sequence differences as the distance measure (0.1 p distance ≈ 22.35 synonymous differences). Synonymous substitutions were analyzed and bootstrap values (100 replications) are given for selected nodes. Human, swine, and avian clades are identified with large brackets. The arrow identifies the position of the 1918 sequences. A distance bar is shown below the tree. Influenza strain abbreviations used in the analyses are described in Materials and Methods and in ref. .

Figure 3

Amino acid changes in three lineages of the influenza virus HA protein segment, HA1. A phylogenetic tree was derived by using paup (17). The tree was then imported into macclade (19), and the ancestral sequences at each node of the tree were inferred. The tree shows the numbers of unambiguous changes between these sequences, with branch lengths proportional to the number of changes. Colors denote the numbers of changes along each branch and are identified by the color key at the lower right.