Rapid Emergence of Highly Pathogenic Avian Influenza Subtypes from a Subtype H5N1 Hemagglutinin Variant

In 2014, novel highly pathogenic avian influenza A H5N2, H5N5, H5N6, and H5N8 viruses caused outbreaks in Asia, Europe, and North America. The H5 genes of these viruses form a monophyletic group that evolved from a clade 2.3.4 H5N1 variant. This rapid emergence of new H5Nx combinations is unprecedented in the H5N1 evolutionary history.


The Study
To determine the evolutionary history of the HA proteins of these novel HPAI subtypes, we collected all HPAI H5 coding region sequences for all subtypes, except H5N1, and then aligned them with 850 H5N1 HA sequences representing all HPAI H5N1 clades (selected from ≈5,000 total sequences) and constructed a phylogenetic tree ( Figure  1). Reassortment events leading to the generation of novel H5Nx subtypes are almost uniquely restricted to a single branch of the tree; the branch contains all isolates of the recent HPAI H5N2, H5N5, H5N6, and H5N8 outbreaks. The only other H5Nx reassortants that have been identified are a limited number of H5N2 subtype isolates that are present in 5 other branches of the tree.
A more detailed analysis ( Figure 2)  Within the subtree shown in Figure 2, the H5N2 viruses are present in 2 branches. N2 of A/duck/Jiangsu/ m234/2012 was derived from an H11N2 virus (14); the N2 of the other viruses in this branch were derived from an avian H3N2 virus (15). In addition, 12 H5N1 reassortants were found to be spread over different branches of the subtree ( Figure 2 (1), suggesting that independent reassortment events have taken place. The H5N8 and the H5N6 viruses have segregated into 2 branches. Analysis of the N6 proteins (data not shown) indicates that the H5N6 viruses (all from southern China) are the result 2 independent reassortment events with avian H6N6 strains. Unfortunately, sequences from the recent H5N6 outbreaks in Vietnam (5) are not yet present in the databases. The 2 different H5N8 virus clusters most likely evolved from a single H5N8 reassortant virus (A/duck/Jiangsu/k1203/2010) that was isolated in China in 2010 (13). Both clusters were identified in Korea in 2014, whereas members of the most evolved cluster were detected later in 2014 in Japan, Germany, the Netherlands, and the United Kingdom.

Conclusion
Since 1996, reassortment events involving H5N1 HPAI viruses have, as far as detected, only rarely led to the generation of new H5Nx subtypes. The 2008 generation of an H5N5 reassortant virus (prototype A/duck/Guangdong/ wy/24/2008) represents the creation of a new HPAI virus that has led to the generation of a range of novel H5Nx reassortants that acquired novel NA proteins (H5N2, H5N6, and H5N8). The H5N6 reassortant became of particular concern after spreading over a wide geographic area in Southeast Asia and causing a fatal human infection in China (5). Meanwhile, the H5N8 subtype spread to Europe in November 2014, resulting in large economic losses in the poultry industry. On the basis of reports from the World Organisation for Animal Health, H5N8 and H5N2 viruses were detected in Canada and the United States in December 2014.
In this study, we exclusively focused on the unique occurrence of new HA-NA combinations. Recent publications have already described the reassortment events of the internal gene segments of several of the viruses mentioned above (6)(7)(8)(11)(12)(13)(14). In contrast to novel HA-NA combinations, novel constellations of internal gene segments are far from unique and have frequently been  (9), we aligned the coding region sequences for 89 HPAI H5 HA subtype viruses, excluding H5N1, with those for 850 H5N1 HA viruses representing all HPAI H5N1 clades (1); the 89 H5 HA sequences were identified in the NCBI Influenza Virus Resource (10) and the GISAID EpiFlu Database (http://www.gisaid.org). A phylogenetic tree was constructed by using the PHYLIP Neighbor Joining algorithm using the F84 distance matrix (http://www.ncbi. nlm.nih.gov/genomes/FLU/DatasetExplorer/fluPage.cgi?pageInclude=References.inc#PHYLIP). The number of sequences present in a branch is indicated between brackets. Stars indicate the branches that contain subtypes other than H5N1. The genotypes (H5N2, H5N5, H5N6, and H5N8) and their numbers of occurrence in a particular branch are indicated at right. Scale bar indicates evolutionary distance (nucleotide substitutions per site). Details for GISAID-derived sequences are shown in the Table. observed for HPAI H5N1 viruses (3). Our analysis indicates that new HPAI viruses have emerged that carry H5 proteins capable of matching with multiple NA subtypes. Whether the formation of new HA-NA combinations confers a selective advantage that contributed to the emergence of these novel subtypes is not known and requires elaborate research. However, the balance between HA (receptor binding) and NA (receptor cleavage) protein activities is known to be critical to cell entry and host tropism and may be an important factor that lead to the emergence of new HA-NA combinations. In contrast to HPAI H5N1, the novel clade 2.3.4.4 viruses, excluding H5N6 viruses, have not caused human infections. However, it is unknown to what extent the repeated acquisition of a new NA proteins could enhance the rate of evolution of the HA protein. Obviously such changes could further affect host and tissue specificity, potentially having serious consequences. Therefore, surveillance is required to monitor further spread, evolution, and potential changes in host range.

Figure 2.
Hemagglutinin protein tree (neighbor-joining, point accepted mutation distance matrix model) of subtypes present in branch descending from highly pathogenic avian influenza A(H5N1) cluster 2.3.4 (see Figure 1). MUSCLE (9) was used to align protein sequences. Subtype group positions are indicated at right. Bootstrap values (n = 1,000) at key nodes are indicated. Scale bar indicates evolutionary distance (nucleotide substitutions per site).