Genetic analysis 'doubles' data on avian flu viruses

Feb 7, 2006 (CIDRAP News) – US researchers recently published preliminary findings from their sequencing of 336 avian influenza (AI) viruses collected around the world, an effort they say has doubled the amount of public genetic data on avian flu viruses.

The researchers from St. Jude Children's Hospital in Memphis also say their analysis has uncovered one possible clue to the reasons for the severe illness and high death rate in humans infected with the H5N1 avian flu virus.

The team, led by senior author Clayton W. Naeve, PhD, used a library of roughly 11,000 flu viruses, including about 7,000 avian flu viruses, collected over three decades by virologist Robert G. Webster, PhD, a member of the group.

The 336 viruses they analyzed were gathered between 1976 and 2004 from ducks, gulls, shorebirds, and poultry in North America, Eurasia, and Australasia. Their analysis covers 2,196 new AI gene sequences and 169 complete AI virus genomes. The samples include all 25 known types of the two flu virus surface proteins, hemagglutinin and neuraminidase.

"This information is a true gold mine, and we are inviting all of the miners to help us unlock the secrets of influenza," Webster commented in a St. Jude news release.

The team supplemented their analysis by including 2,143 AI virus gene sequences already available from GenBank, for a total of 4,339 genes. AI viruses have eight gene segments that code for 11 known proteins, the report notes. The authors developed phylogenetic trees for each of the eight gene segments to show their relationships across different virus isolates.

To examine the variability of the viruses more closely, the authors developed a technique they call proteotyping, in which they identified gene variants with unique amino acid signatures. "Using this approach has allowed us to observe for the first time viruses sharing not only specific genes but genes coding for specific proteotypes," the report says.

Hemagglutinin and neuraminidase are highly variable because they are the main targets of the host immune system and constantly evolve to evade it, the report notes. But the researchers also found that the NS gene, which codes for two nonstructural proteins, called NS1 and NS2, is also highly variable. NS1 appears only in infected host cells, where it regulates many cell functions during infection and, according to considerable evidence, plays a role in virulence by blocking certain host antiviral genes, the article says.

The team identified a particular ligand, or amino acid cluster that binds to other molecules, on one end of the NS1 molecule. By examining 1,196 NS1 sequences from avian, human, swine, and equine flu virus samples, the researchers discovered that all the avian samples had a certain amino acid sequence in this particular ligand, while most of the human samples had a different sequence.

Thirty-eight of the human samples had sequences matching those found in the avian viruses, and most of those 38 were recent H5N1 viruses, the report says. The authors concluded that the human H5N1 cases have involved NS1 proteins with "avian"-type sequences at the identified ligand, whereas flu viruses from low-mortality human pandemics (1957 and 1968) had typically "human" sequences at that location.

The report briefly describes two experiments the authors conducted to compare the ability of the avian and human NS1 proteins to bind to certain proteins, called PDZ domains, involved in many host cell processes. The results indicated that the avian NS1 protein is much more apt to bind to PDZ domains than the human NS1 is.

"While the molecular consequences of these interactions are as yet unknown, it appears that avian NS1 proteins when introduced into human cells have the opportunity to bind to and disrupt many PDZ–domain protein mediated pathways that the human NS1 protein cannot," the authors write.

They suggest that disruption of these pathways may contribute to the high death rates in human H5N1 cases, though other genes and gene products are clearly involved. Further, they suggest, the NS proteins "may prove valuable as targets for antiviral therapy."

Virologist Yoshihiro Kawaoka of the University of Tokyo and Ujiversity of Wisconsin, Madison, said animal experiments are needed to assess the effects of the NS gene variations on pathogenicity, according to a Science news story accompanying the report. But he called the report a good example of using sequence information to develop new hypotheses.

Obenauer JC, Denson J, Mehta PK, et al. Large-scale sequence analysis of avian influenza isolates. Science 2006 Mar 17;311(5767):1576-80 (Epub 2006 Jan 26) [Abstract]

Newsletter Sign-up

Get CIDRAP news and other free newsletters.

Sign up now»


Unrestricted financial support provided by

Bentson Foundation 3M Gilead 
Grant support for ASP provided by

  Become an underwriter»