Army of bird flu viruses decoded
Genetic data highlight a protein that could be key to flu pandemics.
The mass decoding of nearly 170 strains of avian influenza has pinpointed a particular protein region that may help the H5N1 virus to kill with a vengeance.
Human pandemics of influenza are all thought to have stemmed from bird flu viruses, and virologists are desperate to know why some viruses prove more lethal than others. The viruses' genetic codes could reveal such secrets but, until now, scientists have collected only a few complete flu genome sequences.
Now, Clayton Naeve of St Jude Children's Research Hospital in Memphis, Tennessee, and his colleagues have doubled the amount of genetic information available on bird flu viruses in one swoop. The team used a bank of around 7,000 viral samples from ducks, gulls, poultry and other birds from around the globe. These were collected by Robert Webster, an infectious diseases researcher also at the St Jude hospital, over the past three decades.
The team deduced the genetic code of 169 complete flu virus genomes and many more partial sequences, and report their findings in the journal Science1. They compared each sequence with every other sequence, and with previously determined sequences, to pull out key differences that might explain why some viruses are more virulent than others.
The researchers showed that one key viral protein known as NS1, which is manufactured inside infected cells and blocks their anti-viral response, could explain why some viruses are so damaging to humans. This NS1 protein tends to have a specific form in bird flu viruses, and a different form in most human flu viruses.
The team showed that samples of the H5N1 virus that killed humans in Asia between 1997 and 2004 had the 'bird' form of NS1. And the strain that caused the terrible 1918 pandemic had a similar, but unique, bird form of the protein. By contrast, the viruses that triggered milder human pandemics in 1957 and 1968 possessed a more 'human' form.
The group additionally showed that the bird form of NS1 can bind to 30 human proteins, whereas the human form could bind very few of those they tested. The researchers propose that this helps the virus interfere with these proteins and harm the cells.
There are 11 proteins known to exist in flu viruses. Previously, research into virulence had been largely focussed on the haemagglutinin and neuraminidase proteins (this is what the H and N stand for in H5N1), which sit on the surface of the virus and help it to invade cells. Researchers suspect that other proteins are involved; now they have pinned down the important role of NS1.
Albert Osterhaus, a virologist at the Erasmus University in Rotterdam who is planning a similar large-scale sequencing project in the Netherlands, calls the paper "highly significant". "The data and analysis give us a very clear virulence marker which turns out to be quite universal," he says.
The researchers also devised an easier way to compare bird flu viruses with one another, a method they call 'proteotyping'. Instead of comparing raw genetic information, they translated this into information about proteins, and then used this to sort viruses into a kind of evolutionary tree. This helps researchers to track how virulence to humans evolved. "The new methodology of proteotyping is very powerful," Osterhaus says.
The study does not answer the big outstanding question about the current outbreak of H5N1: will the virus acquire the ability to jump easily between people?
But the researchers hope that further scrutiny of the data, which they have placed in databases that are accessible to all scientists, will help to answer this. "We hope other investigators will start to mine this data," Naeve says.
Post a comment to this story by visiting our newsblog.
- Obenauer J.C., et al. Science10.1126 /science.1121586 (2006).