Two simple changes to the virus that caused the 1918 influenza pandemic take away its ability to jump between mammals in the laboratory. Scientists say the resulting virus is more like a bird flu that is unable to spread after taking hold in a mammal's upper airways. The finding does not reveal whether or how exactly the feared H5N1 bird flu virus would make a reverse set of changes and begin hopping between humans, but it underscores the idea that a certain type of biological change may be crucial in causing pandemics.

"You render the most pathogenic influenza virus known to humans nontransmissible—that's fascinating," says bioengineer Ram Sasisekharan of the Massachusetts Institute of Technology, who was not part of the study.

Researchers from the Centers for Disease Control and Prevention (CDC) and their colleagues precisely mutated a protein called hemagglutinin (HA), which sits on the surface of a flu virus and plays a major role in infection by recognizing receptor proteins on host cells. The unaltered hemagglutinin is akin to a key that unlocks a so-called receptor molecule in cells lining the upper airways of humans and other mammals. The changes were designed to make this key fit a new lock—a receptor that is relatively sparse in the upper airways of humans but is found deep in our lungs and in the airways and intestines of birds.

Needing proxies for humans, the researchers infected caged ferrets with either the original 1918 flu virus (reconstructed two years ago) or the altered form. A day later they placed the cages of the infected ferrets next to healthy animals to see if the virus would spread through the air. "What we notice with the 1918 virus with the natural HA on it is the virus spread very efficiently," says Terrence Tumpey, a senior microbiologist at the CDC and first author of a report published online February 1 by Science. In contrast, the group found that the altered virus did not spread at all.

"It clearly demonstrates the importance of the [HA] switch to transmission," Sasisekharan says. Animal models come with some uncertainty, cautions biochemist James Stevens of the Scripps Research Institute in La Jolla, Calif., but he says "it's probably the hemagglutinin switch which is going to be one of the critical events" in a future pandemic.

The result fits with prior research showing that H5N1 tends to infect the deep lungs of its victims, where it has a poor launching pad into the air. Tumpey says the big question is which changes, if any, would allow H5N1 bird flu to morph into a pandemic strain. Not all viruses are alike in their hemagglutinin proteins. "H5" refers to one distinct form; the 1918 virus is an H1.

In earlier work, Stevens and his co-workers altered the equivalent amino acids in H5N1 that Tumpey's group did for 1918 flu, but the virus failed to acquire a preference for the human upper airway receptor. Many public health experts nonetheless fear that H5N1 could become a pandemic strain through other changes. "If we know the routes it might take, we can actually look out for those routes," Stevens says.

Tumpey says it is unclear how exactly the hemagglutinin switch in the 1918 virus cancels its transmission. "There's virus up there in the upper respiratory tract, so the question is why isn't it being transmitted."