Editor’s Note: This is part one of a two-part series on avian flu. Part two will run in Friday’s issue.

As the bird flu virus H5N1 continues to alarm public health officials, as well as many civilians, two new studies shed light onto some of the reasons why it hasn’t yet began human-to-human transmission.

In the March 16, 2006, issue of Sciencexpress - the early online edition of Science - researchers showed that mutations in only two amino acids could enable one particular strain of H5N1 to infect humans more easily. But the change from birds to humans was far less steep than those seen in two other hemagglutinins, H1 and H3, which were the hemagglutinin types on the viruses that caused the 1918 "Spanish flu" and 1968 pandemics, respectively.

And in today’s issue of Nature, a second group of scientists, from the University of Wisconsin-Madison and the University of Tokyo, report a reason why H5N1 currently is not able to spread efficiently from human to human: The cell surface receptors it prefers are so deep in the lungs that a mere sneeze from an infected person will not bring the virus into daylight, or to the nasal passages of the next unsuspecting victim.

In the Sciencexpress paper, the researchers found that introducing mutations in the H5N1 strain’s hemagglutinin at amino acids 226 and 228 caused "an increase in H5N1’s ability to potentially bind to human receptors," first author James Stevens, senior research associate at the Scripps Research Institute, told BioWorld Today.

But Jeffrey Taubenberger, co-author on the paper, pointed out that the shift was "not as clean or dramatic" as it is in the case of H1 and H3 hemagglutinins.

The influenza virus uses its hemagglutinin (the ‘H’ in viral nomenclature) protein to enter cells. During the entry process, the hemagglutinin protein is cut into two pieces by cellular proteases.

The researchers crystallized the cut hemagglutinin of one highly pathogenic strain of H5N1, isolated from a Vietnamese boy who died of the bird flu in 2004, and studied its structure. In 2004, the team had reported the crystal structure of the 1918 pandemic strain’s hemagglutinin, and the scientists compared the two structures, as well as the hemagglutinin structure of a less pathogenic H5N1 strain isolated from a duck in 1997 and a fourth strain containing H3 subtype hemagglutinin.

Though the two H5 hemagglutinins were more similar to each other in terms of amino acid sequence, the pathogenic H5 strain structurally was most similar to the 1918 pandemic strain. In the studies published on the resurrected 1918 pandemic virus last October, the researchers had found that the 1918 strain’s hemagglutinin was responsible for much of the severe lung damage. (See BioWorld Today, Oct 6, 2005.)

Structural Similarities, But Mutational Differences

The researchers, from The Scripps Research Institute in La Jolla, Calif.; Atlanta’s Centers for Disease Control and Prevention; and the Armed Forces Institute of Pathology in Rockville, Md., next made specific substitutions in the H5 hemagglutinin structure to see how easily it could switch from a preference for birds to humans.

H1 through H3 flu virus types can switch from a preference for bird-like to human-like sugars, or vice versa, through changes of only two amino acids each, and the Scripps team tested whether the same was true for H5. The answer: Well, sort of.

Despite the pathogenic H5N1 strain’s structural similarity to the H1 hemagglutinin, amino acid substitutions that lead to a species preference switch for the subtype H1 had no effect on H5.

"I was surprised," Taubenberger said. "Given the structural similarities, I would have predicted that the mutations that work for H1 also work for H5." Instead, "the change that worked for the H3 subtype worked to a lesser degree for H5."

For Species Preference: Location, Location, Location

The influenza virus enters cells by binding to host-cell surface glycoproteins and glycolipids, which are proteins and fats with sugars stuck onto them. In the case of influenza virus, the main player with respect to binding is sialic acid, which can exist in different forms. The exact structure of the sialic acid (as well as other sugars) varies both between species and between cell types, and so whether the virus is easily able to infect a species depends on which sialic acid types it prefers and whether it encounters that specific sialic acid type in the first organ it infects.

Birds are first infected in intestinal cells, and those cells have what is known as a 2,3 linkage. Paulson, co-author on the Sciencexpress paper, stressed that "there may be other cells in the birds that have other linkages, but these are the first the virus encounters."

Cells in the human upper respiratory tract, which is where human-adapted flu virus strains take up residence, have lots of what’s known as 2,6-linked sialic acids. In the Nature paper, the scientists investigated the overall distribution of receptors in the lung, and found that in the lower respiratory tract, there are cells with the 2,3 linkages recognized by the avian-adapted virus. But the location of those cells is sufficiently remote to make it unlikely that even an infected person would spread the virus.

But if the H5N1 virus changes to a preference for 2,6 linkages - as the Sciencexpress team showed it could do, albeit not extremely, with a change of two amino acids - "it would be more easily disseminated from human to human," Paulson said. "The paper addresses this by showing that the cells it would first encounter" - in the event of such a switch - "are very rich in 2,6 linkages."