Mapping Avian Flu Mutations

Mapping Avian Flu Mutations

Scientists are using a variety of tools and techniques to determine how the H5N1 influenza virus is mutating.

By Catherine Shaffer

The spread of influenza A subtype H5N1 beyond its usual reservoirs in Asia has triggered a worldwide flu epidemic—for birds. Since historical flu pandemics are thought to derive from avian viruses, the race is on to learn as much as possible about the molecular determinants of virulence and pathogenicity in the virus and prevent a new pandemic. Scientists trying to track potential mutation pathways for H5N1 have made a number of key findings, including one that calls into question whether a single mutation in H5N1 could lead to a human pandemic. It could turn out that avian influenza is literally for the birds.

The flu virus is a remarkably simple organism, being an RNA virus composed of only eight genes. The two most well known genes are for hemagglutinin (HA) and neuraminidase (NA). The NP gene encodes a nucleoprotein, the M gene encodes two different matrix proteins in different reading frames, and the NS gene encodes non structural proteins NS1 and NS2, also using different reading frames. The three remaining genes, PA, PB1, and PB2, encode the three RNA strands that make up the RNA polymerase.

The flu life cycle begins when the viral envelope, with its two surface glycoproteins, HA and NA, binds to a sialic acid-terminated oligosaccharide on the surface of the host cell via HA. The virus enters the cell through endocytosis. The host cell machinery transports the genetic material to the nucleus where transcription and replication begin. Because viral replication happens rapidly, the cell is quickly overwhelmed and usually dies as a result of the infection. At the end of the cycle, the viral particles are assembled and the surface proteins HA and NA once again mediate transport through the cell membrane, this time from the inside to the outside.

Surface specificity
Ruben Donis, PhD, is chief of the Molecular Virology Branch of the Influenza Division at the Centers for Disease Control (CDC), Atlanta. One of their primary goals with H5N1 is identifying markers of virulence and pathogenicity to be able to perform receptor-binding studies, cellular analysis, and sequence analysis on various viral strains. This would allow the CDC to assess the danger of H5N1 samples without relying on animal manipulation studies using mice and ferrets.

Donis and colleagues extract RNA from viruses grown in chicken eggs using off-the-shelf RNA extraction and amplify it using reverse transcription polymerase chain reaction (PCR). The resulting DNA sequence can then be analyzed and aligned with other viruses, yielding valuable information about phylogenetic relationships of the virus, particularly the highly variable HA gene.

The researchers created a phylogenetic map for the viruses, based on sequence similarities in the HA gene, and divided the virus strains into clades. It was found that strains of virus isolated from human beings, most of who succumbed to the disease, were highly pathogenic in ferrets, whereas strains of H5N1 isolated from chickens were not highly pathogenic in ferrets.

Donis and his group continue to closely monitor changes in the HA binding pocket that would indicate the virus is acquiring the ability to recognize human receptors. The sialic acid linkage on the host cell surface is crucial. Avian-adapted flu viruses preferentially bind to sialic acids that are linked in an α-2,3 configuration with galactose. This is the sugar that is most abundant in the respiratory tract of chickens and ducks. In well-adapted human flu viruses, such as seasonal H1-type influenza, the HA is optimized to bind to an α-2,6-linked sialic acid. The α-2,6 linkage has a bend or a kink in it, compared to the stick-like α-2,3. One of the important differences between avian- and human-adapted influenza viruses is that the two types of HA have different amino acids lining the binding pocket to either accommodate the straight-line linkage or the kinked linkage, respectively.

Donis says there are two amino acid positions that could determine this specific interaction. "We have very good knowledge of what happens in the transmission from ducks to humans with previous pandemic virus. The H3 virus that came into the human population in ‘68 had changes at positions 226 and 228. The virus that came into humans in 1918, the so-called Spanish flu, is a different subtype, H1. That one had changes at 190 and 225.”

Modeling Containment Strategies As part of the National Institute of General Medical Sciences' (NIGMS) Models of Infectious Disease Agent Study (MIDAS), researchers are examining containment strategies for the H5N1 or another subtype of influenza virus. Two groups of researchers started with basically the same computer code and independently developed two models for the spread of flu, one at a point of origin in Southeast Asia and one for an uncontrolled global pandemic. The Southeast Asia study focused on ways to contain the epidemic at its source. The second phase of the analysis tried to predict the outcome of various public health strategies for minimizing the spread of a pandemic in the US that had already become widespread throughout the world. "There are essentially individual people modeled in the computer who go to work, go to school, travel, and so on according to available data about the fine-grainness of the model,” says Jeremy Berg, PhD, the director of NIGMS. The group at the University of Washington in Seattle was led by Ira Longini, PhD; another worked under the direction of Donald Burke, PhD, at Johns Hopkins University, Baltimore. The MIDAS network provided most of the computational resources. Longini says their recent work was published in the Proceedings of the National Academy of Sciences. "It was 281 million people and can be run in six to twelve hours with stochastic realization. A smaller one, for a half million people for Southeast Asia, runs in about twenty minutes.” The model suggested that effective strategies for containment of flu in the US would include aggressively tracing cases and contacts or focused treatment with antiviral drugs within a geographic area. Among the less effective strategies studied were restrictions on travel. Variable parameters within the model included viral replicator time and viral transmission time. "We reviewed literature from the past influenza outbreaks of 1918, 1957, and 1968,” says Burke. "The problem is if we were to have a new pandemic, there is no guarantee that a new pandemic would follow those values, so we do sensitivity analysis around those.”

The poor ability of H5N1 to bind to cell receptors on human host cells may be mechanistically responsible for the different clinical course of the infection from human-adapted flu virus subtypes. "The current thinking is that of all the people that are exposed, very, very few get sick,” Donis says. "If you look at the people that are getting sick, the virus is located not in the upper respiratory tract, but in the alveoli, at the very end of the lungs . . . What happens is that the virus has to get deep, deep, deep to gain a foothold and replicate. Fortunately for us . . . the return trip is equally difficult.”

This means that even though the infected individual may become very ill, he or she is unlikely to transmit the infection to other people. This raises the question of whether an adaptation of the HA gene to accommodate the human conformation a-2,6 receptor would at the same time reduce the massive and deadly infection to something resembling the seasonal flu. "If the virus has to gain the ability to infect the upper respiratory tract, it would lose virulence. If it gains the ability to infect the upper respiratory tract, while retaining the ability to infect pneumocytes, we could have a very nasty virus.”

One theoretical bridge between avian and human flu is the pig. Swine have both α-2,3 and α-2,6 receptors on their cells, and thus are highly susceptible to both avian and human viruses. The theory is that the pig can serve as a host in which the genes of the two different viruses can shuffle and recombine, resulting in a new strain of virus.

Core proteins and beyond
At St. Jude Children's Research Hospital in Memphis, Tenn., some of the world's leading influenza experts are working to characterize the pathogenic potential of recently emerging H5N1 for mammalian species. Their studies have revealed candidates in viral proteins other than HA and NA. In a paper published on January 26 in Sciencexpress, 15 authors, including St. Jude's Robert Webster, PhD, and Erich Hoffmann, PhD, report the results of a massive gene sequencing project, the first of its kind for avian influenza.

Currently, public repositories of genetic information for influenza are heavily skewed toward surface glycoproteins and shorter genes such as M and NS. In order to identify species-specific markers of pathogenicity, the group sequenced 4,339 genes from hundreds of different avian influenza strains. In addition to grouping the virus genomes into clades (a standard approach for phylogenetic analysis), the group also created proteotypes for them based on unique amino acid signatures.

From this analysis, the viral protein NS1 emerged as a potential virulence determinant. NS1's exact function is not known, but it may play a role in mediating cell defenses, such as interferon. This new analysis revealed that NS1 contained a previously unknown PDZ domain ligand. Since PDZ domain proteins are important in many structural and signaling pathways in the cell, it is possible that avian NS1 widely disrupts these cellular activities.

click the image to enlarge Since surface glycoproteins interact with the body's immune system, a human strain of virus with HA and NA genes from avian H5N1 was created using the eight plasmid system and two existing virus strains. (Source: Erich Hoffmann, PhD)

With NS1 on the table as an important determinant of species specificity in influenza virus, it is clear that the search cannot be limited to HA and NA. Hoffmann studied a number of other viral genes to see if there was an effect on virulence. He and his colleagues amplified the viral genes using real time PCR with a standard set of primers for influenza A virus, then inserted the viral cDNAs into plasmids. After confirming that the virus regenerated from MDCK/293 T cells was identical to the parent viruses, the team set about re-assorting the genes by combining the desired plasmids. Initial experiments exchanging the HA and NA genes in one of the same highly pathogenic strains of H5N1 analyzed by Donis et al with a less pathogenic chicken strain revealed that HA and NA were not sufficient to make the chicken strain highly pathogenic. "We swapped H and N, and they didn't have an effect. They didn't determine the virulence...hemagglutinin is not sufficient,” Hoffmann says.

When Hoffmann switched out all three polymerase genes from the harmless chicken virus into the lethal human strain virus, all animals survived, demonstrating that the polymerase genes are important for virulence. The reverse experiment resulted in a chicken virus with three polymerase genes from the human-derived virus which was lethal to mice, but which ferrets survived after marked symptoms. The three polymerase genes, PA, PB1, and PB2,would seem to be important in the virulence of the virus and linked because single-gene swaps did not result in dramatic changes in virulence.

The interrelation of these three genes suggests that changes or mutations affecting one gene may require a complementary change in one of its partners. On a larger scale, the analysis of the St. Jude team revealed that the vast majority of complete genomes in their sample shared at least five genes with one other genome. This means that mutations of highly conserved genes seem to be complemented by changes in other genes at the same time. This calls into question any suggestion that a single mutation in H5N1 could lead to the feared human pandemic.

Shaffer is a freelance writer based in Ann Arbor, Mich.

http://www.birdflubreakingnews.com/templates/birdflu/window.php?url=http%3A%2F%2Fwww.genpromag.com%2FShowPR%7EPUBCODE%7E018%7EACCT%7E1800000100%7EISSUE%7E0606%7ERELTYPE%7ECVS%7EProdCode%7E00000000%7EPRODLETT%7EQ.html