Supercomputers Identify Potential New Drugs to Combat 'Bird Flu'

Computers pick up "hot pockets" where the virus may grow to create medications ahead of the virus.

A supercomputer-generated rendering shows a potential drug docked to the active site of the neuraminidase enzyme of the avian flu virus.
+ More

Like most viruses, the ones that cause flu are sneaky little things armed with an amazing ability to rapidly change their spots—that is, the viral molecules that launch a seek-and-destroy response from the human immune system. In the case of avian or "bird flu," new strains are popping up all the time and rendering old medicines and vaccines practically useless, which is seriously bad news if you're trying to prevent a possible world-wide epidemic.

Now scientists have called upon the power of massive supercomputers to stay ahead of the changes and help keep a fresh supply of anti-virus treatments on hand for public health efforts. One University of California, San Diego team has isolated more than two dozen promising and novel compounds. In some cases, the compounds appeared to be equal or stronger inhibitors than are currently available anti-flu remedies.

"If those resistant strains begin to propagate, that's when we're going to be in trouble, because we don't have any anti-virals active against them," said Rommie Amaro, a UC San Diego chemist and member of the research team. "So, we should have something as a backup, and that's exactly why we're working on this."

Using computers at the San Diego Supercomputing Center and the National Center for Supercomputing Applications in Urbana-Champaign, Ill., the group ran complex programs to mimic the movements of a particularly wiggly protein called neuraminidase 1 (or N1), which the avian flu virus uses to spread infection to new cells. As the proteins changed shape according to physical laws, the computers picked up a "hot pocket" that appeared to be quite dynamic and flexible and therefore a target for medicines aimed at stopping the infection process.

According to the researchers, the computer simulations represent an advance over other types of 3-D studies, such as crystallography, because they are able to capture a protein's constant twitching and jiggling in a sort-of motion picture instead of snap shots of the receptor at rest.

The team conducted a "virtual screen" of 1,883 compounds selected from the National Cancer Institute Diversity Set, using a computational tool called AutoDock that predicts how small molecules, such as drug candidates, bind to a known 3-dimensional receptor. Compounds that most easily bind to the site are considered to be top hits for further study as drug candidates. Five other compounds known to experimentally bind to avian influenza N1 were also screened, including some now being tested in clinical trials.

About 27 compounds showed significant promise, all having potentially the same or stronger bonding affinity with N1 than currently available anti-flu drugs like Tamiflu and Relenza. Several looked like particularly good candidates because they bound to both the regular active site and an additional side pocket that opened during the computer simulation.

"The general idea is that we will be able to make a better drug through the strategic targeting of multiple active site pockets," said Amaro. Binding to more than one site by way of a chemical "bridge" can increase a drug's potency. The research now moves into the lab, where the compounds will undergo testing against the actual virus.

This study appears in the Journal of Medicinal Chemistry and was supported by the National Science Foundation, the Telemedicine and Advanced Technology Research Center, and the National Institutes of Health, and other agencies and organizations.

—By Leslie Fink/NSF

This report is provided by the National Science Foundation, an independent federal agency that supports fundamental research and education across all fields of science and engineering, in partnership with U.S. News and World Report. For more information, go to