A doughboy's lungs yield 1918 flu virus.

On Sept. 19, 1918, an Army private destined for the trenches of World War I France reported to the base hospital at Camp Jackson, S.C. Although otherwise healthy, the 21-year-old complained of chills, fever, headache, backache, and a cough. "Opinion: Influenza," a doctor noted in the medical record.

Within a week, he was dead-1 of 21 million people worldwide who would succumb to the influenza pandemic of 1918. For almost a century, samples of the doughboy's lungs sat in a warehouse run by the Armed Forces Institute of Pathology (AFIP) in Washington, D.C. Hidden in his tissues lay RNA bearing the solution to an enduring mystery, the genetic code for the worst pandemic in human history.

Now, for the first time, the killer has been exhumed and fragments of its genes deciphered, say Jeffery K. Taubenberger and his colleagues at AFIP.

"This is not just a medical detective story," Taubenberger says. "This could happen again. It would be really useful to find out what happened in 1918 and apply that knowledge to protect us against future outbreaks." "This is a tremendous advance," says virologist Robert Webster of St. Jude Children's Research Hospital in Memphis, Tenn. "The 1918 virus represents the ultimate disease-causing agent-in a sense it's like ebola gone mad. We need to understand as much as possible about this virus because the world will get another pandemic, maybe late in this century or early in the next."

One pandemic was one too many. In the 1918 outbreak, nearly 700,000 people died in the United States. Historian Alfred Crosby has written that Washington, D.C., seized two train cars of coffins headed for Pittsburgh so that the capital's undertakers could bury the dead.

Taubenberger and his colleagues began their search for the virus' genes by selecting at random 28 of the 70 pandemic victims whose lung samples are stored at AFIP. Autopsy reports from 1918 disclosed that seven of these servicemen died soon after becoming ill, enhancing the likelihood that lung tissues might contain intact bits of RNA from the virus' unusual eight-strand genome.

People who live longer are less likely to harbor the virus, because the body's defenses eradicate the microbes, Taubenberger says. In such cases, bacterial pneumonia delivers the fatal blow. But in the seven servicemen who died quickly, the immune counterattack might not have had time to wipe out the virus.

The researchers drew a blank in six cases. The private from Camp Jackson, however, was unusual. His left lung had suffered extensive bacterial pneumonia, but his right lung had not. This raised the possibility that the right lung might still harbor the virus. To find out, the researchers removed some tissue from the paraffin in which it was stored. Step by step, they broke it down until only RNA remained.

"The people who preserved this tissue never imagined what might be possible down the road," says team member Ann H. Reid.

Reid made millions of copies of nine RNA fragments of five flu genes. Thomas G. Fanning of AFIP then deciphered the sequences of the fragments and compared them to every other known sequence of the flu gene.

"It's unique," Taubenberger says. The team has also confirmed prior evidence suggesting that the sequences most closely resemble those from swine flu. Their report appears in the March 21 Science.

Researchers disagree on whether it will be possible to rebuild the entire genome of the virus, perhaps yielding clues to its spectacular virulence. Most agree that the work might permit the making of a vaccine, if needed. "If this fossil were to reemerge," says Webster, "we could use this information to get a best-match vaccine that would probably protect us quite well."