The Blog

Remember Anthrax?

Part 2

12:00 AM, Apr 20, 2002 • By DAVID TELL
Widget tooltip
Single Page Print Larger Text Smaller Text Alerts

6) What's going on here? Is it Ames or isn't it?

Insofar as any of them feels sure of the answer, none of the scientists now working with the government will state it unambiguously, in part because they are concerned for the security of a massive ongoing investigation. Even were security not a concern, however, the question whether they are dealing with "the Ames strain" would still be a vexed one, for two reasons. First, "strain" is a word that has no fixed, technical meaning: Any set of microbes with closely related genetic characteristics gets called a "strain," but how closely related they must be--at what level of analysis should a set of microbes be subdivided into "strains"--is a subjective judgment. Is your second cousin "family" or merely distant kin?

Then there is the peculiar nature of anthrax itself. Anthracis is the world's most molecularly homogeneous bacterial species. As recently as 1995, every laboratory isolate ever tested appeared to be genetically identical, and only three of them had been labeled strains. There was Sterne, named after the South African researcher who developed veterinary medicine's still-standard anthrax vaccine. There was Vollum, originally recovered from livestock in England and a staple of the U.S. bioweapons program in the 1960s. And there was Ames, so dubbed by a USAMRIID scientist in 1981 (for the town in Iowa, though the sample actually originated in Texas). The Sterne strain was sui generis; it didn't cause disease. But no one knew exactly why, and the three strains remained genetically indistinguishable.

In 1996, a group of researchers at a Veterans Administration hospital in North Carolina were the first to announce the detection of meaningful differences in the DNA of various anthrax isolates; anthracis, they thought, was a species encompassing five distinct genetic subgroups. Over the next few years, building on this work, Dr. Paul Keim of Northern Arizona University in Flagstaff examined hundreds of anthrax samples from a collection maintained by Louisiana State University veterinarian Martin Hugh-Jones and further refined the bacterium's phylogeny: He found 89 unique genomes--strains, if you dare--in six major families of anthracis. Then, in 1999, Tim Read of the non-profit Institute for Genomic Research in Rockville, Maryland, published a full map of the Ames strain's main chromosome on the Internet--all the DNA letter-strings for nearly 6,000 genes. Using Read's data, Keim has since located at least 50 places along the anthrax genome where various strains diverge, and he believes he has positively fingerprinted more than 100 such strains in his own ballooning archive of bacteria.

Keim and Read are both heavily involved in the current investigation. Read has completed another full sequence of anthracis DNA, this one taken from Bob Stevens, last fall's first fatality. Keim has compared that roadmap with Read's previous, Ames-strain diagram. The two samples are said to be a match.

But they can only be a match within the limits of Keim's existing classification system for anthrax. And back in the fall, before they stopped talking to the press so frequently, he and Read were quite candid about what those limits entail. Bacillus anthracis mutates into separate strains at a glacial pace. It takes maybe a million generations before even a single piece of its DNA is altered. And any given isolate of the bacteria finishes only a few hundred of those reproductive cycles during each active life span, while infecting an animal or human host. Then, if it's lucky enough not to be killed by antibiotics or incinerated with its victim's corpse, an anthrax colony hibernates in sporulated form for decades at a time. It is because their opportunities for genetic development have been so few and far between that many of the anthracis strains Paul Keim believes he has identified are separated by just a handful of DNA nucleotides--out of more than five million in the bacterium's full genome.

Which means, Keim admitted to Science magazine back in November, that many of his putative strains cannot be distinguished from one another outside the margin of error for current DNA sequencing technology: one misread nucleotide in every 100,000 examined. Around the same time, asked by an NPR radio interviewer whether it would be "possible to find out who sent the anthrax, where it came from, [by] doing gene studies of it," Tim Read paused a moment--and said, "I don't think so."

7) Assuming that it were possible to match last fall's anthrax with the Ames strain, would that mean it had to have come from an American laboratory?