From the Promed-mail article of 2/13/02:
The main chromosome of the bacteria contains 5 167 515 DNA letters holding information for 5960 genes. The bacterium also contains 2 small rings of DNA known as plasmids, which carry the genes essential for its virulence. The plasmid's DNA was decoded several years ago by scientists at the Los Alamos Nuclear Laboratory. Dr. Keim's success came from studying a site on the second of these plasmids called [the] poly-A tract. He found that Ames stocks held in different laboratories varied in the number of A's one of the 4 units of DNA they contained in the poly-A tract. The number of A's varied from 8 to 25.
Differences in Anthrax Strain Used
In Attacks Are Found by Scientists
By MARK SCHOOFS and ANTONIO REGALADO
Staff Reporters of THE WALL STREET JOURNAL
Two teams of scientists working with the FBI to track down the anthrax mailer have announced they have found genetic differences among samples of the Ames strain, which is the strain used in last fall's terror attacks.
The breakthroughs -- announced during the past week by researchers at the Institute for Genomic Research, known as TIGR, in Rockville, Md., and Northern Arizona University in Flagstaff -- have raised hopes that it might be possible to trace the anthrax used in the terror mailings back to a sample held in a particular laboratory. That, in turn, would allow investigators to focus their probe on that lab -- a strategy the FBI has said it wants to pursue.
But, TIGR President Claire Fraser cautioned, "There's no guarantee that anything anyone's doing will pinpoint where this came from."
The Ames strain of anthrax "is one of the toughest" organisms to distinguish, says NAU's Paul Keim, because the oldest known sample is just 21 years old and it mutates very slowly. So to compare samples, Dr. Keim looks at specific regions of the bacterium's DNA that are known to mutate the quickest and that are therefore more likely to differ among samples. He has collaborated in this work with researchers at Los Alamos National Laboratory.
The mutable regions that Dr. Keim focuses on repeat a certain array of genetic letters, such as ATT. The more times this array of letters gets repeated, the longer that stretch of DNA. Dr. Keim compares the length of these variable regions with lined-up lengths of string.
Distinguishing among wholly different strains of anthrax is relatively easy and can be done by comparing just eight mutable regions of DNA. But trying to discriminate among samples within the same strain is much harder.
Dr. Keim compared four samples of the Ames strain -- one taken from a goat that died in 1997 and three taken from laboratories. (Last week, he reported he had compared five strains, which he now says was an error.) Dr. Keim says all the strains were in his laboratory before the investigation began, and he won't discuss his work with the FBI.
Dr. Keim compared those four Ames-strain samples by matching up 36 regions of their DNA. He found no differences whatsoever.
But then he looked at another segment of DNA called a "poly A tract," so named because it repeats the letter A of the genetic code. According to Dr. Keim, this stretch of DNA ranges from about 12 to about 35 A's in a row. The four different samples of Ames each had a different number of A's.
If Dr. Keim's method were used to compare the terror anthrax to samples of the Ames strain housed in different laboratories, the best that investigators could hope for is an exact match -- a sample that has the exact same number of A's on that segment of DNA.
But it is likely that the terror strain won't match up exactly with Ames strains held by any lab, due to mutations. Lacking an exact match, scientists would try to draw a kind of family tree, called a phylogenetic tree, showing how closely related the terror isolate of anthrax is to the samples of the Ames strain held by different laboratories.
Scientists can't draw a tree based on just one genetic marker. So Dr. Keim and other researchers thus are trying to find more markers -- and TIGR announced this week that it has made progress in doing just that.
TIGR scientists have sequenced the entire genome of a reference strain of Ames and of the anthrax that killed Bob Stevens, the first person to die in the bioterrorism mailings. They have found some points where the genetic code differs by just one letter. There are more than five million letters in the genetic code of anthrax; if researchers can identify just several that differ among samples of Ames, that might allow them to home in on the source lab.
At a scientific meeting on Sunday, Dr. Fraser presented data showing TIGR had detected 80 differences that she said were reliable and not due to errors in sequencing the genome. In an interview, she said additional work had narrowed that number to "below 80," but she declined to elaborate. Dr. Fraser cautioned, however, that it would be "premature" to conclude that these findings could trace the terror anthrax back to a particular lab.
Write to Mark Schoofs at mark.schoofs@wsj.com and Antonio Regalado at antonio.regalado@wsj.com