The Anthrax Letters: The Attacks That Shocked America (37 page)

Randy Murch put his cup of coffee on the saucer, removed his black suit jacket, and draped it over the back of his chair. He tilted his head forward. We were in a corner of the breakfast room of a hotel in northwest Washington. “I coined the term ‘microbial forensics’ when I was with the FBI,” he said. After 23 years with the bureau, Murch, 50, retired at the end of 2002 to work on terrorism issues at the Institute for Defense Analyses, a think tank in Alexandria, Virginia. Now, 2 months later, in February 2003, he discussed his former role as deputy assistant director in charge of forensic programs in the FBI’s laboratory. He lifted a pair of gold-rimmed reading glasses from his pocket, slipped them on, and sketched a diagram of the bureau’s hierarchy. His position was on a rung parallel to that of the head of counterterrorism, Robert Blitzer, with whom he often worked. They were below the FBI director’s slot, separated from the director by only the deputy director.

Murch spent much of his career on the third floor of the FBI’s Washington headquarters building. On that floor, laboratories are replete with chemical reagents, incubators, centrifuges, and microscopes. “You see the same instruments and skill sets here as at other labs,” Murch said. But an outsider is also struck by a difference. Shelves and tables are filled with items that seem eerily out of place: a dress, a shirt, a smashed headlight. Here a knife, a revolver, there a blood sample, a urine sample. “Substrates,” Murch called them—items that will be analyzed for fingerprints, DNA typing, illicit chemicals, or any such possible evidence.

After receiving a Ph.D. in plant pathology from the University of Illinois in 1979, Murch began working for the FBI. Almost from the outset he was worried about biological terrorism, though few others at the time were contemplating the subject. But in 1996, with increasing concern about terrorism in general, he was able to establish the bureau’s Hazardous Materials Response Unit. “Just in time for the rash of bioterrorism hoaxes that began the following year,” he quipped. Murch spoke softly through a closely clipped mustache and beard. In the period after September 11, 2001, his efforts largely focused on suspects and collaborators connected with the attacks. Through court-ordered electronic surveillance he was locating, tagging, and tracking them.

What did he think of the FBI’s profiling of the likely anthrax mailer? I asked. Murch answered:

Did he think the perpetrator was domestic? “I think it was either of two possibilities: a homegrown guy who had the stuff on a shelf and took advantage of the September 11 timing or, maybe, somebody from outside who sent it to a local operator.”

Murch is inclined toward the former possibility. Whoever did it, he said, knew his way around the United States and the East Coast in particular. “It may not be Hatfill, but I think it is somebody like him.” Murch tilted his head back and said that he predicted years ago that some domestic person would use biological agents as terror weapons. “You don’t need much equipment or an advanced degree to make biological weapons,” he said. “You could fit all the stuff in a garage.”

Murch’s interest in microbial forensics has brought him in touch with many outside scientists through the years. A decade earlier he met Paul Keim and Paul Jackson, who were working on DNA fingerprinting at the Los Alamos National Laboratory. He sensed then that what they were doing could be important to criminal investigations. In the wake of the anthrax attacks, the FBI did indeed turn to Keim and others in the scientific community for help.

On October 5, 2001, the day that Bob Stevens died, the
Journal of Clinical Microbiology
accepted a paper titled “Molecular Investigation of the Aum Shinrikyo Anthrax Release in Kameido, Japan” for publication in its December issue. Aum Shinrikyo was the cult that in 1995 released sarin nerve agent in the Tokyo subway. In 1993 the same cult disseminated anthrax bacteria from the roof of a facility in Kameido, a suburb of Tokyo. Unlike the nerve agent, which killed 12 people and injured more than a thousand, the anthrax had no discernible effect. The reason had been unclear. Now this new study would confirm that the cult’s anthrax had been the Sterne strain, a nonpathogenic bug used to vaccinate animals.

A decade earlier Paul Keim, principal investigator of the study, was struggling to stretch his $50,000 grant to run his laboratory at Northern Arizona University. Things began to change after he developed a technique to differentiate strains of anthrax. By the beginning of October 2001 his grants had climbed to $1.6 million. Before the end of the month the figure had doubled to $3.2 million. The FBI and other government agencies were pouring money into Keim’s laboratory with the hope that he could help identify the source of the bacteria in the anthrax letters. In fact, it was Keim who confirmed that the anthrax bacteria in the letters were of the Ames strain, the same strain as in the infected patients.

Keim received a Ph.D. in 1981 from the University of Kansas, followed by postgraduate work at the University of Utah. Despite his new celebrity, Keim’s attire still consists of blue jeans, running shoes, and, when temperature permits, a T-shirt. His lanky frame and long face are topped with a full crop of blond hair. When I told him that he seems younger than his age, 46, he chuckled, “Let’s see. What do they call me in the press? Oh yes, a young civilian scientist.” He ended the sentence in full-throated laughter.

Keim has amassed 1,350 anthrax strains from around the world, the largest collection anywhere. “Actually, whether they are called strains or isolates is hotly debated in the microbiology community,” he said. He added that the more we learn about the genetic makeup of bacteria, the more challenging it is to classify them. “Bacteria are exchanging DNA all the time,” Keim explained, “and some species do so more frequently than others.”

Escherichia coli
, a common intestinal bacterium, is a good example of the variability problem. To survive and thrive, these bacteria must divide often, producing around 300 generations per year. This frequency of division increases the chances for mutations—alterations of their genes—which explains the large diversity in the genomic structures of
E. coli
. “Two different strains of
E. coli
have been entirely sequenced,” Keim said, “and they only share about 60 percent of their genes in common.”

Bacillus anthracis
is very different. Anthrax spores can lie inactive in the soil for a century or longer. During that period the chances for DNA exchanges are nil. If spores happen to infect a cow or other livestock, they will germinate, reproduce, and become open to the possibility of mutation. But because anthrax outbreaks are infrequent, the long-dormant bacteria remain almost identical to each other.

Before Keim’s innovation, differentiating between strains of anthrax bacteria was very difficult, and identifying distinctions within one strain was almost impossible. But in the mid-1990s Keim began collaborating with Paul Jackson, a colleague who worked at the Los Alamos National Laboratory, to scan the anthrax genome. Using specialized “restriction” enzymes, they cut DNA strands at specific sites and broke up the anthrax genome (which is composed of 5.5 million base pairs) into fragments. Then, Keim explained: “We’d grab restriction fragments out of the genome and compare those same restriction fragments across different strains and see if we could find differences. We looked at thousands. The fragments were pretty short, maybe 250 base pairs.”

In this way, Keim could detect change even if by only a single nucleotide, or DNA base. And this provided the ability to discover the rare genetic inconsistencies even within a single strain. “So we could take about 100 isolates of
Bacillus anthracis
and look at a thousand or more of these restriction fragments. And when we did that, we found only 30 fragments that varied among the strains.” The tiny variation impelled Keim to claim that “
Bacillus anthracis
is the most homogeneous bacterium in existence.”

Keim’s technique is called MLVA, for Multi-Locus VNTR Analysis. The VNTR stands for Variable Number Tandem Repeat. As he heard himself speak the double mouthful, he chuckled, “I’ve been blamed for creating an acronym within an acronym.” So what exactly is it that Keim’s MLVA does that compares anthrax bacteria with each other? “OK. What we’ve done is identify the most informative regions of the
Bacillus anthracis
genome for discriminating among isolates.”

What does he mean by “informative regions”? He responded by way of analogy:

In February 2002, Keim announced at a news conference that he had found distinguishing features among stocks of Ames anthrax that were stored in different laboratories. His findings seemingly would allow for tracing the mailed anthrax to one of the laboratories through his MLVA technique. But on instruction from the FBI, for whom he was doing these assessments, he refused to name the laboratories from which he obtained the bacteria.

A year later, no publicly known information had yet tied the anthrax in the letters to a particular laboratory. I asked if Keim still thought it possible to identify the laboratory of origin. Again citing his agreement with the FBI, he spoke cautiously: “I really can’t answer that other than to say it’s possible.”

As the investigation of the anthrax letters unfolded, Keim himself learned a lot about the source of some of the strains in his collection. When we spoke, early in 2003, he said, “You know, 18 months ago, I didn’t know the Ames strain had come from Texas or when it was discovered.” The strain was first acquired by the army for vaccine research at Fort Detrick, Maryland, after it infected some Texas cows in 1981. The bacteria were so virulent that samples were later sent to other military and civilian laboratories for study. But an early shipment used the return-address label of a government laboratory in Ames, Iowa. Confusion over the strain’s origin arose in 1985 after a scientific paper mistakenly assumed that the return-address was the original source of the bugs.

The U.S. germ arsenal was destroyed after President Nixon’s decision in 1969 to eliminate this country’s offensive biological weapons program. At that time the Army’s anthrax weapons had been made up of the Vollum strain. Thus, as Keim belatedly learned, the Ames bacteria could not have been part of the Army’s earlier biological arsenal. Although the Ames strain was found in the United States, samples had been sent to several laboratories overseas, including the British biological defense establishment at Porton Down. Which leads to the question of whether the source of the material in the anthrax letters might have been outside the United States. “That’s a really interesting question,” Keim said. “But I have totally disengaged my mind from that kind of speculation.” He sees his job as purely scientific. “My role in the investigation is very focused. We look at DNA. We try to identify. We try to discriminate. For me to speculate beyond that would be a distraction from what I’m doing.”