Panic in Level 4: Cannibals, Killer Viruses, and Other Journeys to the Edge of Science

by Richard Preston

  In 1984, Craig Venter had begun working at the NIH, where he eventually developed an unorthodox strategy for decoding bits of genes. At the time, other scientists were painstakingly reading the complete sequence of each gene they studied. This process seemed frustratingly slow to Venter. He began isolating what are called expressed sequence tags, or ESTs, which are fragments of DNA at the ends of genes. When the ESTs were isolated, they could be used to identify genes in a rough way. With the help of a few sequencing machines, Venter identified bits of thousands of human genes. This was a source of unease at the NIH, because it was a kind of skimming rather than a complete reading of genes. Venter published his method in 1991 in an article in Science, along with partial sequences from about 350 human genes. The method was not received well by many genomic scientists. It was fast, easy, and powerful, but it didn’t look elegant; it looked like an application of brute force, and some scientists seemed threatened by it. Venter claimed that two of his colleagues, who eventually became heads of public genome centers, had asked him not to publish his method or move forward with it for fear they would lose their funding for genome sequencing done their way.

  The NIH decided to apply for patents on the gene fragments Venter had identified. James Watson blew his stack over the idea of anyone trying to patent bits of genes. He got into a hostile situation with the director of the NIH, Bernadine Healy, who defended the patenting effort. In July 1991, during a meeting in Washington called by Senator Pete Domenici, of New Mexico, to review the genome program, Watson disparaged Craig Venter’s methods. “It isn’t science,” he said, adding that Venter’s machines “could be run by monkeys.”

  It was a strange moment. The Senate hearing room was almost empty—few politicians were interested in genes then. But Craig Venter was sitting in the room. “Jim Watson was clearly referring to Craig as a monkey in front of a U.S. senator,” another scientist who was there said to me. “He portrayed Craig as the village idiot of genomics.” Venter seemed to almost thrash in his chair, stung by Watson’s words. “Watson was the ideal father figure of genomics,” Venter says. “And he was attacking me in the Senate, when I was relatively young and new in the field.”

  That day in his office in Cold Spring Harbor, James Watson insisted to me that he hadn’t been comparing Craig Venter to a monkey. “It’s the patenting of genes I was objecting to. That’s why I used the word ‘monkey’! I hate it!” he said testily. The patent office turned down the NIH’s application for a patent, anyway. But a few years later, two genomics companies, Incyte and Human Genome Sciences, adopted Craig Venter’s EST method for finding genes, and it became the foundation of their businesses. Those businesses, combined, were worth many billions of dollars on the stock market. Samuel Broder, the chief medical officer at Celera, who was a former director of the National Cancer Institute, said to me, heatedly, “None of the people who severely and acrimoniously criticized Craig for his EST method ever said they were personally sorry. They ostracized Craig and then went on to use his method with never an acknowledgment.”

  James Watson said, “The EST method has proved immensely useful, and it should have been encouraged.”

  Venter was increasingly unhappy at the NIH. He had received a ten million dollar grant to sequence human DNA, and he asked for permission to use some of the money to do EST sequencing, but his request was denied by the Human Genome Project (which James Watson was then running). Venter returned the grant money with what he says was a scathing letter to Watson. In addition, Venter’s wife, Claire Fraser, had been denied tenure at the NIH. Her review committee (which was composed entirely of middle-aged men, she said) explained to her that it could not evaluate her work independently of her husband’s. At the time, Fraser and Venter had separate labs and separate research programs. Fraser considered suing the NIH for sex discrimination.

  Meanwhile, James Watson, as head of the Human Genome Project at the NIH, had gotten himself into continuing scrapes with the head of the NIH, Bernadine Healy. During a press conference at which Healy was present, Watson criticized the NIH’s policy of seeking patents on genes, and he labeled Healy (who was his boss) a “lunatic.” Shortly afterward, Bernadine Healy fired James Watson. She forced him to resign from his position as head of the Human Genome Project. Watson had done himself in with his mouth.

  That summer, Craig Venter was approached by a venture capitalist named Wallace (Wally) Steinberg, who wanted to set up a company that would use Venter’s EST method to discover genes, create new drugs, and make money. “I didn’t want to run a company, I wanted to keep doing basic research,” Venter said. But Wally Steinberg offered Venter a research budget of seventy million dollars over ten years—a huge amount of money, then, for biotech. Venter, along with Claire Fraser and a number of colleagues, left the NIH and founded TIGR, which is a nonprofit organization. At the same time, Wally Steinberg established a for-profit company, Human Genome Sciences, to exploit and commercialize the work of TIGR, which was required to license its discoveries exclusively to its sister company. Thus Venter got millions of dollars for research, but he had to hand his discoveries over to Human Genome Sciences for commercial development. Venter had one foot in the world of pure science and one foot in a bucket of money.

  By 1994, the Human Genome Project was mapping the genomes of model organisms, which included the fruit fly, the roundworm, yeast, and E. coli (a bacteria that lives in the human gut), but no genome of any organism had been completed, except for virus genomes, which are relatively small. Venter and Hamilton Smith (the Nobel laureate and discoverer of DNA “scissor” enzymes, who was then at the Johns Hopkins School of Medicine) proposed speeding things up by using a technique known as whole-genome shotgun sequencing. In shotgunning, the genome is broken into small, random, overlapping pieces, and each piece is sequenced, or read. Then the jumble of pieces is reassembled in a computer that compares each piece to every other piece and matches the overlaps, thus assembling the whole genome. It’s quite a lot like putting together a picture puzzle by matching the edges of the pieces to neighboring pieces, except that a genome can consist of millions of pieces and the task of matching them up to form a whole image of a genome requires superpowerful computers and really hot software.

  Venter and Smith applied for a grant from the NIH to shotgun-sequence the genome of a disease-causing bacterium called Haemophilus influenzae, or H. flu for short. It causes fatal meningitis in children. They proposed to do it in just a year. H. flu has 1.8 million letters of code, which seemed massive then (though the human code is two thousand times as long). The review panel at the NIH gave Venter’s proposal a low score, essentially rejecting it. According to Venter, the panel claimed that an attempt to shotgun-sequence a whole microbe was excessively risky and perhaps impossible. He appealed. The appeals process dragged on. While the appeals dragged on, he went about shotgunning H. flu anyway. Venter and the TIGR team had nearly finished sequencing the H. flu genome when, in early 1995, a letter arrived at TIGR saying that the appeals committee had denied the grant on the ground that the experiment wasn’t feasible. Venter published the H. flu genome a few months later in Science. Whole-genome shotgunning had worked in spite of the objections of a funding committee at the NIH; it was almost as if the NIH wanted to prevent Venter’s method from working. The method worked very well, however; this was the first completed genome of a free-living organism.

  It seems quite possible that Venter’s grant was denied because of politics. The NIH, the National Institutes of Health, is funded by tax dollars. The review panel seems to have hated the idea of giving taxpayer money to TIGR to make discoveries that would be turned over to a corporation, Human Genome Sciences, which could then profit from the discoveries. It turned down the grant in spite of the fact that “all the smart people knew the method was straightforward and would work,” said Eric Lander, one of the leaders of the Human Genome Project.

  Around this time, the venture capitalist Wally Steinberg died of a heart attack, and his deat
h provided a catalyst for a split between TIGR and Human Genome Sciences, which was run by a former AIDS researcher, William Haseltine. Venter and Haseltine were widely known to despise each other. Venter had already sold his stock in Human Genome Sciences because of the rift between them, and after Steinberg died the relationship between the two organizations was formally ended.

  Late in 1997, TIGR was doing some DNA sequencing for the Human Genome Project, and Venter began going to some of the project’s meetings. That was when he started calling the heads of the public project’s DNA-sequencing centers the Liars’ Club, claiming that their predictions about when they would finish a task and how much it would cost were lies. His calling them liars did not win him many friends.

  Francis Collins, a distinguished medical geneticist from the University of Michigan, had become the head of the NIH genome program shortly after James Watson resigned in 1992. In early January 1998, an internal budget projection from Collins’s office somehow made its way to Watson (he seemed able to find out anything that was happening anywhere in molecular biology). This budget projection was supposed to be secret. It was an “eyes only” document intended to be seen by only about eight people, namely the top heads of the Human Genome Project. It contained a graph marked “Confidential” indicating that Francis Collins planned to spend only sixty million dollars per year on direct human-DNA sequencing through 2005. This was peanuts. It also predicted that by that year—when the human genome was supposed to be completed—actually only 1.6 billion to 1.9 billion letters of human code would be sequenced; that is, slightly more than half of the human genome would be done by then. The implication was that the whole human genome wouldn’t get done until maybe 2008 or afterward.

  This upset James Watson. Watson had hoped that his successor, Francis Collins, would aggressively pursue the human genome and get the work done as fast as possible; for one thing, he, James Watson, wanted to be alive to see the human genome. Second, the millions of people around the world who suffered from genetic diseases weren’t getting any younger. He decided to discuss it with Eric Lander, the head of the Human Genome Project’s DNA-sequencing center at MIT. Lander was a voluble, articulate, brilliant man in his forties who projected a high degree of self-confidence. He spoke in a rapid voice. He had a broad face and a mustache, and he owned stock in biotech companies that he advised or had helped to found. Eric Lander had become quite wealthy.

  On January 17, James Watson traveled to Rockefeller University, on the East Side of Manhattan, where Eric Lander was giving the prestigious Harvey Lecture. The two men met after the lecture at the faculty club at Rockefeller. They were dressed in tuxedos, and Eric Lander had been drinking. By long tradition among medical people, the Harvey Lecture is given and listened to under the influence. Watson himself seemed a little tipsy. The scientists continued to drink after the lecture.

  The Rockefeller faculty club overlooks a lawn and sycamore trees and the traffic of York Avenue. Watson and Lander sat down with cognacs at a small table in a dim corner of the room, on the far side of a pool table, where they could talk without being overheard. Also present and drinking cognac was a biologist named Norton Zinder, who was one of Watson’s best friends. Zinder, like Watson, was a founder of the Human Genome Project. One of the older men brought up the confidential budget document with Lander, and both of them began to press him about it. They felt that it provided evidence that Francis Collins did not intend to spend more than sixty million dollars a year on human-DNA sequencing—and this was nowhere near enough money to get the job done anytime soon, they felt.

  Watson evidently believed that Eric Lander had influence with Francis Collins, and he urged him to try to persuade Collins to spend more on direct sequencing of human DNA, and to twist Congress’s arm for more money.

  Norton Zinder was somewhat impaired with cocktails. “This thing is potchkying along, going nowhere!” he said, hammering the little table and waving his arms as he spoke. For him, the issue was simple: he had had a quadruple coronary bypass, and he had been receiving treatments for cancer, and now he was afraid he would not live to see the deciphering of the human genome. The human genome had begun to seem like a vision of Canaan to Norton Zinder, and he thought he wouldn’t make it there.

  Eric Lander did not view things the way the older biologists did. In his opinion, the problem was organizational. The Human Genome Project was “too bloody complicated, with too many groups.” He felt the real problem was a lack of focus. He wanted the project to create a small, elite group that would do the major sequencing of human DNA—shock cavalry that would lead a charge into the human genome.

  The three men downed their cognacs with a gloomy sense of frustration. “I had essentially given up on seeing the human genome in my lifetime,” Zinder said.

  NOT LONG BEFORE Watson and his friends began lamenting the slowness of the public project, the Perkin-Elmer Corporation, which was a manufacturer of lab instruments, had started secretly talking about an ambitious corporate reorganization. It controlled more than 90 percent of the market for DNA-sequencing equipment, and it was developing the Prism machine. The Prism was then only a prototype sitting in pieces in a laboratory in Foster City, California, but already it looked as if it was going to be at least ten times faster than any other DNA-sequencing machine. Perkin-Elmer executives began to wonder just what it could do. One day Michael Hunkapiller, who was then the head of the company’s instrument division, got out a pocket calculator and estimated that several hundred Prisms could whip through a molecule of human DNA in a few weeks, although only in a rough way. To fill in the gaps—places where the DNA code came out garbled or wasn’t read properly by the machines—it would be necessary to sequence the molecule again and again. This is known as repeat sequencing, or manyfold coverage, and he thought it might take a few years. Hunkapiller persuaded the chief executive of Perkin-Elmer, Tony White, to restructure the business and create a genomics company.

  In December 1997, executives from Perkin-Elmer began telephoning Craig Venter to see if he’d be interested in running the new company. He blew them off at first, but a few months later he went to California with a colleague, Mark Adams, to check out the prototype Prism. When they saw it, they immediately understood its significance. They were looking at the equivalent of a supersonic jet in relation to a propellor aircraft. Before the end of that day, Venter, Adams, and Hunkapiller had laid out a plan for decoding the human genome fast. A month later, Norton Zinder, Watson’s friend, flew to California to see the machine. Zinder saw it, too. “It was just a piece of equipment sitting on a table, but I said, ‘That’s it! We’ve got the genome!’” he recalled. Zinder joined Celera as a member of its board of advisers, and received stock in the company, which considerably enriched him. Now he could take a lesson from Eric Lander; he could cash in on the biotech boom. (“And what’s wrong with that?” he asked me. “The chemists have been cleaning up all their life. Now the biologists are starting to get their hands on the money, and people are saying, ‘Whoo, that’s not kosher!’ What’s not kosher about it?”)

  After Norton Zinder got involved financially and scientifically in the Celera effort to race past the Human Genome Project, it led to some strains between him and James Watson. They maintained their friendship but finally had to agree not to speak about Celera with each other. They evidently feared that one or both of them could have a stroke arguing about Craig Venter.

  ONE DAY not long after Norton Zinder saw the Prism machine and realized it was going to revolutionize the reading of DNA, Craig Venter and Mike Hunkapiller walked into the office of Harold Varmus, the director of the NIH, to talk to him about something. Harold Varmus was a Nobel laureate and an expert in genes and DNA. He had won the Nobel Prize in Medicine in 1989 (with J. Michael Bishop) for a theory of cancer-causing genes—a model of how cancer arises from genes embedded in a person’s code. In Varmus’s office that day, Craig Venter wanted to talk about the human code and the ongoing effort to read it. He announced
the pending formation of a corporation, to be led by himself, that was going to decode the human genome. (Celera did not yet have a name.) Venter proposed to Varmus that the company and the public project collaborate, sharing their data and—this point is enormously important to scientists—sharing the publication of the human genome, which meant sharing the credit and the glory for having done the work. This included, of course, the unspoken possibility of the Nobel Prize in Medicine. The Nobel Prize would seem to have been made for the team that first decrypted the human DNA.

  Harold Varmus was skeptical. He suspected that this wasn’t a sincere offer from Craig Venter. He wondered if Venter might be angling for something that would be good for Craig Venter but maybe not so good for the Human Genome Project. He told Venter that he needed time to consider the proposal, particularly to check back with his subordinate Francis Collins (the head of the NIH’s part of the Human Genome Project), to see what Collins thought of this unusual offer of collaboration.

  Later that same day, Craig Venter and Mike Hunkapiller drove to Dulles Airport, where they met Francis Collins at the United Airlines Red Carpet Club. There they personally offered collaboration to Collins.

  Venter recalled later that Collins seemed upset with his offer. Collins recalled that he merely asked Venter for some time to consider it. Extra time was one thing that Craig Venter was not prepared to give Francis Collins.