Mering and Minkowski were mystified: Why had removing an abdominal organ precipitated this odd syndrome? The clue emerged from a throwaway fact. A few days later, an assistant noted that the lab was buzzing with flies; they were swarming on the pools of dog urine that had now congealed and turned sticky, like treacle.I When Mering and Minkowski tested the urine and the dog’s blood, both were overflowing with sugar. The dog had become severely diabetic. Some factor synthesized in the pancreas, they realized, must regulate blood sugar, and its dysfunction must cause diabetes. The sugar-regulating factor was later found to be a hormone, a protein secreted into the blood by those “islet cells” that Langerhans had identified. The hormone was called isletin, and then insulin—literally, “island protein.”
The identification of insulin in pancreatic tissue led to a race to purify it—but it took two further decades to isolate the protein from animals. Ultimately, in 1921, Banting and Best extracted a few micrograms of the substance out of dozens of pounds of cow pancreases. Injected into diabetic children, the hormone rapidly restored proper blood sugar levels and stopped their thirst and urination. But the hormone was notoriously difficult to work with: insoluble, heat-labile, temperamental, unstable, mysterious—insular. In 1953, after three more decades, Fred Sanger deduced the amino acid sequence of insulin. The protein, Sanger found, was made of two chains, one larger and one smaller, cross-linked by chemical bonds. U-shaped, like a tiny molecular hand, with clasped fingers and an opposing thumb, the protein was poised to turn the knobs and dials that so potently regulated sugar metabolism in the body.
Boyer’s plan for the synthesis of insulin was almost comically simple. He did not have the gene for human insulin at hand—no one did—but he would build it from scratch using DNA chemistry, nucleotide by nucleotide, triplet upon triplet—ATG, CCC, TCC, and so forth, all the way from the first triplet code to the last. He would make one gene for the A chain, and another gene for the B chain. He would insert both the genes in bacteria and trick them into synthesizing the human proteins. He would purify the two protein chains and then stitch them chemically to obtain the U-shaped molecule. It was a child’s plan. He would build the most ardently sought molecule in clinical medicine block by block, out of an Erector Set of DNA.
But even Boyer, adventurous as he was, blanched at lunging straight for insulin. He wanted an easier test case, a more pliant peak to scale before attempting the Everest of molecules. He focused on another protein—somatostatin—also a hormone, but with little commercial potential. Its main advantage was size. Insulin was a daunting fifty-one amino acids in length—twenty-one in one chain and thirty in the other. Somatostatin was its duller, shorter cousin, with just fourteen.
To synthesize the somatostatin gene from scratch, Boyer recruited two chemists from the City of Hope hospital in Los Angeles—Keiichi Itakura and Art Riggs—both veterans of DNA synthesis.II Swanson was bitterly opposed to the whole plan. Somatostatin, he feared, would turn into a distraction; he wanted Boyer to move to insulin directly. Genentech was living in borrowed space on borrowed money. Scratched even a millimeter below its surface, the “pharma company” was, in truth, a rented cubicle in an office space in San Francisco with an offshoot in a microbiology lab at UCSF, which, in turn, was about to subcontract two chemists at yet another lab to make genes—a pharmaceutical Ponzi scheme. Still, Boyer convinced Swanson to give somatostatin a chance. They hired an attorney, Tom Kiley, to negotiate the agreements among UCSF, Genentech, and the City of Hope. Kiley had never heard the term molecular biology, but felt confident because of his track record of representing unusual cases; before Genentech, his most famous former client had been Miss Nude America.
Time too felt borrowed at Genentech. Boyer and Swanson knew that two reigning wizards of genetics had also entered the race to make insulin. At Harvard, Walter Gilbert, the DNA chemist who would share the Nobel Prize with Berg and Sanger, was leading a formidable team of scientists to synthesize insulin using gene cloning. And at UCSF, in Boyer’s own backyard, another team was racing toward the gene cloning. “I think it was on our minds most of the time . . . most days,” one of Boyer’s collaborators recalled. “I thought about it all the time: Are we going to hear an announcement that Gilbert has been successful?”
By the summer of 1977, working frantically under Boyer’s anxious eye, Riggs and Itakura had assembled all the reagents for the synthesis of somatostatin. The gene fragments had been created and inserted into a bacterial plasmid. The bacteria had been transformed, grown, and prepped for the production of the protein. In June, Boyer and Swanson flew to Los Angeles to witness the final act. The team gathered in Riggs’s lab in the morning. They leaned over to watch the molecular detectors check for the appearance of somatostatin in the bacteria. The counters blinked on, then off. Silence. Not even the faintest blip of a functional protein.
Swanson was devastated. The next morning, he developed acute indigestion and was sent to the emergency room. The scientists, meanwhile, recovered over coffee and doughnuts, poring through the experimental plan, troubleshooting. Boyer, who had worked with bacteria for decades, knew that microbes often digest their own proteins. Perhaps somatostatin had been destroyed by the bacteria—a microbe’s last stand against being co-opted by human geneticists. The solution, he surmised, would be to add another trick to the bag of tricks: they would hook the somatostatin gene to another bacterial gene to make a conjoined protein, then cleave off the somatostatin after. It was a genetic bait and switch: the bacteria would think they were making a bacterial protein, but would end up (secretly) secreting a human one.
It took another three months to assemble the decoy gene, with somatostatin now Trojan-horsed within another bacterial gene. In August 1977, the team reassembled at Riggs’s lab for the second time. Swanson nervously watched the monitors flicker on, and momentarily turned his face away. The detectors for the protein crackled again in the background. As Itakura recalled, “We have about ten, maybe fifteen samples. Then we look at the printout of the radioimmunoassay, and the printout show[s] clearly that the gene is expressed.” He turned to Swanson. “Somatostatin is there.”
Genentech’s scientists could barely stop to celebrate the success of the somatostatin experiment. One evening, one new human protein; by the next morning, the scientists had regrouped and made plans to attack insulin. The competition was fierce, and rumors abounded: Gilbert’s team had apparently cloned the native human gene out of human cells and were readying to make the protein in buckets. Or the UCSF competitors had synthesized a few micrograms of protein and were planning to inject the human hormone into patients. Perhaps somatostatin had been a distraction. Swanson and Boyer suspected ruefully that they had taken a wrong turn and been left behind in the insulin race. Dyspeptic even during the best of times, Swanson edged toward another bout of anxiety and indigestion.
Ironically, it was Asilomar—the very meeting that Boyer had so vociferously disparaged—that came to their rescue. Like most university laboratories with federal funding, Gilbert’s lab at Harvard was bound by the Asilomar restrictions on recombinant DNA. The restrictions were especially severe because Gilbert was trying to isolate the “natural” human gene and clone it into bacterial cells. In contrast, Riggs and Itakura, following the lead with somatostatin, had decided to use a chemically synthesized version of the insulin gene, building it up nucleotide by nucleotide from scratch. A synthetic gene—DNA created as a naked chemical—fell into the gray zone of Asilomar’s language and was relatively exempt. Genentech, as a privately funded company, was also relatively exempt from the federal guidelines.III The combination of factors proved to be a crucial advantage for the company. As one worker recalled, “Gilbert was, as he had for many days past, trudging through an airlock, dipping his shoes in formaldehyde on his way into the chamber in which he was obliged to conduct his experiments. Out at Genentech, we were simply synthesizing DNA and throwing it into bacteria, none of which even required compliance with the NIH guidelines.”
In the world of post-Asilomar genetics, “being natural” had turned out to be a liability.
Genentech’s “office”—the glorified booth in San Francisco—was no longer adequate. Swanson began scouring the city for lab space for his nascent company. In the spring of 1978, having searched up and down the Bay Area, he found a suitable site. Stretched across a tawny, sun-scorched flank of hillside a few miles south of San Francisco, the place was called Industrial City, although it was hardly industrial and barely a city. Genentech’s lab was ten thousand square feet of a raw warehouse on 460 Point San Bruno Boulevard, set amid storage silos, dump sites, and airport-freight hangars. The back half of the warehouse housed a storage facility for a distributor of porn videos. “You’d go through the back of Genentech’s door and there would be all these movies on shelves,” one early recruit wrote. Boyer hired a few additional scientists—some barely out of graduate school—and began to install equipment. Walls were constructed to divide the vast space. A makeshift lab was created by slinging black tarp across part of the roof. The first “fermenter” to grow gallons of microbial sludge—an upscale beer vat—arrived that year. David Goeddel, the company’s third employee, walked around the warehouse in sneakers and a black T-shirt that read CLONE OR DIE.
Yet no human insulin was in sight. In Boston, Swanson knew, Gilbert had upped his war effort—literally. Fed up with the constraints on recombinant DNA at Harvard (on the streets of Cambridge, young protesters were carrying placards against gene cloning), Gilbert had gained access to a high-security biological-warfare facility in England and dispatched a team of his best scientists there. The conditions in the military facility were absurdly stringent. “You totally change your clothes, shower in, shower out, have gas masks available so that if the alarm goes off you can sterilize the entire laboratory,” Gilbert recalled. The UCSF team, in turn, sent a student to a pharmaceutical lab in Strasbourg, France, hoping to create insulin at the well-secured French facility.
Gilbert’s group oscillated at the brink of success. In the summer of 1978, Boyer learned that Gilbert’s team was about to announce the successful isolation of the human insulin gene. Swanson braced himself for another breakdown—his third. To his immense relief, the gene that Gilbert had cloned was not human but rat insulin—a contaminant that had somehow tainted the carefully sterilized cloning equipment. Cloning had made it easy to cross the barriers between species—but that same breach meant that a gene from one species could contaminate another in a biochemical reaction.
In the narrow cleft of time between Gilbert’s move to England and the mistaken cloning of rat insulin, Genentech forged ahead. It was an inverted fable: an academic Goliath versus a pharmaceutical David, one lumbering, powerful, handicapped by size, the other nimble, quick, adept at dancing around rules. By May 1978, the Genentech team had synthesized the two chains of insulin in bacteria. By July, the scientists had purified the proteins out of the bacteria debris. In early August, they snipped off the attached bacterial proteins and isolated the two individual chains. Late at night on August 21, 1978, Goeddel joined the protein chains together in a test tube to create the first molecules of recombinant insulin.
In September 1978, two weeks after Goeddel had created insulin in a test tube, Genentech applied for a patent for insulin. Right at the onset, the company faced a series of unprecedented legal challenges. Since 1952, the United States Patent Act had specified that patents could be issued on four distinct categories of inventions: methods, machines, manufactured materials, and compositions of matter—the “four M’s,” as lawyers liked to call the categories. But how could insulin be pigeonholed into that list? It was a “manufactured material,” but virtually every human body could evidently manufacture it without Genentech’s ministrations. It was a “composition of matter,” but also, indisputably, a natural product. Why was patenting insulin, the protein or its gene, different from patenting any other part of the human body—say, the nose or cholesterol?
Genentech’s approach to this problem was both ingenious and counterintuitive. Rather than patenting insulin as “matter” or “manufacture,” it concentrated its efforts, boldly, on a variation of “method.” Its application claimed a patent for a “DNA vehicle” to carry a gene into a bacterial cell, and thereby produce a recombinant protein in a microorganism. The claim was so novel—no one had ever produced a recombinant human protein in a cell for medicinal use—that the audacity paid off. On October 26, 1982, the US Patent and Trademark Office (USPTO) issued a patent to Genentech to use recombinant DNA to produce a protein such as insulin or somatostatin in a microbial organism. As one observer wrote: “effectively, the patent claimed, as an invention, [all] genetically modified microorganisms.” The Genentech patent would soon become one of the most lucrative, and most hotly disputed, patents in the history of technology.
Insulin was a major milestone for the biotechnology industry, and a blockbuster drug for Genentech. But it was not, notably, the medicine that would catapult gene-cloning technology to the forefront of public imagination.
In April 1982, a ballet dancer in San Francisco, Ken Horne, visited a dermatologist, complaining of an inexplicable cluster of symptoms. Horne had felt weak for months and developed a cough. He had bouts of intractable diarrhea, and weight loss had hollowed his cheeks and made his neck muscles stand out like leather straps. His lymph nodes had swollen. And now—he pulled his shirt up to demonstrate—reticulated bumps were emerging on his skin, purple-blue of all colors, like hives in a macabre cartoon film.
Horne’s case was not isolated. Between May and August 1982, as the coasts sweltered in a heat wave, similarly bizarre medical cases were reported in San Francisco, New York, and Los Angeles. At the CDC in Atlanta, a technician was asked to fill nine requests for pentamidine, an unusual antibiotic reserved to treat Pneumocystis pneumonia. These requisitions made no sense: PCP was a rare infection that typically afflicted cancer patients with severely depleted immune systems. But these applications were for young men, previously in excellent health, whose immune systems had suddenly been pitched into inexplicable, catastrophic collapse.
Horne, meanwhile, was diagnosed with Kaposi’s sarcoma—an indolent skin tumor found among old men in the Mediterranean. But Horne’s case, and the other nine cases reported in the next four months, bore little resemblance to the slow-growing tumors previously described as Kaposi’s in the scientific literature. These were fulminant, aggressive cancers that spread rapidly through the skin and into the lungs, and they seemed to have a predilection for gay men living in New York and San Francisco. Horne’s case mystified medical specialists, for now, as if to intersect puzzle upon puzzle, he developed Pneumocystis pneumonia and meningitis as well. By late August, an epidemiological disaster was clearly appearing out of thin air. Noting the preponderance of gay men afflicted, doctors began to call it GRID—gay-related immune deficiency. Many newspapers accusingly termed it the “gay plague.”
By September, the fallacy of that name had become evident: symptoms of immunological collapse, including Pneumocystis pneumonia and strange variants of meningitis, had now begun to sprout up among three patients with hemophilia A. Hemophilia, recall, was the bleeding illness of the English royals—caused by a single mutation in the gene for a crucial clotting factor in blood, called factor VIII. For centuries, patients with hemophilia had lived in constant fear of a bleeding crisis: a nick in the skin could snowball into disaster. By the mid-1970s, though, hemophiliacs were being treated with injections of concentrated factor VIII. Distilled out of thousands of liters of human blood, a single dose of the clotting factor was equivalent to a hundred blood transfusions. A typical patient with hemophilia was thus exposed to the condensed essence of blood from thousands of donors. The emergence of the mysterious immunological collapse among patients with multiple blood transfusions pinpointed the cause of the illness to a blood-borne factor that had contaminated the supply of factor VIII—possibly a novel virus. The syndrome was renamed acquired immunodeficienc
y syndrome—AIDS.
In the spring of 1983, against the backdrop of the early AIDS cases, Dave Goeddel at Genentech began to focus on cloning the factor VIII gene. As with insulin, the logic behind the cloning effort was immediately evident: rather than purifying the missing clotting factor out of liters of human blood, why not create the protein artificially, using gene cloning? If factor VIII could be produced through gene-cloning methods, it would be virtually free of any human contaminants, thereby rendering it inherently safer than any blood-derived protein. Waves of infections and deaths might be prevented among hemophiliacs. It was Goeddel’s old T-shirt slogan brought to life—“clone or die.”
Goeddel and Boyer were not the only geneticists musing about cloning factor VIII. As with the cloning of insulin, the effort had evolved into a race, although with different competitors. In Cambridge, Massachusetts, a team of researchers from Harvard, led by Tom Maniatis and Mark Ptashne, were also racing toward the factor VIII gene, having formed their own company, named the Genetics Institute—colloquially called GI. The factor VIII project, both teams knew, would challenge the outer limits of gene-cloning technology. Somatostatin had 14 amino acids; insulin had 51. Factor VIII had 2,350. The leap in size between somatostatin and factor VIII was 160-fold—almost equivalent to the jump in distance between Wilbur Wright’s first airborne circle at Kitty Hawk and Lindbergh’s journey across the Atlantic.