“What has long appeared to be simply the agent of a bothersome plant disease,” wrote Mary-Dell Chilton in 1983, “is likely to become a major tool for the genetic manipulation of plants: for putting new genes into plants and thereby giving rise to new varieties with desired traits.” At the time, coaxing Agrobacterium tumefasciens into adding a new gene, any gene, to a plant, was itself bothersome. Building the vector, the ring of DNA that the bacterium would insert into the plant cell, was a laborious process, involving several painstaking steps. Still, the possibilities excited Roger Beachy, a scientist then in the same department as Chilton at Washington University in St. Louis. “While she was making her vectors,” he recalled in 2003, “I was imagining what I could use them for.”

“Vector” comes from the Latin for “carrier.” Biologists had commonly used it to describe the carrier of a disease—the tsetse fly is the

vector for sleeping sickness; the aphid is the vector for tomato mosaic virus. To Beachy, who had been studying plant viruses since graduate school, Chilton’s new vectors were an opportunity to turn the meaning of the word on its head. He was investigating the phenomenon of cross-protection. “If you infect a tomato plant with a mild strain of tomato mosaic virus, you can protect it from infection by a severe strain,” he explained. “This had been known for more than 20 years.” Why it worked, though, was a mystery. It was like vaccination, except that plants don’t have an immune system. Beachy wondered what part, precisely, of the virus provided the protective effect—and could he make Chilton’s vector carry that vaccine-like bit into a plant cell?

Tomato mosaic virus, like many plant viruses, is a single-stranded slip of RNA coated with protein. Once inside a cell (placed there usually by an aphid as the insect tastes a leaf), the protein coat unwraps and the naked RNA begins interfering with the cell’s protein-making machinery, directing it to make viral proteins instead of its usual plant proteins. Beachy thought the key to cross-protection might be the coat protein.

By 1981 scientists had sequenced the tobacco mosaic virus genome, a close cousin to the tomato mosaic virus. It was a manageable 6,400 nucleotides long. Beachy had earlier mapped the order of this virus’s genes, so he knew more or less where to look for its coat protein sequence. He made DNA clones of it and of other parts of the genome, then approached Chilton with the idea of using her vectors to insert each bit of viral DNA into a plant cell. If the plant carrying the coat protein gene became resistant to the virus and the others did not, he would know the source of cross-protection.

Chilton was game to try, but an offer to work for Ciba-Geigy came before the experiment could be done. Rather than giving up, Beachy sought out one of Chilton’s competitors in the so-called Agrobacterium race. Across town in St. Louis was the headquarters of Monsanto. There, in a laboratory for which both Chilton and Jef Schell from the University of Ghent had consulted, Robert Fraley and Stephen Rogers agreed to take on the project. With Fraley’s and Rogers’s assistance, and with funding from Monsanto, Beachy and a student made a vector. To the coat protein gene they attached a promoter, a DNA sequence that

turns on a gene. A promoter is critical: without it the process by which DNA is transcribed into RNA and translated into protein stops at step one. The promoter acts like a hook: to start transcribing the gene, the enzyme RNA polymerase latches on to the promoter.

Beachy chose the wrong promoter. Although Agrobacterium neatly carried the new gene into the plant cell, where it combined with the plant’s own DNA in a process that came to be known as transformation, the promoter was too weak to be effective, Beachy said. “We could barely detect the coat protein. It was there, but at extremely low levels. The student was terribly disappointed.”

Fraley and Rogers, however, knew a trade secret: Monsanto had a new promoter. Designated 35S for the length of the original RNA whose synthesis it directed, it came from cauliflower mosaic virus, known as CaMV. Patented in 1994 by Fraley, Rogers, and Rob Horsch, all working at Monsanto, the CaMV 35S promoter became the standard gene promoter used in experiments with plants. In the early 1980s, Beachy said, “They wouldn’t tell me much about it. Monsanto did the transformation and we got the plants back. Our first success came with tobacco. Then came tomato, then petunia. We had a lot more coat protein, so we had resistance.” Expressing just the gene coding for the virus’s protein coat was indeed enough to protect the plant against infection.

As with many advances, including smallpox vaccinations in people, the knowledge that it worked preceded a deeper understanding of how it worked. Several years later scientists working in Beachy’s lab discovered that, in plants containing the coat protein gene, the virus was unable to come apart. The viral RNA was not released into the cell, and the cell was not infected. This coat protein-mediated resistance, as they called it, allowed plants to fight off infection by more than one closely related virus. “For example, plants that were made resistant to tobacco mosaic virus,” Beachy explained, “were also resistant to tomato mosaic virus.”

It was discovered recently that virus resistance can also develop by a different biochemical mechanism called post-transcriptional gene silencing. When a gene—whether viral or some other—is expressed at too high a level, a feedback process is triggered that shuts down its function. This process affects the RNA transcript. The gene continues

to be transcribed, but the transcript is destroyed. With no template from which to make the protein, no protein is made. A gene has to be almost identical to the overexpressed gene to be shut down along with it. No other gene in the plant is affected.

Post-transcriptional gene silencing works directly on RNA. Most plant viruses use RNA as their genetic material; their genes are made of RNA, not DNA. When a virus infects a plant cell, it directs the plant cell to make lots of viral RNA, triggering the gene silencing mechanism directly. Post-transcriptional gene silencing has a memory, as does our immune system, though it works in a very different way. When a plant is infected a second time with the same virus or a very closely related virus, the gene silencing system is reactivated, and the invading viral RNA is destroyed as soon as it takes off its coat.

This phenomenon is the cross-protection Beachy was interested in. At the time it seemed logical—perhaps by analogy to the immune system—to suppose that cross-protection worked because the plant recognized the viral coat protein. In some cases, the coat protein is indeed the active agent. But more often this type of gene silencing works by recognizing and destroying the RNA. RNA silencing is as precise as the human immune system, although it works by a completely different principle.

Beachy, Fraley, and Rogers’s successful T-DNA contained the DNA code for the coat protein. It was inserted into the plant’s genome using Agrobacterium. It worked because the strong CaMV 35S promoter caused the gene to be transcribed so often, that is, to produce so much RNA, that the post-transcriptional gene silencing system was triggered. The RNA of an infecting virus was thus recognized and destroyed as soon as it entered the cell.

In 1985 Beachy, Fraley, and Rogers applied for a patent on their technique to protect crops from viral diseases. The patent was finally awarded in 2003. By then, virus-resistant varieties of squash, papaya, tomato, pepper, cucumber, sugar beets, and plums—all created using the coat protein technique—were on the market. Virus-resistant sweet potatoes were in development. And Beachy, by then director of the Donald Danforth Plant Science Center, a private research institute in St. Louis, was at work on virus-resistant rice.

Some of the virus-resistant plants, as well as the technique used to make them, are patented. Patents are widely misunderstood as ownership of an idea or an invention. Yet to get a patent the inventor must make public enough information that someone “skilled in the art” can reproduce the invention (if given a license or other legal right to do so by the patent owner). A patent merely lets the inventor prevent others from using the idea without permission for a limited period of time, usually 17 to 20 years. It is easier to own, to control, an idea if it is kept secret and not patented.

Patents are meant to encourage creativity by letting the inventor get a head start on making a profit. But they are also intended to further scientific progress. Before Monsanto patented the CaMV 35S promoter it was a closely held trade secret. Even a collaborator like Beachy was kept in the dark. Once patented, it was available to researchers around the world, sometimes free, sometimes for a licensing fee, depending on whether the researcher was at a university or other research institute or in a company.

The patent system has been affecting western agriculture since at least 1750. Until the mid-1700s farmers in Europe and America were most influenced by the Latin poet Virgil. According to historian Mauro Ambrosoli, who studied farming handbooks published between 1350 and 1850, much of the advice given was not based on science or on practical experience, but on classical poetry. And much of it was particularly absurd when applied to England: Virgil wrote 2,000 years ago, for instance, “If fat the soil, let sturdy bulls upturn it from the year’s first opening months, and let the clods lie bare till baked to dust by the ripe suns of summer.”

Jethro Tull, a lawyer and Oxford graduate, challenged this status quo. Experimenting on his own Howberry Farm, he advised reducing the quantity of seed sown. Instead of broadcasting seed—flinging it out of baskets by the handful—he invented a seed drill to plant individual seeds neatly in rows. Instead of sowing rye, barley, and clover together, to keep down the weeds, he devised a horse-drawn hoe to handle the weeding. He saved seeds from his own harvests to use the next year. For this he was fiercely attacked by the Private Society of

Husbandmen and Planters, whose leader was gardener and seed merchant Stephen Switzer, the author of “good works on agronomy,” Ambrosoli writes, that were “frankly inspired by” Virgil. The seeds sown in England were, at the time, often grown in Scotland and imported for resale by seed merchants like Switzer. Buying new seed each year prevented blight and smut and cut down on weeds; seeds from the cold north also gave a better yield when planted in the fertile south. “The attack on Tull becomes more understandable,” Ambrosoli continues, “if connected with the wish to maintain control over the seed market and to avoid losing buyers of the illustrated booklets sold with the seeds.”

Jethro Tull was a gentleman farmer. He did not patent his seed drill or his horse hoe, but after 1750, Ambrosoli notes, “every new invention concerning agricultural machinery was jealously defended with patents.” In America by 1836 the head of the U.S. Patent Office, Henry Ellsworth, believed it was his responsibility to encourage the introduction of new plant varieties along with the invention of new machines. The idea that plants themselves could be patented grew stronger in the 1870s, when Louis Pasteur, the inventor of milk pasteurization, attempted to patent a type of yeast. Luther Burbank was among those who argued at the time that if a yeast could be patented so could a plum or a potato. Yet, as his biographer, Peter Dreyer, notes, “The degree to which cultivated plants were human creations was not generally recognized in those days.” No one thought of a potato as intellectual property.

It was Burbank’s posthumous testimony that led to the passage of the first bill in the United States to include plant breeders among the ranks of inventors. As Dreyer writes in A Gardener Touched with Genius, one of the main opponents of the bill was Congressman Fiorello La Guardia, later mayor of New York. “Acting apparently on misunderstood protests he had received, La Guardia successfully blocked passage, until the sponsoring congressman, Fred S. Purnell of Indiana, inquired what he thought of Luther Burbank. ‘I think he is one of the greatest Americans that ever lived,’ the New Yorker replied flamboyantly. Purnell then proceeded to read into the record a letter written by Burbank to Paul Stark shortly before his death. ‘I have been for years in

correspondence with leading breeders, nurserymen, and federal officials, and I despair of anything being done at present to secure to the plant breeder any adequate returns for his enormous outlays of energy and money,’” Burbank had written. The letter went on:

As Dreyer relates, La Guardia immediately withdrew his objection to the bill and it passed both House and Senate.

The Plant Protection Act of 1930 covered cultivated plant varieties, whether they had been developed by a breeder like Burbank, or discovered in the field by a plant explorer, as long as they were reproduced asexually. (Curiously, tubers were excluded: Burbank still could not have patented his famous Idaho potato.) Such a plant patent was intended to prevent anyone except the developer or discoverer from producing the variety for a number of years. The stricture that the plant had to be propagated asexually—through grafting, cuttings, runners, or the dividing of bulbs—arose from the need to preserve the distinctive characteristics of the variety, which sexual reproduction tended to disrupt.

Yet many sexually reproducing plant varieties do maintain their distinctiveness—growers say they breed true—and by the 1960s these were given patent-like protection in some European countries. After a number of unsuccessful attempts, the U.S. Congress did so as well, passing the Plant Variety Protection Act in 1970. To receive a plant patent, the plant variety must be uniform, stable, and distinct from other varieties. To encourage breeders to develop new varieties, the act set up a mechanism to provide exclusive marketing rights by certifying seed.

The 1970 act excluded fungi and bacteria. So when Ananda Chakrabarty, through his employer, General Electric, applied for a

patent on a strain of bacteria that could efficiently degrade crude oil, his application was denied. Crude oil is a complex mixture of hydrocarbons. Several strains of the bacterium Pseudomonas putida were known to produce enzymes that degraded one or another of those hydrocarbons—that is, the bacteria ate the hydrocarbons, nourishing themselves and reproducing. But mixtures of strains that could eat different hydrocarbons weren’t very efficient when faced with the task of degrading crude oil. Some grew better than others. The overall efficiency of converting crude oil to bacterial biomass (which is 75 to 80 percent protein) wasn’t very good.

Chakrabarty discovered that the ability to degrade hydrocarbons was due to genes on plasmids that the bacterium carried. Different strains had different plasmids and so could degrade different hydrocarbons. Chakrabarty devised a way of combining the multiple plasmids in a single strain. Usually, similar plasmids can’t replicate in the same bacterium—they exclude each other—so Chakrabarty used ultraviolet radiation to fuse the multiple plasmids inside of bacterial cells, creating a strain with one large plasmid. It carried genes coding for enzymes that could degrade many different hydrocarbons, and it grew better on crude oil than did a mixture of strains, each carrying a single plasmid. Because Chakrabarty’s bacteria could be useful in cleaning up oil spills, General Electric decided to apply for a patent on the process of constructing bacteria with such fused plasmids, as well as on the organism itself.

The patent application was filed in June 1972. In late 1973 the U.S. Patent and Trademark Office issued its decision. It accepted the process claim, but did not grant a patent on the bacterium, which it said was a product of nature and therefore not patentable. General Electric’s patent attorney appealed the decision, arguing that Chakrabarty’s bacterium was not a product of nature; it had been changed by fusing incompatible plasmids. The appeal board acknowledged this argument, but still rejected the patent claim because it sought to patent a living organism.

General Electric appealed to the U.S. Court of Custom and Patent Appeals. In 1978 the court granted the patent. In its view, as long as a microorganism was useful, novel, and the product of human interven-

tion, it merited patent protection. The decision that living organisms could be patented was appealed to the Supreme Court, which returned it to the Court of Custom and Patent Appeals, where it was reaffirmed. The court argued that patent law should apply to “anything under the sun that is made by man,” and that this included living organisms.

The Chakrabarty case changed how patent law was interpreted. Although Chakrabarty didn’t use molecular techniques in constructing his oil-eating bacteria, by the time the decision was final, most of the recombinant DNA techniques in use today had been invented. The first biotechnology companies had been started, and the first patent applications had been filed.

Based on the Chakrabarty precedent, the patent office board of appeals issued an administrative ruling in 1985 that plants could be patented without following the special provisions of the 1930 Plant Patent Act or the 1970 Plant Variety Protection Act. In 1987 the board ruled that a polyploid oyster, a multicellular animal, could be patented. In issuing this ruling, Donald Quigg, the assistant secretary and commissioner of patents and trademarks, explained that the patentability of living organisms required that “they must be given a new form, quality, properties, or combinations not present in the original article existing in nature.”

One consequence of patenting “anything under the sun that is made by man” is that nearly any DNA sequence can be patented. A cloned DNA sequence—although identical to the sequence in the organism from which it was derived—is not considered to be a “product of nature” because it required human intervention to clone it. Patents have therefore been issued on isolated bits of DNA. This practice is consistent with the longstanding one of issuing patents for chemical compounds purified from naturally occurring mixtures, but it continues to be problematic. For plant breeding, the patenting of DNA has two faces. On the positive side, it brings such trade secrets as the CaMV 35S promoter out into the open. On the negative side, it creates an additional, time-consuming step between discovery and application.

As soon as his patent application was filed in 1985 and he was free to talk about his invention, Roger Beachy called a colleague, Jerry Slightom at UpJohn Company, which then owned Asgrow Seeds. In a few months Asgrow had negotiated the rights to the viral coat protein technology and had begun developing a virus-resistant variety of yellow squash. Later, Asgrow was acquired by a multinational seed company, and a broad technology licensing agreement that included Beachy’s invention was made with Monsanto. One item in that package concerned soybeans. “The access to elite, high-yielding soybean lines was an opportunity,” Beachy recalled, “for a series of experiments that—10 years later—would become Roundup Ready soybeans.”

Slightom, however, did not stop with squash. He placed a call to Cornell University plant pathologist Dennis Gonsalves, an expert on the papaya ringspot virus and a native of Hawaii, and told him about Roger Beachy’s new technique. Gonsalves talked with Beachy, then approached John Sanford, also at Cornell. Sanford had by then applied for a patent on his gene gun (the so-called “cowboy method” of transforming a plant), and was in the process of forming a company, Biolistics Inc. (from biology and ballistics), to market the new tool. He was close to announcing the first transformation of corn. Gonsalves persuaded him to experiment with papayas and the coat protein of the papaya ringspot virus as well.

The papaya ringspot virus is spread by aphids from weeds and other wild plants that generally show no symptoms. Papaya are grown from seed, but don’t begin producing for 9 to 12 months. A fully grown, 30-foot-tall papaya tree is productive for at least 3 years. When infected by papaya ringspot virus, however, the plants are stunted and their fruit is misshapen and tasteless; eventually the papaya tree dies, depriving the farmer of not only one harvest, but several.

In the United States, the Hawaiian islands are the only place papayas will grow. Oahu had been the center of papaya growing until the virus wiped out its papaya farms in the 1950s. Papaya growing moved 19 miles across the water to the Puna district of Hawaii, where the virus was kept under control by burning all infected papaya plants and all wild plants or other crops nearby that could harbor the virus. By the 1990s papaya was being grown by 200 to 300 farmers, each averaging

10 acres of trees; their annual production of over 50 million pounds of fruit was valued at $17 million.

Yet the papaya ringspot virus was only stalled, not defeated. Nineteen miles of ocean is not an uncrossable distance for an aphid. Aphids have traversed the Pyrenees from France into Spain, and the Atlantic Ocean from Europe to North America. The easiest (and most probable) way for them to travel is to hitch a ride on a fruit or twig or flower carried by a person. And once on Hawaii the virus would spread very rapidly, because the island’s climate is very comfortable for aphids. The winters are so mild, the insects do not need to mate and lay eggs to keep the population going through the colder months, and so there are no male aphids in Hawaii. Instead every day each female aphid gives birth to between 8 and 22 live clones of herself. If she carries the papaya ringspot virus, so do they. The population of virus-carrying aphids quickly reaches the millions.

Insecticides provide no relief. It takes an aphid less than a minute to infect a plant. By the time a farmer has noticed the infestation and gotten out the spray, the damage is done. Millions of dollars had been spent on conventional breeding programs to create a papaya that could withstand the ringspot virus, but with little success. No naturally occurring genes for resistance to papaya ringspot virus had been found in any papaya variety.

In 1987 Gonsalves recruited horticulturalist Richard Manshardt at the University of Hawaii and plant physiologist Maureen Fitch from the USDA’s Hawaii Agricultural Research Center; later University of Hawaii plant pathologist Steve Ferreira would join the group. Without knowing exactly how having the papaya ringspot virus’s coat protein gene could protect the papaya from the virus, the team decided to try it. They cloned the coat protein gene from a Hawaiian isolate of the virus, then used Sanford’s gene gun to insert a vector containing the gene into young embryos of a commercial Hawaiian papaya named Sunset. Using tissue culture techniques, they grew the transformed embryos into plants and tested their resistance to the virus. They found the first plant to show resistance, number 55-1, in 1991.

In 1992—coincidentally the year the virus was detected in the Puna district—number 55-1 underwent field trials in Hawaii. It proved to

be remarkably resistant to the virus which, meanwhile, was spreading rapidly. By 1994 almost half of Puna’s papaya acreage was infected, and a number of farmers had gone out of business.

The University of Hawaii researchers crossed 55-1 with commercial cultivars to produce the varieties SunUp and Rainbow. In 1994 the team received a permit from the USDA to conduct a large-scale field trial of these papayas near an abandoned orchard in Puna. The results were striking. Within a year, all the plants without the coat protein gene were infected, while only three of those with the new gene showed symptoms. Better yet, all of the resistant plants were still healthy two years later. The fruit was good, and the farmers approved.

But hurdles remained. Before the papaya seeds could be distributed to farmers—free of charge—licenses had to be obtained from the corporations and universities who held patents on the genes and processes that had been used to make the variety. Then the virus-resistant papaya had to be vetted by the appropriate government agencies. Hawaii’s Papaya Administrative Committee (a group of papaya growers organized by the USDA) began negotiating for licenses from Monsanto, Asgrow Seeds, Cambia Biosystems, and the Massachusetts Institute of Technology. The University of Hawaii and Cornell scientists prepared the documents required by the government agencies. It wasn’t at all clear that these two processes could be accomplished in time to save the papaya industry in Hawaii. Because the problem was so severe, the USDA funded a program to produce seeds of the resistant varieties so that they would be ready for immediate distribution—if and when the permissions came through.

In the United States any crop modified using molecular techniques—and only those crops—must be scrutinized by at least two, and often three, government agencies. The crop is evaluated as a potential toxic substance, pesticide, or plant pest, and sometimes as a food additive as well.

In view of humankind’s long history of tinkering with food plants, this state of affairs is very odd. Over the last century breeders have

learned to churn up plant genomes in many new ways. They cross plants from different species and even different genera. They use tissue culture, chemicals, and radiation to make mutants—plants that might be more resistant to drought, disease, and pests, or that might provide more or tastier food than the unmutated variety. Their techniques have become increasingly complex, and increasingly invasive.

For example, in 1979 Shivcharan S. Maan of Fargo, North Dakota, received United States Patent No. 4,143,486. “The invention,” wrote Maan, “satisfies the long felt need for a relatively simple, commercially feasible method of producing hybrid wheat seeds.” Maan’s method begins with a single cell from the weed Aegilops squarrosa, or goat grass, one of the ancestors of modern wheat. First the nucleus of the cell is removed and discarded. Then a new nucleus, taken out of a cell of bread wheat, Triticum aestivum, is inserted into the goat grass cell. This alloplasmic wheat cell is grown, using tissue culture techniques, into a mature plant, and the seed is harvested. This seed is then exposed to radiation or soaked in a mutagenic chemical. In the example included in his patent, Maan notes that 500 seeds from plants having A. squarrosa cytoplasm and a T. aestivum nucleus were exposed to a mutagenic chemical for 16 hours before being planted in a greenhouse. The seed spikes of the mature plants were covered with plastic bags so that the plants would self-pollinate. “The 45,000 seeds obtained from self-pollination were harvested and planted in the field.” Maan and his colleagues did not know what types of mutations the chemical treatment might have caused. To learn, they grew the plants to maturity and “visually examined” them for “abnormalities,” choosing the 39 plants that could not make pollen to include in further breeding. The thousands of others they simply plowed under.

No federal agency checked to see if these plants were toxic or allergenic. No federal agency needed to be asked before the plants could leave the greenhouse. No federal agency needed to approve the scale-up of the field tests. Asked if he needed to file an environmental impact statement before field-testing his new plants, one long-time breeder burst out laughing. “Oh!” he said, “there’s nothing like that. In none of these other types of breeding are the genes considered foreign. You can make species crosses, generic crosses. Those aren’t foreign. Even

somaclones—somatic hybrids. Even a hybrid between a mouse cell and a tomato cell is not policed. There’s not a legal thing you have to do. Traditionally bred cultivars have no legals.”

Nor has anyone ever suggested that mutagenized or irradiated crop seeds—like Maan’s experimental hybrid wheat—should be planted behind double fences topped with barbed wire and guarded around the clock. Yet that’s how the first genetically modified corn plants were grown at the USDA’s Beltsville Agricultural Research Center in the late 1980s. Unlike the surveillance of Parmentier’s potatoes, growing outside Paris in 1786, the guard was not meant to convince passersby that the plants were valuable, nor even to protect them from being stolen. It was called for only because the rules for field-testing this new variety of corn were based on outsized fears that crop plants modified using molecular techniques were somehow new and might be dangerous.

Those fears date from the early days of recombinant DNA. By 1972 DNA from almost any source could be cloned in bacterial cells by recombining it with plasmid DNA. Some scientists—among them Paul Berg, who would later receive the Nobel Prize for his pioneering recombinant DNA work—began to worry that, by using these techniques, researchers might unwittingly create a new human pathogen. If it escaped from the laboratory, it could never be recalled.

Berg’s lab had developed a way to splice together DNA from different organisms several years before Boyer and Cohen did. But Berg chose not to replicate the molecules in E. coli or other bacteria, fearful of the biological consequences. Conversations at meetings of molecular biologists in the summers of 1972 and 1973 often converged intensely on whether recombinant DNA experimentation could unintentionally create hazardous organisms. By 1974 Berg had persuaded Boyer and Cohen, as well as other prominent molecular biologists, to state their concerns in public. They wrote a letter to the National Academy of Sciences and published it in Science magazine. It was titled: “Potential biohazards of recombinant DNA molecules.”

This self-organized group of concerned scientists formulated four goals: to institute a moratorium on recombinant DNA research, to address public fears about creating new genes, to consider what would happen if such genes got out of the laboratory, and to ask the director

of the National Institutes of Health, Robert Stone, to assemble a committee that would write guidelines, carry out risk assessment, and convene an international group of scientific leaders to discuss the guidelines.

Stone quickly formed a 15-member committee, the first Recombinant DNA Advisory Committee, better known by its colorful—and arguably appropriate—acronym as the RAC. An international meeting was convened at the Asilomar Conference Center in California. Scientists from all over the world argued, sought to persuade each other, and eventually crafted a framework for carrying out experiments.

Not everyone agreed that regulations were necessary. Joshua Lederberg, discoverer of bacterial sex, said that just the act of regulating recombinant DNA research would make people think it was dangerous, whether it was or not. Reflecting on what happened in the following decades, Jim Watson was quite blunt, saying: “And boy, he was right.”

Donald Fredrickson assumed the directorship of the National Institutes of Health (NIH) in mid-1975 and began to assess his responsibilities. He recalls: “Little did I know what was coming.” But from the beginning he was determined to address both the social and the scientific issues. The NIH established an Office of Recombinant DNA Activities and began converting the Asilomar framework into what would become the NIH Guidelines for Recombinant DNA Research. The first version was issued in June 1976. Fredrickson was clear that the guidelines were not regulations. Because they addressed hypothetical, not known, hazards they would evolve as knowledge accumulated and experience grew.

Not unexpectedly, many in the federal government and in communities around the country disagreed. They began to clamor for laws to regulate recombinant DNA research. After all, if scientists themselves had raised questions, surely there was something to worry about. But laws are difficult to undo. The research might prove quite harmless. If it did, laws would impose unnecessary restrictions, interfering with the development of much-needed new pharmaceutical products, such as human insulin, interferon, tissue plasminogen activator, and hepatitis B vaccine. Fredrickson urged President Gerald Ford to extend

the NIH guidelines to all federal and private research. After some delay, the directive was issued in September 1976, and the Federal Interagency Committee on Recombinant DNA Research was formed to look at the regulatory authorities of each agency and to see whether new laws were really necessary. Efforts to pass laws did not stop. Over the next several years, 12 bills to regulate DNA research were introduced in Congress; none was passed.

As soon as the NIH guidelines were issued, it was clear that they needed revising. The Asilomar framework, for instance, was limited by the knowledge of those who had attended the meeting. None of these scientists worked with pathogens and the sometimes deadly toxins they produce. So the first guidelines were dominated by the fear that recombinant DNA could turn the laboratory strain of the common gut bacterium E. coli into a virulent and contagious pathogen. Experts in pathogens, brought in to educate the Recombinant DNA Advisory Committee, explained that a pathogen was more than simply an organism that can produce a toxin. A pathogen needs a complex delivery mechanism to invade another organism and deliver its toxin to the right cells. Scientists who studied pathogens had learned how to grow them carefully and analyze them in ways that did not jeopardize their own health. They pointed out that recombinant DNA techniques could make their research even safer. Instead of working with the pathogen itself, they could study individual pathogen genes—including toxin genes—by expressing them in a crippled laboratory bacterium that had no capacity to deliver its poison.

Each of the other issues that had aroused early fears was examined in equally careful detail. The number of laboratories and companies carrying out recombinant DNA research expanded rapidly—none of the hypothetical disasters materialized. Confidence in the safety of recombinant DNA steadily grew. Revision by revision, the guidelines were relaxed. More kinds of experiments were classified as exempt, as holding so little potential for harm that they could be carried out under ordinary laboratory conditions. Today we know that, far from being inherently hazardous, recombinant DNA technology is among the safest technologies ever developed.

In a memoir Fredrickson concluded: “It is possible that the ‘re-

combinant DNA affair’ will someday be regarded as a social aberration, with the guidelines preserved under glass. Even so, we can say that the beginnings were honorable. Faced with real questions of theoretical risks, the scientists paused and then decided to proceed with caution. That decision gave rise to dangerous overreaction and exploitation, which gravely obstructed the subsequent course. Uncertainty of risk, however, is a compelling reason for caution. It will occur again in some areas of scientific research, and the initial response must be the same.”

Yet even as the guidelines for laboratory research were being relaxed, the next area of concern was emerging. An unfortunate turn of phrase—“release into the environment”—had been chosen to describe field tests of plants and animals, including bacteria and other microorganisms, modified by recombinant DNA techniques. Unlike the familiar term “field test,” “release experiment” had the unintentional and menacing implication that, once released, these plants or animals could not be controlled.

The NIH approved the first formal release in 1983. As biologist Paul Lurquin writes in High Tech Harvest: Understanding Genetically Modified Food Plants, the experiment in which one researcher drank a vial of genetically engineered E. coli bacteria “to demonstrate that they could not survive in the human gut” did not count as a release. Nor did the estimated 100 million recombinant organisms that were escaping the world’s labs every day on the clothes and shoes of each scientist or technician. Henry Miller, a former FDA official now at Stanford University, noted, “A vast and varied unsupervised ‘release experiment’ involving tens of thousands of laboratories and untold millions of discrete new genotypes of recombinant microorganisms has been in progress for three decades, with no known untoward results.”

The first organism officially released was an altered strain of a common bacterium, Pseudomonas syringae. The normal Pseudomonas is a nuisance to farmers. It damages crop plants by causing ice to form on their leaves. Just as clouds form by the condensation of water around particles of dust in the sky, ice forms more readily when nucleated by a particle, even a protein. Certain bacteria on leaves produce a particularly efficient ice-nucleating protein. Pseudomonas syringae contains a

gene, the ice+ gene, that encodes the ice-nucleating protein. Plant pathologist Steven Lindow of the University of California, Berkeley, isolated the gene, deleted much of it, and reintroduced it into the bacterium. Then he sprayed his “ice-minus” strain on strawberry plants to see if the temperature at which ice formed could be reduced by a crucial few degrees, letting farmers protect their strawberries from late spring frosts.

Because of state regulations the researcher who did the spraying, Julianne Lindemann, wore a “moon suit”—a full-body white coverall, with helmet and breathing apparatus. Her picture became an icon for protesters against recombinant DNA technology, in spite of the fact that it implied danger where none existed. It was taken, as one biologist points out, “by unprotected photographers standing some 10 feet away”—photographers who saw no danger to themselves in being exposed to ice-minus bacteria. (Ice-minus variants of P. syringae are, in fact, common in nature, comprising about half of the population.) Afterwards, activists made a practice of tearing out Lindow’s strawberry plots, and work on ice-minus was brought to a halt. “The unbelievable irony,” writes Lurquin, “is that the bacteria he modified to prevent ice formation in plants are used today in large amounts in their normal configuration (that is, containing the ice nucleation gene) to generate snow on ski slopes during warm winter weather. I wonder sometimes if the people who destroyed Lindow’s test plots strapped on their skis right after their acts of vandalism.”

The NIH was reprimanded by a federal court for approving the release of ice-minus bacteria without filing a formal environmental assessment, and in early 1984 a Cabinet Council Working Group was formed to bring the federal agencies’ representatives together to work out the regulatory issues. It proposed a Coordinated Framework for the Regulation of Biotechnology, published in 1986. The agencies’ efforts to avoid new legislation began a process with outcomes that few people—except perhaps Lederberg—foresaw.

The regulation of plants and plant foods fell to three federal agencies. Under the Federal Plant Pest Act of 1957 and the Plant Quarantine Act of 1912, the USDA has the power to decide if a new plant variety or microorganism is likely to become a pest. Its Animal and

Plant Health Inspection Service (APHIS) keeps a list of actual plant pests, on which can be found both Agrobacterium tumefasciens and cauliflower mosaic virus. Consequently, all plant varieties created using Agrobacterium-mediated transformation or the CaMV 35S promoter—which includes almost all of those that have come to be labeled genetically modified—were defined as “regulated articles.” The term was coined by the USDA to capture, for case-by-case regulation, not only bona fide plant pathogens, but organisms that “may be” plant pests.

The USDA further extended that definition to include any organism that contains DNA from different genera. This extended definition covers all plant varieties created through recombinant DNA techniques and Agrobacterium-mediated transformation. Plants produced using wide crosses between genera or the techniques Maan used to create hybrid wheat are not, however, regulated. A breeder using molecular modification techniques must obtain a permit for field tests and, after several years of tests, petition APHIS to “deregulate” the new crop, that is, to take it off the list of regulated articles. The petition must describe the genes, regulatory sequences, and transformation procedure used, analyze the plant’s genetic and agronomic traits, and provide data on any environmental consequences of growing the plant. APHIS then performs an environmental assessment. If the agency decides that a new variety poses an environmental risk, it can halt all further development.

The second federal agency, the EPA, likewise identified two existing statutes that could apply. These were the Toxic Substance Control Act (TSCA) and the Federal Insecticide, Fungicide, and Rodenticide Act (FIFRA). TSCA was passed in 1976 to allow EPA to track industrial chemicals. FIFRA, passed in 1947, lets the EPA assess the toxicity of chemicals and living organisms used to control fungi, insect, and animal pests and set “tolerances,” the amounts of such chemicals that can persist in the environment or in foods without creating a health hazard. To stretch these statutes to cover plants and animals modified by recombinant DNA techniques, the EPA found it necessary to define such organisms as “new.” An organism would be “new” if significant human intervention had been used to develop it. (In practice, only

those modified by molecular techniques are considered “new,” although it would seem that the amount of intervention Maan used to develop hybrid wheat, for instance, is “significant.”)

Under these two acts the EPA regulates plants, such as Bt corn or virus-resistant papaya, modified to protect themselves from insects or viruses. It regulates microorganisms, such as the ice-minus bacterium or one, under development, that would allow a plant other than a legume to fix nitrogen. And it is concerned with new uses of existing pesticides, such as the pairing of Roundup (glyphosate) with Roundup Ready soybeans—but only if the herbicide-resistant plant is produced using molecular techniques. Herbicide-resistant crops created by chemical mutation or by the somaclonal variation that results from conventional tissue-culture techniques—and several such crops are now on the market—are not regulated. The EPA examines data provided by the plant breeder on the inserted genes and their products. It reviews the risks and benefits to the environment. It can require the breeder to prepare a resistance management plan, so that insects or diseases do not quickly become resistant to what it used to call a plant pesticide and now defines with the less disquieting name of plant incorporated protectant or PIP. It checks if the protectant is toxic to animals or humans and, if it is, sets a tolerance level for residues allowed in food.

The mandate of the third agency, the FDA, comes from the Federal Food, Drug, and Cosmetic Act, signed into law in 1938. By the 1940s the FDA had realized that “the vast research efforts needed to assure that all food chemicals were safe was clearly beyond all foreseeable resources,” according to a history published by the agency. In the 1950s three amendments to the law “fundamentally changed the character of the U.S. food and drug law: the Pesticide Amendment (1954), the Food Additives Amendment (1958), and the Color Additives Amendment (1960),” the history explains. “With these laws on the books, it could be said for the first time that no substance can legally be introduced into the U.S. food supply unless there has been a prior determination that it is safe. By requiring the manufacturers to do the research, a problem of unmanageable size was made manageable.” The onus is on the seller to ensure that the food is safe.

For a crop created through molecular techniques, the FDA recommends that it be compared to a standard variety to determine whether it is substantially equivalent. If the food’s nutrient content has been changed—as in Golden Rice, for example, but not the new Hawaiian papaya varieties—the new crop requires a more stringent review, as do any crops that contain an antibiotic-resistance marker gene, which is treated as a food additive. (Most plant developers are now using new marker genes that do not rely on antibiotics.) As its guidelines state, “The FDA considers, based on agency scientists’ evaluation of the available information, whether any unresolved issues exist regarding the food derived from the new plant variety that would necessitate legal action by the agency if the product were introduced into commerce. Examples of unresolved issues may include, but are not limited to, significantly increased levels of plant toxicants or anti-nutrients, reduction of important nutrients, new allergens, or the presence in the food of an unapproved food additive.”

Through legal action the FDA can take a crop or the food derived from it off the market—no matter how it was developed—if it is found to be unsafe. As an FDA publication noted in 2000, “First and foremost, the law simply forbids the marketing of unsafe food. Anyone who violates this provision may be held criminally liable, the food may be seized and destroyed, and the establishment can be required to cease doing business until it complies with the law.” For this reason the FDA’s guidelines, unlike those of the other two agencies, were not originally mandatory. It was “prudent practice,” the agency stated, “for developers of new varieties to consult with the agency on safety and regulatory questions, especially with regard to products developed through new technology.”

Through this stretching of definitions familiar crop plants—corn, wheat, rice, cotton—long subjected to a variety of genetic manipulations came under the regulatory purview of the USDA, the EPA, and FDA, but only if the genetic technique used to modify them was molecular.

This process-based definition makes no biological sense. As the Council of the National Academy of Sciences pointed out in a 1987 publication, it sets one kind of genetic modification apart from all

those that breeders have used for decades. The Council had asked a small group of experts in molecular techniques, ecology, evolution, and plant pathology to examine the various kinds of environmental and health problems that might arise from modifying plants, animals, and microorganisms by recombinant DNA (R-DNA) techniques. The committee reported that the many thousands of plants that had been made using these methods had not revealed unexpected hazards. Indeed, the problems were familiar ones. They were the same as the problems of plants modified by the many other genetic techniques in use. The committee concluded: “Assessment of the risks of introducing R-DNA-engineered organisms into the environment should be based on the nature of the organism and the environment into which it is introduced, not on the method by which it was produced.”

But it was already too late. The double standard was firmly in place. Despite wide adoption of this language, and the recognition that regulation should not be process-based, it is.

It took from 1995 to 1997 for the USDA, EPA, and FDA to approve the virus-resistant papaya varieties SunUp and Rainbow even, as Gonsalves writes, “with excellent cooperation from these agencies.” Licensing agreements with the several owners of the patented technologies used were reached in April 1998. Papaya seeds were distributed to Hawaiian farmers—free of charge—in May 1998, and the first SunUp and Rainbow papayas were sold in U.S. markets that year. Almost miraculously, the Hawaiian papaya industry bounced back, with production in 2000 nearing levels that hadn’t been seen for half a decade. By 2003 75 percent of the papaya grown in Hawaii were Rainbow or SunUp. Under the headline, “Stalked by Deadly Virus, Papaya Lives to Breed Again,” the New York Times in July 1999 credited the genetically engineered papayas with having saved the Hawaiian papaya industry.

In spite of this positive publicity, Hawaiian papaya seeds, created by university and government scientists and given away free, are not the genetically modified food that first comes to consumers’—or activists’—minds. As tracked by the International Service for the Acqui-

sition of Agri-biotech Applications (or ISAAA), the genetically modified food grown in 2002 by more than 5 million farmers on 145 million acres in 16 countries worldwide is the product of five large, multinational corporations: Aventis, BASF, DuPont, Monsanto, and Syngenta. Some 90 percent of the market is Monsanto’s alone.

The 145 million acres (more than 95 million acres in the U.S.) are planted in varieties of soy, canola, cotton, and corn that can tolerate an herbicide (the list includes the glyphosate in Roundup; the glufosinate in Basta, Challenge, Rely, and Finale; and the sethoxydim in Poast) or that make their own Bt pesticide, a crystalline protein from the bacterium Bacillus thuringiensis. Two crops (cotton and corn) can do both. The world leader, with 62 percent of the area farmed, is herbicide-tolerant soy.

So few acres are planted in virus-resistant papaya that they do not register on the ISAAA’s charts. Because of their market predominance, Bt corn and Roundup Ready soy, canola, and cotton, all sold by Monsanto, have come to mean GM or GMO in the minds of most protestors against genetically modified foods. Monsanto has been dubbed “Monsatan” by activists in the United Kingdom and has become a scapegoat for the industry. Debating Lord Peter Melchett, an organic farmer and former U.K. Labour minister in 1999, Monsanto C.E.O. Robert B. Shapiro was “surprisingly contrite.” He sounded, wrote Michael Specter in the New Yorker magazine, “like one of those Chinese leaders who during the Cultural Revolution were made to walk through the streets in a dunce cap. ‘Our confidence in this technology and our enthusiasm for it has, I think, widely been seen, and understandably so, as condescension or indeed arrogance,’ he said. ‘Because we thought it was our job to persuade, too often we forgot to listen.’”

For every other crop technology, from hybrid corn to the latest Intellicoat Early Plant corn—each seed encased in a hot-pink polymer coating that monitors the soil temperature, keeping the seed from sprouting when spring comes late—the company’s job has solely been to persuade. Those who needed persuading were farmers, not consumers. Seeds of new GM varieties bought from seed companies cost more than those not modified by molecular techniques. Farmers must agree not to save seeds for replanting the next year. Most farmers do not save

seeds of corn, soy, canola, or cotton anyway. The practice hasn’t been common in America since the 1930s, because saved seeds often transmit diseases. But the idea of signing a contract to that effect rubs some farmers the wrong way.

Despite the high costs and the contracts, the acreage planted in GM varieties of corn, soy, canola, and cotton continues to grow. From 2001 to 2002 it increased about 10 percent in the U.S., 12 percent worldwide. The reason, according to one analyst, is that farmers found genetically modified seeds “an attractive commercial option. Farmers planting herbicide-tolerant GM soybeans in the United States, for example, could gain roughly $6 per acre in reduced herbicide costs, despite technology fees and no change in yields.”

Farmers have not been uniformly satisfied, as with any new technology. Some have run afoul of the intellectual property rules. By saving or selling seed, or even by failing to weed out those seeds that escaped the harvest and sprouted as volunteers the next season, they can bring on Monsanto’s private investigators (clued in via the company’s toll-free tip line by a law-abiding neighbor). As with teenagers nabbed for pirating music over the Internet, the farmers caught tend to expostulate loudly about freedom. Monsanto, like the rock stars who object to their music being “shared” without royalties being paid, comes off looking like a bully.

Protesters have picked up on these intellectual property disagreements between multinational companies and family farmers to argue that genetic engineering will give control of our seed—our food—to heartless corporations. Yet the protesters too “forgot to listen.” Some genetically engineered crops—like the virus-resistant papaya—were designed specifically to solve the problems of small, family farmers. They make no profit for Monsanto, even though Monsanto controls the patent on the coat protein technology; it was donated. The fact that there are so few of these kinds of genetically modified crops is not the result of any failure in the technology. According to many scientists, the reason a handful of companies dominates the market is the cost of complying with federal regulations. Gonsalves and his colleagues persevered; many lack the resources and simply give up.

Ingo Potrykus hopes that the Golden Rice he and Peter Beyer pat-

ented will be given away free, and that farmers will be encouraged to save the seeds. He noted in 2003 that activists have “not completely unjustified concerns that big companies might dominate the seed markets and food production. But this has nothing to do with the technology,” he said. “If following the regulations were easier, small companies could come up with comparable products. I very much dislike this concentration process,” he added, referring to the fact that the vast majority of genetically modified crops being grown were developed by large profit-making companies and not by universities or nonprofit agricultural research institutes. “The opposition is against it, too,” Potrykus continued, “but they are the cause of it. They’re the ones who have made it so expensive.”

To Roger Beachy the regulations are “so impositional that we are really working hard to exclude the public sector, the academic community, from using their skills to improve crops.” Noting that the number of field trials run by universities has greatly decreased since 2000, he said, “It’s not surprising, because the cost of product development is so high that there is little chance that an academic research team can develop new varieties for commercial release.” And it is the universities, the academic researchers, who have historically made a point of breeding local varieties, like the Hawaiian papayas, that are meant to benefit small farms.

“When we brought out the coat protein technology,” Beachy said, “APHIS, the health inspection service of the USDA, was interested in it. After a significant review, they said that the new crop varieties represented little or no risk. In fact, at one time they considered that coat protein genes could be deregulated as a class of genes. Vegetables have viruses all the time. We’ve been eating these viruses and their coat proteins for many, many years.” A new papaya that expresses a viral coat protein gene contains, in fact, less of the coat protein than an ordinary papaya from a virus-infected grove. And yet the USDA was not the only federal agency involved. Beachy recalled, “The EPA took a different view. They said, A virus is a pest. Therefore, the gene that stops a virus must be considered to be a pesticide. In some ways the EPA has begun to regulate the process not the product. The USDA, which was once a stalwart of product-based regulations, is now in danger of be-

coming more process-related in its policies.” Plant breeders, he said, are becoming “seriously discouraged.”

The effect of the regulations has been, as Joshua Lederberg foresaw, to make people think the technology is dangerous, whether it truly is or not.