IN THE SPRING of 1987 California scientists removed the lids from a few petri dishes, scraped out yellowishwhite blotches that smelled like sweaty feet, diluted the material in water, and sprayed it on a tiny patch of potatoes and another of strawberries. With the spraying, genetically engineered microbes made their much delayed debut in the environment.
The blotches were colonies of bacteria, nicknamed “ice minus,” that had been altered in a laboratory so that they would protect plants from mild frosts. Boosters of biotechnology predicted that ice minus would soon be joined by microbes genetically altered to kill caterpillars and weevils, supply nitrogen to the roots of plants, and eliminate the increasing and undisputed menace to the environment of chemical pesticides and fertilizers.
Alternatives to farm chemicals are hardly a new dream. In her landmark 1962 book Silent Spring, Rachel Carson urged the nation to take “the other road,” to seek “biological solutions, based on understanding of the living organisms they seek to control, and of the whole fabric of life to which these organisms belong.“ The dream, however, remains largely unfulfilled. Researchers have identified some 1,500 microorganisms or microbial toxins with the potential to control insects. But in their naturally occurring forms most have proved too short-lived, too slow to kill, and too narrowly specific in their effects to oust chemicals from the farm. Only nineteen microorganisms—eleven bacteria, four viruses, three fungi, and one protozoan—have been registered as pesticides by the Environmental Protection Agency, as compared with some 640 chemicals, used in 24,000 pesticide products. In some areas the introduction of wasps and other predators has successfully controlled insect pests. Altogether, however, only one percent of the world’s 10,000 crop pests can now be controlled effectively by natural biological methods.
With the advent of genetic-engineering techniques, in the mid-1970s, biologists held out the hope that many of the drawbacks to pest control by microbes could be overcome. But in the early 1980s, when agricultural researchers began trying to enhance the efficiency of biological agents through genetic engineering, not everyone perceived it as the fulfillment of Carson’s dream. Indeed, when the plant pathologist Steven Lindow first sought permission, in 1982, to field-test ice minus, he aroused the nation’s misapprehensions, not only about genetic engineering but also about technology in general. It took five years of congressional scrutiny, public hearings, lawsuits, ever-shifting regulatory hurdles, and even vandalism of the fields before Lindow and Advanced Genetic Sciences, a small Oakland, California, biotech firm that had license to use his concept, were allowed to spray their microbes outdoors.
To those of us who gathered for the events, near Brentwood, in California’s Central Valley, and at Tulelake, California, on the Oregon border, the signs of the nation’s anxiety were unmistakable. Both patches of land were surrounded by metal towers and stakes laden with EPA air-monitoring equipment. By regulatory fiat the researchers wore “moonsuits” and respirators as they sprayed, creating an ominous tableau for the television cameras.
These dramatic precautions were abandoned within months. Regardless, only half a dozen live, genetically engineered microorganisms have followed ice minus out of the lab. None is available to farmers. The technique remains, as one biotech executive put it, in an “intellectual cul-de-sac,” trapped there by public perceptions and a formidable regulatory scheme.
Just what is this ice-minus bacterium? Where does it come from? How does it work? How do we know whether it poses any threat at all to the environment?
FARMERS IN THE northern half of the United States spend millions of dollars a year on frost protection: smudge pots to heat the air, sprinklers to wet the crop, wind machines and helicopters to mix the cold ground layer with warmer air aloft, and artificial fogs or foam blankets to prevent heat loss from the ground. Still, they lose as much as $3 billion a year in direct damage from mild frosts. Perhaps more important, the threat of frost limits both the growing season and the northern range of crops.
Traditional plant breeders, who engineer the genes of crop plants simply by selecting successful strains, have pushed back some of these limits. Since the turn of the century they have extended the northern range of corn on this continent by 200 miles and brought wheat to the northern Canadian prairies. Plant biologists are now trying to pinpoint the genes that confer frost-hardiness or rapid maturation, in order eventually to enhance them or to endow more-vulnerable crops with their powers.
Genetic engineering of hardier crops seems to farmers I have spoken with to be a natural extension of plant breeding. Designing bacteria to control frost, however, strikes them as bizarre and unnatural. Yet this concept, too, had its beginnings in a cornfield. It was the serendipitous spinoff, in an odd, backwards sort of way, of a traditional plantbreeding experiment.
The story begins with Paul Hoppe, a U.S. Department of Agriculture corn pathologist at the University of Wisconsin, who was laboring during the 1940s and 1950s to breed new strains of fungus-resistant corn. To ensure that his plants were challenged by fungal epidemics, he would grind up infected corn leaves from the previous year’s crop and sprinkle them on the field. He did this as usual in June of 1961. Two weeks later a late freeze hit the Midwest. The plants sprinkled with dried corn leaves were killed. The rest survived.
When Steven Lindow arrived at Wisconsin twelve years later to work on his doctorate, “this weird thing about corn plants,” as he now describes it, was presented to him. Was there something about the infected leaves that enhanced a plant’s sensitivity to cold? He pursued the matter first as a biochemistry problem, making active extracts from the old leaves and searching for a chemical that made young leaves vulnerable. Then one day he added a bactericide to the extracts and the frost-enhancing activity stopped. He began isolating and testing dozens of strains of bacteria from his leaf extracts, trying to find which ones somehow increased a plant’s susceptibility to low temperatures. In early 1975 his thirty-first isolate gave him an answer.
Lindow had a refrigerator loaded with test tubes full of water and colonies of bacteria that he had isolated from the corn leaves. When he went to the refrigerator one day, he found that the temperature had dipped a little below freezing. Every tube of bacterium No. 31 had frozen. None of the others had.
Water doesn’t necessarily turn to ice at 32°F. Small amounts of pure distilled water can remain liquid at temperatures as low as -40°, a phenomenon called supercooling. Even large amounts of pure water readily supercool to 14°. Drop a catalyst—a nucleating agent that helps orient the water molecules into an ice-like lattice—into a flask of supercooled water and the water will instantly, dramatically, crystallize into ice. Or lower the temperature further and random groupings of water molecules will trigger this phase shift spontaneously.
Lindow soon established that bacterium No. 31, a strain of common leaf bacteria called Pseudomonas syringae, has a protein in its cell membrane that can nucleate ice. In fact, this so-called ice-plus microbe has turned out to be one of the most effective ice nucleators known.
Plants are 90 percent water, and water in plant tissues can supercool. Lindow began to wonder: do the ice-nucleating bacteria that live on plants’ stems and leaves limit supercooling and allow damaging ice crystals to form inside the plants’ tissues at warmer temperatures? In a greenhouse he grew plants under aseptic, or microbe-free, conditions, and found they could tolerate temperatures as low as 23°. When he sprayed the plants with ice-nucleating P. syringae, they froze at 28°. Lindow began washing bacteria from the leaves of dozens of plant species, both crops and weeds; he found ice-nucleating strains on leaf surfaces everywhere. Anywhere from a tenth of a percent to ten percent of the bacteria on a leaf surface may be capable of nucleating ice, he discovered. Of all the ice-plus strains, P. syringae turned out to be the most common. “You probably eat it in every salad,” he told me.
In growth chambers and field plots Lindow began testing various ways of manipulating the microbial populations on corn leaves: killing off all microbes with antibiotics, for example, or coating the leaves with detergents and heavy metals, such as zinc and copper, to inhibit ice nucleation. He also tried something else: spraying plants with bacteria that are naturally ice-minus— that is, bacteria that are essentially neutral and do not possess an ice-nucleating capability—to make sure that plant surfaces were too full to support later-arriving ice-nucleating cousins.
The results confirmed that frost damage declines as the ice-plus population shrinks. But detergents, heavy metals, and antibiotics either are toxic to plants or provide frost protection too short-lived to be widely practical on the farm. And the natural ice-minus bacteria competed poorly with the icenucleating ones. Not until he had completed his Ph.D., in 1978, and accepted a faculty post at the University of California-Berkeley, did Lindow begin to consider the possibility of genetic manipulation of P. syringae.
For every protein an organism makes, it must carry a gene that encodes the blueprint for it. Knock out the socalled ice gene and the bacterium would be unable to make the ice-nucleating protein. Lindow’s team made its first ice-minus mutants not by genetic engineering but by the more traditional (and still largely unregulated) technique of exposing bacteria to mutagenic chemicals. A decade ago, with no fanfare or need for federal approval, they sprayed these mutant bugs on potatoes in a held near Tulelake.
But chemicals — and radiation, which is also sometimes used—are crude. They mutate multiple genes, not just the targeted trait, and leave the bugs “sick in subtle ways,” Lindow explained to me. So in 1981, when genetic-engineering techniques became available for manipulating P. syringae, his team set out to locate the ice gene and remove it precisely, without causing other, haphazard damage. They started with no idea where among the bacteria’s 3,000 genes the ice gene might be, or what its protein product looked like. In the lab one day Douglas GurianSherman, a graduate student, walked me through the process they used.
“THINK OF THE bacteria as plastic bags filled with DNA JL and other stuff,” he said. “You want to break them open and then separate the DNA from the other components.” The terminology breaking, loading, cutting, splicing, inserting—suggests a dramatically mechanical, almost surgical operation. In truth, I found, generic engineering is a chemical and enzymatic process, to the naked eye a seemingly endless mixing and separating of clear liquids. First enzymes, detergents, and acidic compounds are added to a bacterial suspension to lyse, or eat away, the outer membranes of the bacteria. The resulting liquid is spun in a centrifuge to separate out the DNA, which differs in weight and density from other organic molecules inside the bacterial cell.
Once the DNA of P. syringae was isolated, the next step was to chop it into fragments, one of which would carry the ice gene. The chopping was done with restriction enzymes, popularly referred to as “chemical scissors.” Like most of the tools of genetic engineering, these enzymes had originally been isolated from microbes, which use them to disable invading viruses or to cut and splice their own DNA for replication and repair. Researchers now buy dozens of kinds of these enzymes off the shelf.
It is the specificity of the various restriction enzymes that gives genetic engineering its precision. DNA is formed in a linear sequence composed of four chemical bases—adenine, thymine, guanine, cytosine (ATGC)— strung along two complementary strands. A is always found paired with T, G with C. The sequence of bases GAATTC on one strand, for example, will always be paired with CTTAAG on the other, the pairs linked like the teeth of a zipper, the whole zipperlike sequence twisted into a double helical shape. When a restriction enzyme called EcoR1 recognizes the sixbase-pair pattern GAATTC and its complement, it will make a staggered cut in the double strands, slicing each strand between the A and the G. This leaves an AATT tail on one cut end and a TTAA tail on the other—socalled “sticky ends,” single-stranded and eager to link up with a complementary match.
“Often the whole thing is done in a test tube about this big,” Gurian-Sherman said. He was holding up a oneinch vial. “You might have a solution of DNA that’s half the tube, and then you might add a tiny bit of this enzyme. A tiny bit of another. And usually a buffer solution.” Once the DNA is chopped, the researchers generate thousands of copies of each fragment by inserting each into a bacterium for cloning. The first step toward cloning is splicing each fragment onto a small segment of viral DNA called a cosmid. “From a supplier you buy what are called right and left arms,” Gurian-Sherman said. “You just take your fragments and your cosmid arms and mix them. Both need to have been cut with the same restriction enzymes, so that they have sticky ends, ready to link up.”
To begin to create an ice-minus version of P. syringae, Lindow’s team took these cosmid-linked fragments of P. syringae DNA and presented them to tiny viruses, called phages, that attack bacteria and use the bacteria’s cellular machinery to produce multiple new viruses. “The phage sees the cosmid DNA, recognizes it, and pulls it in just as if it were winding string onto a spool,” Gurian-Sherman said. The phages were then allowed to infect Escherichia coli, a bacterium that has become the white mouse of genetic engineering. Inside E. coli the viral DNA multiplied freely, making thousands of clones of the P. syringae fragments with the cosmid DNA.
Which bacterial colonies carried the fragments with the ice gene? E. coli are naturally ice-minus, and so Lindow’s team picked out the ones that now nucleated ice. Then, to isolate the gene further, the researchers did what is called subcloning. Essentially, they removed the fragment and started over, repeating the whole process several times, using ever smaller lengths of DNA. In mid-1982 the team announced the isolation of the ice gene itself—the smallest fragment of P. syringae DNA that could turn an ice-minus bug ice-plus.
Once they had the gene, the researchers gutted it by cutting a chunk out of the center with restriction enzymes and rejoining the ends. Their next move was to insert this crippled gene into the parent strains of P syringae that still carried full copies of the ice gene. This was done by encouraging mating and “conjugal transmission” of the crippled gene from the E. coli to P. syringae. Putting the two strains of bacteria together on a petri dish for several hours is encouragement enough for mating to occur.
By manipulating the environment in the petri dish, Lindow also made it impossible for these P. syringae to survive unless they not only accepted the crippled ice gene but also inserted it into their own chromosomes next to their good ice gene. For the few that survived this test, a final manipulation awaited. Lindow shifted the environment of the dish again, relieving the pressure on the bacteria to keep the new gene. He hoped that in a few of them the new gene would detach itself from the chromosome a little sloppily, taking along a chunk of the bug’s own ice gene.
A few colonies of ice-minus P. syringae, identical to their parents except for a single gene, were all that remained: yellowish-white blotches that smelled like sweaty feet and were no longer able to nucleate ice.
SINCE ITS inception, amid a barrage of frightening headlines in the mid-1970s, genetic-engineering research had been subject to guidelines laid down by the National Institutes of Health. Strict containment of all lab-altered creatures was the goal. No crises occurred, and gradually both public anxiety and federal oversight relaxed. Then, in 1982, Lindow proposed to field-test ice minus.
Congress called hearings. The Washington activist Jeremy Rifkin, an opponent of virtually every application of genetic engineering, filed the first in a series of procedural lawsuits. Ecologists emphasized the problems created by non-native species such as gypsy moths and kudzu vines. In 1983 the EPA entered the fray by declaring ice minus a pesticide under the jurisdiction of the Federal Insecticide, Fungicide and Rodenticide Act. The rationale went like this: ice-plus bacteria encourage plants to freeze and thus are pests; ice minus, if it arrives first, blocks this pestiferous activity and thus is a pesticide. At the same time, the EPA ended its practice of exempting from scrutiny small-scale (under ten acres) field trials of microbial agents. To get permission to spray ice minus on a quarter-acre potato patch, Lindow faced an enormous array of lab tests and paperwork previously required only for products at well-advanced stages of commercialization.
By the time his experiment was approved, Lindow had filed more than 1,300 pages of applications and data with federal and state agencies. In labs and greenhouses he had tested ice minus for pathogenicity on seventy-five plants, from zinnias to sugar beets. He had run an extensive battery of “product identity” tests and scrutinized the bug’s life-style preferences, all to make sure that it had no novel powers that might cause it to run amok.
Two seasons in the field—1987 and 1988—confirmed the bacterium’s staid behavior, and also its ability to protect crops: potato seedlings coated with ice minus received only a third as much frost damage as unprotected plants. Lindow chose not to conduct a third field test last summer. Faced with regulatory uncertainty, ice minus’s future on the farm is far from assured.
A few other lab-altered creatures did go to the fields last summer. BioTechnica Agriculture, of Overland Park, Kansas, tested several genetically modified strains of rhizobia, nitrogenfixing bacteria that have long been used to foster the growth of alfalfa, soybeans, and other legumes. Crop Genetics International, of Hanover, Maryland, tested bacteria designed to protect against rice stem borers and European corn borers. These bacteria, which live inside the vascular systems of plants, had been outfitted with toxin genes taken from another bug, Bacillus thuringiensis, or BT, a natural microbial pesticide sold in the United States since 1961.
None of the tested bugs have ventured away from their release sites or persisted in the environment. Undoubtedly they have changed for a time the mix of microbial populations in the soil or on leaves, but so do chemical pesticides and herbicides, and farm practices such as tilling, irrigation, and crop rotation. In May of 1988 a Congressional Office of Technology Assessment report concluded, “With adequate review none of the small-scale field tests proposed or probable within the next several years are likely to result in an environmental problem that would be widespread or difficult to control.”
THERE IS still no consensus, however, on what constitutes “adequate review.” A draft proposal for revising the EPA regulations was circulated in the last months of the Reagan Administration, but it drew criticism from all sides and finally died at the Office of Management and Budget. A key provision would have created a number of local environmentalbiosafety committees to assume some unspecified oversight role. The EPA also proposed to broaden the scope of the Toxic Substances Control Act to oversee work with naturally occurring organisms. Industry argued that the EPA should move in the opposite direction, exempting all or most smallscale research with microbes. Environmentalists oppose such an exemption but also don’t want the EPA squandering its regulatory energies on bugs commonly found in compost heaps or sewage ponds.
The EPA and the White House are currently working out the Bush Administration’s position. No one, however, expects any dramatic shifts in the regulations. With major environmental legislation like the Clean Air Act under review, genetically engineered microbes have not been a priority on anyone’s agenda this year.
Many in the biotech industry echo the sentiments of Jerry Caulder, the president of Mycogen Corporation, who believes that the current regulations would work well “if we didn’t have people coming along playing what-if to the nth degree on every release.” He adds, “The EPA is in a terrible position. The regulations are supposed to be scientifically reasonable, and yet the public wants an assurance of absolute safety. You really can’t do both.” The EPA has no equivalent of the medical-research community’s clinical trial, in which new therapies are evaluated in relation to standard ones, allowing the better of two often imperfect alternatives to be determined.
The 1988 OTA report stated,
In evaluating the potential risks associated with these new technologies, the appropriate question is not “How can we reduce the potential risks to zero?” but “What are the relative risks of the new technologies compared with the risks of the technologies with which they will compete?” Furthermore, What are the risks posed by over-regulating, or failing to develop fully the new technologies? How do we weigh costs and benefits? How much review is enough?
Until such questions have been answered satisfactorily and the answers guide our regulatory strategies, we will have little choice but to keep relying on the chemical pesticides and fertilizers whose harmful effects are already apparent.
— Yuonne Baskin