The Wizard and the Prophet2
How much grain is required to produce a pound of beef, pork, or chicken? The small farm down the street from me as I write provides one answer: zero. It has fifteen cows and a dozen pigs, all fed by grazing fallow land and eating scraps (uneaten or damaged produce, pulled weeds, spent pea and bean vines, and so on). Industrial farms, the source of the vast majority of meat on grocery-store shelves, provide a different—and, alas, more complicated—answer. Beef farmers in the Midwest buy 650-pound steer that have been raised on pasture and feed them leavings in the form of silage (mowed grass and clover, wheat, and maize plants that are cut after harvest) and distillers’ grains (maize, rice, or barley from which the starches have been stripped to make products like beer, ethanol, or high-fructose corn syrup). The feedyards where steer are fattened before slaughter surround ethanol and corn-syrup plants like moons around a planet. As a result, the animals form a critical component of the overall grain industry—but actually eat less grain themselves than one might think.
Every uptick in meat consumption is associated with a bump in grain production. But the precise amount of the increase is not straightforward. For beef, it is affected by a host of other factors, including the subsidies for ethanol, the price of corn syrup (and the sugar for which it is a substitute), and the demand for leather, bone, fat (an ingredient in airplane lubricants), keratin (extracted from hooves and used in fire-extinguishing foam), and other meat by-products. Matters are just as complex for pork, chicken, and farm-raised fish. Almost no matter what the scenario, though, if tomorrow’s newly affluent billions are as carnivorous as Westerners today, the task facing tomorrow’s farmers will be huge. Between 1961 and 2014, the world’s meat production more than quadrupled. Simply reproducing that jump could easily require doubling the world’s grain harvest.
To double grain output by 2050, arithmetic indicates, harvests would have to rise by an average of 2.4 percent per year. Unluckily, yields are nowhere close to keeping up. A widely cited study from 2013 demonstrated that average global increases in wheat, rice, and maize production have been between 0.9 percent and 1.6 percent per year, about half of what is needed. And in some areas harvests aren’t increasing at all. “Basically, the breeders have been pulling rabbits out of their hats for fifty years,” Kenneth G. Cassmann, a yield specialist at the University of Nebraska, told me years ago. “Well, they’re starting to run out of rabbits.”
Logically speaking, only two paths to increasing harvests exist. One is to lift actual yields—the yields produced by farmers, some of whom are better at their work than others. If they are provided better equipment, materials, and technical advice, farmers can bring their harvests closer to the theoretical maximum. The other is to increase the potential yield—the theoretical maximum—which should bring up the actual yield with it.*7
Both approaches were employed in the Green Revolution. By planting more land, deploying more irrigation, and pumping in more synthetic fertilizer, farmers increased their actual yields. At the same time, Borlaug and his successors at IRRI and CIMMYT dramatically increased the potential yield of wheat and rice by breeding high-yielding dwarf varieties. Channeling the energy of photosynthesis and the nutrition provided by fertilizer into grain, these varieties had a “harvest index”—the percentage of the plant’s mass that is grain—of about 50 percent, almost twice the previous figure. (For maize, dwarfing didn’t work, because the shorter plants shaded themselves too much. Instead scientists bred plants that could tolerate being packed closer together.) The sum of the two methods was the Green Revolution.
The situation is different today. Farmers can’t plant much more land; in Asia, almost every acre of arable soil is already in use. Indeed, as cities expand into the countryside the supply of farmland may be decreasing. Nor can fertilizer be increased; it is already being overused everywhere (except some parts of Africa). Irrigation, too, cannot readily be expanded. Most land that can be irrigated is already irrigated. Some increase in actual yield is certainly possible. But most scientists believe they must raise the potential yield—which brings us back to rubisco.
Nature, as one recalls, has not been able to develop more-efficient rubisco. But evolution has produced a work-around: C4 photosynthesis. Named, prosaically, after a molecule with four carbon atoms that is involved in the process, C4 involves a wholesale reorganization of leaf anatomy. The change is almost invisible to the naked eye, but has profound consequences for the plant.
Ordinary photosynthesis is a cycle with two main stages. In the first stage, chloroplasts trap solar energy and use it to break apart water molecules into hydrogen and oxygen atoms. This sequence is known as the “light” reactions because it makes use of sunlight. The hydrogen is plugged into the second stage; the oxygen filters out of the cell and into the air.*8 In the second stage—the “dark” reactions—the hydrogen from the first stage combines with the carbon from the carbon dioxide grabbed by rubisco. The result is a compound called G3P that other cellular mechanisms break down and rebuild into the sugars, starches, and cellulose that make up plants. In ordinary photosynthesis, both stages—the light and dark reactions—take place in a layer of cells right below the surface of the leaf. The excess gases and the sugars, starches, and cellulose produced in this layer of photosynthetic cells pass into the interior of the leaf. The gases filter up through spaces in the cells to small holes in the leaf surface while the other materials are passed down into interior cells and then into veins and the rest of the plant.
By contrast, C4 plants split photosynthesis in half. The light reactions—the reactions in which chloroplasts use captured solar energy to break apart water molecules—take place near the leaf surface, as in ordinary photosynthesis. But something different occurs with the dark reactions, those that incorporate carbon dioxide. When carbon dioxide comes into a C4 leaf, it is grabbed not by rubisco but by a different enzyme that uses it to form a compound known as malate (this is the molecule with the four carbon atoms). The malate is then pumped into special cells in the interior of the leaf called “bundle sheath” cells.
Credit 43
Bundle sheath cells are deep inside the leaf, wrapped in a living layer around the veins. In C4 photosynthesis, bundle sheath cells are where rubisco acts. Because they are deep in the leaf, oxygen from the air doesn’t easily slip into them. Meanwhile, carbon dioxide is released from the imported malate. Almost without oxygen and boosted with carbon dioxide, each bundle sheath cell is a microscopic replica of the ancient atmosphere in which photosynthesis evolved. More than 3 billion years ago the atmosphere had a hundred times as much carbon dioxide as it does now and almost no oxygen. Rubisco’s inability to distinguish carbon dioxide and oxygen was not a problem, because oxygen was rare. Bundle sheath cells have little oxygen, and rubisco is denied the opportunity to mistake it for carbon dioxide. C4 photosynthesis is thus much more efficient. Barely 3 percent of the flowering plants are C4, but they are responsible for about a quarter of all the photosynthesis on land.
The impact of C4 is evident to anyone who has looked at a recently mowed lawn. Within a few days of mowing, the crabgrass in the lawn springs up, towering over the rest of the lawn (typically bluegrass or fescue in cool areas). Fast-growing crabgrass is C4; lawn grass is ordinary photosynthesis. The same is true for wheat and maize. Plant them on the same day in the same place and soon the maize will overshadow the wheat—maize is C4, wheat is not. In addition to growing faster, C4 plants also need less water and fertilizer, because they don’t waste water on reactions that lead to excess oxygen, and because they don’t have to make as much rubisco. And they better tolerate high temperatures—C4 is especially common in the tropics.*9
Remarkably, C4 photosynthesis has arisen independently more than sixty times. Maize, tumbleweed, crabgrass, sugarcane, and Bermuda grass—all these very different plants evolved C4 photosynthesis on their own. When many different species develop the same traits, the implication is that a lot of plants are “pre-adapted” to create that trait. Somewhere in the
ir DNA, very likely, are genetic switches that promote it.
Further evidence for this idea is that a few species are intermediate—some parts of the plant use ordinary photosynthesis, some use C4 photosynthesis. One of these in-between species is maize: its main leaves are C4, whereas the leaves around the cob are a mix of C4 and ordinary photosynthesis. If two forms of photosynthesis can be encoded from the same genome, they cannot be that far apart. Which in turn implies that people equipped with the tools of molecular biology might be able to transform one into another.
In the botanical equivalent of a moonshot, an international consortium of almost a hundred agricultural scientists is working to convert rice into a C4 plant—a rice that could grow faster, require less water and fertilizer, withstand higher temperatures, and produce more grain. Funded largely by the Bill & Melinda Gates Foundation, the C4 Rice Consortium is the world’s biggest genetic-engineering project. But the term “genetic engineering” does not capture the ambition of the project. What the researchers are trying to develop bears the same resemblance to typical genetically modified organisms that a Boeing 787 does to a paper airplane.
The genetic engineering that appears in news reports largely involves big companies like Monsanto sticking individual packets of genetic material, often taken from other species, into crops. The paradigmatic example is Monsanto’s Roundup Ready soybean, in which DNA from a bacterium found in a California waste pond is inserted into soybeans, making them assemble a chemical compound in their leaves and stems that resists Roundup, Monsanto’s widely used herbicide. The foreign gene lets farmers spray Roundup over their fields, killing weeds but leaving soy plants unharmed. Except for making this one tasteless, odorless, nontoxic substance—a protein with the unwieldy moniker of 5-enolpy-ruvylshikimate-3-phosphate synthase—Roundup Ready soy plants are, in theory, wholly identical to ordinary soy plants.
The C4 Rice Consortium is trying something different in scale and process. Rather than companies tinkering with individual genes to sell branded goop, the consortium is trying to refashion the most fundamental process of life—with the intent of giving away the result. And instead of slipping genes from other species into rice, the initiative is hoping to switch on chunks of the DNA already in rice to create, in effect, a new, more productive species—common rice, Oryza sativa, will become something else, Oryza nova. (Or, possibly, the team may use genes from related species that are similar to rice genes but are for technical reasons easier to manipulate.)
Although plant breeding has advanced since Borlaug’s day, it remains a long, labor-intensive process. At IRRI, rice seedlings are sprouted in climate-controlled tanks (bottom, left), then manually sown in greenhouses (top). Much as Borlaug and his team did at Sonora and Chapingo, IRRI staffers still sort through the harvested grain by hand (bottom, right). Credit 44
When I visited IRRI, scores of people were doing what science does best: breaking a problem into individual pieces, then attacking the pieces. Some were sprouting rice in petri dishes. Others were trying to find chance variations in existing rice strains that might be helpful. Still others were studying a model organism, a C4 species called Setaria viridis. Faster growing than rice and not needing to be raised in paddies, Setaria is easier to work with in the lab. There were experiments to measure variations in photosynthetic chemicals, in the rate of growth of different varieties, in the transmission of biochemical markers. Twelve women in white coats were sorting rice seeds on a big table, grain by grain. More workers were in fields outside, tending experimental rice paddies. All the appurtenances of contemporary biology were in evidence: flat-screen monitors, humming refrigerators and freezers, tables full of beakers of recombinant goo, Dilbert and XKCD cartoons taped to whiteboards, a United Nations of graduate students a-gossip in the cafeteria, air-conditioners whooshing in a row outside the windows.
The cell’s photosynthetic machinery is programmed by scores or even hundreds of genes. As the project began, it was plausible to doubt that so much DNA could be altered in a controlled fashion. Multiple techniques for genetic engineering have been invented, but at the time the most common for cereals like rice, wheat, and maize involved shotgunning thousands of microscopic particles of gold or tungsten at plant embryos (the first precursor cells for leaf, root, and stem, which in this case have been pulled out of the seed and grown in petri dishes). The particles are coated with snippets of DNA that contain desirable genes. In a process that astonished biologists when they discovered it could occur, a few of these particles slam through the cell walls and hit the cell nucleus in just the right way that the nucleus—or, more exactly, the DNA in the nucleus—incorporates the new bits of DNA. And every now and then the DNA from the particles is transferred in a way that allows the new genes to be switched on. The method was clumsy; because it inserted the DNA randomly, its effects were unpredictable. And it could insert only one gene at a time. Nobody knew whether it was possible to alter multiple groups of genes in this way and end up with a coherent result. In 2012 scientists at Harvard and Berkeley unveiled a new method of gene editing called CRISPR that promised more precise control. The C4 Rice Consortium went on alert.
The project has two main goals: (1) locating and switching on the precursor genes that will create the physical structures of C4 photosynthesis (the bundle sheath cells as well as a network of extra veins for them to wrap around), and (2) locating and switching on the precursor genes that create the substances involved in C4 photosynthesis (the malate-producing enzyme and other molecules that are involved in the reaction). In a sense, they want to create both the arena and the players in the arena. Initial research suggests that about a dozen genes play a major part in the leaf structure; another ten genes, perhaps, have an equivalent role in the biochemistry. Alas, properly altering them to create a C4 organism, difficult as that is, would be only the first step. The next, possibly more arduous, would be breeding varieties that can channel the extra growth provided by photosynthesis into grain, rather than roots or stalk. All the while, the new varieties must be disease-resistant, easy to grow, and palatable to their intended audience of several billion Asians.
“I think it can happen, but it might not,” Langdale, the project leader, told me. She was quick to point out that even if C4 rice runs into insurmountable obstacles, it is not the only biological moonshot. Nitrogen-fixing maize, wheat that can grow in saltwater, enhanced soil microbial ecosystems—the list of possibilities is as long as imagination allows.*10 The odds that any one of them will succeed may be small. But the odds that all of them will fail are equally small.
Kantian Interlude
The protests were a surprise. Not many activists come to Brentwood, a small town about thirty miles east of San Francisco. Nobody expected that they would slip into the strawberry patch at night and rip out 2,200 seedlings. The company that had planted the strawberries was Advanced Genetic Sciences, of Oakland, California. Its researchers discovered that most of the vandalized plants were still alive and replanted them. The next night, guards watched the plot from a white van. Protesters snuck up on the van and slashed its tires. The next day, April 24, 1987, technicians in moon suits sprayed the replanted strawberries with bacteria.
The bacterium was Pseudomonas syringae. In ordinary circumstances P. syringae sits on plant leaves and obtains nutrients from dust and rainfall. Like all bacteria, it has a protective coating on its surface. The coating contains a protein that interacts with water in a fashion most unfortunate for farmers. Liquid water doesn’t turn easily into solid ice; it must be cooled well below freezing to crystallize spontaneously. But water molecules will transform quickly into ice if they have an object—a nucleus, scientists say—to crystallize around. The protein on the outer coat of P. syringae is just the right size and shape for an ice nucleus. As a result, bacteria-coated plants freeze more readily than those without the bacteria. Estimates of the cost to U.S. farmers of P. syringae–induced freezes at the time ranged up to $1.5 billion a year. Researchers at the Universi
ty of California at Berkeley used genetic-engineering techniques to incapacitate the gene that produces the offending surface protein. Advanced Genetic Sciences, based in Oakland, wanted to turn these altered bacteria into a product that farmers could use to protect their fields. In 1983 it announced plans to test its “ice-minus” bacteria by spraying them on plots of strawberries and potatoes in rural California. The hope was that the engineered bugs would crowd out their natural, ice-causing cousins, protecting their host plants from frost. It would be the first release of genetically modified organisms into the wild—and the opening skirmish in a battle that continues to the present day.
The experiment had been approved by the Recombinant DNA Advisory Committee, a semi-official body at the National Institutes of Health formed after a scientific conference in 1975. The conference was itself the culmination of a series of meetings that had begun when scientists realized that they were now able to manipulate DNA to transfer genes between species. Held at the Asilomar center on California’s Monterey Peninsula, the conference was attended by 145 people from thirteen nations, along with sixteen journalists who had agreed to defer publication until the meeting ended. In three days of debate, the group did what it could to assess risks and then set out measures to avoid them. Before closing, the conference issued a declaration that, as conference organizer Paul Berg put it, could be summed by a single sentence: “With reservations, some form of experiments should proceed; some, however, should not.” Later a longer, official Asilomar statement was published. Universities and governments around the world used it as a basis for biotechnology regulation. The United States, for instance, formed the Recombinant DNA Advisory Committee in response.