Jay Keasling carries a picture in his mind of a place he has never seen, of children he has never met. But it is an image that radically changed how he felt about his research in chemistry—and has set Keasling, a Berkeley chemical engineering professor, on a quest far grander than he ever imagined.
The picture in Keasling’s mind is of a medical clinic in Malawi, filled with thin, feverish children lying in cribs. Some die, succumbing to a virulent strain of cerebral malaria that leaves them deeply unconscious or convulsing as if they were in a devastating car accident. But the lucky ones—the ones who reached the clinic in time—receive a dose of an expensive drug called artemisinin. A day or two later, these children are on their feet, playing in their cribs.
“Playing!” Even now, almost two years after Keasling heard the searing description, he shakes his head in wonder at the magic of a chemical compound that pulls children back from the shadow of a horrible death. At the time, he had not expected to be jolted. He had just finished giving a lecture about his work, this one at Michigan State University. His slides described the science and laid out the case for the research: artemisinin drugs were too expensive for most people in developing countries. Afterward, Keasling visited with a few researchers doing related work, including one who spent half her year teaching at Michigan and the rest in Malawi, performing autopsies on child victims of malaria. Her descriptions swept away the academic tidiness of Keasling’s work. To her, artemisinin was almost a mirage: a compound derived from a plant with tremendous healing power yet out of reach of those who needed it most. As he listened to her, Keasling realized that his work wasn’t just about science anymore. If all went right, he might make the price of a dose of artemisinin plummet to the value of a U.S. postage stamp. He could save lives.
Keasling, 42, is a biological and chemical engineer. He has fused chemistry and biology in a novel way to design microbes to churn out specific products. He has helped catalyze a unique collaboration of academia, private industry, and philanthropy to move his research from the laboratory into real products. And he has found a passion that runs even deeper than the satisfaction that he has drawn from pushing the boundaries of academic science. His research has taken on the shape of those Malawi children rising, shaking off death, and playing—playing!—in their cribs.
“This is what I’m doing,” he says, with an unblinking gaze. “My plan is to make sure this gets out—and to do it to the very end.”
Students who have worked with Jay Keasling at Berkeley marvel at his steady temper, his optimism. Nothing seems to unnerve the matter-of-fact scientist with the square jaw, compact build, and fondness for Italian-made shoes. When complicated pieces of machinery are irrationally balky, or when experiments that took days to set up yield nothing, Keasling shrugs and carries on.
“As bad as the day gets,” he confides, “it’s still better than shoveling pig manure.” He means it. Keasling grew up on a farm in Nebraska, the only son of a farmer who raised cattle, soybeans, and at least 200 pigs. Keasling doesn’t like to talk about pigs. “They’re smart, but they’re mean,” he says curtly. If he never saw another pig—except as a side of bacon on his breakfast plate—Keasling would be a happy man.
But he learned a lot of lessons on the farm that have served him since the day he dusted the dirt off his jeans and headed to the University of Nebraska to study biology. The ethos of farming—the ceaseless battle between farmer and nature, each trying to bend the other to its will—was deeply etched into Keasling. Nature wasn’t something that you just admired; it was something you worked with, battled, tended, nurtured—in short, did whatever it took to get the results you wanted. “Biology is pretty robust,” Keasling says. “You don’t have to worry too much about tinkering with it.”
And farm work was continual, too, filling every moment of the day from dawn until dusk, and afterward. Keasling took to heart the core lesson of farm life: If you want to have a prayer of getting things done, you’d better get up early in the morning.
When Keasling took his first genetics class at the University of Nebraska in 1983, biotechnology was gaining momentum. The first flush of biotech companies—including the likes of Genentech and Amgen—were still refining how to manipulate the genetic material of different organisms. Genes were the key. Scientists had decoded how genes served as instruction manuals for making just about anything, including proteins, which in turn are the catalysts for so much of what our bodies do—growing, moving, carrying oxygen, fending off infections, and so on. Genetic engineering promised to take just the right snippet of DNA, put it into an organism that acts like a natural copying machine, and let it rip. The resulting protein could then be given to people whose bodies failed to make enough of the right protein on their own.
It was fascinating—but to the restless young man from the farm, too much of biology still seemed devoted to observing nature rather than taming it. Keasling opted for engineering.
When he reached the Ph.D. program in chemical engineering at the University of Michigan, Keasling had to catch up on engineering with a few undergraduate courses. A class in process controls dazzled him. He marveled at how engineers could isolate one element of a system, then fine-tune it to do a job. Twist the red knob to add heat; twist the blue knob to release pressure.
Biologists, by contrast, were working with the equivalent of half-built Rube Goldberg machines. They had a powerful tool: the ability to add or delete a gene. But a single change could trigger a cascade of unanticipated reactions. Sometimes researchers got what they wanted; more frequently they did not. Worse, they couldn’t reliably predict what might happen.
Keasling envied the controls of the process engineers. “I wanted to manipulate a cell like an engineer does a chip,” Keasling says. Electrical engineers string together components—transistors, resistors, capacitors—to build a system that behaves in predictable ways. Keasling wanted to do the same thing inside a cell. “I wanted to create the equivalent of oscillators, pulse generators—subsystems that can turn reactions off and on—anything that would give us finer control.”
After a postdoc at Stanford University, Keasling joined Berkeley in 1992, intent on devising biological subsystems. Senior colleagues had doubts. One quietly advised him not to bother: “He told me, ‘Jay, we have all the tools we need. Work on modeling,’” (predicting what a cell might spew out when it got a new gene) rather than on trying to tinker with the mechanisms themselves.
But Keasling remained stubbornly focused on trying to build tools that would let him do chemistry inside a cell, say, catalyzing a series of reactions just as the “metabolic pathways” in plants or microbes do when they encounter different chemicals. His timing was pitch-perfect. Elsewhere, scientists were sequencing the genomic structure of organisms from humans to corn—then putting the data on the Internet or into electronic databases, available with a few clicks of a computer mouse. To Keasling, these databases became like extensive lists of ingredients he could make or use in the reactions he catalyzed within cells.
Other researchers also were trying to bring an engineering mind-set to biology. At the Massachusetts Institute of Technology, for instance, longtime computer science pioneer Tom Knight was building a new team aimed at building biological circuitry. Knight and collaborators, including Drew Endy, popularized the term “synthetic biology” to describe this emerging discipline of catalyzing precise reactions within living organisms to create novel biological products or systems. Some work on “software” for cells—ways of instructing a cell to do something differently than it naturally would do, such as blinking off or on. Others, notably Keasling, work on the equivalent of rewiring cells. For instance, inside the cells of a plant, hundreds of enzymes work to convert carbon into a more complex molecule. The results of one chemical reaction trigger another, which triggers another—and so on—until a rose produces a sweet scent or a maple tree leaks sap. Synthetic biologists may isolate a snippet of those reactions—a sort of recipe called a metabolic pathway—and then try to re-create it in an entirely different organism. If they pick the right recipes and figure out how to ensure that the reactions dovetail nicely (instead of catalyzing unanticipated results), then they might be able to get the second organism to do something beyond what nature intended.
By the late 1990s, Keasling was developing a collection of tools, reliable metabolic reactions that he could depend on. Biotechnologists had long used “promoters” like switches to turn genes off or on. Keasling developed what he nicknamed a “bio-rheostat”—like an individual dimmer switch—which could gradually turn on a microbe’s genes, one by one. He began plucking the design of metabolic pathways from various organisms and trying them in different microbes—designing new biological circuits and trying to get the organisms to do something new.
At the same time, Keasling built an eclectic laboratory team, looking for talented people in diverse fields. Among them: Vincent Martin, a postdoctoral microbiologist from the University of British Columbia, with a deep interest in using engineered organisms to combat pollution created by the pulp and paper industry; and a Ph.D. student, Neil Renninger, who had degrees in environmental and chemical engineering from MIT. Later, Jack Newman, who earned his Ph.D. in microbial systems from the University of Wisconsin, Madison, joined, too. A full year before Newman finished his degree, he had been won over by a lecture he heard Keasling give. “That was it,” Newman declares. “I didn’t even apply anywhere else for my postdoc.”
As they developed their tools, Keasling’s team looked for ways to use them by customizing microbes to clean up waste and by creating new products. For instance, the researchers added new pathways to bacteria so that the microbes could absorb heavy metals out of wastewater and then deposit the contaminants on their cell walls. The water could be cleaned by precipitating out the metal-loaded microbes. Other microbes got pathways that made nerve agents look like nutritious snacks. Even so, Keasling and his team knew the work was unlikely to leave the laboratory. “The public won’t let you clean the environment with microbes,” he says tersely.
Making products with microbes is a different story. The biggest category of byproducts created by plants is “terpenoids”—chains of hydrocarbons that become plants’ fragrances, colors, resins, flavors, and other byproducts. Some help a plant reproduce, attracting the insects or creatures that pollinate it. Other terpenoids are rudimentary defense systems, warding off hungry pests or choking weeds. Scientists have identified more than 50,000 terpenoids, many of which show tremendous medicinal promise, such as the extract from the Pacific yew tree, which is used to make the anti-cancer drug Taxol. But because chemists can’t easily brew them and plants make them only sparingly, many terpenoids are not cheap.
Other researchers had tried sliding a plant gene into microbes, but even the most successful attempts produced only minuscule amounts of relevant material. “The genetics of plants are not well suited for microbes,” notes Martin, the microbiologist.
Keasling’s approach: add another pathway or two to create a chain of reactions that would ultimately lead to the terpenoids he wanted to produce. For starters, he took a chemical pathway typically found in yeast—which happened to be triggered by a chemical common in bacteria—and wove it into his colonies of E. coli.
That combination, in turn, produced so much particulate byproduct that it threatened to drown the bacteria—until Keasling added a third subsystem, a pathway from a plant, which could convert that stream of byproducts into a chemical that was a step closer to a terpenoid. To Keasling, that third step—the pathway from a plant—gave the system its particular character—would it make one terpenoid or another? Even better: Others could be swapped in its place. All he had to do was to pick a manageable terpenoid.
He considered trying to make carotenoids — a group of well-characterized compounds that give salmon its pink glow, make carrots orange, and tomatoes red, and so are an important flavor, fragrance, and colorant. “They make very pretty microbes,” Keasling observes.
Fate intervened. A member of Keasling’s team pointed out an article describing how another scientist had cloned the first gene in amorphadiene—a compound that led to a drug called artemisinin. Typically, artemisinin was derived from a Chinese weed called sweet wormwood (a close relative of American sagebrush). But artemisinin had two intriguing characteristics: It could cure malaria and was of little interest to the big-league pharmaceutical companies.
Artemisinin was a relative newcomer to the crusade against malaria. In the late 1960s, during China’s Cultural Revolution, the Chinese government launched an ambitious program to investigate the properties of plants used in traditional herbal medicines. Among them was the weed qing hao—in Latin, Artemisia annua, or sweet wormwood. Herbalists used teas made of qing hao for centuries to treat hemorrhoids and reduce fevers. (Europeans used oil from a local variant, wormwood, to flavor vermouth and absinthe.)
About five years later, Chinese scientists announced that they had created a new remedy for malaria by soaking qing hao leaves in ethyl ether and extracting the active chemical. Western scientists were skeptical, in part because the Chinese gave scant details of their techniques. But by the mid-1980s, international scientists confirmed the results.
It was none too soon. Malaria was surging through Africa and other parts of the developing world. Decades of haphazard use of chloroquine, the cheapest remedy against malaria, had incubated a disaster: Throughout much of Africa, the parasite that causes malaria grew resistant to chloroquine and other mainstream remedies. By the early 1990s, malaria had grown fierce, wiping out an estimated 1.5 million people a year (particularly children and pregnant women), and collectively costing African nations as much as $12 billion a year.
But artemisinin was tough to get. Crops of sweet wormwood were low and unpredictable. Manufacturers used diesel fuel as the solvent to extract the active chemicals, hardly a benign process. Unscrupulous traders hoarded supplies, driving up prices. Chloroquine treatment could be had for about 10 cents a treatment; a comparable regime of artemisinin officially costs $2.40—or, on the black market, as much as $27, reported experts in Africa.
To Keasling, artemisinin sounded like a grand target. “We like to do things in a big way,” he says. “And artemisinin was big.”
So much real science happens in quiet moments. There was no magical instant when Keasling’s team realized that their techniques and theory would work. There were long, unmarked hours through days and evenings, when the incubator shakers in Keasling’s lab hummed softly, swishing glass flasks filled with mocha-colored solution. There were countless times when one of the researchers would draw a few milligrams of the pungent, yeasty liquid from the flasks with a long syringe, then squeeze a few drops into tiny vials for additional testing. And there were afternoons when they would put the vials into gas chromatography machines—hulking machines that whirred and blinked red lights as they baked the samples until they turned to gas, before spitting out charts striped with faint bars. The researchers would squint a bit and debate—was it there? Yes, there was a trace: Their microbes were making amorphadiene, a precursor to artemisinin. Could the microbes do better? Could they produce more? They designed new experiments, switched microbes. The data came steadily back: Yes, they could make amorphadiene. Yes, they could produce a lot more.
By the spring of 2002, Keasling was confident that his techniques could reliably churn out amorphadiene. The science was invigorating: What should they do with it? Martin urged Keasling to think broadly: What about designing a company around making terpenoids?
Soon, Keasling and his team, including Martin, Newman, and Renninger, were putting in even longer hours. Once a week, after a day in the laboratory, they would meet at Keasling’s house with yellow notepads to list questions and ideas about building a company around their science. None of them had ever worked in a company, much less started one. Newman and Renninger signed up for seminars about entrepreneurship and managing your own business.
After a year of late-night pizzas, take-out Chinese food, and wine from Keasling’s cellar, they incorporated Amyris (at the suggestion of Martin’s wife), named for a plant that produces fragrant oil, frequently used as a substitute for sandalwood. They knew they wanted to be in the business of producing artemisinin but were unsure how to fund the work. Malaria had no shortage of victims, but those victims lacked money to pay for drugs. No money meant no market, no eager venture capitalists, in short, no funding for a startup.
Keasling reached out to his peers within the university. A colleague who had wangled an interview with senior officials at the Bill and Melinda Gates Foundation to discuss his work added three slides describing Keasling’s science at the tail end of his presentation. It was just enough. Word came back to Keasling: Make a pitch to the Gates Foundation.
The Gates Foundation had already galvanized the once-moribund field of malaria research. In 2001, the Gateses began supporting research on combating malaria—and were horrified to realize that they had effectively doubled the worldwide funding for such work. Funding would grow: By the end of 2004, the foundation would pump $345 million into malaria research.
These were no simple handouts. Gates himself was too much of a pragmatist. He wanted the foundation to act like a seed fund, catalyzing big change with carefully tended investments. The program managers hired by the foundation to oversee its programs were experienced public health officers, many of whom had spent years in Asia and Africa fighting malaria and other pandemic illnesses. They wanted to fight malaria on every front they could—with medicines for treating patients with malaria, with research on vaccines, and ultimately with programs that supported low-technology interventions such as bed nets soaked in insecticides. And they set “milestones,” or objectives that the programs they financed had to meet for funding to continue.
Another Berkeley colleague in public health introduced Keasling to an unusual San Francisco-based entrepreneur by the name of Victoria Hale, who was running a not-for-profit pharmaceutical company. Hale, who was awarded a MacArthur “Genius” grant in September, had worked for the Food and Drug Administration, and had seen pharmaceutical companies abandon work on drugs desperately needed to fight illnesses in developing countries because they were likely to be unprofitable. She hoped to take the most promising formulations for treating neglected diseases through the final stages of regulatory approval, then make them available at cost in developing nations.
Hale’s startup, the Institute for OneWorld Health, was in the midst of reinvigorating work on a drug for curing visceral Leishmaniasis, a deadly disease transmitted by the bite of a sand fly. She jumped at the chance to work on malaria. With prodding from the Gates Foundation, Keasling, the fledgling Amyris, and Hale fused their programs into an unusual collaboration, as unique as Keasling’s scientific approach of integrating the metabolisms of multiple organisms: Keasling’s university lab would continue to work on the scientific foundations for using microbes to make precursors to artemisinin. Amyris would take that research and develop industrial-scale processes for manufacturing the chemicals in bulk. And Hale’s OneWorld Health would ensure that those chemicals would wind up as real medicines—whether manufactured by Amyris or by another partner. The University of California agreed to give Amyris royalty-free rights to develop Keasling’s technology for making artemisinin, which would be a component in medicines sold at cost in more than 80 developing countries. But Amyris agreed to pay the university royalties if it used the technology.
Even so, the science was still new, the ideas demonstrated only in a university laboratory tended by researchers as anxious as first-time mothers. The scientists who anonymously reviewed Keasling’s proposal to the Gates Foundation peppered their comments with words such as “risky” and “ambitious.”
“They were right,” Keasling concedes. “But it wouldn’t be worth doing if it wasn’t scientifically risky.” He pauses. “Malaria is killing a child every 30 seconds. Someone has to take the risk.” In December 2004, the Gates Foundation agreed to back the Keasling-Amyris-OneWorld Health partnership with a five-year grant of $42.6 million. OneWorld Health would oversee the effort, helping Amyris weigh manufacturing decisions such as how much product it would need to make. What would be the optimal price for the artemisinic acid that Amyris would produce? How fast would it be able to drive down costs? How should the final drugs be packaged and distributed? How will they be protected against rogue companies that try to counterfeit the products? And what will be the political implications of manufacturing artemisinin-based drugs?
Amyris will have to tiptoe around some explosive issues over the price and supply of artemisinin. Current supplies are tight, pushing spot prices for artemisinin up to $600 or even $1,300 per kilogram. Such prices are persuading more farmers in China and even in Africa to plant sweet wormwood. Manufacturing artemisinin should drive prices down—ideally as low as $100 per kilogram—a blow to farmers even as it is a boon to malaria victims.
In the meantime, Keasling still had some science ahead.
Soon after winning the Gates grant, Keasling began plotting out what was needed to finish the journey to artemisinin. Step one: Add the key genes—Keasling believed there were three—from sweet wormwood to his microbes so that they would produce artemisinic acid. If microbes could produce enough artemisinic acid, the last step to the final conversion to artemisinin itself would be classic chemistry. Keasling once again figured he needed to graft a new specialist into his group. He recruited a postdoctoral researcher, a plant biochemist from the University of British Columbia, Dae-Kyun Ro, for the job.
Ro had his doubts about joining an engineering lab. He decided to take a different job—one at the University of California, Riverside, in a highly regarded plant biology program. Ro hopped a flight to Riverside to join a conference and accept the job, grabbing a copy of Time magazine from a newsstand along the way. An article stopped him cold: “Death by Mosquito.” Millions of people were dying from malaria, noted the article. Why wasn’t more effort going into finding a cure? When Ro got off the plane, he politely told the researchers at Riverside that he had decided to join Keasling’s laboratory.
When Ro joined in late 2004, Keasling suggested a classic chemist’s approach for sifting out which genes from sweet wormwood triggered the production of artemisinic acid. The genes proved more elusive than a four-leaf clover.
Ro abandoned Keasling’s approach and began thinking about the problem like a botanist instead of a chemist. What other plants might be like sweet wormwood? After three months and some hefty database analysis of sunflower and lettuce, relatives of sweet wormwood, Ro found a four-leaf clover: one short DNA sequence. Ro copied the gene that produced the sequence and plunked it into yeast.
One night in March a little after 8:00 PM, Ro put the first samples of his newly concocted material into a gas chromograph. Because he had only one of what Keasling estimated to be three key genes, Ro simply hoped to see traces of artemisinic alcohol—a few inches closer to the ultimate goal. Instead, he saw artemisinic acid itself. It was a bull’s-eye that would have impressed Robin Hood.
“It was shocking: It did everything we needed it to do,” Keasling says. “Ro’s approach was shorter and more innovative than the one I had suggested,” he adds. And better: Ro’s success meant Keasling could speed up the research program by about six months.
Keasling still gets up before sunrise every day. Although he teaches classes in the spring semester, meetings fence in his autumn-semester days. He also now leads a multi-university research program funded by the National Science Foundation called SynBERC (Synthetic Biology Engineering Research Center). And in his Berkeley laboratory on Potter Street, some 50 researchers and students work on projects that include bioremediation and biofuels. Still the biggest single group—more than 20—is the artemisinin team.
Keasling’s laboratory isn’t tackling the science solo this time, though. On Monday mornings, Keasling’s team meets with researchers from Amyris to discuss progress on the last significant scientific hurdle: picking the right microbial “engine” for brewing artesmisinic acid. Biologists classically use either yeast or E. coli to copy genetic sequences. Typically, E. coli produces a precursor in higher concentrations. But yeast is naturally more compatible with the kinds of pathways that Keasling’s process introduces. So which will it be—E. coli or yeast? “There’s a sort of horse race going on,” Keasling confides. He isn’t placing any bets. Interesting scientific results will come from the work on both platforms, he says. “If artemisinic acid production in either E. coli or yeast gets over a particular threshold, the choice will be clear,” adds Newman, who is now vice president of research at Amyris. “Either could be a fine choice.”
“They’ve made really tremendous scientific progress,” says Thomas Brewer, senior program officer at the Gates Foundation who oversees the artemisinin grant. “They’re a year ahead of the milestones we optimistically hoped they’d meet.” Even so, Brewer is realistic, saying, “There’s still a long way to go to get the project to meaningful volume”—meaning a few tons a year of artemisinic acid.
Figuring out how to turn what Newman calls the “dreamy science” that happens in laboratory flasks into production in 100,000-liter vats falls squarely on the back of Amyris.
Along with Newman, Renninger threw in his lot with Amyris and now works as the vice president of development for the three-dozen-person firm, housed on one floor of a modest brick building just across from the Emeryville train station, a few miles from Keasling’s lab. They persuaded another postdoc from Keasling’s lab, Kinkead Reiling, to join Amyris as president and co-founder. Vincent Martin returned to Canada and took up an assistant professor post at Concordia University in Quebec. (Ro also has moved on, joining the University of Calgary as an assistant professor.)
Friday mornings at 7:30, Keasling drives over to Emeryville to talk through big strategic questions with the Amyris team. Early on, the founders agreed that Amyris would aim to be a specialty chemical company, but not a pharmaceutical maker. Still, they must plot out how quickly Amyris will be able to make product—and what else the company should do.
“We’ve got to get the artemisinic acid out as soon as possible,” Keasling insists. “There are rogue pharmaceutical companies making artemisinin-based drugs.” Some sell fake drugs; others make pure artemisinin-based products—a development that health experts warn could lead to a new generation of parasites that develop immunity to artemisinin. “We’ve got to beat them,” Keasling says. “We’ve got to put those rogue manufacturers out of business.” But it’s a race in slow motion: If everything goes right, the earliest Amyris could hope that products based on its chemicals could be available would be by late 2009 or early 2010.
And Amyris is broadening its horizons, too, weighing the commercial products it could make as the artemisinin project matures. “All the founders share one belief,” Newman says. “We don’t want to make trivial products”—not even if the potential market looks lush. “Green chemistry” interests the group—that is, finding non-polluting techniques to make widespread products. Biofuels also keenly interest Keasling and the Amyris executives, who have signed a deal with Kleiner Perkins Caulfield & Byers, a well-known Silicon Valley venture capital outfit.
Yet with all the products that his chemical tools may yield, Keasling remains haunted by the images of the children in Malawi, and so fixated on artemisinin. He pauses and says, “This may be the most important thing I ever do in my life.”