Biology's Master Programmers
Photographs by Mark Ostow
George Church is an imposing figure—over six feet tall, with a large, rectangular face bordered by a brown and silver nest of beard and topped by a thick mop of hair. Since the mid-1980s Church has played a pioneering role in the development of DNA sequencing, helping—among his other achievements—to organize the Human Genome Project. To reach his office at Harvard Medical School, one enters a laboratory humming with many of the more than 50 graduate students and postdoctoral fellows over whom Church rules as director of the school’s Center for Computational Genetics. Passing through an anteroom of assistants, I find Church at his desk, his back to me, hunched over a notebook computer that makes him look even larger than he is.
Church looms especially large these days because of his role as one of the most influential figures in synthetic biology, an ambitious and radical approach to genetic engineering that attempts to create novel biological entities—everything from enzymes to cells and microbes—by combining the expertise of biology and engineering. He and his lab are credited with many of the advances in harnessing and synthesizing DNA that now help other researchers modify microörganisms to create new fuels and medical treatments. When I ask Church to describe what tangible impact synthetic biology will have on everyday life, he leans back in his chair, clasps his hands behind his head, and says, “It will change everything. People are going to live healthier a lot longer because of synthetic biology. You can count on it.”
Such grandiosity is not uncommon among the practitioners of synthetic biology. Ever since Church and a few other researchers began to combine biology and engineering a dozen years ago, they have promised it would “change everything.” And no wonder. The very idea of synthetic biology is to purposefully engineer the DNA of living things so that they can accomplish tasks they don’t carry out in nature. Although genetic engineering has been going on since the 1970s, a rapid drop in the cost of decoding and synthesizing DNA, combined with a vast increase in computer power and an influx into biology labs of engineers and computer scientists, has led to a fundamental change in how thoroughly and swiftly an organism’s genetics can be modified. Church says the technology will eventually lead to all manner of breakthroughs: we will be able to replace diseased tissues and organs by reprogramming cells to make new ones, create novel microbes that efficiently secrete fuels and other chemicals, and fashion DNA switches that turn on the right genes inside a patient’s cells to prevent arteries from getting clogged.
Even though some of these applications are years from reality, Church has a way of tossing off such predictions matter-of-factly. And it’s easy to see why he’s optimistic. The cost of both decoding DNA and synthesizing new DNA strands, he has calculated, is falling about five times as fast as computing power is increasing under Moore’s Law, which has accurately predicted that chip performance will double roughly every two years. Those involved in synthetic biology, who often favor computer analogies, might say it’s becoming exponentially easier to read from, and write into, the source code of life. These underlying technology trends, says Church, are leading to an explosion in experimentation of a sort that would have been inconceivable only a few years ago.
Up to now, it’s proved stubbornly difficult to turn synthetic biology into a practical technology that can create products like cheap biofuels. Scientists have found that the “code of life” is far more complex and difficult to crack than anyone might have imagined a decade ago. What’s more, while rewriting the code is easier than ever, getting it right isn’t. Researchers and entrepreneurs have found ways to coax bacteria or yeast to make many useful compounds, but it has been difficult to optimize such processes so that the microbes produce significant quantities efficiently enough to compete with existing commercial products.
Church is characteristically undeterred. At 57, he has survived cancer and a heart attack, and he suffers from both dyslexia and narcolepsy; before I visited him, one of his colleagues warned that I shouldn’t be surprised if he fell asleep on me. But he has founded or taken an advisory role in more than 50 startup companies—and he stayed awake throughout our time together, apparently excited to describe how his lab has found ways to take advantage of ultrafast sequencing and other tools to greatly speed up synthetic biology. Among its many projects, Church’s lab has invented a technique for rapidly synthesizing multiple novel strings of DNA and introducing them simultaneously into a bacterial genome. In one experiment, researchers created four billion variants of E. coli in a single day. After three days, they found variants of the bacteria in which production of a desired chemical was increased fivefold.
The idea, Church explains, is to sort through the variations to find “an occasional hopeful monster, just as evolution has done for millions of years.” By mimicking in lab experiments what takes eons in nature, he says, he is radically improving the odds of finding ways to make microbes not just do new things but do them efficiently.
A DNA Turn-On
In some ways, the difficulties researchers have faced making new, more useful life forms shouldn’t come as a surprise. Indeed, a lesson of genome research over the last few decades is that no matter how elegantly compact the DNA code is, the biology it gives rise to is consistently more complex than anyone anticipated. When I began reporting the early days of gene discovery 30 years ago, biologists, as they often do, thought reductively. When they found a gene involved in disease, the discovery made headlines. Scientists said they believed that potent new medicines could soon deactivate malfunctioning versions of genes, or that gene therapy could be used to replace them with healthy versions in the body.
The early biotech companies also employed a one-gene-at-a-time approach. Companies would locate the gene that made a particular protein, such as insulin; then, using gene-splicing technology first developed in the 1970s, they would snip open the DNA of a bacterium and slide in the insulin-making gene. It’s a technology that has led to today’s biotech industry.
Yet some scoffed at the idea that such techniques involved any real engineering. “To us, it was no more engineering than changing a red light bulb with a green light bulb,” says James Collins, a Boston University bioengineer who is credited with helping create the field of synthetic biology in 2000. “Many of us thought that working at the single-gene level was just a starting point, that we really needed to figure out how all these newly identified genes arising from the Human Genome Project fit into networks, pathways, and circuits inside the cell.” By comparison, says Collins, “synthetic biology is genetic engineering on steroids.”
I met Collins on a rainy winter day in his office on the BU campus. He’s an enthusiastic storyteller, full of details, digressions, and gossip. And at 46, he has lived through the conception and birth of synthetic biology. Collins told me how he and other engineers in the late 1990s felt left out of what appeared to be the most important science of the time, the sequencing of the human genome. It seemed that every other cover of the journal Science was hailing some new gene breakthrough. But with a slew of unanalyzed DNA data piling up in computer databases, it was becoming clear that biologists didn’t know how genetic parts worked together. Collins says, “We had felt like we were kids outside a candy store. So we figured, ‘How can we get in?’”
Collins wanted to study cellular processes by constructing gene networks rather than taking them apart. As a first step, he built a biological toggle circuit. A toggle is a mechanism with two possible states—in the case of a light switch, on or off. In the switch he and his colleagues built from DNA, two genes next to each other in a bacterial genome both produced proteins when they were “on.” But Collins set things up so that each protein would block production of the other—if gene 1 was on, it would keep gene 2 off, and vice versa. With the aid of chemicals or a thermal pulse, Collins could flip between the two states.
The DNA toggle switch was analogous to an electronic transistor, able to store a single bit of information. It was also an engineered example of the kind of feedback loop that often determines whether cells grow, divide, or die. “The idea that you could build a circuit out of biological parts helped launch the field of synthetic biology,” says Collins. The results were published in January 2000.
Soon Collins’s toggle was joined by an expanding list of DNA circuits, including biosensors, oscillators, bacterial calculators, and similar molecular gadgetry. Researchers even established a Registry of Standard Biological Parts: 7,100 different DNA structures are available to order. Scientists were excited by the idea that biology might be modular and predictable, like something made with Lego blocks or computer code. Many scrambled to found companies that they hoped would commercialize the technology to produce fuels, drugs, or other products.
While comparisons to computer programming inspired many researchers, however, these tended to oversimplify biology, which has not proved entirely predictable. Furthermore, the claims that some synthetic-biology companies made now appear to have been overly optimistic, Collins says.
Indeed, Collins believes the rush to commercialize has been a mistake. “The companies are sucking the oxygen out of the field,” he says, noting that they have hired scores of geneticists from university labs. They’re “scooping up our seed corn, the young researchers who should be staying in academic labs working through new ways to engineer biology.” He worries that the race to apply the new technology means “there’s going to be a number of biotech carcasses on the side of the road in the near future.”
Not even George Church has been immune: Codon Devices, a company he cofounded in 2004, was forced to shut down. Codon, in Church’s words, was established to be the Intel of the bioengineering industry, building ready-made synthetic-biological modules that researchers could use to redesign, say, a yeast. The company “burned through its cash,” he laments.
Church hopes his latest enterprise avoids a similar fate. Called Warp Drive Bio, the company combines computer science, chemistry, and genetic engineering in ways that would not have been possible until recently. It aims to use ultrafast DNA sequencing and synthetic-biology techniques, some of which Church pioneered, to hunt for potential medicines by scouring the DNA of millions of environmental samples that drug companies have collected and stored over several decades. Warp Drive is, in effect, searching for genetic parts that nature has already programmed to make particularly potent, useful chemicals. Church’s technology will be used to generate copies of those parts, incorporate them into bacteria, and optimize their performance. Then the bacteria can be used to produce chemicals that, if all goes according to plan, have new and interesting therapeutic properties.
Warp Drive, which was launched in January, employs fewer than a dozen full-time staffers and occupies only about 1,000 square feet of office and lab space in Cambridge, Massachusetts. But the startup, which has raised $125 million in investments, has already formed a strategic partnership with the French pharmaceutical company Sanofi. If Warp Drive Bio meets certain milestones, it has the option to demand that Sanofi purchase the company for $1 billion or more. The deal was struck after Warp Drive’s principal founder, Harvard biochemist Gregory Verdine, was invited to Paris last May and gave a two-hour presentation that had Sanofi’s head of research, Elias Zerhouni, and several other staffers crowding around his laptop.
Zerhouni, a former head of the U.S. National Institutes of Health, immediately grasped the novelty and potential of Warp Drive’s idea for sifting through nature’s existing stockpile of DNA parts. “We’ve been plagued by a lack of creativity,” he says. “It made sense to give them the resources they need.”
Verdine’s insight is that nature is particularly adept at creating chemicals that act safely and precisely on a desired biological target. He says that half the small-molecule drugs developed over the last 30 to 35 years have been natural products or derivatives of such products. “It struck me that probably something useful in evolution helped tailor properties in these compounds that made them better suited to work in complex cell systems like the human body,” he says. “Nature seemed to have already engineered in complexities that drug chemists don’t understand.”
I interviewed Verdine in a spare room, no bigger than a walk-in closet, just outside Warp Drive’s lab in a converted book factory in Cambridge. If Church and Collins are intent on creating new synthetic parts and bioengineering techniques, Verdine is hoping to use many of the same techniques to unwrap the mysteries of how nature does it. Over the decades, he explained to me, pharmaceutical researchers have collected and stored tens of thousands, and more likely millions, of environmental samples, including dirt and pond scum. The idea was to discover some potent chemical in these mixtures by dripping extracts onto cancer cells or into petri dishes of bacteria. But that process is laborious and subject to chance. Most drug companies have scaled back such research.
The answer, Verdine decided, was to search for DNA instead. Given the plummeting cost of DNA sequencing, it’s now feasible to simply decode all the genetic material present in, say, a drop of pond water teeming with microörganisms. Verdine says many of the natural drugs that have already been identified have similar DNA signatures—clusters of genes that often occur together in a microbe’s genome. The trick, he adds, is to scan the samples’ DNA to locate familiar-looking clusters that might be recipes for synthesizing a natural product—ideally, an important one that hasn’t been found before.
Once identified, the DNA sequences will need to be engineered into a bacterium so that the company can produce the chemical and study the potential drugs. This is where the synthetic-biology techniques developed by Church will be crucial: in transforming the code into actual compounds. “We use genomics and informatics to find a gene cluster. But that’s an information unit,” Verdine says. “We have to get the molecule. Synthetic biology involves coaxing the cluster into biosynthetic factories, which then produce the molecules. If we don’t have the molecule, the cluster is useless.”
The idea of resorting to nature’s stockpile for parts, Church says, is “ironic and interesting” given synthetic biology’s interest in producing entirely new DNA circuits and, ultimately, generating entire organisms from scratch. Researchers today may alter, copy, and paste DNA with increasing ease, but they still struggle when it comes to actually composing DNA that does anything useful. They are still editing nature’s code and learning from it. It turns out that for now, nature is still the best programmer.
Michael Waldholz is a former managing editor at Bloomberg News, where he oversaw the publication’s coverage of health and science. Previously, he was a reporter and news editor at the Wall Street Journal. His books include Curing Cancer and Genome, which was published in 1990.
Become an MIT Technology Review Insider for in-depth analysis and unparalleled perspective.Subscribe today