One day in 1998, Randy Rettberg ‘70 went to visit an old friend, Tom Knight ‘69, SM ‘79, PhD ‘83, a research scientist in the Artificial Intelligence Lab at MIT. Rettberg, who had worked in the computer industry for 30 years, was surprised to see that Knight, a self-described “nerdy semiconductor designer” who had participated in the development of ARPAnet, had set up a biology lab in the middle of the AI Lab, then located in Tech Square.
“His electronic equipment was gone, the logic analyzers were gone, and he was showing me bottles and telling me how cool this bottle was—it had this nice top on it that wouldn’t drip,” Rettberg recalls. “And he had this nice incubator so he could grow things, and an autoclave.”
Knight had talked to Rettberg about the idea of applying engineering principles to biology: taking cells apart, figuring out how they work, and reassembling the parts (in this case, genes) to do something new. In the 1990s, Knight had decided to take the plunge. He spent five years taking nearly every MIT course in molecular biology, absorbing as much information as he could about a field in which he had almost no background.
Rettberg, who was looking for a career change, decided to leave his job as chief technical officer of a Sun Microsystems division and join Knight. He, too, was starting from scratch when it came to biology. “I got all of the biology books and chemistry books that I could,” he says. “I went to the Coop and got a stack about a foot tall. That’s about $800 worth of books. I read and read and read until I couldn’t go anymore because I didn’t know how to pronounce the words. I realized I better learn the right way to pronounce things or I was going to look really stupid.” So he came to MIT as an unpaid research affiliate in Knight’s lab and started taking biology courses.
Why would electrical engineers spend all that time learning about the inner workings of cells? Knight and Rettberg, who is now a principal research engineer in MIT’s Department of Biological Engineering, wanted to see whether biology is sufficiently modular—and sufficiently well understood—to let researchers design, build, and test biological systems. Might they one day be able to treat cells as living circuit boards, letting genes stand in for electrical components like resistors and capacitors? They wondered if they could ultimately redesign living cells by assembling biological “circuits” from a set of standardized “parts” (genes), just as an engineer can build circuits to control electronic devices by combining the right components. If so, they could treat biology as a manufacturing technology, programming cells to produce things they wouldn’t normally make—for example, drugs, fuels, or plastics. “Biology just happens to be in the business of making more copies,” Knight says. “But we can subvert that. We can use it to make just about anything.”
This new approach, known as synthetic biology, initially aroused skepticism among biologists, recalls Ron Weiss, SM ‘94, PhD ‘01, who was a grad student of Knight’s in the late 1990s. In those early days, “it was rare to find a biologist that would understand or care about what we were doing,” he says. Synthetic biology goes further than genetic engineering, which usually involves adding a single gene to a cell so that it will do something it wouldn’t normally do. It is also different from metabolic engineering, which uses the techniques of genetic engineering to maximize cells’ production of commercially useful products, such as insulin. Assembling a given set of genes in novel ways enables synthetic biologists to accomplish highly specific and sophisticated tasks that they wouldn’t be able to achieve by modifying cells one gene at a time, a process that doesn’t always make it possible to control their function.
Now an associate professor of biological engineering, Weiss joined the MIT faculty in 2009 to launch a new synthetic-biology research initiative at MIT—the Center for Integrative Synthetic Biology. The center is slated to open this fall in Technology Square and will include Rettberg and about a dozen faculty members from departments all over MIT, including biological engineering, biology, chemical engineering, and electrical engineering and computer science. (Knight, now on leave from MIT and working at Gingko Bioworks, a synthetic-biology company he cofounded, is expected to join when he returns to the Institute as a senior research scientist in electrical engineering and computer science.)
One of a handful of synthetic-biology programs in the world, the new center aims to make synthetic biology as convenient as possible by integrating it with systems biology—a computational approach to figuring out the complex biological interactions that determine a system’s behavior (for example, a cell’s response to a particular hormone). By unraveling these systems and figuring out ways to reëngineer them, the researchers hope to advance research in biofuels and synthesis of biological molecules—and to develop new ways to treat cancer, diabetes, and other diseases.
Knight worked with Marvin Minsky in MIT’s AI Lab as a high-school student, spent most of the 1970s as an MIT research staffer working on major hardware projects such as the Lisp machine (the first commercial single-user workstation), and then got his PhD in 1983, specializing in integrated-circuit design. After spending a lot of time thinking about the limitations of Moore’s Law—the idea that computer processing speeds should double approximately every two years—he found an unlikely source for an idea about where chip makers might turn to overcome them. In the late 1980s, he read a paper by Yale physicist Harold Morowitz, who proposed studying a type of bacteria known as mycoplasmas, identifying the function of each of their genes and proteins, and pinning down how they interact. Mycoplasmas are among the simplest bacteria, with only about 500 to 700 genes. Knight realized that biology wasn’t as hopelessly complex as he’d imagined; there were living systems so simple that one could plausibly tease out how their parts work—and work together. He started wondering whether he could use those bacteria as tiny factories, stripping out any genes that seemed unnecessary and adding genes for desired traits—traits that could help them produce drugs, biofuels, or computer chips. (In fact, researchers at the J. Craig Venter Institute used a modified version of a mycoplasma genome to create the first synthetic cell and to assemble an entire synthetic genome. See “TR10: Synthetic Cells,” p. 56.)
By the time Knight began working with bacteria, in the 1990s, it had become clear that most mycoplasmas are too pathogenic for his lab’s safety rating. So he settled on another simple bacterium, Mesoplasma florum. He got researchers at the Broad Institute to sequence it for him and has since figured out that it can still function even after many of its genes have been deleted. He’s now working on “refactoring the genome,” which he describes as “tearing it apart to pieces that we understand, taking out the pieces we don’t understand, and recoding for simplicity the pieces that are essential.”
Ever the engineer, Knight is striving to make his system as simple as possible. “There’s this cultural difference between the engineering community and the scientific community, which is the reaction to complexity,” he says, telling an old joke to illustrate his point: “The biologist goes into the lab in the morning, does an experiment, discovers that a system she’s looking at is twice as complicated as she thought it was, and says, ‘Great! I get to write a paper!’ The engineer goes into the lab, does the same experiment, gets the same result, and says, ‘Damn, how do I get rid of that?’ ”
Getting rid of complexity will help would-be cell designers realize another advantage of electrical engineering: the ability to design, test, and build as quickly as possible. “The efficiency of engineering often is determined by how rapidly one can go around that [design-test-build] loop,” says Knight. “If you’re a software engineer, that loop is very, very quick. It can be two minutes … If you’re a biologist, that loop, at the moment, is a week to a month. You screw around trying to figure out how to put these pieces of DNA together, and when you’re done, maybe you have a good way of testing it, and maybe you don’t.” He adds that the scarcity of good tools for determining what’s going on in cells limits the efficiency of the design process as well.
The sometimes glacial pace of traditional molecular biology discouraged Knight when he first started trying to design modified organisms. “I realized that every time I wanted to do an experiment, it turned into two experiments,” he says. “There was the experiment I wanted to do, and there was another experiment associated with building the piece of DNA that I needed. The frustrating thing from an engineering standpoint was that every time someone did that, they would do it in a different way.” One reason had to do with the enzymes used to snip DNA at specific points to extract a desired gene: researchers “would be driven by the accidents of what restriction enzyme sites were present in pieces of natural DNA,” he says. “They would be driven by what enzymes they happened to have in the freezer, or which ones their mentor had shown them how to use five years ago.”
That frustration led Knight to develop the concept of BioBrick parts—standardized pieces of DNA that can be joined in different combinations and introduced into a host bacterium so that it will perform a specific task. The collection of these genes, known as the Registry of Standard Biological Parts, is modeled on a 1,000-page catalogue called the TTL Data Book, which lists hundreds of circuit components. Electrical engineers who want to build TTL (transistor-transistor logic) circuits can refer to the book and pick out the elements they need to achieve a particular function. Knight and Rettberg hoped the same principle could be applied to biological design. As Rettberg put it, “Can simple biological systems be built from standard interchangeable parts and operated in living cells? Or is biology simply so complicated that every case is unique?” Now, he says, “we know that you can, sometimes; and no, biology is not always too complicated. Sometimes it is—sometimes you get tricked by something you didn’t think of—but the same thing happens in writing a [computer] program.”
As Knight laid out the concept in a 2003 paper, each BioBrick is a piece of DNA that includes a gene associated with a specific trait. To make one, you enter the sequence of the desired gene into a DNA synthesis machine, which strings together nucleotides in the correct order. The BioBrick is capped at both ends with DNA sequences that allow it to be connected to other parts. Then it’s integrated into a circular piece of DNA called a plasmid, which can be inserted into a bacterial cell. The BioBricks are designed so they can easily be combined into larger circuits, or series of genes that induce a bacterial cell to perform complex functions such as glowing when exposed to a certain chemical. Knight’s company, Ginkgo Bioworks, is now designing new BioBricks, automating assembly of DNA, and pursuing applications such as fuel production.
The best evidence that this approach works, Rettberg says, is that undergraduate teams can successfully use it to design a wide array of projects in a matter of months for the International Genetically Engineered Machine competition, or iGEM. The competition, now held annually at MIT, grew out of a January 2003 IAP course dreamed up by Rettberg, Knight, Professor Gerald Sussman ‘68, PhD ‘73, and Drew Endy, a former MIT assistant professor now at Stanford.
During that first IAP, the students came up with intriguing designs but didn’t finish building their “machines” because it took too long to synthesize the DNA. Still, the course was repeated the following year, and the first official competition was held in summer 2004, drawing five teams. The event has expanded steadily ever since: last November, 130 teams presented projects. Undergraduate teams have built an arsenic detector, bacteria that can detect and clean up environmental pollutants such as toluene, and a vaccine against Helicobacter pylori, a bacterium that can cause ulcers. None of these projects have advanced to the point of commercial viability, but a company called Lumin Sensors plans to test the arsenic detector, built by a team from the University of Edinburgh, for use in India.
More lighthearted projects have included bacteria that glow in the dark or smell like mint. In just months, the teams can build systems that might have taken years using traditional genetic-engineering techniques. “Nobody’s been able to do any of this stuff before,” says Rettberg. “The kids all know they are making something that is new and that their teachers and their parents and their older brothers had no idea that anybody could do.”
In just seven years, iGEM students have contributed thousands of parts to the Registry of Standard Biological Parts, which now has more than 7,000 entries. That registry is one of several synthetic-biology standards in existence, but Knight says it’s more important to follow standards than to try to make everyone use the same one. “By spending a small amount of time up front to standardize the pieces of DNA,” he says, “you put yourself into a position where the assembly of pieces of DNA is completely straightforward, thought-free, automation-friendly, and does not become an experiment in itself.”
Ron Weiss, who studied computer science at MIT, was drawn to synthetic biology by the prospect of developing new medical treatments. As a graduate student, he created some of the first biological circuits (whose parts would later be entered in the registry), including some that allow cells to communicate with their neighbors. He also developed circuits to demonstrate Knight’s concept of a “biological inverter”—comparable to an electronic inverter, which takes an input signal and produces the opposite output. In cells, an inverter can be created using a repressor protein, which binds to DNA and blocks transcription of a specific gene.
After finishing his PhD, Weiss joined the faculty at Princeton, where he began working on mammalian cells. That’s when his research started catching the eye of biologists. “When I started to give talks about results we were getting in the mammalian work, people had such an easy time connecting to the work and understanding why it is that we want to be able to do these things,” he recalls. “When I say I can program stem cells to differentiate into particular cell types using these elaborate, sophisticated rules and programs, they could actually see why that might be relevant.”
Weiss’s lab at MIT is now working on genetically programming stem cells to turn into pancreatic beta cells, the insulin-producing cells that type 1 diabetics lack. “We have a complex genetic program that steps these cells through a variety of phases to mimic what happens in embryogenesis,” Weiss says. “We’ve done it in mouse embryonic stem cells, it looks like it’s working, and now I have a postdoc who’s working on a human-embryonic-stem-cell version of that.”
His lab is also designing cells that would detect and kill pathogens, forming something like an artificial immune system. And it is working on genetic circuits, perhaps to be incorporated into cells by harmless viruses, that could detect and kill tumor cells. These are very long-term goals, however. “I think it’s going to take longer than people expect,” Weiss says. “These are all very complicated systems. But I do think it’s going to have a significant impact on our ability to address medical problems. This notion that we can potentially engineer cells in our bodies so that we can treat disease or medical conditions in a programmable fashion to me is really exciting.”
Synthetic biologist Christopher Voigt, whom MIT recently recruited from the University of California at San Francisco, will codirect the new center. Voigt’s research involves programming E. coli cells to act like sensors that respond to touch, light, and odors; he has already created versions that respond to light by changing color, allowing him to generate images in a petri dish of bacteria.
Another member, assistant professor Timothy Lu ‘03, MNG ‘03, PhD ‘08, is pursuing both industrial and medical applications for synthetic biology. As a graduate student in the Harvard-MIT Division of Health Sciences and Technology, Lu worked with Boston University professor James Collins to engineer a bacteriophage that can attack films of bacteria that accumulate on surfaces. The bacteriophage (a type of virus that targets bacteria) kills 99.997 percent of the cells in these biofilms, which are very difficult to eradicate using traditional antibiotics.
Novophage, the company Lu cofounded with Collins and others to commercialize the technology is developing industrial applications for their biofilm-fighting virus. They are also collaborating with the U.S. Army to engineer viruses that could kill antibiotic-resistant bacteria in soldiers returning from such places as Iraq and Afghanistan. “These guys are coming home with blast wounds that are contaminated with bugs that are very resistant to antibiotics,” says Lu. Of particular concern is a bacterium called Acinetobacter baumannii, which can cause pneumonia and infections of the bloodstream and urinary tract.
Lu says undergraduate interest in synthetic biology is growing rapidly, and he hopes the new center will help draw even more students into the field. “It’s an interesting discipline because students don’t generally come to MIT exposed to it, and then they kind of hear about it as they go through their classes,” he says. “That’s why the center is a really great idea, to try to increase the visibility of the discipline on campus. We’re hopeful this is going to grow into something pretty cool.”
DeepMind’s cofounder: Generative AI is just a phase. What’s next is interactive AI.
“This is a profound moment in the history of technology,” says Mustafa Suleyman.
What to know about this autumn’s covid vaccines
New variants will pose a challenge, but early signs suggest the shots will still boost antibody responses.
Human-plus-AI solutions mitigate security threats
With the right human oversight, emerging technologies like artificial intelligence can help keep business and customer data secure
Next slide, please: A brief history of the corporate presentation
From million-dollar slide shows to Steve Jobs’s introduction of the iPhone, a bit of show business never hurt plain old business.
Get the latest updates from
MIT Technology Review
Discover special offers, top stories, upcoming events, and more.