CAMBRIDGE, Mass. Leaders of a new movement are kicking off the first Synthetic Biology 1.0 conference at the Massachussets Institute of Technology this week. "Synthetic biology" is the blanket term for a multidisciplinary attempt to identify a class of standard operational components that can be assembled into functioning molecular machines.
Central to that effort is the ability to isolate discrete biomolecular mechanisms and define standard interfaces for them so that they can be assembled in much the same way as electronic circuits. This confluence of computer science and biology is so remarkable that this new movement rises to the level of moon shot initiatives: to reverse-engineer life itself.
Rummaging through an upright freezer in the Stata Center, the new home for MIT's Computer Science and Artificial Intelligence Laboratory, Tom Knight extracts a plastic box so cold that wisps of water vapor encircle it, giving a B-movie spin to the scene. Pulling the lid off, he presents a visitor with a matrix of small vials sealed with yellow stoppers. The vials contain carefully engineered segments of DNA that code for metabolic functions in the bacterium Mesoplasma florum.
"Just pour the contents of one of these vials into a standard reagent solution, and the DNA will transform itself into a functional component of the bacteria," Knight explained. Each vial codes for a standard biochemical part in a catalog of so-called BioBricks that Knight, a senior research scientist, and his collaborators in MIT's synthetic-biology initiative are compiling.
Programs under way at universities and other labs around the world now address various aspects of this novel idea of applying engineering principles to biological systems. There are practical reasons to pursue engineered biology.
Pharmaceutical companies are finding that naturally occurring metabolic pathways in bacteria that produce useful drugs are not as efficient as re-engineered pathways.
Nanotechnologists are finding that naturally occurring biological functions can be redirected to tasks such as building molecular circuits.
And for MIT's Knight, engineered organisms represent the quickest route for a true nanotechnology that could manufacture materials and systems on a molecular scale. His dictum: "Biology is the nanotechnology that works."
Knight views the emerging field through the lens of electrical engineering and circuit design. "Why are engineers good at doing this, why are we the right people?" he asks. "I would argue that we have a set of tools and an intellectual approach suited to this task."
No other engineering discipline has developed the capability of designing, building and debugging systems consisting of billions of parts. That capability rests in a system of functional abstraction so deft that engineers can design complex circuits or computers on a workstation and then hand off the design to another group of experts-process engineers-for transfer to silicon.
As the catalog of standard biochemical parts expands, Knight expects to see the same development take place in synthetic biology. Each part will need a standard interface and well-defined function, so that someone at a workstation could assemble working systems at the abstract level described in the catalog and pass the design to then another group of biochemists for synthesis in a biological manufacturing system.
Pioneers for profit
Even though the field is in its infancy, entrepreneurial opportunities are appearing. Genetics engineer John Mulligan, for example, has founded Blue Heron Biotechnology (Bothell, Wash.) to offer synthesized DNA strands to pharmaceutical companies and biological researchers. At a secure Web site, clients submit a sequence of DNA amino acid units (adenine, cytosine, thymine, guanine); from it, Blue Heron will synthesize a strand of the molecule which is guaranteed to be accurate to at least one part in a million.
Blue Heron also has a unique packaging system for delivering its product. The DNA strands are built in a circular configuration called a plasmid, which is then injected into a standard bacterium. The circular strand does not interact with the bacterium's cellular processes, but when it reproduces itself, it also builds an exact copy of the plasmid.
That allows Blue Heron's clients to produce unlimited quantities of the DNA by culturing the bacterium.
DNA synthesis is perhaps the simplest type of synthetic biology; researchers routinely copy and splice natural DNA segments in the lab using the organisms they are working with. However, after helping to set up a gene-sequencing lab at Stanford University and participating in the human genome project, Blue Heron's Mulligan began to see an opportunity.
"We were able to increase the throughput of processing DNA by several orders of magnitude and that creates new opportunities," he said. Blue Heron's business model is simply that the cost to a lab for synthesizing DNA itself will be greater, in terms of the raw materials and labor, than having it done at Blue Heron.
DNA synthesis begins with oligonucleotides, which biotech companies now supply at the rate of about $400 million per year. Mulligan estimates that worldwide, biotech researchers and pharmaceutical companies spend about $1 billion synthesizing DNA, so there is a substantial market for this simple first step in synthetic biology.
More to the point, Mulligan's DNA-manufacturing line resembles Knight's concept of a structured design system. Workstations connected to an Oracle database of DNA data set up a program of biochemical processes required to produce a given strand. That program is then fed to the processing line, which heavily depends on robots with some human intervention at critical steps.
"To maintain our cost rationale for our clients we have to automate as much as possible, but unlike conventional manufacturing, every part that comes off of our line is unique," Mulligan said.
Mulligan is finding that his company's software-level biomanufacturing system is turning into a kind of design automation tool for bioresearchers. Researchers are more and more often using supercomputers to simulate biological experiments, a trend that goes by the term "in silico" experimentation.
For example, it is possible to design a protein with a specific configuration on a computer and then access Blue Heron's software system to construct the DNA sequence that would produce it inside a cell. The protein and DNA may not actually exist in nature, but the company could supply the protein researcher with a strand of synthesized DNA that would produce it through the standard RNA synthesis processes common to all cells.
Several decades away
Will the world soon see biomanufacturing lines turning out artificial beings? "It will be several decades, maybe a century, before anyone synthesizes an organism," Mulligan said. Although projects such as sequencing the human genome have spawned a new generation of supercomputers and microarray technologies for tackling biosynthesis, even greater leaps in technology will be needed to decode a full, functioning organism.
Or at least, that is the view of the typical biologist, says MIT's Knight. "Here is the difference between a biologist and an engineer," he said: "A biologist goes into the lab, studies a system and finds that it is far more complex than anyone suspected. He's delighted, he can spend a lot of time exploring that complexity and writing papers about it. An engineer goes into the lab and makes the same finding. His response is: 'How can I get rid of this?'"
Engineers excel at eliminating irrelevant complexity in order to build something that works and is fully understood. And that is Knight's strategy for reverse-engineering Mesoplasma florum, an organism that, even by bacterial standards, is very simple.
It is being attacked from two directions. In the bottom-up effort, metabolic functions are being identified and abstracted, and listed in the BioBrick catalog. This listing requires that each entry have standard chemical inputs and outputs, the DNA sequence that synthesizes it and a functional description of its biochemical operation.
The other direction of attack conforms to the "how can I get rid of this" approach to complexity. Segments of DNA are selectively removed from Mesoplasma specimens until the organism fails, a technique that is helping the bioengineers to remove complexity that is irrelevant.
They hope the two approaches will meet somewhere in the middle, and that a functional organism, one that is fully understood and can be synthesized from standard parts, will emerge.
Knight also expects practical fallout along the way. One example is a project by one of his graduate students to use protein arrays to achieve highly uniform doping in semiconductors. At current design dimensions, there are so few atoms in doped regions that statistical variations become disruptive.
Since proteins that interlock in regular arrays can be designed with BioBrick circuits, it is possible to attach a dopant atom to each one and have them self-assemble on the surface of silicon to eliminate the random variation in conventional doping processes.
And there will no doubt be surprises along the way as bioengineering takes hold. "We are used to systems that, once built, stay put. You turn them on and they always perform the same way," Knight said. "That is not the way biological systems operate. They reproduce something that is not now part of engineering and they mutate.
That may be viewed as a bad thing and we will have to find some way to eliminate it, or it may turn out to be a good thing." In fact, mutation, combined with selection of the fittest, has been nature's sole bioengineering tool.
Thus, tools that arose out of the VLSI revolution for managing complexity through abstraction and modularization are now being turned toward engineered systems on the nanoscale.
Fortunately, nature has delivered working examples that will provide critical information about how to do that kind of engineering. And perhaps later in the 21st century, EEs will be jockeying bio-EDA systems as they design nanomanufacturing systems inspired by living cells.