I, bacterium: how bugs can become robots
(Filed: 04/05/2005)

Engineers are trying to program cells like tiny computers so they can be used to repair tissues or detect toxins. Roger Highfield reports

These abstract patterns are the result of efforts to improve on four billion years of evolution and turn bacteria into living robots.

Patterns: bacteria

Engineers have programmed the innocuous bacteria that inhabit the human gut to communicate with each other and produce various motifs. The colourful feat by the bacterium Escherichia coli is reported by a team from Princeton University in New Jersey in the latest issue of the journal Nature and represents a milestone in an emerging field known as "synthetic biology".

Proponents aim to make living cells function as if they were tiny computers so they can be harnessed as workhorses that detect toxins by glowing, help to build things, or repair tissues and organs within the body.

Scientists have been tinkering with genes for three decades and the new field is an extension of conventional genetic engineering. However, traditional GM offers nothing like the precision of a mature engineering discipline. If that can be achieved by synthetic biology, it will offer a sophisticated new way to explore old questions about how life got started on Earth, how a fertilised egg turns into a community of billions of cells in a body, and even what life might be like elsewhere in the cosmos.

The more enthusiastic supporters of synthetic biology hope to build living things that can behave predictably, design "plug and play" genetic circuitry, even use an expanded genetic code that would allow a bug to perform feats that no natural organism can. They want, for example to design a cell that will swim around blood vessels and digest fatty deposits on artery walls, customise bacteria that warn of pollutants or grow plants that change colour when they detect a whiff of nerve gas.

"We think now of cells as programmable devices. We are really moving beyond the ability to program individual cells to programming a large collection - millions or billions - of cells to do interesting things," said Dr Ron Weiss of Princeton, who reported the bacterial art. "We want to achieve sophisticated behaviours that depend on complicated sets of instructions."

Life runs on "wetware", with many protein signals floating from one part of a cell to another, or communicating between cells. Some proteins blend in new combinations. Others cleave one protein into two. Many interfere with the way genes - the instructions used to make proteins - are used to create feedback loops of activity, where the presence of a protein speeds up or damps down a metabolic pathway. In all, there are about 4,000 proteins that run the network of living circuits in an E coli.

Five years ago, Drs Michael Elowitz and Stanislas Leibler, then at Princeton, assembled three interacting genes (dubbed a repressilator) in a way that gave the E coli a green twinkle so they can blink on and off like Christmas-tree lights.

Meanwhile, James Collins, Charles Cantor and Timothy Gardner of Boston University made a genetic switch, endowing each modified bacterium with a rudimentary digital memory. Using the switch, they have wired up bacteria so that they get together to form communities called biofilms under the glow of ultraviolet light.

Dr Weiss has taken a similar approach to create a "Goldilocks" genetic circuit, one that lights up when a target chemical is present but only when the concentration is not too high and not too low. Earlier this year, working with Sara Hooshangi and Stephan Thiberge, he altered bacteria to make them respond to patterns of protein inputs, so that high or low levels or an output protein could be produced, like the 1s and 0s of a digital electronic circuit. The bacteria, for example, could be made to perform basic mathematical logic functions that are more commonly associated with silicon chips.

The new Nature paper applies similar techniques to a large population of cells. "Here we're showing an integrated package where the cells have an ability to send messages and other cells have the ability to act on these messages," said Dr Weiss, who worked with Subhayu Basu and colleagues in the California Institute of Technology, Pasadena.

They found a way to program E coli bacteria to emit red or green fluorescent light in response to a signal emitted from another set of

E coli. In one experiment, the cells glowed green when they sensed a higher concentration of the signal chemical and red when they sensed a lower concentration. Together, in a Petri dish, they formed a bull's-eye pattern - a green circle inside a red one.

Bacteria could help to detect an anthrax attack

The creation of patterns this way may be useful as a tool for other scientists, particularly developmental biologists who are trying to understand how the cells of an embryo arrange themselves into patterns that become the body parts of a mature organism. In fruit fly embryos, for example, the first cells are thought to differentiate into the head, abdomen and other segments in response to chemical signals that are emitted from the ends of the embryo. "We can make three segments like this," he said. "The genetic circuit we are using has similarities to one of those in the fly, which lays down a pattern along an axis."

As well as demonstrating how patterns are formed in living things, the bacteria could be turned into a sensing system for the detection of chemicals or organisms in laboratory tests. "The bull's-eye could tell you: This is where the anthrax is," said Dr Weiss.

The creation of such patterns is also a key step in one of Dr Weiss's eventual goals, which is to have the cells secrete materials that build physical devices such as antennas or transmitters in places that are hard for humans to reach, such as minefields and toxic dumps. Programmed cells could also be used to control the repair or construction of tissues within the body, possibly guiding stem cells to the locations where they are needed for the growth of new nerve, muscle or bone cells in a process Weiss called "programmed tissue engineering".

In addition to lab experiments, Dr Weiss and his colleagues are creating computer models of their engineered organisms, which allow them to study how small genetic modifications can influence how an organism behaves. So far, he said the experimental results have matched the computer models fairly closely, but the goal is to have a mathematically exact description of how each component works. The faithful reproduction of the protein circuits of E coli in a computer is one of the grand challenges of biology.

"If we build a network from scratch, we should be able to model all the important details," he said. At some point, a computer program will be able to create a genetic circuit to accomplish a given task. "Then we can do an experiment to see if the community of cells is behaving as we desire. That is going to have a tremendous number of applications."

But it will not be straightforward. Last month, Dr Elowitz - now at Caltech - and colleagues published a study that shows how genes blink on and off in living cells. The time lapse movies reveal a fundamental difference between cellular and electronic circuits when it comes to random fluctuations, or "noise". In a nutshell, the noise is much bigger and slower in cells than in electronics, an insight that will have to be taken on board when attempting to understand and reproduce the machinery of life.

Despite the worthy goals of synthetic biology, some are uneasy at the thought of scientists creating new kinds of bugs, of private labs creating genetic codes to run bacteria and viruses, and of scientists working out how to use bacteria to build structures, gather up plutonium or attack cancerous cells.

"We have to be careful what we do in the lab," agreed Dr Weiss. "It is mind-boggling and incredibly complex, but we do go to great lengths to understand and predict what these organisms will do."

Next story:  'Magnificent' superjumbo takes to the sky without a hitch