Designing 'Gene Circuits' that Control Cell Populations with Killer Genes

September 17, 2005

Durham, N.C. -- Lingchong You's Duke University research team makes and programs circuits, although not the kind that work in electronics devices. His are "synthetic gene circuits" that can regulate cell populations with molecular signaling and intentional extermination.

Such biocircuits have great potential for applications in biotechnology, computation, environmental engineering and medicine. For example, a "suicide" biocircuit could potentially be programmed into bacteria used to clean up pollution, making the microbes die off once their job was done. Synthetic gene circuits could also be engineered into bacteria to enable them to react to the presence of cancer cells by killing or disabling the tumor.

However, realizing such applications is a major challenge, said You, an assistant professor of biomedical engineering at Duke's Pratt School of Engineering and the university's Institute for Genome Sciences and Policy.

He created his first circuit, genetically engineered into the common gut bacterium E. coli, as a California Institute of Technology postdoctoral researcher working in Frances Arnold's laboratory. It was a "population control circuit" that used a technique called "quorum-sensing” to keep the densities of the bacteria in check. He and his colleagues described their achievement in an article in the April 4, 2005, issue of Nature.

The Caltech team created the circuit by introducing foreign genes into the E. coli. One foreign gene prompted the bacterial cells to release signaling molecules that accumulated as cell populations increased. When the signaling molecule concentrations grew high enough, a balancing act between two other imported genes in the altered bacteria tipped in a way that released a "killer" protein. Some of the E. coli cells then started dying.

"It was like the cells were all whispering to each other," he said in an interview. "At low cell densities, the chemical concentration of the signaling molecule was low. But, as cell densities increase, it was like a room getting more crowded. Each cell continued to whisper, but the overall noise level grew very high.

"That induced some unbearable consequences as some of the cells began getting killed off. As they died, the overall cell population started to decrease, then eventually built back up."

Under this process, cell numbers were maintained at relatively steady levels for extended periods. "We found that the steady state could last more than 30 hours, which was quite remarkable," he recalled. But, ultimately, each experimental run ended when the bacteria began mutating to produce offspring that lacked the control genes.

You then scaled down and modified this experiment so it could unfold within six ultra-tiny arrays called "microchemostats," each holding only billionths of a liter. Caltech graduate student Frederick Balagadde, working in the laboratory of Stephen Quake, fabricated the arrays -- complete with ultra-tiny bacterial growth chambers, "microfluidic" pumps and nutrient supply channels -- on a single dime-sized chip.

"The idea was to shrink the bacterial population by many orders of magnitude," You said. "We wanted to see if the cell-to-cell communications would still be effective in regulating the size of that smaller population. We hypothesized that with the much smaller size the process by which the population would be taken over by mutants would be delayed."

Balgadde, You and others reported on the microchemostat array July 1, 2005, in the journal Science, about 10 months after You came to Duke. "We did dramatically slow down the rate at which mutants took over," he said. "So we were able to observe circuit dynamics for a much longer time, sometimes for over 500 hours. But mutants still appeared faster than we expected."

You plans to send students and staff from his Laboratory of Biological Networks to Stanford,where Balagadde and Quake have both relocated, to train on the microchemostats in preparation for his next challenge: engineering two different cell populations to communicate with each other.

"We're trying to engineer a synthetic predator-prey system," he said. Under this arrangement, one cell population -- the predator -- will have a gene introduced that orders the release of a killer protein. But a gene in the other cell population -- the prey -- simultaneously signals the production of an "antidote" that turns off the predator's killer gene.

The predator cell population is thus rescued from self-destruction by the prey -- as long as the density of the prey cell population is high enough. "The predator relies on the prey to grow; the prey is acting like the food for the predator," You said, explaining his reasoning for the two populations' nicknames.

But if the predator cell population grows comparatively large, it will signal a separate killer gene to turn on in the prey, thus unleashing a protein that destroys some of the prey population.

He plans to introduce other genes to both cell populations that would make them glow in two different colors so their anticipated interactions and oscillations can be better traced. "This is much more complex than the population control circuit," he said. "It involves two populations and two different signals, like two different languages. Hopefully, if we build this circuit we could set up as a model system to look at how natural ecological interactions work."

Looking ahead, he envisions using similar concepts to program a "death pathway" into microbes engineered for environmental bioremediation efforts -- such as toxic waste cleanup. "One concern is, now that its job is done, how do you get rid of these bacteria," he said.

"Another tremendously interesting project is thinking about ways to make bacteria do computations using genetic circuits," he added.

And a "Holy Grail" goal is designing synthetic gene circuits for medical purposes, he said. "For example, if I designed some circuit that will sense some 'signature' of a cancer cell and do something that will either kill that cell or produce a gene to fix the problem."

You, the first faculty member to work on gene circuits at Duke, spoke Aug. 20, 2005, at a special "Life Engineering" symposium at the University of California, San Francisco, sponsored by the National Academies Keck Futures Initiative and other institutions.

He became involved with gene circuits as an outgrowth of his interest in mathematical modeling. "I did most of my Ph.D. work in computation, on the ability of mathematical models to analyze how a virus develops," he said.

His group has also developed a mathematical modeling simulation software package called Dynetica for use in the study of cellular networks.