Energy from E. coli

From left: Jay Keasling with Francesco Pingitore and Chris Petzold.

Courtesy of Lawrence Berkeley Nat'l Lab - Roy Kaltschmidt, photographer

Jay Keasling watches as 700 billion Escherichia coli swish around inside a benchtop bioreactor in the brand-spanking new fermentation room of the Joint BioEnergy Institute in Emeryville, Calif. Seven copper pipes line the wall with a ready supply of nitrogen, oxygen, water, and other essentials, while an automated controller-looking like a souped-up frozen yogurt machine-regulates the temperature, pH, and oxygenation of the cloudy solution brewing within this one liter tank. This isn't just any E. coli multiplying inside, Keasling says proudly, "This is a strain we engineered and now it's producing biodiesel."

If anyone can marshal in the new era of alternative energy, it may well be Keasling, a bioengineer at the University of California, Berkeley, and the CEO of JBEI, which is a US Department of Energy-sponsored partnership between three California universities and three national laboratories. In the last five years, Keasling has coaxed yeast to synthesize the antimalarial compound, artemisinin. With the help of the Gates Foundation in 2004, his company Amyris Biotechnologies nudged the price down from $2.40 per dose to an expected $0.25. That project wrapped up in December 2007, and while Keasling remains the head of the scientific advisory board, he's not part of management. He has been too busy pondering the next great challenge for synthetic biology. And there's no question that repurposing his genetic constructs to produce isooctane-the key molecule in unleaded gasoline and a compound no known organism manufactures in nature-would be a step beyond traditional genetic engineering.

Growing up on a corn farm in Nebraska, Keasling never understood the logic of turning feed into ethanol and rebuilding our entire energy infrastructure, such as gas stations and pipelines, to accommodate it. "Ethanol is not a great choice," he says. Keasling's plan, along with a team of 125 scientists, is to farm a tall grass as the substrate and let plant-degrading microbes collected from tropical rainforests break down the cellulose into sugars. Then, bioengineers will feed those sugars to homegrown yeasts or bacteria like E. coli, which synthesize the hydrocarbons found in gasoline, diesel, and jet fuel. These organisms naturally produce compounds called isoprenoids that are the building blocks for artemisinin and other hydrocarbons.

According to Keasling, it took about 50 genes and control elements to get these organisms to crank out artemisinin and a few more tweaks to increase the efficiency. Although the isooctane molecule poses a special challenge to biology, Keasling expects to produce a variety of other components of gasoline and diesel fuel with no greater difficulty.

Stepping into the sleek, modern Emeryville laboratory, passing the robotic liquid-handlers and bench after bench of analytical chemistry equipment, the last thing you expect to find is greenery. But sure enough, the Nebraskan scientist, in his jeans and plaid shirt, ducks down one hallway, past a cart full of potting soil and into a nook where under the glare of full-spectrum UV lights, Arabidopsis, rice, switchgrass, and tobacco burst forth like a scene from the sequel to the movie Wall-e. All this greenery will eventually find its way back to the fermentation room.

As the bacteria secrete oil, molecule-by-molecule, it floats to the top of the chamber-another practical advantage over ethanol, which is both toxic to the bacteria that produce it and must be distilled from the solution. The amount of biodiesel produced today is tiny, and Keasling says that the process would have to be scaled up one million times and decreased in cost by over a 100-fold in order to meet the energy needs of the United States. Keasling is not troubled by this consideration, since E. coli are already used to produce large quantities of ethanol and other chemicals. By the time the biofuel process is industrialized, Keasling will probably be on to another project and have licensed the technology. "I don't want my folks getting the yields up to the next little iota," he says, "We should not be looking at the last decimal of Pi. We should be looking at the first decimal."