Last spring, on a research vessel cruising through the North Sea, Swiss scientists examined tiny vials of bacteria mixed with seawater for hints of fluorescent light. By analyzing how brightly the bacteria glowed, and with which colors, they were able to diagnose and characterize the early aftermath of an oil spill.

"We were actually very happy that we could do this, and that it turned out so well," says Jan Van der Meer, an environmental microbiologist at the University of Lausanne, in Switzerland. He announced his team's results last week at the Society for General Microbiology's autumn meeting in Dublin.

Living biosensors like these bacteria, which are engineered to glow a particular color in response to a given chemical, have graced petri dishes in research laboratories for decades. But it is only recently that they are being put to practical use, as scientists adapt and deploy them to test for environmental contaminants. Sensor bacteria give faster and cheaper--if somewhat less precise--results than traditional chemical tests do, and they may prove increasingly important in detecting pollutants in seawater, groundwater, and foodstuffs.

In preparation for their research expedition, Van der Meer and his team created three different strains of bacteria, each tailored to sense a particular kind of toxic chemical that leeches into seawater from spilled oil. They began with different strains of bacteria that naturally feast upon these chemicals, each releasing specialized enzymes when they come in contact with their chemical of choice. By hooking up the gene for a fluorescent or bioluminescent protein to the cellular machinery that makes those enzymes, the scientists effectively created a living light switch: whenever the chemical was present, the bacteria would glow.

For each class of toxic chemical, Van der Meer used a different color protein, so that he could easily determine which chemicals were present based on the wavelength of emitted light. And whenever possible, he transferred the entire switch mechanism into another strain of bacteria more suited to a highly controlled lab life than its exotic, oil-eating cousins.

The research team, working in concert with several other European labs, obtained permission from the Dutch government to create a small, artificial oil spill in the waters of the North Sea. They sampled seawater at various time points after the spill, using a luminometer to measure whether sensor bacteria added to each sample had detected the corresponding chemical. Unlike traditional chemical analyses, which can take weeks and require large, expensive instruments, the biosensor test could be performed on site in a matter of minutes.

"Analytical methods can potentially take a long time and a lot of processing," says Ruth Richardson, a bioenvironmental engineer at Cornell University. "It certainly isn't something you can do remotely."

Van der Meer adds that bacterial sensing, which is inexpensive compared with chemical methods, could be particularly useful for routine monitoring. "The extreme simplicity of this is that the heart of the sensor is the bacterial cell, and that the cell is a multiplying entity," says Van der Meer. "It's extremely simple to reproduce them, and then you have enough for thousands of tests."