DURHAM, N.C. -- A team of scientists at Duke University Medical Center has used a component of the "nose" of common bacteria to create a genetically-engineered protein biosensor that can identify the presence of a particular chemical in its surroundings.
The group's prototype biosensor can detect the presence of maltose, a sugar compound, in blood, and can switch on a molecule that produces fluorescent light to signal its catch. With further modifications, the researchers believe an array of biosensors can be created to instantly sample a wide variety of "target" molecules, from hormones in blood to pollutants in river water.
Based on the work, published in the April 29 issue of the Proceedings of the National Academy of Sciences, primary author Homme Hellinga predicts the first useful application will be a portable "bedside laboratory" that physicians can use to receive on-the-spot analysis of a patient's blood. Such a kit would reduce the time and expense of sending samples out for laboratory analysis, and would help doctors in diagnosing particular disorders and in precisely regulating medications used during cancer chemotherapy or other complexes diseases.
The study was funded by the Whitaker Foundation, which awarded Hellinga a three-year Whitaker Fellow grant. Hellinga is an assistant professor of biochemistry at the medical center. Working with him on the study were graduate students Jonathan Marvin and Ethan Corcoran, undergraduate student Neil Hattangadi, and technical assistants Jian Zhang and Sheryl Gere.
The idea of exploiting proteins to use as molecular diagnostic devices has been around for a long time, and has been used successfully in a few instances, such as measuring blood glucose for the management of diabetes. The secret in creating such devices lies in finding a way to have the protein "signal" that it has recognized the molecule of interest. But engineers have had to rely on happenstance to find naturally occurring proteins that combine recognition of the desired molecule with just the right signaling properties, Hellinga said, "and such combinations of circumstances are very rare." So he and his team set out to change the odds by trying to genetically engineer a protein to build in a signaling function, which is called a "reporter."
The group "borrowed" a protein from E coli bacterium that senses the presence of maltose, a disaccharide sugar molecule that this microbe sometimes feeds on.
Using advance computer systems, they studied the protein to see how its shape changes when maltose becomes attached to it. The researchers were looking for a special spot on the molecular surface of the protein where they could hook up the reporter molecule. They planned to use a fluorophore reporter, a molecule that changes its emitted fluorescent light when the shape of its immediate surroundings is altered. The trick, said Hellinga, was to locate the spot where the reporter feels the movement in the structure of the protein when maltose becomes attached to it. Using their computers, the team created a special sequence of events: the protein binds to maltose, and starts to change. These structural changes affect the fluorophore which then, itself, changes. It emits light, thus providing the desired signal of these molecular events.
The researchers then set about to build the designed protein in the laboratory to see whether it would work as predicted. Using genetic engineering methods, they created a mutant protein with a specialized chemical hook that allows the reporter group to be attached precisely in the right place. They produced the protein and tested it with maltose, and it worked. "It was a very exciting moment to see that our computer models actually worked," Hellinga said.
But this was only the first step in the goal of rationally creating biosensors. Next the team had to demonstrate that they could change the binding properties of the protein without destroying the properties of the reporter. They further re-engineered the prototype to see whether the response of the sensor could be "tuned" to measure different levels of maltose. Again using genetic engineering, they made a series of changes in the structure of the binding docket of the protein, each of which was expected to respond to different levels of maltose. Once more they were successful, and now had created a small family of engineered proteins which, when combined together, can measure maltose over a very wide range of levels. Using genetic engineering they had therefore "tuned" the response of the protein to maltose, analogous to an amplifier being adjusted to a particular sound level.
With completion of this milestone, the team predicts they can create "a suite" of different sensors within three years. "We want to create chemosensitive surfaces that respond in a very specific manner to whatever it is we are interested in, whether checking PCBs in the environment or for sugar levels in the blood of diabetics," Hellinga said. "We will have to keep engineering the protein so that it will behave itself, but it is quite doable."