For the past few days in my office, a small fruit fly has flown past me whenever I eat a banana. I've resisted swatting the fly (which I imagine is an escapee from a nearby lab) because, when you think about it, the fly has just performed an impressive feat. It has intercepted hundreds of different odorant molecules coming off the fruit, used a few thousand neurons in its tiny brain to turn the chemical information into electrical signals, interpreted the signals as food, and altered its flight path to attempt a fly-by snack attack.
How the fly does all this is just beginning to be unraveled, in large part because of research in Dr. Richard Axel's lab. Dr. Axel, University Professor of Biochemistry & Molecular Biophysics, and his graduate students and postdocs have identified about 1,000 odor receptors in rodents and about 100 in Drosophila, the fruit fly, inspiring labs around the world to study how the brain interprets odor.
Studies of the fly receptors have shown that the structure of the fly's olfactory system is remarkably similar to the mouse's (and probably our own) but haven't been able to reveal how the system actually works. Recently, two members of the lab have developed cutting-edge imaging techniques to peer into the fly's sandgrain-sized brain to see smell in action.
The olfactory system, even in flies, is complex. Unlike the visual system, where three receptors for blue, red, and green light are sufficient to produce hundreds of different hues, the olfactory system employs hundreds or thousands of odor receptors whose signals combine to produce a potentially infinite array of smells.
The odor receptors are arranged in the fly's antennae that lie in the front of its head, between its two large compound eyes. The receptors are contained within sensory neurons that run from the antennae to the antennal lobe, the first relay station for smell processing in the fly's brain. The antennal lobe looks a bit like cauliflower, with the florets called glomeruli.
Sensory neurons contain odor receptors in the antennae, and anatomical studies show that each sensory neuron appears to contain just one type of odor receptor. With incredible precision, all the different neurons that contain that one receptor type wind their way to just one glomerulus within the antennal lobe.
The anatomy of the fly's system, says one of the lab's graduate students, Allan Wong, logically leads to the idea that smell is first represented in the brain as a "map" of active glomeruli on the antennal lobe. A peach odor, which probably activates several different receptors, might map to, say, glomeruli 1, 2, and 10; a different odor would have a different pattern, activating, say, glomeruli 1, 2, and 21.
Until now, the tiny size of the fly's brain kept researchers from using imaging techniques to confirm what the anatomy suggested and to see just how different odorants mapped out onto the antennal lobe.
Dr. Jing Wang, a postdoc in the lab, thought it could be done with two-photon microscopy, a new technique that uses infrared laser light to see into living organisms but without cooking the tissue like confocal microscopy tends to do. Dr. Axel agreed, and Dr. Wang built a scope from scratch.
The scope alone, however, cannot distinguish between neurons that are firing and quiescent ones. For that, Dr. Wang teamed up with Mr. Wong to perfect the use of a fluorescent dye in the fly brain that would light up a neuron when the neuron fired. When they were done, the whole system produced an unheard of 30-to-1 signal-to-noise ratio, about 10 times better than the next best system. "It's like fMRI for the fly," Dr. Wang says.
The two researchers then focused their new "fly fMRI" on the miniscule antennal lobe to capture the most detailed images yet of fly brains in the act of smelling. They wafted 16 different odors representing various scents over the fly's antennae and watched the neurons light up.
The movies reveal that every fly is hard-wired to produce the same map in response to a given odor molecule. When flies encounter the ripe banana smell of isoamyl acetate, for example, neurons in the same four glomeruli fire in each fly. The mushroomy smell of 1-octen-3-ol, on the other hand, causes neurons in only one glomerulus to fire. (Movies of glomeruli in action are available at http://www.cell.com/cgi/content/full/112/2/271/DC1/)
"We think this is what the fly brain sees," Mr. Wong says, "but we haven't demonstrated yet that the animal actually makes use of these patterns." To be sure, they'll have to alter an odor molecule's map and see if the change alters fly behavior.
The ultimate goal, the researchers say, is to use the imaging system in combination with electrical recordings from individual neurons to trace the path of smell through the entire brain. "We essentially understand now how odor is deconstructed," Dr. Wang says. "Now we need to figure out how it's reconstructed in the brain."