Research Interests

Molecular genetic & physiological approaches
to study innate behaviors in the fly

Animals in the wild encounter many types of external stimuli such as threat and food, and must exhibit appropriate responses for survival. How does a brain recognize such stimuli with sensory systems, create internal representations of these external stimuli, and then elicit appropriate behavioral responses? To address this problem, I have been taking a systems approach and using the fruit fly, Drosophila melanogaster, because of a wealth of genetic tools available, a relatively simple brain, and a complex, interesting behavioral repertoire. Rapidly emerging tools also permit relatively facile identification of neural substrates. During my postdoctoral work, my focus has been to identify neurons that subserve a particular innate behavior, apply functional imaging and electrophysiology to probe their activity, and therefore define precisely contributions of each set of neurons to the behavior.

We have developed a novel behavioral paradigm in Drosophila in collaboration with Seymour Benzer. This paradigm involves avoidance of a substance, called Drosophila Stress Odorant (or dSO) emitted by flies subjected to mechanical stress or electrical shock. Most naive flies chose the fresh tube when given a choice in a T-maze between a fresh tube and a conditioned tube in which emitter flies were previously stressed. Through Gas Chromatography & Mass Spectometry (GC/MS) analyses, I identified CO2 as one component of dSO. Consistent with this, flies exhibited avoidance to CO2 in a dosage dependent manner. We next identified a single pair of glomeruli in the antenna lobe (AL) innervated by Gr21a+ expressing olfactory neurons that were specifically activated by CO2 through functional imaging. And this paired glomeruli (the V glomeruli) were necessary for avoidance to CO2 in collaboration with Dr. Richard Axel’s laboratory (Suh et al., 2004). Furthermore, activation of Gr21a+ expressing neurons using recently developed Channelrhodopsin-2 elicited avoidance response to its stimulus, the blue light (Suh et al., 2007). These experiments together indicate that Gr21a+ expressing neurons are dedicated to detecting CO2, and that avoidance to CO2 is likely mediated by a dedicated circuit. It is worth emphasizing that the CO2 neuron is a rare example of “object detector” found in the peripheral nervous system.

These studies described above demonstrate that we can map neural substrates to a particular behavior using molecular genetic approaches combined with functional imaging and electrophysiological techniques. In my laboratory, I would like to use similar approaches to address a slightly different problem. Stimuli such as CO2 elicit the same stereotypic response- avoidance- whether flies were starved or were altered in their circadian rhythm. Conversely, other stimuli generate behavioral responses largely 70 influenced by the internal state of animal. I have observed that avoidance or attraction (or valence) to appetitive odor in a simple T-maze is dependent on the satiety or hunger state of flies. Starved flies are attracted to appetitive odor whereas satiated flies are repelled by the same odor. This switch (or “satiety induced valence switch”) is specific to appetitive odor as many chemical odors tested thus far did not render flies to undergo the switchboth starved and satiated flies exhibited the same response to each chemical odor, whether it is attractive or aversive (Unpublished data). This result suggests that although appetitive odor is a complex mixture of compounds, it presents as “unitary percepts” to the fly. These observations suggest the hypothesis that the fly has a group of neurons or “grandmother” cells essential for recognition of the important stimulus and their activity is modulated by the satiety state.