The development of imaging systems involving calcium-sensitive fluorescent dyes has provided an unprecedented opportunity to observe the activity of neurons and circuits in real time. In a report in Cell, Wang et al. describe how they have used a dye called G-CaMP to study the relationship between structure and function in the Drosophila olfactory system.

A fly antennal lobe, in which G-CaMP — a calcium-sensitive green fluorescent protein — is expressed only in projection neurons that innervate the lobe. The high signal-to-noise ratio of G-CaMP provides a representation, at cellular resolution, of a defined population of neurons in the brain as the fly is stimulated by odorants at physiological concentrations. Different odours elicit different patterns of activation in the antennal lobe. Courtesy of J. W. Wang, Center for Neurobiology and Behavior, Columbia University, New York, USA.

In Drosophila, each olfactory sensory neuron expresses one of around 80 different odorant receptor subtypes. Projections from neurons that express the same receptor converge in structures called glomeruli in the antennal lobe. The glomeruli are innervated by the dendrites of projection neurons, which relay information to the mushroom bodies and protocerebrum. The fly olfactory system has become a popular model for studying olfactory coding because it is much simpler and more accessible than that of vertebrates, yet the glomerular anatomy of the primary relay centres is strikingly similar.

Wang et al. imaged the heads of flies in which either the projection neurons or the sensory neurons expressed the G-CaMP protein. The fluorescent intensity of this protein reflects the intracellular calcium level (a signature of electrical activity), and the authors detected the fluorescence using two-photon microscopy. This sensitive detection system enabled them to generate a high-resolution map of the glomeruli that were activated by different odours at concentrations that the fly would encounter in its natural environment.

The authors showed that each odour activated a specific combination of glomeruli. The response patterns were highly reproducible, not only between different trials in the same fly, but also between different flies. Interestingly, imaging of sensory and projection neurons produced the same odour-evoked patterns of glomerular activity, indicating that the pattern generated by the stimulation of sensory neurons is transmitted intact to higher processing centres in the brain.

Wang et al. also used this imaging technique to examine the molecular basis of olfactory coding in the fly. Sensory neurons that express the or43a receptor gene project to the DA4 glomerulus, and the authors identified a range of odours that activate DA4, but not another glomerulus, VA1lm. However, when they expressed or43a ectopically in the neurons that project to VA1lm, this glomerulus now responded to the same range of odours as DA4. This implies that the response patterns of individual glomeruli are probably determined by single receptor subtypes.

By combining calcium imaging with two-photon microscopy, Wang et al. have generated a model to test various principles of olfactory coding in flies, many of which might also be relevant to vertebrates. In addition to providing new insights into olfaction, this imaging technique is also likely to have more general applications for measuring neuronal activity in the fly brain in relation to various behaviours.