What we ‘like’ is often dependent on how we are behaving at any given moment. Odours and gasses seem particularly subject to such changes in hedonic valence (i.e. preference). The smell of sweat, for example, normally repulsive, suddenly doesn't seem so bad once one is out on the basketball court or halfway through a 5 km run. In a recent article published in Current Biology, Sara Wasserman, Alexandra Solomon and Mark Frye have made inroads into understanding the neural bases for state-dependent changes in sensory perception by studying what fruit flies think of carbon dioxide (CO2).

Previous work has shown that fruit flies will run away from a CO2 source. This is puzzling given that CO2 is a by-product of rotting fruit (a favourite fly food). Wasserman and colleagues first decided to test whether flies are repulsed or attracted by CO2 once they leave the ground. The team tethered individual flies within flight simulators that allowed animals to fly in place but also rotate freely. When the team exposed these animals to plumes of CO2 they found that the flies always turned upstream into the CO2 regardless of starting orientation. Fruit flies do exactly the same thing when they are attracted to odours on the wing. However, when given a chance to walk on a small glass slide, the flies, as expected, tried to walk their way out of the gas stream. Flies are clearly attracted to CO2 while flying, but are then repulsed by the gas as soon as they hit the ground running.

What receptors allow flies to sense and track CO2 molecules during flight? The team knew that the receptors that mediate aversion to CO2 during walking reside in a segment of the fly antenna. Flies were not able to track CO2 plumes in the flight simulator when this segment was covered over with glue. The team then used genetic targeting techniques to selectively inhibit synaptic release in a subset of antennal neurons that contain identified CO2 receptors. Surprisingly, this did not stop the little fliers from tracking CO2 plumes. The team decided to look for alternative receptors that might be involved, looking at two receptor pathways thought previously to be dispensable for sensing CO2. They found that mutations in both carbonic acid (a CO2 metabolite) receptor and co-receptor involved in odorant sensing prevented the airborne flies from tracking CO2. The two pathways come online during flight and cooperate to mediate responses to CO2.

To investigate what causes the switch in pathways, the team genetically inhibited the synaptic release of octopamine, a neuromodulator upregulated during flight. They found that these flies also actively avoided CO2 while flying. Reducing octopamine release essentially made a flying fly behave as if it were walking in the presence of CO2. This suggests that octopaminergic signalling modulates and reconfigures CO2 detection circuitry as flies shift between different locomotor modes.

The work of Wasserman and colleagues resolves the paradox of why flies would be put off by an environmental cue that could lead them to food. When flies are on the ground, CO2 is repulsive (perhaps because it is a cue that is proportional to overcrowding). But once in the air, CO2 is attractive to flies, presumably because it can lead them over long distances to food sources. Overall, this work shows that a single molecule can trigger exactly opposite behavioural responses depending on the neuromodulatory state of an animal and reminds us of the extent to which neural circuits can be completely and utterly reconfigured by neuromodulation.

M. A.
Drosophila tracks carbon dioxide in flight
Current Biology