Obviously, the search for safe, edible food is a lifelong pursuit of all animals. Although animals are born with innate preferences for certain foods,the ability to modulate these preferences dynamically is clearly adaptive. Put simply, any starving animal (including humans) will tolerate normally unpalatable food. Surprisingly, this complex-seeming behavioural trait is tractable to simple genetics, and studies in Drosophila have now implicated two signalling pathways that are conserved from fly to human:neuropeptide Y and insulin-like proteins.

Drosophila larvae must feed voraciously in order to gain sufficient weight for pupation. It is possible to score larval feeding behaviour quantitatively by feeding larvae a tasty liquid diet of yeast and sugar with a green dye and measuring how their bodies change colour. When quinine (the bitter antimalarial drug found in Indian tonic water) is added to their diet, larvae find it disgusting – perhaps because they have yet to discover gin! However, larvae starved for 40 or 120 min overcome their innate aversion to quinine and start feeding again. This assay provides a robust baseline from which to look for modulation of risk-taking behaviour in feeding.

Mammalian neuropeptide Y (or its Drosophila homologue NPF) is thought to modulate feeding behaviour in both fly and human, but is it involved specifically in the decision to overrule avoidance of unpalatable food? To find out, the experimenters produced lines of transgenic flies in which the promoters of either NPF or its receptor (NPFR) drove expression of the yeast transcription factor GAL4. In such fly lines, any transgene downstream of the UAS promoter (which binds GAL4) will be switched on in any cells where GAL4 is switched on. This potent and widely used Drosophila technology thus provides a toolbox for expressing transgenes of choice in very specific populations of cells – in this case, just those cells expressing NPF or NPFR.

In this study, the authors used this approach to drive expression of a temperature-sensitive allele of the synaptic protein shibire, known(ironically for a study of feeding) as shits, in cells expressing NPF or NPFR. This is an excellent way of disrupting synaptic transmission: the mutant dynamin protein encoded by shitsperforms normally at 23°C but becomes dysfunctional in just a few seconds at 30°C. These manipulations didn't affect the behaviour of non-deprived larvae, but food-deprived larvae responded very differently. At the permissive temperature (23°C), the transgenic flies behaved like wild-type flies,feeding on aversive food, but at the restrictive temperature (30°C) they avoided it. These results suggest that the NPF/NPFR neuronal circuit is involved in suppressing avoidance of unpalatable food in hungry animals. To confirm this result, the authors used RNAi interference; by driving UAS-controlled constructs encoding double-stranded RNA to either NPF or NPFR,they were able to take down expression of each gene specifically. Again, they were also able to make hungry flies avoid aversive food. By contrast, flies over-expressing NPFR ate even more of the quinine diet after food deprivation. These results further show that the control of avoidance depends not just on the NPF neuronal circuit but also on signalling through NPF. Similar studies implicated the insulin receptor and DILP2, one of the seven insulin-like proteins encoded by the Drosophila genome.

So it seems as if the decision to overrule an innate aversion to noxious food when starving is attributable in Drosophila to the NFP and insulin-like signalling pathways, and to a relatively small neuronal circuit. Interestingly, neuropeptide Y in humans is associated with suppressing anxiety and fear, leading to a plausible anthropomorphic explanation for the behaviour. It will be interesting to see if this model is applicable beyond insects.

Wu, Q., Zhao, Z. and Shen, P. (
). Regulation of aversion to noxious food by Drosophila neuropeptide Y-and insulin-like systems.
Nat. Neurosci.