Neural circuit evolved to process pheromone differently in two species of fruit fly

The males of two closely related species of fly respond differently to a female pheromone. It emerges that this difference is due to alterations in the activity of an evolutionarily conserved neural circuit in the brain.
Nicolas Gompel is in the Faculty of Biology, Ludwig Maximilian University, Munich, 82152 Planegg-Martinsried, Germany.

Search for this author in:

Benjamin Prud’homme is at the Aix-Marseille University, CNRS, Developmental Biology Institute of Marseille (IBDM), 13288 Marseille, France.

Search for this author in:

Particular sensory signals — a certain odour or taste, for instance — can trigger opposing responses in different species. What changes in neural circuits could elicit such contradictory behaviours? In a paper in Nature, Seeholzer et al.1 compare the circuits that control courtship in males in two closely related species of fruit fly that behave differently when exposed to a particular pheromone. They find an explanation for these opposing behaviours that few neuroscientists would have predicted.

While wandering on a rotten apple, a male fruit fly is likely to encounter a variety of other flies, only a few of which will be females of the same species. Identifying and courting these conspecific females is a crucial step in securing the male fly’s reproductive success. But this is not always an easy task. For instance, two sister species of fruit fly — Drosophila melanogaster and Drosophila simulans — are often found on the same fruits. Males of these species therefore rely heavily on sensing pheromones to identify conspecific females2.

To test whether a female fly is of the same species, the male touches her abdomen with his foreleg, sampling her pheromones using sensory neurons. D. melanogaster females produce the pheromone 7,11-HD, which promotes courtship behaviour when sampled by D. melanogaster males2. By contrast, D. simulans females do not produce 7,11-HD, and D. simulans males react to the pheromone (on either D. melanogaster females or D. simulans females doused with 7,11-HD) by aborting courtship behaviour. Hence, 7,11-HD is detected by males from both species, but elicits contrary behavioural responses in each.

What differences in the nervous systems of the male flies could account for the opposing role of 7,11-HD in regulating courtship decisions in these two species? Seeholzer et al. set out to answer this question by comparing the architecture and function of the neuronal circuits that govern courtship behaviour in both. To achieve this, the authors used a combination of sophisticated techniques: the genetic manipulation of specific populations of neurons; the activation of particular neurons with light using a technique called optogenetics; and the observation of neuronal activity in real time using functional imaging.

In D. melanogaster males, the decision to initiate courtship is made by the activation of a region in the brain called P1, which contains neurons that integrate multiple sensory signals from outside the brain3. Seeholzer et al. found that P1 has the same role in D. simulans. Moreover, the neural pathway that detects 7,11-HD and propagates the sensory input to P1 is structurally similar in D. melanogaster and D. simulans (Fig. 1).

Figure 1 | Changes in the activity of a courtship-promoting neural circuit. a, Female fruit flies of the species Drosophila melanogaster produce the pheromone 7,11-HD. Seeholzer et al.1 report that D. melanogaster males identify females of their species through neurons in the foreleg that express the protein Ppk23. These ppk23 neurons activate a population of neurons called vAB3 that, in turn, activates neurons in the P1 brain region, promoting courtship behaviour. vAB3 neurons also activate the mAL neuron population, which inhibits P1. In this species, vAB3 activation overrides mAL inhibition to trigger P1 activity, leading to courtship. b, The authors find that this neural circuit is evolutionarily conserved in males of a sister species, Drosophila simulans. However, mAL-mediated inhibition overrides vAB3 activation, so that P1 is inhibited and courtship behaviours are aborted. Whether this difference reflects changes in the activity of mAL neurons or changes in how the inputs are detected by P1 neurons remains unclear.

The authors showed that 7,11-HD is detected by a few taste-perceiving neurons in the male foreleg of each species that are characterized by expression of the protein Ppk23. These neurons activate a population of neurons called vAB3 that, in turn, directly excites P1 neurons4. However, the researchers demonstrated that vAB3 neurons also activate a population of neurons called mAL that directly inhibits P1 neurons4. The pheromone-processing pathway therefore bifurcates into excitatory and inhibitory routes that converge on P1. This circuit presumably provides a mechanism for setting pheromone responses and for tightly controlling the excitation of P1.

If the structure of the pheromone-processing pathway is evolutionarily conserved between species, how does 7,11-HD elicit a net excitation of P1 to trigger courtship in D. melanogaster, but not in D. simulans? Seeholzer et al. found that it’s simply a matter of balance. The mAL-mediated inhibition of P1 neurons completely masks their excitation by vAB3 neurons in D. simulans, resulting in a net inhibition of P1. But mAL activity only dampens the activation of P1 by vAB3 in D. melanogaster.

Seeholzer and colleagues’ work therefore suggests that it is not the architecture of the circuitry that has changed during the evolution of the two species, but the relative weight of the inhibitory and excitatory pathways that converge on P1 neurons. This is unexpected, because the idea that changes outside the brain can explain behavioural diversity has long dominated this field510. But the authors have shown clearly that evolutionary changes can be embedded more deeply in brain circuits, where no one has looked so far. This is a provocative new conceptual framework that explains how a behaviour can change between closely related species.

The most interesting aspect of Seeholzer and colleague’s work is undoubtedly the unexpected nature of the findings. But the study’s combination of state-of-the art techniques also makes it a technical triumph, for two reasons. First, using neurogenetic tools designed for the model organism D. melanogaster in the non-model species D. simulans was a challenge. Second, comparing the structure and function of individual circuit components between species, in the haystack of neurons that makes up the brain, is uncharted territory.

Of course, several questions remain. For instance, what caused the change in balance between the inhibitory and excitatory branches of the courtship pathway? Perhaps it was a progressive change that emerged from genetic variation between individuals belonging to the last common ancestor of these species. Under sexual selection, initial variations that caused two groups of males to respond differently to 7,11-HD, as well as two groups of females to produce different levels of the pheromone, might have gradually built a barrier between the incipient species. The genetic basis of the difference in neural-circuit activity between the two species, however, remains to be defined.

It is also unclear how fine-tuning of the circuit differs between the species. Seeholzer et al. showed that vAB3 and mAL neurons are active in both species, so there must be a difference in the way their signalling is propagated to P1. Perhaps there are changes in the levels of signalling by mAL neurons, or maybe P1 neurons respond differently to inhibitory inputs in each species.

Finally, the study highlights that integrative nodes in neural circuits such as P1 might be poised to accommodate evolutionary changes that underlie variations in behaviour. This concept is analogous to the evolution of particular organismal shapes. In such a setting, particular genes at key nodes in gene-regulatory networks are poised to accommodate changes that lead to the generation of forms11. Future work, following the approach pioneered by Seeholzer et al., will reveal the extent to which the evolution of behaviour is targeted to particular nodes in neural circuits.

Nature 559, 485-487 (2018)

doi: 10.1038/d41586-018-05595-y
Nature Briefing

Sign up for the daily Nature Briefing email newsletter

Stay up to date with what matters in science and why, handpicked from Nature and other publications worldwide.

Sign Up


  1. 1.

    Seeholzer, L. F., Seppo, M., Stern, D. L. & Ruta, V. Nature 559, 564–569 (2018).

  2. 2.

    Billeter, J. C., Atallah, J., Krupp, J. J., Millar, J. G. & Levine, J. D. Nature 461, 987–991 (2009).

  3. 3.

    Kimura, K., Hachiya, T., Koganezawa, M., Tazawa, T. & Yamamoto, D. Neuron 59, 759–769 (2008).

  4. 4.

    Clowney, E. J., Iguchi, S., Bussell, J. J., Scheer, E. & Ruta, V. Neuron 87, 1036–1049 (2015).

  5. 5.

    Schneiderman, A. M., Hildebrand, J. G., Brennan, M. M. & Tumlinson, J. H. Nature 323, 801–803 (1986).

  6. 6.

    McGrath, P. T. et al. Nature 477, 321–325 (2011).

  7. 7.

    McBride, C. S. et al. Nature 515, 222–227 (2014).

  8. 8.

    Prieto-Godino, L. L. et al. Nature 539, 93–97 (2016).

  9. 9.

    Karageorgi, M. et al. Curr. Biol. 27, 847–853 (2017).

  10. 10.

    Prieto-Godino, L. L. et al. Neuron 93, 661–676 (2017).

  11. 11.

    Stern, D. L. & Orgogozo, V. Evolution 62, 2155–2177 (2008).

Download references