The circuit basis of olfactory mate recognition and localisation

Abstract

The circuit basis of olfactory mate recognition and localisation Finding a mate has two essential steps: recognising an appropriate partner and locating it. Sex pheromones are used in solving both tasks; they carry qualitative information to recognise, and positional information to locate a mate. How does the brain extract both a qualitative and a quantitative message from a single signal, like a pheromone smell? We used the pheromone sensing circuit of Drosophila melanogaster as a model to understand how such a signal is processed in the brain and how that instructs sexual behaviours. We used connectomics, functional imaging, and behavioural tests to investigate the anatomy and function of a circuitry dedicated to processing cis-vaccenyl acetate (cVA) a male pheromone signal. We discovered a previously unknown second order projection neuron class (PN) that senses cVA, and relays this signal to a neural hub for sexual behaviours - including female receptivity. We imaged the activity of sensory neurons and second order PNs in the cVA pathway while presenting the female with a male fly at defined positions and asked what features of the stimulus can be decoded from neuronal activity. We show that both sensory neurons and PNs carry information about male position by cell type-specific distance tuning properties and lateralisation ability. Additionally, we found an interneuron that is only activated by contralateral input and is well placed in the circuit to create mutual inhibition between the two sides. We propose that this mechanism underlies improved lateralisation in PNs compared to sensory neurons. To further understand cVA processing we characterised the responses of two third-order neurons. We found that one is selectively tuned to an approaching male, and the response magnitude is strongly dependent on approach speed. A second, previously described, sexually dimorphic third-order neuron receives taste input besides cVA, and integrates these signals at its dendrites. Following the flow of information downstream of the taste and cVA integrating neurons we discovered a layered axo-axonic circuit; this layout suggests a hierarchical gating of different sensory modalities. Through connectomics and functional experiments, we showed that signals from the abdominal ganglion - controlling reproductive organs - and leg muscle load information enter this axo-axonic motif, and that manipulating any of these neurons during behaviour affects female receptivity bidirectionally. These findings together extend our understanding of cVA’s function as a pheromone that signals positional information. Supporting this function we propose an olfactory lateralisation mechanism previously unknown in Drosophila. Our observations about third order neurons suggest that quantitative and qualitative signals are separated at the third layer of cVA processing and that integrating other relevant modalities plays a role in this separation.