The interactions between animals, including humans, and the microbes surrounding them are governed by a subtle balance. To understand the mechanisms underlying this equilibrium, the insect Drosophila melanogaster serves as a powerful model, as microorganisms are essential for the host's development, survival, nutrients, or making certain food sources digestible (Storelli et al., 2011). Some beneficial microbes occupy ecological niches that would otherwise be filled by harmful microorganisms if they were absent (Kim et al., 2010). Conversely, certain pathogenic species must be detected, avoided, or even actively combated by animals (Charroux et al., 2020, Chen et al., 2024, Tleiss et al., 2024). The ability of animals to detect, identify, and respond appropriately to the microbial world is therefore essential for their homeostasis and sustainability.

Recent studies have begun to explore the hypothesis that essential bacterial components, such as lipopolysaccharides (LPS) and peptidoglycan (PGN), which are well-characterized ligands for receptors expressed by immune cells, might also modulate host neuronal activity. Such molecular dialogue could allow the host to “sample” its surroundings via its sensory system and adopt behaviors that limit contact with pathogenic microbes or, in case of infection, reduce its impact on the host and its offspring (Kobler et al., 2020, Yanagawa et al., 2017). For instance, PGN detection by a few specific neurons can modulate oviposition behavior in infected females (Kurz et al., 2017, Masuzzo et al., 2022). In this case, the detection of a universal bacterial cell wall component by certain neurons in the brain acts as a molecular alarm, triggering a behavioral change in infected females, reducing their oviposition rate, and potentially conserving energy for the energy-demanding immune response. Other studies demonstrate that bacterial molecular patterns are directly detected by the fly's sensory system. The detection of LPS by bitter taste neurons enables flies to avoid LPS in feeding and oviposition assays (Soldano et al., 2016). We previously reported that by activating the bitter gustatory network, bacterial PGN can suppress the appetitive effect of a sucrose solution, demonstrating that PGN can be perceived as a bitter substance by flies (Montanari et al., 2024)

In this study, we used the proboscis extension reflex assay (Shiraiwa and Carlson, 2007) and calcium imaging to assess whether a PGN solution could also activate the sweet taste circuit in flies. Our data reveal that PGN can elicit Gr5a+ neuron-dependent activation within the sweet taste network, demonstrating a dual response of the fly's taste system to PGN. This response is attractive at high concentrations and repulsive at lowers, differing from the dual response to salt or hexanoic acid observed in other gustatory neurons (Jaeger et al., 2018, McDowell et al., 2022, Pradhan et al., 2023, Zhang et al., 2013). This dichotomous response suggests that Drosophila shares a need to detect and differentiate between high and low concentrations of specific compounds present in food substrates, engaging different receptors and neuronal cells with distinct coding logic. Furthermore, the detection of bacteria by flies requires integrating inputs from multiple subsets of sensory neurons to generate appropriate responses to bacterial PGN.