Bacterial Uracil in the Larval Gut Primes Adult Flies to Taste Microbial Danger

In the crowded microbial world that animals inhabit from birth, the ability to sense danger before it strikes can mean the difference between survival and infection. A new study published in PLOS Biology by Romane Milleville, Gérard Manière, Martine Berthelot-Grosjean, Olivier Lamotte, Yaël Grosjean, Julien Royet and C. Léopold Kurz has now traced, molecule by molecule and neuron by neuron, how an encounter with bacteria during larval life permanently rewires the taste system of the fruit fly Drosophila melanogaster. The team identifies uracil, a simple nucleobase released by certain bacteria, as the chemical cue that primes developing larvae, and shows that this signal travels through a chain of reactive oxygen molecules and a single specialized sensory neuron to leave an enduring imprint on adult feeding behavior.

The work builds on the group’s earlier discovery that adult flies whose larvae were raised alongside certain bacteria reject food laced with peptidoglycan, a core component of bacterial cell walls. Flies reared in germ-free conditions, by contrast, happily extend their proboscis toward sucrose solutions containing peptidoglycan, even though their taste system remains fully functional for other aversive compounds such as caffeine. Intriguingly, only some bacterial species could restore this sensitivity in germ-free larvae: the pathobiont Levilactobacillus brevis succeeded, while the symbiont Lactiplantibacillus plantarum failed. Two questions therefore loomed. What bacterial molecule carries the priming signal, and which host cells and proteins receive and transmit it?

To answer the first question, the researchers exploited a mono-association protocol in which sterilized eggs are deposited onto food containing a chosen bacterial strain or chemical, and the adults that later emerge are tested with the proboscis extension reflex, a standard readout of gustatory response. Live L. brevis primed the larvae, but heat-killed bacteria and bacterial supernatant did not, indicating that a living metabolic process was at work. A clue came from earlier studies showing that bacterial uracil metabolism governs the transition from harmless commensal to damaging pathogen in the fly gut. When the team turned to Erwinia carotovora, a bacterium for which uracil-metabolism mutants were already available, the picture sharpened dramatically. Two mutant strains unable to release uracil lost the ability to prime larvae, while simply adding purified uracil at a concentration of one millimolar to the food was sufficient to trigger priming in completely germ-free animals. Cytosine and thymine, close chemical relatives of uracil, could not do the same, underscoring the specificity of the signal.

Timing proved equally critical. The team had previously identified a developmental window between egg laying and 48 hours after oviposition during which bacterial exposure must occur. The new experiments narrowed this window further: larvae exposed to uracil only 24 hours after hatching failed to acquire the primed state, confirming that the earliest moments of larval life are decisive. Calcium imaging of adult proboscis neurons later revealed the cellular consequence. The ppk23-expressing taste neurons that directly detect peptidoglycan in adult flies responded robustly to the bacterial molecule only in animals whose larvae had encountered uracil, demonstrating that the larval experience genuinely alters neuronal responsiveness in adulthood rather than merely changing gross behavior.

With the bacterial cue identified, attention turned to the host machinery. Prior research had established that uracil activates the enzyme Duox in the Drosophila intestine, triggering production of reactive oxygen species that attack invading pathobionts. The team tested whether the same pathway underlies priming. Using RNA interference to reduce Duox specifically in enterocytes, the absorptive cells of the gut lining, they abolished the adult aversion to peptidoglycan. A temperature-sensitive genetic switch allowed them to silence Duox at precise life stages: blocking the enzyme during larval development suppressed priming, while blocking it in adults had no effect. Direct biochemical measurements confirmed that hydrogen peroxide accumulates in larvae within 24 hours of uracil exposure, and that this signal disappears when catalase, the hydrogen peroxide-degrading enzyme, is overexpressed in gut cells or when the antioxidant vitamin C is added to the food.

Yet reactive oxygen species alone could not reproduce the effect. Supplementing larvae with hydrogen peroxide at several concentrations and exposure schedules failed to prime them, even though vitamin C reliably blocked priming when bacteria or uracil were present. This paradox suggests that the spatially and temporally precise burst of ROS generated inside the gut lumen matters in ways that global exposure cannot mimic. One attractive hypothesis, which the authors discuss, is that locally produced ROS oxidize lipids or other biomolecules to form reactive electrophilic metabolites. In vertebrates, lipid peroxidation products such as 4-hydroxynonenal are known activators of TRPA1 channels, which sense oxidative stress through covalent modification of conserved cysteine residues. Whether a similar chemistry operates in the fly intestine remains to be demonstrated, but it would elegantly explain why uracil works while exogenous hydrogen peroxide does not.

The next step was to find the receptor that senses the ROS signal. The fly’s TrpA1 ion channel, an ortholog of the mammalian pain and irritant sensor TRPA1, is activated by heat, reactive oxygen species, ultraviolet light and irritating chemicals, and exists as five alternatively spliced isoforms. Using knock-in flies expressing single isoforms and knock-out flies lacking specific combinations, the researchers showed that isoforms B, C and E, but not A or D, are competent to mediate priming. Mapping the expression patterns of these isoforms revealed a striking convergence: all three are co-expressed in a single isolated neuron at the very front of the larva, within the terminal organ, a peripheral sensory structure. This neuron also expresses Gr66a, a marker of bitter-sensing gustatory neurons. Silencing TrpA1 in these cells during the larval stage, but not the adult stage, abolished priming, and restoring TrpA1-B expression in Gr66a-positive neurons of otherwise TrpA1-deficient flies fully rescued the phenotype. TrpA1-A could not substitute, proving that the isoforms are not interchangeable for this function.

Calcium imaging then connected the pieces. When larvae expressing the calcium indicator GCaMP were stimulated with hydrogen peroxide, exactly one of the three anterior Gr66a-positive neurons responded strongly, and it was the same neuron that responded to quinine, a bitter compound. The TrpA1-B-positive terminal organ neuron was likewise activated by hydrogen peroxide, and this response vanished when TrpA1 was knocked down in that cell. Most remarkably, the team showed that medium in which uracil-treated larvae had been incubated could activate TrpA1-B-positive neurons in completely naive recipient larvae, and that this activity disappeared when Duox was silenced in the donor larvae’s enterocytes. This implies that a diffusible, Duox-dependent signal, likely hydrogen peroxide or a derivative, is released from the gut into the environment, where it can act at a distance. Such a mechanism raises the tantalizing possibility that infected larvae broadcast information about microbial presence to nearby conspecifics, potentially shaping behavior at the population level.

Finally, the team confirmed that the phenomenon extends beyond the laboratory confines of the proboscis extension reflex. In a two-choice free-feeding assay using 96-well plates of colored agar, female flies allowed to roam and feed freely strongly preferred plain sucrose over sucrose mixed with peptidoglycan, with preference indices exceeding 0.8. Flies raised on antibiotics throughout larval life, and TrpA1 knock-out mutants, particularly those lacking isoforms B, C and E, lost this discrimination and consumed both foods indiscriminately. The findings establish a complete signaling arc: bacteria release uracil, uracil triggers Duox-dependent ROS production in the larval gut, ROS activate a single TrpA1-B-positive terminal organ neuron, and the experience is somehow carried across the wholesale destruction and rebuilding of metamorphosis to endow the adult taste system with sensitivity to bacterial peptidoglycan. How that memory survives the radical remodeling of larval tissues and circuits remains the study’s central open question, with candidates ranging from persistent neural circuits and re-specified neurons to epigenetic marks and distributed transcriptional states. Whatever the answer proves to be, the work makes clear that the foundations of what an adult animal can taste are laid, quite literally, in the microbial company it keeps as a child.

Subject of Research: How larval gut exposure to bacterial uracil primes adult Drosophila gustatory aversion to peptidoglycan via Duox-dependent ROS and TrpA1-B neuron signaling

Article Title: Uracil-driven ROS signaling and larval TrpA1-B neuron activation are required for priming the adult Drosophila gustatory response to bacteria

Article References: Milleville, R., Manière, G., Berthelot-Grosjean, M., Lamotte, O., Grosjean, Y., Royet, J., & Kurz, C. L. (2026). Uracil-driven ROS signaling and larval TrpA1-B neuron activation are required for priming the adult Drosophila gustatory response to bacteria. PLOS Biology, 24(10), e3004042. https://doi.org/10.1371/journal.pbio.3004042

Image Credits: AI Generated

DOI: 10.1371/journal.pbio.3004042

Keywords: Drosophila, uracil, reactive oxygen species, Duox, TrpA1, gustatory neurons, peptidoglycan, larval priming, microbiota, terminal organ, gut immunity, sensory plasticity