Introduction

The ability to sense and discriminate between chemical components in food is fundamental for the fitness and survival of all animals ¹. Species across diverse phyla, including vertebrates and arthropods, exhibit conserved appetitive responses to essential nutrients like sugars, while generally demonstrating avoidance of potentially toxic bitter compounds ^1–3. For amino acids and salt (NaCl), behavioral responses are often concentration-dependent: low concentrations of these key nutrients are attractive, whereas high concentrations frequently elicit aversion ^4–8.

In mammals, the perception of the five basic taste modalities (sweet, sour, bitter, salty, umami) is largely thought to adhere to a ℌlabeled lineℍ coding strategy ^1,9,10. In this model, specific taste modalities are mediated by distinct taste receptor cells expressing respective taste receptor - proteins^1,10. In contrast, the gustatory system of the fruit fly, Drosophila melanogaster, is organized in a more complex fashion. While sweet and bitter tastes are generally encoded by distinct populations of gustatory receptor neurons (GRNs), individual GRNs can exhibit multimodal sensitivity ^11–13. For instance, attractive nutrients such as sugars, low salt and some amino acids are sensed by sweet-sensing neurons, whereas chemicals to be avoided, such as bitter compounds, high salt, high acid, and certain amino acids are detected by bitter-sensing GRNs ^12–14. Furthermore, water, acids, and aversive high concentrations of salt are detected by yet other GRNs, distinct from both sweet and bitter GRN populations ^15–17. This multimodal nature is conferred by two types of chemoreceptors, often co-expressed within the same GRNs: about 60 Gustatory Receptors (GRs), which primarily encode receptors for sweet and bitter tasting chemicals, and about 64 Ionotropic Receptors (IRs), which are critical for sensing salts, acids, fatty acids and amino acids ^11–13,18. GR as well as IR based taste receptors are multimeric complexes, most likely hetero tetramers composed of two to four different GRs and IRs, respectively. Most IR based complexes include IR25a and IR76b, which are largely co-expressed and thought to function as core subunits. They are complemented with one or two additional IR subunits that confer specificity for different types of ligands (referred to as tuning receptors) ^7,19–22. For example, in sweet GRNs, IR25a and IR76b form complexes with IR56b to mediate low salt detection ^23–25 and with IR56d to mediate fatty acid responses ^20,26. In contrast, in bitter GRNs, IR25a and IR76b associate with IR51b to detect several amino acids ⁷ and with and IR94e to sense glutamate, respectively ²⁷. And in a third set of GRNs expressing the pheromone receptor Ppk23 and the neurotransmitter glutamate (named Ppk23^glut), IR25a and IR76b combine with IR7c to form a receptor for high concentrations of monovalent salts ²⁸. These observations established that in addition to the multimodal nature of GRNs (expressing both GRs and IRs), specific tuning receptors combine with core subunits to form tetrameric IR complexes that are dedicated to the detection of subgroups of chemical ligands.

The natural environment presents food chemicals as mixtures rather than single-modality isolates. For instance, low salt and amino acids naturally co-occur in the fruit fly’s preferred food sources: rotting, yeast-fermented fruits ²⁹. Consequently, taste systems must be capable of integrating signals from different taste modalities to guide foraging and feeding decisions ¹³. This integration occurs in the suboesophageal ganglion (SOG) and the Nucleus of the Solitary Tract (NTS), the primary taste processing center of insects and vertebrates, respectively, and plays a critical role in coordinating nutrient consumption ^13,30. Furthermore, signal integration has been reported at the peripheral taste cell level ^31,32. Specifically, while the labeled-line logic largely prevails in mammalian taste cells, significant ℌcross-talkℍ occurs within the taste bud ³². For example, ATP released from sweet- or bitter-sensing Type II cells can activate serotonin release in neighboring Type III cells, effectively processing and integrating information before it reaches afferent nerve fibers ^33,34. Similarly, in flies, bitter compounds and high salt solutions suppress sweet responses at the level of GRNs within the same taste sensilla, whereas acids can alleviate bitter-induced suppression of sweet taste ^19,35,36. However, despite these advances in understanding cellular and circuit-level processing, the possibility of integration occurring directly at the level of the gustatory receptor complex has remained largely unexplored.

In this study, we report that the IR20a functions as a multimodal tuning receptor essential for sensing both the amino acid arginine and low NaCl salt. We demonstrate that this dual sensitivity is achieved through the combinatorial assembly of a receptor complex: IR20a requires the co-receptor IR25a for arginine detection while the additional recruitment of IR76b is critically for low salt detection. Crucially, the IR20a/IR25a complex confers higher responses to amino acid-salt mixtures than each of the ligands alone. Heterologous expression of IR20a and IR25a in the pheromone sensing neurons and S2 cells reveals that this enhancement is synergistic rather than additive, indicating that the increased feeding preference arises from receptor synergy. Finally, we show that the IR20a pathway provides a distinct, state-independent layer of nutrient regulation that operates in parallel to the known starvation-modulated salt sensing pathway. Together, our findings uncover a fundamental molecular logic where the combinatorial architecture of peripheral receptor complexes provides a mechanism for integrating distinct, yet co-occurring, nutritional signals.

Results

IR20a-GAL4 labels a novel population of GRNs responding to amino acids and salts

Previous studies have implicated IR20a in amino acid taste in tarsi and the modulation of salt sensitivity in the labellum ³⁷. However, this study did not identify the specific tarsal GRNs for amino acid detection, nor did it establish whether IR20a is endogenously expressed in salt-sensing neurons to modulate salt taste in labellar GRNs. To clarify the cellular function of IR20a, we characterized its expression in the gustatory system and the brain using an IR20a-GAL4 driver ³⁷. Despite prior reports that IR20a is absent from the labellum ³⁸, our confocal imaging uncovered approximately 18 weakly labeled GRNs in this organ (Fig. 1A), corroborating a proposed role for IR20a in high-pH detection ³⁹. Beyond the labellum, IR20a marked two tarsal GRNs on each of the fore-, mid-, and hindlegs (Fig. 1B). In the pharynx, we detected two GRNs in the labral sense organ (LSO) and two in the ventral cibarial sense organ (VCSO) (Fig. 1D3 - D4), consistent with earlier findings ³⁸; however, no signal was evident in the brain or ventral nerve cord (Fig. 1C). Co-expression analysis using IR20a-GAL4, IR25a antibodies and IR76b-QF revealed that IR20a is co-expressed with both IR25a and IR76b (Fig. 1D - E). However, we found no overlap between IR20a-GAL4 and GR64f-LexA or GR66a-LexA (Fig. 1F - G). Thus, IR20a is co-expressed with the core receptors IR25a and IR76b in a population of neurons that is distinct from both sweet and bitter GRNs.

Figure 1.

To characterize the response profiles of IR20a-expressing GRNs, we performed single-cell calcium imaging of tarsal GRNs by expressing GCAMP6 under the control of IR20-GAL4 ⁴⁰. We focused on tarsal neurons because IR20a-GAL4 labels GRNs very weakly in the labellum, precluding reliable detection of GCaMP signals. Based on previous reports implicating IR20a in amino acid detection ³⁷, we first examined responses to 19 proteinogenic amino acids (Fig. 1H). Calcium imaging revealed that stimulation with 100 mM L-arginine or L-proline elicited robust calcium responses in IR20a GRNs, whereas other amino acids failed to induce significant activity (Fig. 1H). Dose-response analyses showed that only high concentrations of L-arginine (50 and 100 mM) and L-proline (100 mM) evoked statistically significant responses (Fig. 1I–K). We next examined salt sensitivity and found that IR20a tarsal GRNs responded strongly to 50 mM NaCl and KCl, but not to LiCl (Fig. 1L). NaCl-evoked responses were concentration dependent and reached a plateau at approximately 50 mM (Fig. 1M–N). We note that lysine was excluded from imaging experiments due to its intrinsic color, which interferes with fluorescence measurements; instead, its role was assessed behaviorally using the proboscis extension reflex (PER) assay in wild type and IR20a mutant flies, which revealed that IR20a is not required for lysine attraction (Fig. S1A - B).

Together, these results indicate that tarsal IR20a GRNs are multimodal sensory neurons that respond to both specific amino acids (arginine and proline) and monovalent salts (Na⁺ and K⁺).

IR20a is required for arginine and low NaCl salt attraction

To determine whether IR20a is functionally required for cellular responses to amino acid arginine, proline, and monovalent salt Na⁺ and K⁺, we examined IR20a mutants using calcium imaging. Ca²⁺ responses to L-arginine, L-proline and KCl were preserved and not statistically different from those of wild type flies (Fig. 2A - C). In contrast, Ca²⁺ responses to 50 mM NaCl were completely abolished in IR20a mutants, a defect fully rescued by reintroducing UAS-IR20a (Fig. 2A - C).

Figure 2.

We next assessed the behavioral relevance of IR20a using two-choice feeding assays. Although our imaging data indicated that IR20a is only necessary for detecting low concentration of NaCl but not arginine, IR20a mutants exhibited significantly reduced attraction to both 100 mM arginine and 50 mM NaCl, a phenotype that was fully rescued by including UAS-IR20a driven by IR20a-GAL4 (Fig. 2D - E). In contrast, attraction to 100 mM proline and avoidance of 50 mM KCl were unaffected in IR20a mutants (Fig. 2F - G).

We further examined the role of IR20a using the proboscis extension reflex (PER) assay. Consistent with previous reports, tarsal stimulation with arginine failed to elicit PER in wild-type flies ⁴¹ (Fig. 2H). In contrast, labellar stimulation with arginine induced a strong PER that was markedly reduced in IR20a mutants and fully rescued by re-expression of IR20a (Fig. 2I). Together with the two-choice feeding assays, these results establish that IR20a is essential for arginine attraction. To determine whether IR20a is selectively required for arginine sensing, we examined PER responses to all other amino acids using distinct amino acid pools. No significant differences were observed between wild-type controls and IR20a mutants for any other amino acid, demonstrating that IR20a specifically mediates arginine attraction rather than amino acid attraction in general (Fig. S1C - D). We next tested PER responses to salt. Wild-type flies displayed weak PER upon tarsal stimulation and robust PER upon labellar stimulation with 50 mM NaCl (Fig. 2J - K). Both responses were significantly reduced in IR20a mutants and were restored by IR20a-GAL4–driven IR20a expression (Fig. 2J - K). Together, these findings indicate distinct roles for IR20a in arginine detection between labellar and tarsal neurons. Although IR20a is not required for arginine-evoked calcium responses in tarsal GRNs, it is critical for arginine-driven feeding preference and the labellar PER. In contrast, IR20a is essential for low NaCl salt detection at both the cellular level and behavioral attraction. UAS-hid mediated genetic ablation of IR20a neurons further confirms the role of IR20a in arginine and low salt attraction (Fig. S1E - F). Thus, our data establish IR20a as a key component of both arginine and low NaCl salt attraction in Drosophila.

IR20a functions with IR25a to detect arginine, while IR76b confers low NaCl salt responses

We next tested the requirement for IR25a and IR76b in IR20a-GAL4 expressing tarsal GRNs for arginine and low-salt detection. Calcium imaging revealed that IR25a is essential for both arginine and low NaCl salt responses (Fig. 3A - C). In contrast, IR76b is dispensable for arginine detection but is required for NaCl responses (Fig. 3D - F). This suggests that IR20a forms a complex with IR25a to detect arginine, whereas low-salt sense requires the additional recruitment of the co-receptor IR76b.

Figure 3.

To determine if IR20a/IR25a is sufficient for arginine detection and IR20a/IR25a/IR76b for low-salt detection, we misexpressed IR20a in Ppk23-positive tarsal GRNs, which function in pheromone sensing and do not respond to ligands such as arginine or low NaCl salt (Fig. 3I) ⁴². Immunostaining confirmed that Ppk23-GAL4 labels a pair of GRNs in tarsal 5a sensilla. All of these GRNs express IR25a, but only one pair also expresses IR76b (Fig. 3G - H). Calcium imaging of the Ppk23 GRN in the 5a sensillum revealed that misexpression of IR20a conferred sensitivity to both arginine and NaCl (Fig. 3I - J). Crucially, mutation of IR25a abolished both the arginine and NaCl sensitivity acquired by IR20a misexpression (Fig. 3K). However, a mutation of IR76b only affected the conferred NaCl response but leaving arginine responses in tact (Fig. 3L). These data are consistent with our loss-of-function analysis and suggest that IR20a cooperates with IR25a to detect arginine, while the addition of IR76b is required to confer low-salt sensitivity.

To rigorously define the molecular mechanism, we characterized receptor ligand specificity by expressing different combinations of IR20a, IR76b, and IR25a in Drosophila S2 cells and performed calcium imaging. Expression of IR20a, IR25a, or IR76b alone, as well as co-expression of IR20a with IR76b or IR25a with IR76b, failed to elicit responses to arginine (Fig. 4A - F). In contrast, co-expression of IR20a with IR25a, or co-expression of all three genes elicited robust responses to 100 mM L-arginine (Fig. 4G - H). By contrast, neither IR20a nor IR25a alone, nor their co-expression, elicited responses to NaCl, whereas expression of IR76b alone introduced robust NaCl responses, and co-expression with either IR25a or IR20a abolished these responses (Fig. 4A, I -N). This is consistent with previous reports that IR76b forms a sodium channel in cultured cells ⁴, and further suggests that addition of either IR20a or IR25a into a IR76a-only complex suppresses NaCl channel activity. However, triple expression of IR20a, IR25a, and IR76b restored sensitivity to NaCl (Fig. 4A, O). To confirm the specificity of these responses, we tested all basic amino acids except for lysine at their maximum water-soluble concentrations using pooled stimulation paradigms. Neither the IR20a/IR25a complex nor the IR20a/IR25a/IR76b complex responded to any basic amino acid other than arginine (Fig. S2). Together, these results demonstrate that IR20a forms distinct receptor complexes with different co-receptor compositions, enabling multimodal detection of arginine and sodium.

Figure 4.

IR20a confers synergistic responses to an arginine and NaCl mixture

Given that amino acids and salt often occur together within a natural food source, the multimodal detection of arginine and NaCl salt by IR20a prompted us to test whether IR20a confers synergistic responses to their combination. PER assays with labellar stimulation revealed that a low concentration mixture (10 mM each) of arginine and NaCl elicits a stronger, IR20a-dependent response than either alone (Fig. 5A - B). To determine whether IR25a and IR76b are required co-receptors, we misexpressed IR20a in Ppk23 tarsal 5a associated GRNs. Calcium imaging revealed that control neurons were unresponsive to a 10 mM mixture of arginine and NaCl (Fig. 5C), while IR20a misexpression conferred a robust response (Fig. 5C - D). This response was abolished in IR25a mutants but remained intact in IR76b mutants (Fig. 5E - F). Heterologous expression in S2 cells further confirmed the synergistic response. Specifically, cells co-expressing IR20a and IR25a did not respond to arginine or NaCl alone at concentrations of 10, 25, or 50 mM, but elicited a strong response to a 50 mM mixture (Fig. 5G - I and Fig. S3). However, the calcium response to the mixture was not further enhanced in the triple expression condition (Fig. 5 J - K). These results reveal that IR20a functions with IR25a to integrate arginine and low NaCl salt taste information to mediate synergistic responses, thereby enhancing feeding preference.

Figure 5.

IR20a provides a tonic arginine and low NaCl salt signal parallel to the state-modulated IR56b pathway

Amino acid and salt preferences in flies are strongly influenced by internal nutritional state ^17,25,41,43, but the sensory mechanisms underlying this modulation remain poorly understood. Having established the requirement of IR20a for arginine and low NaCl salt detection, we asked whether it mediates state-dependent shifts in preference. Two-choice feeding assays revealed that wild-type flies exhibit increased preference for arginine and low NaCl salt following amino acid deprivation and salt deprivation, respectively. However, these state-dependent changes occur independently of IR20a (Fig. 6A - B). Calcium imaging in IR20a tarsal GRNs confirmed that IR20a GRN responses to salt do not change between salt-fed and deprived conditions (Fig. 6C - D). In contrast, two-choice feeding assays showed that the state-dependent increase in salt preference is dependent on another low-salt tuning receptor, IR56b, which co-express with IR25a and IR76b in sweet GRNs but does not co-express with IR20a (Fig. 6E). Calcium imaging further confirmed IR56b is essential for state-dependent low salt sensitivity (Fig. 6F–G, and Fig. S4E). As a control, IR56b GRNs showed no change in sucrose and H2O sensitivity after salt deprivation (Fig. S4 A - D). These results suggest that IR20a provides a tonic arginine and low NaCl salt signal to maintain the stable detection of these essential nutrients, while the IR56b pathway adjusts sensitivity based on homeostatic need.

Figure 6.

Discussion

In this study, we identify a specific population of taste sensory neurons in Drosophila that are broadly tuned to both low NaCl salt and specific amino acids arginine, implying that these essential nutrients are perceived through a multimeric channel containing IR20a, which, in combination with IR25a and IR76b, results in receptors with distinct preference for these ligands. To our knowledge, this represents the first evidence of signal integration between distinct taste modalities occurring at the level of the receptor complex itself.

Refinement of Multimodal Coding in Drosophila

The complexity of the Drosophila gustatory system has long been underappreciated. While early models proposed a labeled-line organization similar to mammals, specifically humans and mice, comprehensive genetics, electrophysiological and calcium imaging studies have revealed that gustatory receptor neurons (GRNs) often exhibit broader tuning profiles than previously thought. For instance, sweet GRNs have been implicated in the detection of fatty acids and low salt, challenging the notion of exclusive modality coding ^17,20.

Our identification of IR20a-expressing neurons significantly expands this perceptive range. Contrary to previous reports suggesting IR20a expression overlaps with amino-acid sensing sweet neurons ³⁷, our intersectional genetic analysis reveals that IR20a marks a distinct population of non-sweet, non-bitter GRNs. While Ganguly et al suggested IR20a mediates detection of various amino acids (including glycine, serine, and phenylalanine) based on broad IR76b-GAL4 imaging and weak behavioral phenotypes ³⁷. Our single-cell resolution imaging demonstrates that IR20a GRNs are narrowly tuned to arginine and proline, as well as monovalent salts (NaCl, KCl). Furthermore, while we find that tarsal responses to arginine are maintained in IR20a mutants, our behavioral assays establish that IR20a is strictly essential for the labellar-mediated preference for both arginine and low salt. This defines a novel, multimodal “savory” channel that is distinct from canonical single-modality sensory pathways.

Combinatorial Assembly of IR Complexes

The IR gene family expanded during invertebrate evolution, presumably from a common ancestor of the iGluR gene family ^44,45. The functional diversity of these receptors appears to rely on the combinatorial assembly of distinct subunits ^11,12,18,46. Based on the work presented here, we propose that IR20a serves as a dual-function tuning receptor that dictates ligand specificity through differential complex formation.

Heterologous expression in S2 cells allows us to dissect this molecular logic. We observed that IR76b alone forms a constitutive Na⁺ channel, consistent with previous studies ⁴. However, co-expression with IR20a or IR25a abolishes this constitutive activity, suggesting that heteromeric assembly alters channel gating properties. We propose that the IR20a/IR25a complex forms the functional receptor for arginine, sufficient for ligand detection even in the absence of IR76b. In contrast, low NaCl salt detection requires the assembly of a tripartite complex: IR20a/IR25a/IR76b. In vivo, where these three subunits are co-expressed, it is likely that distinct stoichiometric populations, such as IR20a/IR25a dimers and IR20a/IR25a/IR76b multimers, coexist within the same dendritic membrane, thereby endowing the neuron with sensitivity to both amino acids and salts.

Receptor-Level Signal Integration

Perhaps the most profound finding of this study is the mechanism of signal integration. Animals feeding in the wild encounter complex mixtures, and our behavioral data show that flies exhibit a synergistic preference for mixtures of arginine and NaCl. While synergy is typically attributed to central neural processing, our data indicate this integration occurs already peripherally at the gustatory receptor level, although additional modulation of this synergy in the brain is possible.

We demonstrate that the sensitivity to arginine of the IR20a/IR25a complex, which does not respond to NaCl alone, can be potentiated by addition of NaCl. This suggests that a possible allosteric mechanism where the binding of arginine to IR20a/IR25a exposes a latent binding site for Na⁺, or conversely, that Na⁺ stabilizes the ligand-bound conformation of the channel. This molecular integration offers a highly efficient mechanism for weighting and prioritizing nutrient signals; the preference for arginine over other amino acids is likely rooted in its unique physiological status, as it is one of the ten essential amino acids (EAAs) that Drosophila must acquire through diet, serving as the indispensable precursor for nitric oxide (NO) signaling and a limiting factor for oogenesis and fecundity ^47–49. In the ecological niche of Drosophila, arginine and NaCl are concurrently enriched in yeast, their primary natural food source ²⁹. By evolving a receptor complex that acts as a coincidence detector, the fly can specifically amplify the sensory output only when these two distinct nutrient classes are present together. This ensures that the IR20a pathway does not just signal the presence of any salt or any nitrogen source, but rather identifies high-quality, ecologically relevant resources that provide both essential amino acids and necessary electrolytes, thereby optimizing the fly’s foraging efficiency and reproductive success.

Segregation of Tonic and State-Dependent Pathways

Finally, our results clarify the relationship between sensory detection and homeostatic regulation. Sodium salt preference in Drosophila is highly plastic, increasing during deprivation ¹⁷. We provide evidence that IR20a provides a tonic, stable signal for low NaCl salt and arginine that remains unchanged between salt-fed and salt-deprived conditions. In contrast, the state-dependent increase in NaCl salt preference is mediated by the previously identified IR56b pathway ^23,25.

This parallel organization, one pathway (IR20a) ensuring the consistent recognition of vital nutrients and the other (IR56b) modulating sensitivity based on physiological need, represents a sophisticated survival strategy. It ensures that the fly never loses the ability to identify nutrient sources, even while dynamically adjusting its behavioral drive to consume them. In summary, IR20a and its co-receptors IR25a and IR76b represent a versatile molecular hub for multimodal nutrient detection and integration.

Materials and Methods

Key resources table

Fly husbandry and strains

Flies were reared on a standard cornmeal diet at 25°C and 70% humidity under a 12 hr light/12 hr dark cycle. The standard diet was prepared by dissolving the following ingredients in 10 L of water: 72.53 g agar, 520 g corn meal, 1100 g malt extract, 275 g yeast, 31.27 g propionic acid, and 14.07 g tegosept. A complete list of fly strains and source reagents used in this study is provided in the Key Resources Table.

Molecular biology and transgenic flies

To generate UAS-IR20a, genomic DNA isolated from w1118 flies served as the PCR template. The full genomic sequence of IR20a was amplified with primers listed in the Key Resources Table and inserted into the pUAST vector. All constructs were sequence-verified prior to embryo microinjection.

Behavioral Analysis

Two-choice feeding assays

Assays were performed as previously described¹⁹ with minor modifications. Flies were 8–10 days old at the time of testing. For standard assays, newly eclosed flies were raised on standard food. For salt-manipulation experiments, newly eclosed flies were kept on standard food for 2–4 days and then transferred for 5 days to either salt-deficient food (0.7% agar, 563 mM sucrose) or salt-excess food (0.7% agar, 563 mM sucrose, 50 mM NaCl). 563 mM sucrose was selected to match the caloric density of standard food. For amino acid (AA) manipulation, flies were transferred after 2–4 days to either AA-deficient food (standard diet without yeast extract, supplemented with 16.5% maltose for isoenergetic balance) or AA-excess food (standard diet supplemented with 10% yeast extract and 4.16% maltose) for 5 days. Assays were conducted in Petri dishes (60 mm x 60 mm) containing 0.5% agarose mixed with the indicated tastants.

To visualize feeding, one half of the dish contained blue dye (brilliant blue FCF, 0.125 mg/ml) and the other contained red dye (sulforhodamine B, 0.2 mg/ml); dyes were swapped between conditions to control for color bias. Approximately 30 flies per vial were starved for 18–20 hours on wet filter paper, chilled on ice, and transferred to assay dishes in the dark for 2 hours. Feeding preference was scored by examining abdominal coloration under a stereomicroscope. The Preference Index (PI) was calculated as: PI = (N_color1 – N_color2) / N_total-fed. Trials where >50% of flies failed to feed were excluded.

Proboscis Extension Reflex (PER)

PER assays were adapted from previous protocols ¹⁹. For leg stimulation, flies were anesthetized on ice and mounted dorsally on glass slides using double-sided tape, leaving appendages and the proboscis free. For labellar stimulation, flies were immobilized inside a white pipette tip with only the head and labellum exposed. 20–30 flies were mounted per slide and allowed to recover for 1 hour at 25°C (70% humidity). Prior to testing, flies were water-satiated until they ceased drinking. Stimulation was performed by touching the legs or labellum with the indicated tastant (NaCl or amino acids) for 3–5 seconds. A positive response was recorded if full proboscis extension occurred. Each fly was also tested with 20 mM sucrose as a positive control. Two trials were conducted per tastant with a > 20-minute inter-trial interval. Responses were scored as 100% (extension in both trials), 50% (extension in one trial), or 0% (no extension). Flies failing to respond to the sucrose control in both trials were excluded from the analysis.

Immunohistochemistry

Dissection and staining were performed on 6–10 day-old flies. Flies were CO2-anesthetized, and labella and legs were dissected using a razor blade; legs were severed between the third and fourth tarsal segments. Tissues were immediately immersed in 4% paraformaldehyde and briefly centrifuged (1000 rpm, 30 s) to ensure complete submersion. Following fixation on ice for 1 hour, tissues were washed three times (15 min each) in PBST (1x PBS + 0.3% Triton X-100) at room temperature. Blocking was performed in PBST containing 5% horse serum for 2 hours at room temperature. Tissues were incubated with primary antibodies in blocking solution overnight at 4°C, followed by three 15-minute washes in PBST. Tissues were then re-blocked in PBST + 5% horse serum for 30 minutes and incubated with secondary antibodies for 3 hours at room temperature. After three final washes in PBST, samples were mounted in antifade medium and sealed with nail polish. Images were acquired on a Zeiss LSM 800 confocal microscope using 20X or 40X objectives.

Primary antibodies

Chicken anti-GFP (1:2000), rabbit anti-RFP (1:2000), mouse anti-HA (1:500), mouse anti-nc82 (1:5000), and rabbit anti-IR25a (1:200). Anti-IR25a was generated against synthetic peptides (SKAALRPRFNQYPATFKPRF and DVAEANAERSNAADHPGKLVDGV) as described in Benton et al ⁴⁴.

Secondary antibodies

Goat anti-chicken Alexa 488 (1:2000), goat anti-rabbit Alexa 488 (1:200), goat anti-mouse Alexa 568 (1:2000), goat anti-rabbit Alexa 568 (1:2000), and goat anti-mouse Alexa 647 (1:2000).

Calcium Imaging

Single GRN calcium imaging

single tarsal GRN calcium imaging was performed as previously described ⁴⁰. UAS-GCaMP6m was driven by specific GAL4 lines in 7–10 day-old females. Legs were prepared by severing at the tibia-femur junction. The cut end was sealed with grease and attached to double-sided tape on a glass-bottom dish (35 x 35 mm), leaving the 4th and 5th tarsal segments suspended. Warm agarose (50–60°C) was applied to the tape to stabilize the preparation. Imaging was performed on a Nikon Ni-U upright microscope with a 40X water immersion objective. Fluorescence changes were monitored in cell bodies relative to adjacent background regions. Recording began 10 seconds prior to tastant application and continued for 2–3 minutes. The response was quantified as DF/F (%) = (F - F₀) / F₀, where F₀ is the mean baseline fluorescence of the 5 seconds preceding stimulation. The maximum DF/F value was used for analysis.

S2 cell culture and imaging

S2 cell culture and imaging were performed by adapted protocols from previous ²⁴. Drosophila S2 cells were cultured in Sf-900 II SFM supplemented with 10% FBS and 0.5% penicillin/streptomycin at 25°C. Cells were transfected with pAc5.1 vectors encoding IR20a (genomic), IR25a (cDNA), IR76b (cDNA), and GCaMP6m. Imaging was conducted using a Nikon Ni-U microscope (10× objective). Prior to imaging, culture medium was replaced with isotonic bath buffer (10 mM HEPES, 10 mM glucose, 250 mM NMDG-Cl; pH 7.4; 520 mOsm). After a 2-minute equilibration, the buffer was removed to a minimal volume. Following a 10-second baseline recording (F₀), cells were stimulated with ligand solutions containing the test tastant (NaCl or arginine). To maintain constant osmolarity (520 mOsm) across all conditions, the concentration of NMDG-Cl in the stimulation buffer was adjusted inversely to the tastant concentration. Fluorescence was recorded for 120 seconds, and background-subtracted peak DF/F (%) responses were quantified.

Quantification and Statistical Analysis

Normality of data distribution was assessed prior to test selection. For non-normally distributed data, comparisons between two groups were performed using the Mann-Whitney test, while multiple group comparisons were analyzed using the Kruskal-Wallis test followed by Dunn’s post-hoc test. Data are presented as mean standard error of the mean (SEM). The value n represents the number of preparations for immunostaining, the number of neurons for calcium imaging, the number of flies for PER, or the number of plates for two-choice feeding assays. Specific statistical details are provided in Dataset S1 - S8.

Figure supplements

Figure S1.

Figure S2.

Figure S3.

Figure S4.

Data availability

All data generated or analyzed during this study are included in the manuscript and supporting files.

Acknowledgements

We thank Dr. Minjiao Liu for making the pUAST-IR20a construct. We thank Drs. Hui Xiao, Guoqiang Zhang, the Bloomington Drosophila Stock Center, the Tsinghua Fly Center, and VDRC for fly strains and plasmids, and the Animal Center of Yunnan University for antibodies. We thank Prakash Pandey for his critical comments on the manuscript. This work was supported by grants from the National Natural Science Foundation of China (grant no. 32070997 to Y.C), Applied Basic Research Foundation of Yunnan Province (grant no. 2019FY003018 to Y.C.), Research Startup Funding of Yunnan University (to Y.C.) and the NIH (1RO1GMDC05606-01 to H.A.)

Additional information

Authors contribution

B.W. performed the immunostaining, leg calcium imaging, S2 cell culture and imaging, Proboscis extension reflex assays and two-choice feeding assays, G.Q. performed immunostaining, leg calcium imaging, and two-choice feeding assays, X.W. performed Proboscis extension reflex assays. B.W., G.Q., and Y.C. analyzed data, H.A. and Y.C. conceived and supervised the project, designed experiments, and wrote the paper.

Funding

MOST | National Natural Science Foundation of China (NSFC) (32070997)

• Yan Chen

Applied Basic Research Foundation of Yunnan Province (Yunnan Province Applied and Basic Research Foundation) (2019FY003018)

• Yan Chen

HHS | National Institutes of Health (NIH) (1RO1GMDC05606-01)

• Hubert Amrein

Additional files

Data S1. Genotypes and statistical details for Fig.1.

Data S2. Genotypes and statistical details for Fig.2.

Data S3. Genotypes and statistical details for Fig.3.

Data S4. Genotypes and statistical details for Fig.4.

Data S5. Genotypes and statistical details for Fig.5.

Data S6. Genotypes and statistical details for Fig.6.

Data S7. Genotypes and statistical details for Fig.S1.

Data S8. Genotypes and statistical details for Fig.S2.

Data S9. Genotypes and statistical details for Fig.S3.

Data S10. Genotypes and statistical details for Fig.S4.