Molecular Basis of Hexanoic Acid Taste in Drosophila melanogaster

Abstract

Animals generally prefer nutrients and avoid toxic and harmful chemicals. Recent behavioral and physiological studies have identified that sweet-sensing gustatory receptor neurons (GRNs) in Drosophila melanogaster mediate appetitive behaviors toward fatty acids. Sweet-sensing GRN activation requires the function of the ionotropic receptors IR25a, IR56d, and IR76b, as well as the gustatory receptor GR64e. However, we reveal that hexanoic acid (HA) is toxic rather than nutritious to D. melanogaster. HA is one of the major components of the fruit Morinda citrifolia (noni). Thus, we analyzed the gustatory responses to one of major noni fatty acids, HA, via electrophysiology and proboscis extension response (PER) assay. Electrophysiological tests show this is reminiscent of arginine-mediated neuronal responses. Here, we determined that a low concentration of HA induced attraction, which was mediated by sweet-sensing GRNs, and a high concentration of HA induced aversion, which was mediated by bitter-sensing GRNs. We also demonstrated that a low concentration of HA elicits attraction mainly mediated by GR64d and IR56d expressed by sweet-sensing GRNs, but a high concentration of HA activates three gustatory receptors (GR32a, GR33a, and GR66a) expressed by bitter-sensing GRNs. The mechanism of sensing HA is biphasic in a dose dependent manner. Furthermore, HA inhibit sugar-mediated activation like other bitter compounds. Taken together, we discovered a binary HA-sensing mechanism that may be evolutionarily meaningful in the foraging niche of insects.

Keywords: attraction; aversion; gustatory receptor; hexanoic acid; ionotropic receptor; noni.

Fig. 1. Toxicity and biphasic activations of hexanoic acid (HA) in a dose-dependent manner.

(A) Survival rate of control flies fed with 1% sucrose alone, 1% sucrose with the indicated amounts of HA (0.1%, 0.5%, 1%, and 2%), and 1% agar only (n = 4). (B) Dose-response curve of HA tip-recordings from S6, I8, and L6 sensilla (n = 10-12). (C) Tip recording in the presence of 0.1% HA after inhibiting different GRNs (Gr64f-GAL4 [sweet-sensing], Gr66a-GAL4 [bitter-sensing], ppk23-GAL4 [calcium-sensing], and ppk28-GAL4 [water-sensing]) by expressing UAS-kir2.1 under the control of the indicated GAL4s on L6 sensilla (n = 10-15). (D) Tip recording in the presence of 1% HA after inhibiting above-mentioned GRNs by expressing UAS-kir2.1 under the control of the indicated GAL4s on S6 sensilla (n = 10-16). (E) Diagrammatic representation showing the dual mechanism of HA sensation on sweet gustatory receptor neuron (GRN) and bitter GRN. (F) Proboscis extension response (PER) analysis of indicated neuron-ablated flies using above-mentioned GAL4s to 0.1% HA and Gr64f-GAL4-ablated flies to 2% sucrose (n = 6). (G) PER response of neuron-ablated flies of above-mentioned GAL4s to 1% HA (n = 6). All error bars represent SEMs. Single-factor ANOVA coupled with Scheffe’s post hoc analysis was performed to compare multiple sets of data. Asterisks indicate statistical significance compared with the control (*P < 0.05, **P < 0.01).

Fig. 2. Genetic screens using electrophysiology with 0.1% hexanoic acid (HA) and the behavioral assay.

(A) Tip recordings from all labellar sensilla of control flies (n = 10) by stimulation with 0.1% HA. (B) Tip-recording analyses from L6 sensilla to 0.1% HA for control and 31 Ir mutants (n = 10-15). (C) Tip-recording analyses from L6 sensilla to 0.1% HA for control and nine sweet Gr mutants (n = 10-16). (D) Representative sample traces of control and candidate mutants (Gr64d¹ and Ir56d¹) from (B) and (C). (E) Tip recordings with dose responses from L6 and S6 sensilla to 0% to 2% HA for control, Gr64d¹, and Ir56d¹ (n = 10-20). (F) Recovery experiments using tip-recording assays from L6 sensilla for Gr64d¹ and Ir56d¹ defects. Genetically recovered flies were driven by crossing each wild-type gene with Gr64f-GAL4 and Ir56d-GAL4, respectively (n = 10-18). (G) Proboscis extension response (PER) analyses showing the defect and rescue response from labellum for Gr64d¹ and Ir56d¹ defects (n = 6). All error bars represent SEMs. Single-factor ANOVA coupled with Scheffe’s post hoc analysis was performed to compare multiple sets of data. Asterisks indicate statistical significance compared with the control (**P < 0.01).

Fig. 3. Genetic screens using electrophysiology with 1% hexanoic acid (HA).

(A) Mapping analyses using tip recordings from all 31 sensilla to 1% HA (n = 10). (B) Tip-recording assays from S6 sensilla of control and 31 Ir mutants (n = 10-23). (C) Screens with control and 26 Gr mutants to 1% HA using electrophysiology (n = 10-20). (D) Representative sample traces of control, ∆Gr32a, Gr33a¹, and Gr66a^ex83 from (C). (E) Genetic rescues of ∆Gr32a, Gr33a¹, and Gr66a^ex83 deficits in the neuronal responses to 1% HA aversion using its own GAL4/UAS systems (n = 10-16). (F) Heat map analyses representing dose-dependent responses of control, ∆Gr32a, Gr33a¹, and Gr66a^ex83 using tip recordings from S6 sensilla to indicated concentration of HA (n = 10-20). (G) Tip-recording analyses from S6 sensilla to 1% HA for the control and nine sweet Gr mutants (n = 10). All error bars represent SEMs. Single-factor ANOVA coupled with Scheffe’s post hoc analysis was performed to compare multiple sets of data. Asterisks indicate statistical significance compared with the control (*P < 0.05; **P < 0.01).

Fig. 4. Behavioral analysis by stimulation of labellum using 1% hexanoic acid (HA).

(A) Behavioral rescues of ∆Gr32a, Gr33a¹, and Gr66a^ex83 deficits to the 1% HA aversion using the expression of each UAS transgene driven by the respective GAL4 (n = 6-11). (B) Concentration-dependent proboscis extension response (PER) responses by the stimulus to the labellum of control, ∆Gr32a, Gr33a¹, and Gr66a^ex83. The mixture of different concentrations of HA (0%, 0.1%, 0.5%, and 1%) and 2% sucrose was provided as experimental stimulus, and 2% sucrose only as control stimulus (n = 6-9). All error bars represent SEMs. Single-factor ANOVA coupled with Scheffe’s post hoc analysis was performed to compare multiple sets of data. Asterisks indicate statistical significance compared with the control (**P < 0.01; n.s., indicates non-significance).

Fig. 5. Sugar inhibition by hexanoic acid (HA).

Tip-recording analyses from L6 sensilla with 2% sucrose only or mixture of the indicated concentrations of HA (0%, 0.1%, 0.5%, and 1%) and 2% sucrose (n = 14-20). All error bars represent SEMs. Single-factor ANOVA coupled with Scheffe’s post hoc analysis was performed to compare multiple sets of data. Asterisks indicate statistical significance compared with the control (**P < 0.01).