Figure 5. Acetic acid activates sugar- and bitter-sensing neurons.

(A–H) Calcium imaging of taste sensory neurons reveals that acetic acid (AA) activates sugar-sensing neurons (labeled with Gr64f-Gal4; panels A-D) and bitter-sensing neurons (labeled with Gr66a-Gal4; panels E-H) in both fed and two-day starved flies. (A, E) Spatial maps of GCaMP activation by each stimulus for individual flies (A, fed; E, starved). (B, F) Example ∆F/F₀ traces for individual trials in the same flies shown in A and E, respectively. (C, G) Average GCaMP activation across all trials in all flies of each group. Gray bars indicate stimulus delivery (2 s). (D, H) Peak response to each stimulus averaged across all trials for each group. Circles represent individual fly averages. Within each group, responses to each stimulus were compared to responses to water, and fed and starved groups were also compared for each stimulus (*p<0.05, **p<0.01, ***p<0.001, two-way ANOVA followed by Bonferroni post-tests). Data shown in this figure represent n = 15–19 trials, six flies per group (sugar neurons, sucrose stimuli), n = 39–43 trials, 14 flies per group (sugar neurons, AA stimuli), n = 30 trials, 10 flies per group (bitter neurons, lobeline stimuli), or n = 54 trials, 18 flies (bitter neurons, AA stimuli). n values are larger for AA stimuli because we combined the results of two different datasets, but only the more recent dataset included the appropriate sucrose and lobeline stimuli. Only the more recent dataset is shown in panels C and G, whereas peak responses to AA stimuli for the combined dataset are analyzed in panels D and H. See also Figure 5—figure supplements 1–7.

Figure 5—figure supplement 1. Calcium imaging setup for taste neuron imaging.

(A) Side view of fly. Wings and legs were taped to allow unobstructed stimulation of the labellum (arrow). Tastant droplets were delivered to the labellum via a glass microcapillary and removed by a vacuum line, both of which were controlled by MATLAB software. (B) Top view of fly. Red box shows area of cuticle removed to expose the ventral brain, which includes the SEZ. (C) Still frames from video showing tastant delivery during imaging.

Figure 5—figure supplement 2. GCaMP responses in individual flies.

(A–B) Example GCaMP traces from sugar-sensing (A) or bitter-sensing (B) neurons of individual fed or starved flies for three trials of stimulation with each tastant. Sugar neurons were labeled with Gr64f-Gal4 and bitter neurons were labeled with Gr66a-Gal4. Each row represents responses from a different fly. Vertical scale bars (left) = 100% ∆F/F₀. Within-fly variability often appeared greater for acetic acid responses than sugar or bitter responses, but some flies showed more trial-to-trial variability than others (e.g. compare top and bottom rows of each panel). The third row of panel A and first, third, and fourth rows of panel B are from two-day starved flies; all other rows are from fed flies. (C–D) Within-fly variability was quantified by calculating the range of peak responses (maximum - minimum) for each fly to each stimulus across three trials. For each fly, the range of response to each stimulus is plotted versus the fly’s mean response to the stimulus (blue = acetic acid, red = sugar or bitter) since the range tends to increase with the mean response. Lines represent a linear fit for acetic acid responses (blue) or sugar or bitter responses (red). For a given mean response, acetic acid tended to induce a greater range of responses than sugar or bitter, suggesting increased within-fly variability.

Figure 5—figure supplement 3. Acetic acid responses of sugar-sensing neurons in sugar receptor mutants are not affected.

(A) Average GCaMP activation of sugar-sensing neurons in fed control flies (∆8Grs/+) and mutant flies lacking all eight sugar Grs (∆8Grs/∆8Grs) in response to water, 100 mM sucrose, or acetic acid (AA). Both groups of flies were heterozygous for Gr61a-Gal4 and UAS-GCaMP6m to enable expression of GCaMP6m in sugar-sensing neurons. Gray bars indicate stimulus delivery (4 s). (B) Peak response to each stimulus averaged across all trials for each genotype; circles represent individual fly averages. Only the response to sucrose differed significantly between genotypes (***p<0.001, two-way ANOVA followed by Bonferroni post-tests). Not all flies responded to acetic acid, but there was no difference between genotypes: 6 of 9 control and mutant flies responded to at least one concentration of acetic acid. n = 36–37 trials, nine flies per genotype.

Figure 5—figure supplement 4. Acetic acid activates a subset of bitter-sensing neurons.

(A–D) GCaMP responses of bitter neuron subsets were imaged using Gal4 lines that label each of the four subclasses of bitter neurons (S–a, S–b, I–a, I–b) with complete or partial specificity, as noted in each panel. Fed flies were stimulated with water, bitter compounds (1 mM lobeline, 10 mM caffeine), and acetic acid (AA, 1% or 5%) for 3 s (gray bars). Relatively weak GCaMP expression in neurons labeled by Gr47a-Gal4 may account for their lower response magnitudes. (E) Peak GCaMP responses of each of the four bitter neuron subsets shown in panels A-D. Asterisks indicate responses that are significantly higher than the response to water (**p<0.01, one-way ANOVA followed by Dunnett's post-tests). Only neurons labeled by Gr22f-Gal4 showed a significant peak response to either concentration of acetic acid. Neurons labeled by Gr98d-Gal4 did not show a significant difference in the peak response to acetic acid versus water, but the GCaMP traces for 1% acetic acid and water show clear differences in their timecourse (A), suggesting that these neurons may also respond to acetic acid. n = 12–15 trials, 4–5 flies per genotype.

Figure 5—figure supplement 5. Responses of sugar- and bitter-sensing neurons to additional acetic acid concentrations.

(A–B) GCaMP responses of sugar neurons (A) or bitter neurons (B) in two-day starved flies were tested with a range of acetic acid (AA) concentrations in ascending order (n = 24–32 trials, eight flies). Responses were analyzed across all trials (top graphs) or for all trials after excluding the first trial of each AA stimulus (bottom graphs), since the initial exposure to acetic acid often evoked a much stronger response than subsequent trials, and this may obscure dose-response relationships because acetic acid concentrations were always tested in ascending order. Excluding the first trials revealed a clearer dose dependence in sugar neurons at 0.01–1% AA but not in bitter neurons. In all panels, line graphs show average GCaMP traces (gray bars indicate stimulus delivery, 2 s) and bar graphs show average peak responses. Asterisks indicate responses that were significantly higher than the response to water (*p<0.05, **p<0.01, one-way ANOVA followed by Dunnett's post-tests). Sugar neurons were labeled with Gr64f-Gal4 and bitter neurons were labeled with Gr66a-Gal4. 100 mM sucrose or 1 mM denatonium (‘bitter’) was used as a positive control for sugar and bitter neurons, respectively.

Figure 5—figure supplement 6. Water-sensing neurons are activated by acetic acid only in accordance with its osmolarity.

(A–B) Water-sensing neurons labeled with ppk28-Gal4 showed GCaMP responses to various taste stimuli (gray bar, 2 s), including water, sucrose (suc), lobeline (lob), and acetic acid (AA). Responses decreased with increasing tastant molarity (B). (C–D) Adding PEG (3350 g/mol) to each tastant solution (10% wt/vol) strongly diminished the response to both water and acetic acid (p<0.001, two-way ANOVA). Responses to acetic acid in PEG were reduced to the same level as that of PEG alone (p>0.05, two-way ANOVA). **p<0.01, ***p<0.001, two-way ANOVA followed by Bonferroni's post-tests. n = 24–27 trials, eight flies.

Figure 5—figure supplement 7. GCaMP-expressing flies show hunger-dependent changes in PER to acetic acid and sucrose.

(A–D) Fed and starved flies expressing GCaMP6f in sugar-sensing neurons (A–B) or bitter-sensing neurons (C–D) were tested for PER to acetic acid (A, C) and sucrose (B, D). Both genotypes showed hunger-dependent changes in the level of PER to acetic acid as well as sucrose. Fed and starved flies were compared using two-way ANOVA (repeated-measures for panels A and C) followed by Bonferroni’s post-tests (*p<0.05, **p<0.01, ***p<0.001; orange or red asterisks correspond to one- or two-day starved flies, respectively). n = 3–4 sets of flies.