A molecular and neuronal basis for amino acid sensing in the Drosophila larva

Abstract

Amino acids are important nutrients for animals, reflected in conserved internal pathways in vertebrates and invertebrates for monitoring cellular levels of these compounds. In mammals, sensory cells and metabotropic glutamate receptor-related taste receptors that detect environmental sources of amino acids in food are also well-characterised. By contrast, it is unclear how insects perceive this class of molecules through peripheral chemosensory mechanisms. Here we investigate amino acid sensing in Drosophila melanogaster larvae, which feed ravenously to support their rapid growth. We show that larvae display diverse behaviours (attraction, aversion, neutral) towards different amino acids, which depend upon stimulus concentration. Some of these behaviours require IR76b, a member of the variant ionotropic glutamate receptor repertoire of invertebrate chemoreceptors. IR76b is broadly expressed in larval taste neurons, suggesting a role as a co-receptor. We identify a subpopulation of these neurons that displays physiological activation by some, but not all, amino acids, and which mediate suppression of feeding by high concentrations of at least a subset of these compounds. Our data reveal the first elements of a sophisticated neuronal and molecular substrate by which these animals detect and behave towards external sources of amino acids.

Figure 1. Drosophila larvae display innate, stimulus-specific responses to amino acids.

(A) Schematic of the larval gustatory choice assay. Groups of ~30 experimentally-naïve larvae are placed at the centre of a Petri dish divided in four quadrants, which contain control (agarose alone) or test (agarose + tastant) substrates. The Preference Index is calculated by counting the number of larvae on each substrate and dividing the difference by the total number of larvae; positive values reflect behavioural attraction towards a tastant-containing substrate. (B) Preference Indices of wild-type larvae (w¹¹¹⁸) for substrates containing individual L-amino acids (all at 50 mM, except for tyrosine, which was used at 2 mM because of its lower solubility). Essential amino acids are shown in blue. Grey boxes indicate amino acids that produce a Preference Index significantly different from 0 (t-test with Bonferroni correction, p < 0.0025). For boxplots in this and subsequent figures, the boxes represent the upper and lower quartiles, the thick line is the median value and the whiskers show the 5^th and 95^th percentiles. Circles outside the whiskers are outliers. (N ≥ 18 groups of ~30 larvae for each tastant). (C) Venn diagram (adapted from ref. 57) summarising the main physico-chemical properties of amino acids, indicated by single letter codes. Large black letters indicate those that are attractive to larvae; the sole aversive amino acid is shown in white.

Figure 2. IR76b, but not IR25a, is required for behavioural attraction to amino acids.

(A) Schematic of a larval head, illustrating the main olfactory (Dorsal Organ) and gustatory (Terminal Organ, Ventral Organ, Dorsal Pharyngeal Sense Organ, Ventral Pharyngeal Sense Organ, Posterior Pharyngeal Sense Organ) structures. (B) Immunofluorescence with anti-IR25a (magenta) and anti-GFP (green) on IR76b-Gal4;UAS-mCD8:GFP animals revealing expression in the indicated gustatory organs. The merged channels are overlaid on a brightfield image. Scale bars: 20 μm. (C) Preference Indices of wild type (w¹¹¹⁸), IR25a mutants (IR25a²) and IR25a rescue animals (IR25a², CH322-32C20) for the indicated amino acids (Kruskal-Wallis, p > 0.05; N ≥ 20 groups of ~30 larvae per genotype and per tastant). (D) Preference Indices of wild type (w¹¹¹⁸), IR76b heterozygous (IR76b¹/+; where “+” represents Canton S-derived chromosomes) and homozygous (IR76b¹) mutants, and IR76b rescue animals (IR76b-Gal4, UAS-IR76b;IR76b¹) for the indicated amino acids (Kruskal-Wallis, *p < 0.05, ***p < 0.001; N = 14 groups of ~30 larvae per genotype and per tastant).

Figure 3. IR76b-Gal4 neurons respond physiologically to amino acids.

(A) Schematic of the calcium imaging assay. A larval head is dissected, leaving the brain attached, and embedded in agarose. Calcium responses in the subesophageal zone (SEZ) are measured while a tastant solution is applied to the peripheral chemosensory organs. (B) Projections of IR76b-Gal4 neurons to the SEZ (IR76b-Gal4;UAS-mCD8:GFP). The neuropil marker nc82 is shown in grey. Scale bar: 20 μm; A: anterior, P: posterior. (C) Left: False-coloured representation of GCaMP3 fluorescence intensity in one hemisphere of the SEZ of a UAS-GCaMP3;IR76b-Gal4 animal, before and after presentation of 50 mM arginine. Scale bar: 20 μm; L: lateral, M: medial. Right: example time course of the calcium response of IR76b-Gal4 neurons to arginine presentation (arrowhead) from the region of interest indicated by a dashed circle on the left panel. (D) Peak ΔF/F value observed in IR76b-Gal4 neurons after stimulation with 19 individual amino acids (concentration: 50 mM) or water solvent. (Mean response ± SEM is shown; N ≥ 4 animals per stimulus). The valence of the behavioural responses towards amino acids (from Fig. 1B) is indicated by the colour of lettering.

Figure 4. Identification of a single neuronal population responsive to amino acids.

(A) Expression patterns of nine IR-Gal4 lines (visualised with a UAS-mCD8:GFP reporter) that label subsets of gustatory neurons in the Terminal Organ (genotypes of the form: UAS-mCD8:GFP;IRxx-Gal4). Left: soma and dendrites of IR-Gal4 neurons in the Terminal Organ (anti-GFP, green, overlaid on a brightfield image). Right: projection pattern of IR-Gal4 neuron axons in the SEZ (anti-GFP, green), overlaid on the neuropil marker nc82 (magenta). Scale bars: 20 μm. (B) Peak ΔF/F values of calcium responses measured in the indicated IR-Gal4 neurons (genotypes of the form: UAS-GCaMP3;IRxx-Gal4) to presentation of glutamine (200 mM; preliminary observations prompted our use of higher concentrations than in Fig. 3, to ensure detection of responses in the sparse innervations of individual populations of IR-Gal4 neurons). IR7b-Gal4 and IR7d-Gal4 were excluded because the basal GCaMP3 fluorescence with these drivers was too low to perform imaging experiments. (Mean response ± SEM is shown; N ≥ 5 animals per stimulus). (C) Top: False-coloured representation of GCaMP3 fluorescence intensity in one hemisphere of the SEZ of a UAS-GCaMP3;IR60c-Gal4 animal, before and after presentation of 200 mM glutamine. Scale bar: 20 μm. Bottom: example time course of the calcium response of IR60c-Gal4 neurons to glutamine presentation (arrowhead) from the region of interest indicated by a dashed circle on the upper panels.

Figure 5. IR60c-Gal4 neurons are sensitive to a subset of amino acids.

(A) Peak ΔF/F value observed in neurons of UAS-GCaMP3;IR60c-Gal4 animals after stimulation with 19 individual L-amino acids (200 mM). (Mean response ± SEM is shown; N ≥ 5 animals per stimulus). (B) As in (A) after stimulation with 10 individual D-amino acids (200 mM). (Mean response ± SEM is shown; N ≥ 4 animals per stimulus). (C) As in (A) after stimulation with water, caffeine (50 mM) or sugars (200 mM for maltose, glucose and sucrose; 1 M for fructose). (Mean response ± SEM is shown; N ≥ 3 animals per stimulus).

Figure 6. IR60c-Gal4 neurons are not required for behavioural attraction to amino acids.

Preference Indices for four appetitive amino acids (50 mM) of larvae in which IR60c-Gal4 neurons are silenced with Kir2.1 (UAS-Kir2.1:GFP/+;IR60c-Gal4/+), together with genetic controls (UAS-Kir2.1:GFP/+ and IR60c-Gal4/+; “+” represents Canton S-derived chromosomes). (Kruskal-Wallis, p > 0.05; N ≥ 39 groups of ~15 larvae per genotype and per tastant).

Figure 7. IR60c-Gal4 neurons and IR76b mediate feeding suppression by amino acids.

(A) Schematic of the feeding suppression assay. (B) Absorbance (at 625 nm) of extracts from homogenised UAS-VR1/+;IR60c-Gal4/+ larvae and controls (UAS-VR1/+ and IR60c-Gal4/+; “+” represents Canton S-derived chromosomes) after animals have fed on 50 mM sucrose containing 100 μM capsaicin and 0.4% Brilliant Blue for 30 minutes. (Kruskal-Wallis, ***p < 0.001; N = 11 groups of 30 larvae per genotype). (C,D) Absorbance of extracts from homogenised UAS-Shibire^ts1/+;;UAS-Shibire^ts1/IR60c-Gal4 larvae and controls (UAS-Shibire^ts1/+;;UAS-Shibire^ts1/+ and IR60c-Gal4/+; “+” represents Canton S-derived chromosomes) after animals have fed on 50 mM sucrose containing 0.4% Brilliant Blue for 30 minutes at the restrictive (C) or permissive (D) temperature. (Kruskal-Wallis, ***p < 0.001; N = 6 groups of 30 larvae per genotype). (E) Absorbance of extracts from homogenised larvae of the same genotypes as in (C) after animals have fed on 50 mM sucrose containing 0.4% Brilliant Blue for 30 minutes at the restrictive temperature. (Kruskal-Wallis, p > 0.05; N ≥ 12 groups of 30 larvae per genotype). (F–G) Absorbance of extracts from homogenised larvae of the same genotypes as in (C) after animals have fed on 50 mM sucrose containing 500 mM alanine and 0.4% Brilliant Blue for 30 minutes at the restrictive (F) or permissive (G) temperature. (Kruskal-Wallis, ***p < 0.001; N ≥ 12 groups of 30 larvae per genotype). (H–I) Absorbance of extracts from homogenised larvae of the same genotypes as in (C) after animals have fed on 50 mM sucrose containing 500 mM lysine (H) or 500 mM glycine (I) and 0.4% Brilliant Blue for 30 minutes at the restrictive temperature. (Kruskal-Wallis, ***p < 0.001; N ≥ 12 groups of 30 larvae per genotype). (J) Preference Indices of larvae in which IR60c-Gal4 neurons are silenced with Kir2.1 (UAS-Kir2.1:GFP/+;IR60c-Gal4/+), together with controls (UAS-Kir2.1:GFP/+ and IR60c-Gal4/+; “+” represents Canton S-derived chromosomes) for 50 mM sucrose versus 50 mM sucrose +500 mM alanine. (Wilcoxon Signed Rank Test, **p < 0.01, ***p < 0.001; N ≥ 69 groups of ~15 animals per genotype). (K) Absorbance of extracts from homogenised wild-type (w¹¹¹⁸), homozygous IR76b¹ mutant and heterozygous IR76b¹/+ (“+” represents w¹¹¹⁸ derived chromosomes) larvae after animals have fed on 50 mM sucrose containing 500 mM alanine and 0.4% Brilliant Blue for 30 minutes. (Kruskal-Wallis, ***p < 0.001; N = 16 groups of 30 larvae per genotype).

Figure 8. IR60c-Gal4 neurons define a functionally and anatomically distinct aversion pathway.

(A) Absorbance (at 625 nm) of extracts from homogenised UAS-Shibire^ts1/+;;UAS-Shibire^ts1/IR60c-Gal4 larvae and genetic controls (UAS-Shibire^ts1/+;;UAS-Shibire^ts1/+ and IR60c-Gal4/+; “+” represents Canton S-derived chromosomes) after animals have fed on 50 mM sucrose containing 50 mM caffeine and 0.4% Brilliant Blue FCF for 30 minutes. (Kruskal-Wallis, p > 0.05; N ≥ 12 groups of 30 larvae per genotype). (B) Immunofluorescence with anti-GFP (green) and anti-CD8 (magenta) on GR66a-LexA/+;LexAOP-rCD2:GFP/UAS-mCD8;IR60c-Gal4/+ animals revealing that these aversive pathways define distinct populations of neurons in the Terminal Organ (top) but have partially overlapping axonal innervations in the SEZ (bottom). Grey background shows the brightfield image on the top panel and the neuropil marker nc82 on the bottom panel. Scale bars: 20 μm.