Odour receptors and neurons for DEET and new insect repellents

Abstract

There are major impediments to finding improved DEET alternatives because the receptors causing olfactory repellency are unknown, and new chemicals require exorbitant costs to determine safety for human use. Here we identify DEET-sensitive neurons in a pit-like structure in the Drosophila melanogaster antenna called the sacculus. They express a highly conserved receptor, Ir40a, and flies in which these neurons are silenced or Ir40a is knocked down lose avoidance to DEET. We used a computational structure-activity screen of >400,000 compounds that identified >100 natural compounds as candidate repellents. We tested several and found that most activate Ir40a(+) neurons and are repellents for Drosophila. These compounds are also strong repellents for mosquitoes. The candidates contain chemicals that do not dissolve plastic, are affordable and smell mildly like grapes, with three considered safe in human foods. Our findings pave the way to discover new generations of repellents that will help fight deadly insect-borne diseases worldwide.

Figure 1. DEET is detected by Ir40a+ sacculus neurons

a, Schematic of the NFAT (CaLexA)-based method to label neurons activated by DEET. b, Confocal micrographs of olfactory organs from flies stimulated with 10% DEET or solvent (acetone). c, Quantification of GFP+ antennae (Top) and mean numbers of GFP+ cells in chamber I. n=35 (blank), n=30 (solvent), n=20 (10%DEET), n=20 (100%DEET). P < 0.0001,1-way ANOVA with Tukey's post hoc test. d, GFP+ axonal termini in antennal lobes of flies treated as indicated. e,f Expression of GFP+ in the labellum, labral sense organ (LSO), the sub-esophageal ganglion (SOG). Anti-GFP (green) and anti-nc82 (red). For SOG, dorsal is top.

Figure 2. Ir40a neurons detect DEET and are required for repellency

a, Images of calcium activity in Ir40a-Gal4/+;UAS-GCaMP3/+ neurons color-coded as indicated (right). Measurements taken from areas in dashed circles: cells (white), background (red). b, Mean fluorescence intensities for 6 different cells. Red arrowhead indicates onset of ∼2-sec puff of DEET. c, Mean percentage change in fluorescence intensity after application of ∼2-sec indicated stimulus; genotypes were Ir40aGal4/+;UAS-GCaMP3/+ (control) and Ir40aGal4/Ir40aGal4;UAS-GCaMP3/UAS-Ir40aRNAi(2) (Ir40a-RNAi). n=10-13. **P < 0.01, Student's t-test. d, Schematic (left) and results (right) for DEET-treated trap assays for indicated genotypes. n=6 trials, 20 flies/trial for each genotype. Letters indicate statistical significance, P ≤ 0.008, 1-way ANOVA with Tukey's post hoc analysis. Error bars=S.E.M.

Figure 3. Ir40a is required for DEET avoidance

a, Set-up for behavioural 2-choice assay. b,c Mean preference index of indicated genotypes for DEET in 2-choice assays using b, elav-Gal4, and c, Ir40a-Gal4. n=6 trails (20 flies/trial) except elav-Gal4/+;Ir40aRNAi(2) n=10 trials and RNAi experiments with Ir40a-Gal4 n=12 trials each. d, Genotype and schematic for post-developmental knockdown and recovery of Ir40a. e, Mean DEET preference index of flies derived from indicated treatments in 2-choice assays. n=6 trials for all conditions, with 20 flies/trial. For b-e, P < 0.001, 1-way ANOVA with Tukey's post hoc analysis. Error bars=S.E.M.

Figure 4. Chemical informatics prediction of new repellents

a, Cheminformatics discovery pipeline to identify novel DEET-like repellents. b, Hierarchical cluster analysis of 201 training set odorants using optimized descriptors to calculate distances in chemical space. c, Receiver-operating-characteristic curve (ROC) representing computational validation of repellent predictive ability from 20 independent 5-fold cross validations. AUC=Area under the curve. d, DEET, Picaridin, and two unapproved repellents. e, Representative predicted repellents from >400,000 odorant library (Left) and computationally determined values for 1000 top-ranked predicted repellents (Right). f, Representative predicted repellents from >3,000 natural odour library (Left) and computationally determined values for 150 top-ranked predicted repellents (Right). Colour arrowheads indicate values for DEET and selected odours shown in g.

Figure 5. Predicted repellents activate Ir40a neurons and are strong repellents for Drosophila

a, Images of antenna of elav-Gal4/LexAop-CD8-GFP-2A-CD8-GFP; UAS-mLexA-VP16- NFAT, LexAop-CD2-GFP/+ flies exposed to indicated stimuli for 24 hrs. b, BA-activated GFP+ neurons in indicated tissues. c, Mean changes in fluorescence intensity in Ir40aGal4/+;UAS-GCaMP3/+ cells after ∼2-sec application of indicated odorants. n=9-17. d, Mean responses of flies to predicted repellents in 2-choice olfactory and gustatory trap assays measured at 24 and 48 hrs. n=3-10 trials (24 hours) and 7-10 (48 hrs); 10 flies/trial, trials with <40% participation were excluded. e, Quantification of flies of indicated genotypes entering repellent-treated traps. n=6 trials for each genotype, ∼20 flies for each trial. P < 0.001, 1-way ANOVA with Tukey's post hoc test. For c-e, error bars=S.E.M.

Figure 6. A new class of mosquito repellents with desirable safety profiles

a, Arm-in-cage assay to measure repellency in mosquitoes. b, Mean percentage of female A. aegypti present for >5 sec on top net (Left=10% DEET, Right=solvent). Solvent controls performed separately (dark gray). c, Average time on net for each landing event in b. d, Mean percentage of female A. aegypti present for >5 sec on top net in non-contact assay. e, Cumulative repellency summed across minutes 2-5 of indicated non-contact treatment (10%) in comparison to appropriate solvent control. 40 mosquitoes/trial, n=5 trials/treatment for b,c,d, and e. f, Mean weight of vinyl pieces following submersion in indicated compounds for indicated amount of time. n=3, ***P < 10^-5, Student's t-test. Error bars=S.E.M. g, Properties of new repellents. h, Model for DEET detection and processing in Drosophila.