Abstract
DEET (N,N-diethyl-meta-toluamide) is a synthetic chemical, identified by the United States Department of Agriculture in 1946 in a screen for repellents to protect soldiers from mosquito-borne diseases1,2. Since its discovery, DEET has become the world’s most widely used arthropod repellent3, and is effective against many invertebrates, including biting flies4, honeybees5, ticks6, and land leeches4,7. In insects, DEET acts on the olfactory system5,8-14 and requires the olfactory receptor co-receptor orco9,11-13, but its specific mechanism of action remains controversial. Here we show that the nematode Caenorhabditis elegans is sensitive to DEET, and use this genetically-tractable animal to test repellent hypotheses from insects to understand how this synthetic compound is able to affect the behaviour of invertebrates separated by millions of years of evolution. We found that DEET is not a volatile repellent, but interfered selectively with chemotaxis to a variety of attractant and repellent molecules, and induced pausing to disrupt chemotaxis to some odours but not others. In a forward genetic screen for DEET-resistant animals, we identified a single G-protein-coupled receptor, str-217, which is expressed in a single pair of DEET-responsive chemosensory neurons, ADL. Both engineered str-217 mutants and a wild isolate of C. elegans carrying a deletion in str-217 are DEET-resistant. DEET interferes with behaviour in an odour-selective manner by inducing an increase in average pause length during chemotaxis and exploration, and this increase in pausing requires both str-217 and ADL neurons. Finally, we found that ADL neurons are activated by DEET and that optogenetic activation of ADL increased average pause length. This is consistent with the “confusant” hypothesis, in which DEET is not a simple repellent but modulates multiple olfactory pathways to scramble the behavioural response to otherwise attractive stimuli12,13. Our results suggest a consistent motif for the effectiveness of DEET across widely divergent taxa: an effect on multiple chemosensory neurons to disrupt the pairing between odorant stimulus and behavioural response.
We used standard chemotaxis assays15-17 (Fig. 1a) to explore whether and how C. elegans nematodes respond to DEET. There are currently three competing hypotheses about the mechanism of DEET based on work in insects: “Smell-and-repel” —DEET is detected by olfactory pathways that trigger avoidance5,10,14,18, “masking” — DEET selectively blocks olfactory pathways that mediate attraction8-10, and “confusant” —DEET modulates multiple olfactory sensory neurons to scramble the perception of an otherwise attractive stimulus12,13. Inspired by these hypotheses, we tested how DEET may interfere with olfactory behaviours in nematodes to identify similarities and differences with work in insects.
To test the smell-and-repel hypothesis, we presented DEET as a volatile point source. DEET was not repellent alone (Fig. 1b), similar to previous results in Drosophila melanogaster flies9 and Aedesaegypti mosquitoes13. To address the possibility that DEET could be masking responses to attractive odorants8,9 or directly inhibiting their volatility10, we presented DEET alongside the attractant isoamyl alcohol, both as point sources, and found that it had no effect on attraction (Fig. 1c). In considering alternate ways to present DEET, we mixed low doses of DEET uniformly into the chemotaxis agar and presented isoamyl alcohol as a point source (Fig. 1d).In this configuration, DEET-agar reduced chemotaxis to isoamyl alcohol in a dose-dependent manner (Fig. 1e). To ask if DEET has a general effect on chemotaxis, we tested two additional attractants, butanone and pyrazine, as well as the volatile repellent 2-nonanone. Behavioural responses to butanone requires the same pair of primary sensory neurons as isoamyl alcohol (AWC), while pyrazine and 2-nonanone require two different pairs of primary sensory neurons (AWA and AWB, respectively)16,19. DEET eliminated both attraction to butanone and avoidance of 2-nonane, indicating that it can affect responses to both positive and negative chemosensory stimuli (Fig. 1f). In contrast, DEET-agar had no effect on chemotaxis toward the attractant pyrazine (Fig. 1f). We conclude that DEET chemosensory interference is odour-selective, can affect both attractive and repulsive stimuli, and is not a result of non-specific or toxic effects of DEET.
Identifying genes required for DEET-sensation has been of interest for some time. A forward genetic approach in Drosophila melanogaster flies yielded an X-linked DEET-insensitive mutant20 and a population genetics approach in mosquitoes identified a dominant genetic basis for DEET-insensitivity21, but neither study identified the genes underlying these changes. Reverse genetic experiments in Drosophila flies and three mosquito species have identified the insect odorant receptors as a molecular target of DEET9,11-14. However, this chemosensory gene family is not found outside of insects22,23, raising the question of what pathways are required for DEET-sensitivity in non-insect invertebrates. To gain insights into the mechanisms of DEET repellency in C. elegans, we carried out a forward genetic screen for mutants capable of chemotaxing toward isoamyl alcohol on DEET-agar plates (Fig. 1g). We obtained 5 DEET-resistant animals, three of which produced offspring that consistently chemotaxed toward isoamyl alcohol on DEET-agar plates (Fig. 1h, and data not shown). Whole genome sequencing allowed us to identify candidate causal mutations in these 3 strains24, and we chose to focus on LBV003, which maps to str-217, a predicted G protein-coupled receptor (Fig. 2a). In the course of mapping str-217, we discovered that a divergent strain of C. elegans isolated in Hawaii, CB4856 (Hawaiian), is naturally resistant to DEET (Fig. 2b). This Hawaiian strain contains a 138-base pair deletion in str-217 (str-217HW) that affects exons 2 and 3 and an intervening intron, leading to a mutant strain with a predicted frame shift indel and early stop codon (Fig. 2a). To confirm that DEET resistance maps to str-217HW, we tested three near-isogenic lines with a single, homozygous genomic segment of Hawaiian chromosome V introgressed into a wild-type (Bristol N2) background25 (Fig. 2b). Only the ewIR74 line contains str-217HW and, like the parent Hawaiian strain, is DEET-resistant (Fig. 2b). To provide further confirmation that str-217 is required for DEET sensitivity in these strains, we generated two additional genetic tools: an engineered predicted null mutant produced by CRISPR-Cas9 genome-editing (str-217-/-) (Fig. 2a), and a rescue/reporter plasmid that expresses both wild-type str-217 and green fluorescent protein (GFP) under control of the predicted str-217 promoter (Fig. 2c). The LBV003 strain (Fig. 2d), Hawaiian introgressed strain ewIR74 (Fig. 2e), and the str-217-/- engineered mutant strain (Fig. 2f) all showed chemotaxis on DEET-agar (Fig. 2d-f). Expression of the str-217 rescue/ reporter construct in these three strains rendered all three DEET-resistant mutants fully sensitive to DEET, in that none chemotaxed to isoamyl alcohol on DEET-agar (Fig. 2d-f).
We next turned to the neuronal mechanism by which DEET disrupts chemotaxis in C. elegans. In insects, DEET interacts directly with chemosensory neurons and the odorant receptors that they express9,11-14. Isoamyl alcohol is primarily sensed by the AWC chemosensory neuron19. To ask if DEET modulates primary sensory detection of isoamyl alcohol, we used in vivo calcium imaging to monitor AWC activity in the presence and absence of DEET. AWC responded to the addition of DEET with a rapid increase in calcium that decreased to baseline over the course of 11 min of chronic DEET stimulation (Extended Data Fig. 1a). In the presence of DEET, AWC responses to isoamyl alcohol decreased in magnitude, but there was no observed difference in AWC activity between wild-type and str-217-/- mutants in the presence or absence of DEET (Extended Data Fig. 1b-c). This suggests that AWC sensory neurons are not the primary functional target of DEET.
To identify the functionally relevant neurons, we determined where str-217 is expressed by examining the str-217 rescue/reporter strain, and found GFP expression in a single pair of chemosensory neurons, called ADL (Fig. 3a). ADL is not required for chemotaxis to isoamyl alcohol, suggesting an indirect role for ADL in DEET chemosensory interference26. To ask if ADL neuronal function is required for DEET-sensitivity, we expressed tetanus toxin light chain, which inhibits chemical synaptic transmission by cleaving the synaptic vesicle protein synaptobrevin, in ADL27,28. These animals showed the same level of DEET-resistance as str-217 mutants (Fig. 3b). Since both str-217 and ADL function are required for DEET-sensitivity, we used calcium imaging to ask if ADL responds to DEET, and if this requires str-217 (Fig. 3c). Both wild-type and str-217-/- mutants carrying the rescue/reporter plasmid, but not str-217-/- mutants, showed calcium responses to DEET (Fig. 3d-e,h). In control experiments, we showed that the known ADL agonist, the pheromone C9 (ref. 27), activated ADL in both wild-type and str-217-/- mutant animals (Fig. 3f-h) This suggests that the str-217-/- mutation has a selective effect on ADL function. From these data, we conclude that disrupting either ADL activity or str-217 is sufficient to confer DEET-resistance in C. elegans. We note that neither str-217 mutants nor ADL-deficient animals return fully to wild-type levels of chemotaxis (Fig. 3b), suggesting that additional genes and neurons contribute to DEET sensitivity in C. elegans. To ask if str-217 is a direct molecular target of DEET, we carried out calcium imaging experiments in HEK293T cells. DEET did not activate HEK293T cells expressing str-217 (Extended Data Fig. 2), although we cannot exclude the possibility that this nematode receptor is non-functional in mammalian tissue culture cells. It is also possible that str-217 is not the direct in vivo target of DEET, but is involved indirectly in signalling or modulation of DEET-specific responses in ADL.
We next explored how ADL activity can interfere with chemotaxis. Population chemotaxis assays report the location of the animal at the end of the experiment, and do not reveal the details of navigation strategy. To investigate which aspects of chemotaxis are affected by DEET, we tracked the position and posture of individual animals on DEET-agar or solvent-agar plates (Fig. 4a-g). Wildtype, but not str-217-/- mutant animals (Fig. 4d), showed a dramatic increase in average pause length on DEET-agar.
To determine if the increase in average pause length occurs only in the context of chemotaxis to isoamyl alcohol, or as a consequence of DEET alone, we tracked wild-type, str-217-/- mutant (Fig. 4e), and ADL::Tetanus toxin (Fig. 4f) animals on DEET-agar and solvent-agar plates with no additional odorants. Only wild-type animals had a higher average pause length on DEET-agar (Fig. 4e-f). Average pause length was unaffected by DEET in str-217-/- and ADL::Tetanus toxin animals (Fig. 4e-f). Consistent with our prior observation that chemotaxis to pyrazine was unaffected by DEET, wild-type animals showed no increase in average pause length when chemotaxing to pyrazine on DEET-agar (Fig. 4g). This suggests the interesting possibility that pyrazine chemotaxis is not only str-217- and ADL-independent, but can overcome the effect of DEET on average pause length. To test if ADL activity alone is sufficient to increase average pause length, we carried out an optogenetic experiment by expressing the light-sensitive ion channel ReaChR29 in ADL in wild-type animals, and tracking locomotor behaviour on chemotaxis plates. We observed an increase in average pause length when ADL was activated (Fig. 4h-i). From these data, we conclude that ADL mediates the increase in average pause length seen on DEET-agar, and speculate that the increase in long pauses is one mechanism by which DEET interferes with chemotaxis.
In this study, we add the nematode C. elegans to known DEET-sensitive animals, uncover a new neuronal mechanism for a DEET-induced behaviour, and identify a molecular target that is required for complete DEET-sensitivity in an engineered mutant and in a wild-isolate of C. elegans. This work opens up C. elegans as a system to test new repellents in vivo for both interference in chemotaxis and toxicity, and also for discover of additional genes and neurons that respond to DEET. The molecular mechanism by which the str-217 mutation renders ADL DEET-insensitive and worms DEET-resistant remains to be understood. str-217 is a G protein-coupled receptor with no known ligand and that is evolutionarily unrelated to DEET-sensitive odorant receptor proteins previously described in insects. Although we found no evidence that DEET can activate str-217 in heterologous cells, it is conceivable that in the right milieu that str-217 is indeed a DEET receptor. Alternatively, str-217 could act indirectly in concert with an as-yet unknown DEET receptor in ADL. Interestingly, pyrazine chemotaxis is unaffected by DEET in any of our assays, consistent with our model that DEET is not a simple repellent, but a modulator of pausing behaviour to interfere with chemotaxis to some but not all odorants.
These results are reminiscent of the “ confusant” hypothesis in insects, although the molecular and neuronal details by which DEET acts differ markedly between species. In insects, DEET alters responses of individual sensory neurons to attractive odorants9,12, thereby interfering with behavioural attraction. Our data in C. elegans are consistent with a mechanism where DEET can inhibit responses to some stimuli but not others by decreasing avoidance of 2-nonanone, decreasing attractiveness to multiple odorants, and leaving pyrazine behavioural responses intact. We speculate that its promiscuity in interacting with multiple molecules and chemosensory neurons across vast evolutionary scales is the key to the broad effectiveness of DEET.
Methods
Nematode culture and strains
C. elegans strains were maintained at room temperature (22-24°C) on nematode growth medium (NGM) plates (51.3 mM NaCl, 1.7% agar, 0.25% peptone, 1 mM CaCl2, 12.9µM cholesterol, 1mM MgSO4, 25mM KPO4, pH 6) seeded with Escherichia coli (OP50 strain) bacteria as a food source30,31. Bristol N2 was used as the wild-type strain. The CB4856 (Hawaiian) strain, harbouring WBVar02076179 (str-217HW) (http://www.wormbase. org/db/get?name=WBVar02076179;class=variation) and Hawaiian recombinant inbred strains for chromosome V were previously generated25. Generation of extra-chromosomal array transgenes was carried out using standard procedures32, and included the transgene injected at 50 ng/mL, the fluorescent co-injection marker Pelt-2::GFP at 5 ng/ml (with the exception of LBV004 and LBV009, which did not include a co-injection marker), and an empty vector for a total DNA concentration of 100 ng/ml. CRISPR-Cas9-mediated mutagenesis of str-217 was performed as described, using rol-6 as a co-CRISPR marker33. The resulting str-217 mutant strain [LBV004 str-217(ejd001)] results in a predicted frame-shift in the first exon [indel: insertion (AAAAAAA), deletion (CTGCTCCA), final sequence GCGTCGAAAAAAAATTTCAG; insertion is underlined]. The str-217 rescue construct (Pstr-217::str-217::SL2::GFP) used a 1112 nucleotide length fragment 56 nucleotides upstream 5’ of the translation start of str-217.
Microscopy and image analysis
L2-adult stage hermaphrodites were mounted on 1% agarose pads with 10 mM sodium azide (CID 6331859, Sigma-Aldrich, catalogue #S2002) in M9 solution (22 mM KH2PO4, 42mM Na2HPO4, 85.6 mM NaCl, 1µM MgSO4, pH 6). Images were acquired with an Axio Observer Z1 LSM 780 with Apotome a 63X objective (Zeiss), and were processed using ImageJ.
Chemotaxis assays
Chemotaxis was tested as described17, on square plates containing 10 mL of chemotaxis agar (1.6% agar in chemotaxis buffer: 5 mM phosphate buffer pH 6.0, 1mM CaCl, 1mM MgSO)34.logue #360473), 10mg/µL pyrazine (CID: 9261, Sig-ma-Aldrich, catalogue #W401501), 1:10 2-nonanone (CID: 13187, Sigma-Aldrich, catalogue #W2787513). Assays were carried out for 60-90 min at room temperature (22-24°C) between 1pm – 8pm EST. Plates were scored as soon as possible, either immediately or, if a large number of plates was being scored on the same day, plates were moved to 4°C to immobilize animals until they could be scored. The assay was quantified by counting animals that had left the origin in the centre of the plate, moving to either side of the plate (#Odorant,Additions of either ethanol (solvent-agar) or 50% DEET (CID: 4284, Sigma-Aldrich, catalogue #D100951) in ethanol (DEET-agar) were added after agar cooled to <44°C and just before pouring. A total volume of 300 µL ethanol or DEET in ethanol was added to each 100 mL of agar mixture for all experiments except Figure 1b-c and Figure 4h-i. Plates were poured on the day of each experiment, and dried with lids off for 4 hours prior to the start of the assay. 1 ml 1 M sodium azide was added to two spots on either side of the plate just before beginning the experiment to immobilize animals that reached the odorant or ethanol sources. Three days prior to all chemotaxis experiments, 4-6 L4 animals were transferred onto NGM plates seeded with E. coli (OP50 strain). The offspring of these 4-6 animals were then washed off of the plates and washed twice with 24S-Basal buffer (1 mM NaCl, 5.74 mM K2HPO4, 7.35 mM KH PO, 5 µg/mL cholesterol at pH 6-6.2)30 to remove younger animals, and once with chemotaxis buffer. Immediately before the start of the experiment, two 1 l drops of odorant diluted in ethanol, or ethanol solvent control, were spotted on each side of the plate on top of the sodium azide spots. 100-300 animals were then placed into the centre of the plate in a small bubble of liquid. The excess liquid surrounding the animals was then removed using a Kimwipe. Odorants diluted in ethanol were used in this study: 1:1000 isoamyl alcohol (CID: 31260, Sigma-Aldrich, catalogue #W205702), 1:1000 butanone (CID: 6569, Sigma-Aldrich, catalogue #360473), 10mg/μL pyrazine (CID: 9261, Sigma-Aldrich, catalogue #W401501), 1:10 2-nonanone (CID: 13187, Sigma-Aldrich, catalogue #W2787513). Assays were carried out for 60-90 min at room temperature (22-24°C) between 1pm . 8pm EST. Plates were scored as soon as possible, either immediately or, if a large number of plates was being scored on the same day, plates were moved to 4°C to immobilize animals until they could be scored. The assay was quantified by counting animals that had left the origin in the centre of the plate, moving to either side of the plate (#Odorant, #Control) or just above or below the origin (#Other), and calculating a chemotaxis index as [#Odorant -#Control]/ [#Odorant + #Control + #Other]. A trial was discarded if fewer than 50 animals or more than 250 animals contributed to the chemotaxis index and participated in the assay.
Mutant screen
About 100 wild-type (Bristol N2) L4 animals were mutagenized in M9 solution with 50 mM ethyl methanesulfonate (CID: 6113, Sigma-Aldrich, catalogue #M0880) for 4 hours with rotation at room temperature. Mutagenized animals were picked to separate 9 cm NGM agar plates seeded with E. coli (OP50 strain) and cultivated at 20°C. ∼5,000 F2 animals were screened for DEET resistance on 20.3 cm casserole dishes (ASIN B000LNS4NQ, model number 81932OBL11). Five animals across three assays were more than ∼2 cm closer to the odour source than the rest of the animals on the plate and were defined as DEET-resistant. This phenotype was heritable in three strains, and each strain was backcrossed to OS1917 for 4 generations. Whole-genome sequencing was used to map the mutations to regions containing transversions presumably introduced by the EMS, parental alleles of the N2 strain used for mutagenesis, and missing alleles of the wild-type strain OS1917 used for backcrossing35,36. LBV003 mapped to a 5 Mb region on chromosome V, that was further mapped to str-217. LBV002 mapped to a 6.8 Mb region on chromosome V, which was further narrowed down to a likely candidate gene, nstp-3(ejd002). In LBV002, nstp-3(ejd002) contains a T>G transversion of the 141st nucleotide in the CDS, which is predicted to produce a Phe48Val substitution in this sugar: proton symporter. We were unable to map the DEET-resistant mutation(s) in LBV001.
str-217 heterologous expression
HEK-293T cells were maintained using standard protocols in a Thermo Scientific FORMA Series II water-jacketed CO2 incubator. Cells were transiently transfected with 1 μg each of pME18s plasmid expressing GCaMP6s, Gqα15, and str-217 using Lipofectamine 2000 (CID: 100984821, Invitrogen, catalogue #1168019). Control cells excluded str-217, but were transfected with the other two plasmids. Transfected cells were seeded into 384 well plates at a density of 2 x 106 cells/ml, and incubated overnight in FluoroBrite DMEM media (ThermoFisher Scientific) supplemented with foetal bovine serum (Invitrogen, catalogue #10082139) at 37°C and 5% CO2. Cells were imaged in reading buffer [Hanks’s Balanced Salt Solution (GIBCO) + 20 mM HEPES (Sigma-Aldrich)] using GFP-channel fluorescence of a Hamamatsu FDSS-6000 kinetic plate reader at The Rockefeller University High-Throughput Screening Resource Centre. DEET was prepared at 3X final concentration in reading buffer in a 384-well plate (Greiner Bio-one) from a 46% (2 M) stock solution in DMSO (Sigma-Aldrich). Plates were imaged every 1 sec for 5 min. 10 μl of DEET solution in reading buffer or vehicle (reading buffer + DMSO) was added to each well containing cells in 20 μl of media after 30 sec of baseline fluorescence recording. The final concentration of vehicle DMSO was matched to the DEET additions, with a maximum DMSO concentration of 7.8%. Fluorescence was normalized to baseline, and responses were calculated as max ratio (maximum fluorescence level/baseline fluorescence level).
ADL calcium imaging
Calcium imaging and data analysis were performed as described37, using single young adult hermaphrodites immobilized in a custom-fabricated 3 x 3 x 3 mm polydimethylsiloxane (PDMS) imaging chip. GCaMP5k was expressed in ADL neurons under control of the sre-1 promoter27 and was crossed into str-217-/and the str-217-/-rescue strain. Animals were acclimated to the imaging room overnight on E.coli (OP50 strain) seeded plates. All stimuli were prepared the day of each experiment, and were diluted in ethanol to 1000X the desired concentration before being further diluted 1:1000 in S-Basal buffer. Young adult animals were paralyzed briefly in (-)-tetramisole hydrochloride (CID: 27944, Sigma-Aldrich, catalogue #L9756) at 1 mM for 2-5 min before transfer into the chip to paralyze body wall muscles to keep animals stationary during imaging. All animals were pre-exposed to light (470+/-40nm) for 100 sec before recording to attenuate the light response of ADL38. Experiments consisted of the following stimulation protocol: 20 sec S-Basal buffer, followed by 3 repetitions of 20 sec DEET (0.15% DEET and 0.15% ethanol in S-Basal) and then 20 sec S-basal buffer.
GCaMP signals were recorded with Metamorph Software (Molecular Devices) and an iXon3 DU-897 EMCCD camera (Andor) at 10 frames/sec using a 40x objective on an upright Zeiss Axioskop 2 microscope. Custom ImageJ scripts17 were used to track cells and quantify fluorescence. In Figure 3d and f, all frames in 20 sec before the DEET pulse were averaged and subtracted from the average of the frames during the 20 sec DEET or C9 pulse to calculate δF. In Figure 3e and g, traces were bleach corrected using a custom MATLAB script and then the 5% of frames with the lowest values were averaged to create F0. δF/F0 was calculated by (F – F0)/F0 and then divided by the maximum value to obtain. δF/Fmax39. The heatmap traces in Figure 3e and g were also smoothed by 5 frames, such that each data point n is the running average of n-2, n-1, n, n+1, and n+2.
AWC calcium imaging
Calcium imaging of freely moving worms and subsequent data analysis were performed as described39, using a 3mm2 microfluidic PDMS device with two arenas that enabled simultaneous imaging of two genotypes with approximately 10 animals each. We used an integrated line (CX17256) expressing GCaMP5a in AWCON neurons under control of the str-2 promoter crossed into str-217-/- animals. Adult hermaphrodites were first paralyzed for 80-100 min in 1 mM (-)-tetramisole hydrochloride and then transferred to the arenas in S-Basal buffer. The stimulus protocol was as follows: In S-Basal, three pulses of 60 sec in buffer and 30 sec isoamyl alcohol, followed by 120 sec in buffer. Next, the animals were switched to S-Basal with 0.15% ethanol (solvent buffer) and three pulses of 60 sec in buffer and 30 sec in isoamyl alcohol in solvent buffer followed by 120 sec in solvent buffer before a switch to S-Basal with 0.15% ethanol and 0.15% DEET (DEET buffer). In DEET buffer, animals were given 6 pulses of 60 sec in DEET buffer and then 30 sec in isoamyl alcohol in DEET buffer, followed by 120 sec in DEET buffer before switching to solvent buffer. In solvent buffer, the animals received three pulses of 60 sec in buffer and 30 sec in isoamyl alcohol in solvent buffer followed by 120 sec in solvent buffer before a switch to S-Basal. In S-Basal, the animals received three pulses of 60 sec in buffer and 30 sec isoamyl alcohol, followed by 60 sec in buffer.
Each experiment was repeated 3-4 times over 2-3 days and pooled by strain for analysis (wild-type: 31 animals, 4 experiments, 3 days; str-217-/-: 23 animals, 3 experiments, 2 days). Images were acquired at 10 frames/sec at 5X magnification (Hamamatsu Orca Flash 4 sCMOS), with 10 msec pulsed illumination every 100 msec (Sola, Lumencor; 470/40 nm excitation). Fluorescence levels were analysed using a custom ImageJ script that integrates and background-subtracts fluorescence levels of the AWC neuron cell body (6×6 pixel region of interest). Traces were normalized by subtracting and then dividing by the base-line fluorescence, defined as the average fluorescence of the last 2 sec of the first three isoamyl alcohol pulses. The traces in Extended Data Figure 1 were also smoothed by 5 frames, such that each data point n is the running average of n-2, n-1, n, n+1, and n+2. The response magnitudes were calculated by taking the mean of the last 2 sec of an isoamyl alcohol pulse, subtracting the mean of the 2 sec before the isoamyl alcohol pulse (F0), and dividing by this F0. The response magnitudes were calculated for the 5th (0.15% ethanol in S-Basal buffer), 8th (0.15% DEET and 0.15% ethanol in S-Basal buffer), and 14th (0.15% ethanol in S-Basal buffer) isoamyl alcohol pulses. We also quantified the response magnitude of the transition from S-Basal buffer with ethanol to S-Basal buffer with DEET. We took the mean of the first 2 sec after switching to DEET buffer, subtracted the mean of the 2 sec before switching (F0), and divided by this F0.
Chemotaxis tracking and analysis
8-20 adult hermaphrodites were first transferred to an empty NGM plate and then 4-15 were transferred to an assay plate to minimize bacterial transfer. Animals were then placed in the centre on either a 0.15% DEET-agar or solvent-agar plate, and their movement was recorded for 60 min at 3 frames/sec with 6.6 MP PL-B781F CMOS camera (PixeLINK) and Streampix software. Assays were carried out at room temperature, between 12-8pm EST, and lit from below. Worm trajectories were extracted by a custom Matlab (MathWorks) script17, and discontinuous tracks were then manually linked. Tracks were discarded if the animal moved less than two body lengths from its origin over the course of the 60 min trial. If an animal came within 1cm of the isoamyl alcohol stimulus, the track was truncated to remove information from animals immobilized at the odour source because of the addition of sodium azide.
ADL optogenetic stimulation
L4 animals expressing an Psrh-220::ReaChR29 array or array-negative animals from the same plate were raised overnight in the dark on an NGM plate freshly seeded with 100 µL of 10X concentrated E. coli (OP50 strain) with or without 50 µM all-trans retinal (CID: 720648, Sigma-Aldrich, catalogue #R2500), which is required for ReaChR-induced activity. The next day, adult hermaphrodites were first transferred to an empty NGM plate and then 4-15 animals were transferred to the 10 cm circular assay plate to minimize bacterial transfer. Videos were recorded for 26 min at 3 frames/sec with a 1.3 MP PL-A741 camera (PixeLINK) and Streampix software. Blue light pulses were delivered with an LED (455 nm, 20 µW/mm2, Mightex) controlled with a custom Matlab script17,40. Animals were exposed to normal light for 120 sec, before exposure to 12 pulses of blue light (455 nm, 10 Hz strobing) for 120 sec, followed by 120 sec of recovery. This should activate ADL neurons only in retinal-fed animals expressing ReaChR. Worm trajectories were extracted by a custom Matlab script40. Pausing events were extracted, and all pauses ≥3 frames (1 sec) were used for further analysis. Pauses were classified as “ON” if any frame included light illumination. A pause that began just before illumination began, but remained paused while the illumination occurred, was considered an ON pause, just as a pause that occurred in the middle of a light illumination time frame was considered ON. All other pauses were classified as “OFF” pauses. In the analysis in Figure 4h, we took an average pause length for all ON pauses and all OFF pauses for each animal, and pooled all of the animals on each plate. To control for any baseline differences between animals and experiment-to-experiment variation, we examined the increase in average pause length in Figure 4i.
Statistical Analysis
R v3.3.2 was used for all statistical analysis. Inclusion and exclusion criteria were pre-established for all experiments, and in behaviour experiments positions were pseudo-randomized. All scripts and raw data with the exception of raw video files are available in Supplemental Data File 1. Scripts to analyse these data are also available at this link: http://github.com/VosshallLab/ Dennis-Emily_2017
Author Contributions
E.J.D. and L.B.V. developed the concept and designed the experiments. E.J.D performed all experiments and analysis unless noted. X.J. carried out imaging experiments in Figure 3 along with E.J.D., and performed cell identification. M.D. performed AWC calcium-imaging experiments in Extended Data Figure 1. L.B.D. performed HEK cell experiments in Extended Data Figure 2. C.I.B. provided guidance, reagents, experimental design advice, identified cells, and interpreted data. P.S.H. made the original observation that DEET interferes with chemotaxis in C. elegans, and initiated genetic screens for DEET-resistant mutants in his laboratory. E.J.D. and L.B.V. together interpreted the results, designed the figures, and wrote the paper with input from the other authors. The authors declare no competing financial interests.
Acknowledgements
We thank Michael Crickmore, Kevin Lee, Aakanksha Singhvi, Nilay Yapici, and members of the Vosshall Lab for discussion and comments on the manuscript. Shai Shaham advised and Wendy Wang assisted with chemical mutagenesis. Heeun Jang provided guidance on chemotaxis behaviour and imaging. Alejandro Lopez-Cruz and Elias Sheer provided advice on tracking behaviour. Sagi Levy shared the Pstr-2::GCaMP5a strain and Elias Sheer shared the Psrh-220::ReaChR plasmid. We thank Anh Nguyen for her contributions to the early analysis of DEET-resistant mutants in the Hartman laboratory Some C. elegans strains used in this paper were obtained from the CGC, which is funded by NIH Office of Research Infrastructure Programs (P40 OD010440). This work was supported by an NIH grant to E.J.D. (F31 DC014222). L.B.V. is an investigator of the Howard Hughes Medical Institute.