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 invertebrates separated by millions of years of evolution, 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 study its mechanism of action. We found that DEET is not a volatile repellent, but interferes selectively with chemotaxis to a variety of attractant and repellent molecules. DEET increases pause lengths 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. Misexpression of str-217 in another chemosensory neuron conferred strong responses to DEET. Both engineered str-217 mutants and a wild isolate of C. elegans carrying a deletion in str-217 are DEET-resistant. We found that DEET can interfere with behaviour by inducing an increase in average pause length during locomotion, and show that this increase in pausing requires both str-217 and ADL neurons. Finally, we demonstrated 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 behavioural responses12,13. 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 for the mechanism of DEET: “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.
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 Aedes aegypti mosquitoes13, but in contrast to results from Culex quinquefasciatus19. 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 DEET 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 determine if DEET has a general effect on chemotaxis, we tested three additional attractants, butanone, diacetyl, and pyrazine, as well as the volatile repellent 2-nonanone. DEET decreased attraction to butanone and avoidance of 2-nonanone, indicating that it can affect responses to both positive and negative chemosensory stimuli (Fig. 1f). DEET also affected chemotaxis to diacetyl, but not pyrazine, which is notable because both odorants require the same primary sensory neurons, AWA (Fig. 1f). A similar effect is seen in D. melanogaster, where DEET can affect responses to some odours and not others, even in a single chemosensory neuron20.
In D. melanogaster flies and Ae. aegypti mosquitoes, DEET inhibits attraction to complex odour blends of food and host odours21,22. When we provided the food odour of bacterial suspensions of OP50 E. coli to C. elegans, we found that DEET eliminated chemotaxis to this complex odour blend (Fig. 1g). Remarkably, supplementing bacterial odour with pyrazine, but not isoamyl alcohol, restored chemotaxis (Fig. 1g). To exclude the possibility that pyrazine is able to overcome the effect of DEET due to a higher effective concentration, we carried out dose-response experiments with isoamyl alcohol (Fig. 1h) and pyrazine (Fig. 1i) and found that at all concentrations tested, DEET interfered with isoamyl alcohol chemotaxis but not pyrazine chemotaxis. From these data we conclude that DEET chemosensory interference is odour-selective, and can affect both attractive and repellent stimuli.
Identifying genes required for DEET-sensation has been of interest for some time. A forward genetic screen in D. melanogaster yielded an X-linked DEET-insensitive mutant23, and a population genetics approach in mosquitoes identified a dominant genetic basis for DEET-in-sensitivity24, but neither study identified the underlying genes. Reverse genetic experiments in D. melanogaster and three mosquito species identified orco, the co-receptor for insect odorant receptor genes, as a molecular target of DEET9,11–14. However, this chemosensory gene family is not found outside of insects25,26, raising the question of what pathways are required for DEET-sen-sitivity 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. 2a). We obtained 5 DEET-resistant animals, three of which produced offspring that consistently chemotaxed toward isoamyl alcohol on DEET-agar plates (Fig. 2b, and data not shown). We identified candidate causal mutations in two strains using whole genome sequencing27, and focused on LBV003, which maps to str-217, a predicted G protein-coupled receptor (Fig. 2c and d). 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. 2e). 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 and early stop codon (Fig. 2c and d and Supplemental Data File 1). This is not a unique phenomenon. Using published data from the C. elegans Natural Diversity Resource (CeNDR)28, we discovered that 119 of 247 sequenced strains contain predicted changes in the STR-217 protein when compared to N2 wild-type (Supplemental Data File 1). One of these mutant strains shares the str-217HW deletion and 13 have a missense mutation that introduces a stop codon after the fourth amino acid. The remaining 105 strains have one or more of 30 high-confidence non-synonymous substitutions. We further explored DEET resistance in the Hawaiian deletion by testing three near-isogenic lines with a single, homozygous genomic segment of Hawaiian chromosome V introgressed into a wild-type (Bristol N2) background29 (Fig. 2e). Only the ewIR74 line contains str-217mv and, like the parent Hawaiian strain, is DEET-resistant (Fig. 2e). To provide further confirmation that str-217 is required for DEET sensitivity in these strains, we generated an additional genetic tool, a rescue/ reporter plasmid that expresses both wild-type str-217 and green fluorescent protein (GFP) under control of a predicted str-217 promoter (Fig. 2f). The LBV003 mutant strain (Fig. 2g) and the Hawaiian introgressed strain ewIR74 (Fig. 2h) both showed chemotaxis on DEET-agar. The str-217 rescue/reporter construct rendered both DEET-resistant mutants fully sensitive to DEET, in that neither chemotaxed to isoamyl alcohol on DEET-agar (Fig. 2g and h).
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 AWC chemosensory neurons30. To investigate if DEET modulates primary sensory detection of isoamyl alcohol, we used in vivo calcium imaging to monitor AWC odour responses 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 (Fig. 3 a). In the presence of DEET, AWC responses to isoamyl alcohol decreased in magnitude, but there were no significant differences in AWC activity between wild-type and str-217−/-, a predicted null str-217 mutant produced by CRISPR-Cas9 genome-editing (Fig. 3a-c). This suggests that AWC sensory neurons are not the primary functional target of DEET, even though they are affected by it. The polymodal nociceptive neurons ASH also responded to DEET (Fig. 3d and e), but animals lacking ASH are fully DEET-sensitive (Fig. 3f), suggesting that these sensory neurons are also not the primary target of DEET.
To identify such 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. 3g). This was unexpected because ADL is not required for chemotaxis to isoamyl alcohol31, suggesting an indirect role for ADL in DEET chemosensory interference. To investigate if str-217 is only required in ADL, we expressed str-217 in ADL under control of the srh-220 promoter in str-217−/- worms. ADL-specific rescue of str-217 rendered this mutant sensitive to DEET (Fig. 3h). To ask if ADL neuronal function is required for DEET-sensitivity, we inhibited chemical synaptic transmission by expressing the tetanus toxin light chain in ADL32,33. These animals showed the same level of DEET-resistance as str-217 mutants (Fig. 3i), demonstrating that ADL neurons are required for DEET-sensitivity.
Since both str-217 and ADL function are required for DEET-sensitivity, we used calcium imaging to see if ADL responds to DEET, and if this requires str-217 (Fig. 4a). Both wild-type and str-217−/- mutants carrying a rescue plasmid, but not str-217−/- mutants, showed calcium responses to DEET, albeit with some variability in response (Fig. 4b-d). To exclude the possibility that DEET affects ADL indirectly by interacting with other sensory neurons that subsequently activate ADL, we measured calcium responses of ADL in genetic backgrounds that disrupt chemical synaptic transmission between neurons. ADL neurons showed normal responses to DEET in both unc-13 and unc-31 mutant animals, which are deficient in synaptic vesicle fusion34 or dense-core vesicle fusion35, respectively (Fig. 4f-h). In control experiments, we showed that a known ADL agonist, the pheromone C932, activated ADL in both wild-type, str-217−/- mutant, and rescued animals (Fig. 4j-l), suggesting that the str-217−- mutation has a selective effect on ADL responses to DEET.
Since str-217 is necessary for ADL to respond to DEET, we asked if it str-217 is sufficient to confer DEET responses when heterologously expressed. When we expressed str-217 in HEK293T cells we failed to see activation by DEET (Supplemental Data File). We therefore misexpressed str-217 in another pair of C. elegans chemosensory neurons, AWB, and found that str-217 significantly increased the DEET-sensitivity of this cell (Fig. 4m-o). These gain-of-function data are consistent with the hypothesis that str-217 itself encodes a DEET receptor, or cooperates with other proteins in vivo that respond to DEET.
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, but 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. 5a-c). Wild-type, but not str-217−/- mutants (Fig. 5d), showed a dramatic increase in average pause length on DEET-agar. Although str-217−/- mutants are resistant to DEET compared to wild type, their chemotaxis is still reduced to some degree by DEET (Fig. 3h). This is evident in the end-point position of str-217−/- animals, many of which never make it to the odour source (Fig. 5e). Chemotaxis indices did not increase for wild-type animals when we prolonged the assays (Fig. 5f). This indicates that there are additional effects of DEET on chemotaxis in addition to increased pause duration.
To determine if the increase in average pause length occurs only in the context of chemotaxis to isoamylalcohol, or as a consequence of DEET alone, we tracked wild-type, str-217−/- mutant (Fig. 5g), and ADL::Tetanus toxin (Fig. 5h) 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. 5g-h). 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. 5i). This suggests that pyrazine chemotaxis 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 ReaChR36 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. 5j and k). 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 and uncover a new neuronal mechanism for a DEET-induced behaviour. We identify str-217 as a molecular target that is required for 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 and also for discovery 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 encodes a G protein-coupled receptor with no known ligand and that is evolutionarily unrelated to DEET-sen-sitive odorant receptor proteins previously described in insects. Our results are consistent with the hypothesis that str-217 is a DEET receptor, or that it interacts with additional proteins to make ADL DEET-sensitive. Inter-estingly, 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 behaviour that interferes with chemotaxis to some but not all odorants.
Although str-217 has no clear orthologues outside of nematodes, the str-217-dependent mechanism of action of DEET in nematodes we have discovered is reminiscent of the “confusant” hypothesis in insects. In insects, DEET alters responses of individual olfactory sensory neurons to attractive odorants9,12, thereby interfering with behavioural attraction. In C. elegans, DEET inhibits attraction to some odours by activating neurons that induce competing behaviours like pausing. 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.
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. M.D. performed AWC and ASH calcium-imaging experiments in Figure 3 and the AWB calcium-imaging experiments in Figure 4. X.J. carried out imaging experiments in Figure 4 along with E. J.D., and performed cell identification in Figure 3. L.B.D. performed HEK293T heterologous expression experiments. C.I.B. provided guidance and experimental design advice 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.
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, 12.9 μM cholesterol, 1mM MgSO4, 25mM KPO4, pH 6) seeded with Escherichia coli (OP50 strain) bacteria as a food source37,38. 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 generated29. Generation of extra-chromosomal array transgenes was carried out using standard procedures39, and included the transgene injected at 50 ng/mL, the fluorescent co-injection marker Pelt-2::G-FP 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. CRIS-PR-Cas9-mediated mutagenesis of str-217 was performed as described, using rol-6 as a co-CRISPR marker40. 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 0bserver 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, 1 mM CaCl2, 1 mM MgSO4)41. 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 and c,Figure 5j and k. 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 S-Basal buffer (1 mM NaCl, 5.74 mM K2HPO4, 7.35 mM KH2PO4, 5 μg/mL cholesterol at pH 6-6.2)37 to remove younger animals, and onc with chemotaxis buffer. Immediately before the start of the experiment, two 1 mL 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 Kim-wipe. 0dorants diluted in ethanol were used in this study at these concentrations unless otherwise noted: 1:1000 isoamyl alcohol (CID: 31260, Sigma-Aldrich, catalogue #W205702), 1:1000 butanone (CID: 6569, Sigma-Aldrich, catalogue #360473), 10 mg/μL pyrazine (CID: 9261, Sigma-Aldrich, catalogue #W401501), 1:10 2-nonanone (CID: 13187, Sigma-Aldrich, catalogue #W2787513). For bacterial chemotaxis assays, 20 μL of either LB media, or OP50 bacterial suspension grown overnight and diluted in LB media to 1.0 OD at 600nm was applied instead of or in addition to odorants. Assays were carried out for 60-90 min at room temperature (22-24°C) between 1pm – 8pm EST with the exception of Figure 5f, which were quantified after either 55-65 minutes (1 hour), 115-125 minutes (2 hours), 175-185 minutes (3 hours), or 235-245 minutes (4 hours). 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 819320BL11). 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 sequencing27 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 backcrossing42,43. LBV003 mapped to a 5 Mb region on chromosome V, which 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(e-jd002) contains a T>G transversion of the 141st nucleotide in the CDS, which is predicted to produce a Phe48Val substitution in this predicted sugar:proton symporter. We were unable to map the DEET-resistant mutation(s) in LBV001.
str-217 heterologous expression in mammalian tissue culture cells
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, Gq15, 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 (Invit-rogen, 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 s 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 s 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) (Supplemental Data File 1).
ADL calcium imaging
Calcium imaging and data analysis were performed as described44, 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 promoter32 and was crossed into str-217−/- and the str-217−/- rescue strain. Imaging of unc-13 and unc-31 mutant strains was performed by crossing ADL::GCaMP5k expressing animals to the unc- strains and selecting for fluorescent, uncoordinated animals. 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-Ba-sal 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+/-40 nm) for 100 s before recording to attenuate the light response of ADL45. Experiments consisted of the following stimulation protocol: 20 s S-Basal buffer, followed by 3 repetitions of 20 s DEET (0.15% DEET and 0.15% ethanol in S-Basal) and then 20 s S-basal buffer. ADL responses desensitize rapidly31, so only the first of the three DEET pulses was analysed.
All GCaMP signals were recorded with Metamorph Software (Molecular Devices) and an iXon3 DU-897 EMCCD camera (Andor) at 10 frames/s using a 40x objective on an upright Zeiss Axioskop 2 microscope. Custom ImageJ scripts17 were used to track cells and quantify fluorescence. In Figure 4b, f, and j, all frames in 20 s before the DEET pulse were averaged and divided by the average of the frames during the 20 s DEET or C9 pulse to calculate ΔF. In Figure 4c, g, and k, 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/Fmax46. The average value during the stimulus was calculated for each animal and plotted as a single dot in Figure 4d, h, and l. The heatmap traces in Figure 4c, g, and k 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, ASH, and AWB calcium imaging
Calcium imaging of freely moving worms and subsequent data analysis were performed as described46, using a 3 mm2 microfluidic PDMS device with two arenas that enabled simultaneous imaging of two genotypes with approximately 10 animals each. For AWC imaging, we used an integrated line (CX17256) expressing GCaMP5a in AWCON neurons undei 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 s in buffer and 30 s isoamyl alcohol, followed by 120 s in buffer. Next, the animals were switched to S-Basal with 0.15% ethanol (solvent buffer) and three pulses of 60 s in buffer and 30 s in isoamyl alcohol in solvent buffer followed by 120 s 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 s in DEET buffer and then 30 s in isoamyl alcohol in DEET buffer, followed by 120 s in DEET buffer before switching to solvent buffer. In solvent buffer, the animals received three pulses of 60 s in buffer and 30 s in isoamyl alcohol in solvent buffer followed by 120 s in solvent buffer before a switch to S-Basal. In S-Ba-sal, the animals received three pulses of 60 s in buffer and 30 s isoamyl alcohol, followed by 60 s 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/s at 5X magnification (Hamamatsu Orca Flash 4 sCMOS), with 10 ms pulsed illumination every 100 ms (Sola, Lumencor; 470/40 nm excitation). Fluorescence levels were analysed using a custom ImageJ script that integrates and background-sub-tracts fluorescence levels of the AWC neuron cell body (6×6 pixel region of interest). Traces were normalized by subtracting and then dividing by the baseline fluorescence, defined as the average fluorescence of the last 2 s of the first three isoamyl alcohol pulses. The traces in Figure 3a-b 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 in Figure 3c were calculated by taking the mean of the last 2 s of an isoamyl alcohol pulse, subtracting the mean of the 2 s 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) and 8th (0.15% DEET and 0.15% ethanol in S-Basal buffer isoamyl alcohol pulses.
ASH calcium imaging was performed similarly with the following exceptions. For ASH imaging, we used a strain (CX10979) expressing GCaMP3 in ASH neurons under control of the sra-6 promoter. The stimulus protocol used was as follows: 60 s in S-Basal, 60 s in 0.15% ethanol in S-Basal buffer, 60 s in S-Basal, 60 s in 0.15% DEET in S-Basal, and finally 60 s in S-Basal buffer. Each experiment was repeated over 2 days and pooled for analysis (wild-type: 15 animals in 2 experiments on 2 different days).
AWB calcium imaging was performed similarly to AWC imaging with the following exceptions. For AWB imaging, we used a control strain (CX17428) expressing GCaMP5a in AWB neurons under the str-1 promoter and a test strain (CX17660) expressing GCaMP5a under the str-1 promoter as well as expressing str-217 in AWB neurons under the str-1 promoter. Adult hermaphrodites were first similarly paralyzed in 1 mM (-)-tetramisole hydrochloride for 80-100 minutes, but the first 65-75 minutes was in S-Basal buffer and the last 15 minutes was 1 mM (-)-tetramisole hydrochloride in ethanol-buffer. The stimulus protocol used was as follows: 60 s in 0.15% ethanol in S-Basal buffer, 10 s in 0.15% DEET and 0.15% ethanol in S-Basal buffer, and 60 s in 0.15% ethanol in S-Basal buffer. A 4x4 pixel region of interest was used during tracking of the neurons. Baseline fluorescence was defined as the median fluorescence of the 10 s preceding the DEET pulse. Response and peak magnitudes were calculated using traces smoothed by 5 frames and identifying the maximum value within the DEET pulse. Five sets of experiments were conducted over 3 days for a total of 41 wild-type animals and 49 animals expressing str-217.
In Figure 4n, 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)/F046. The heatmap traces in Figure 4n 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. Peak ΔF/F0 (%) in Figure 4o reflects the maximum value of ΔF/F0 (%) during the DEET pulse.
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/s with 6.6 MP PL-B781F CMOS camera (Pix-eLINK) 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 1 cm 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::ReaChR36 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/s with a 1.3 MP PL-A741 camera (PixeLINK) and Streampix software. Blue light pulses were delivered with an LED (455 nm, 45 μW/mm2, Mightex) controlled with a custom Matlab script17,47. Animals were exposed to normal light for 120 s, before exposure to 6 repetitions of blue light (10 Hz strobing) for 120 s, and 120 s of recovery (LED OFF). Worm trajectories were extracted by a custom Matlab script47. Pausing events were extracted, and all pauses >3 frames (1 s) 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, as well as any pauses that began in the during light illumination considered ON. All other pauses were classified as “OFF” pauses. In the analysis in Figure 5j, 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 5k.
Phylogenetic Analysis of str-217 in wild isolates
Data were obtained from CeNDR3628 and plotted in R. Only predicted deletions in exons or missense changes of high confidence were included (Supplemental Data File 1).
Statistical Analysis
R v3.3.2 was used for all statistical analysis. Inclusion and exclusion criteria were pre-established for all experiments, and plate positions were pseu-do-randomized in behaviour experiments. Additionally, qqPlots were evaluated before performing ANOVAs. For the analysis of optogenetic experiments, a Levene’s Test identified heteroskedasticity in these data that was addressed with a boxcox translation. AWC imaging data were similarly boxcox translated and transformed to adjust for the rightward skew. All data necessary to re-create these plots are available in Supplemental Data File 1. Data, scripts to analyse these data, and all statistical analyses are available at GitHub: http://github.com/VosshallLab/Dennis-Emily_2017
Strains
Detailed genotypes of all C. elegans strains and their sources are in Supplemental Data File 1.
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.