Identification of multiple odorant receptors essential for pyrethrum repellency in Drosophila melanogaster

Pyrethrum extract from dry flowers of Tanacetum cinerariifolium (formally Chrysanthemum cinerariifolium) has been used globally as a popular insect repellent against arthropod pests for thousands of years. However, the mechanistic basis of pyrethrum repellency remains unknown. In this study, we found that pyrethrum spatially repels and activates olfactory responses in Drosophila melanogaster, a genetically tractable model insect, and the closely-related D. suzukii which is a serious invasive fruit crop pest. The discovery of spatial pyrethrum repellency and olfactory response to pyrethrum in D. melanogaster facilitated our identification of four odorant receptors, Or7a, Or42b, Or59b and Or98a that are responsive to pyrethrum. Further analysis showed that the first three Ors are activated by pyrethrins, the major insecticidal components in pyrethrum, whereas Or98a is activated by (E)-β-farnesene (EBF), a sesquiterpene and a minor component in pyrethrum. Importantly, knockout of Or7a, Or59b or Or98a individually abolished fly avoidance to pyrethrum, while knockout of Or42b had no effect, demonstrating that simultaneous activation of Or7a, Or59b and Or98a is required for pyrethrum repellency in D. melanogaster. Our study provides insights into the molecular basis of repellency of one of the most ancient and globally used insect repellents. Identification of pyrethrum-responsive Ors opens the door to develop new synthetic insect repellent mixtures that are highly effective and broad-spectrum.


Introduction
Pyrethrum is a botanical insecticide extracted from dry flowers of Tanacetum cinerariifolium (also known as Chrysanthemum cinerariifolium). This plant is grown commercially in many parts of the world, particularly in East Africa and Australia, for extraction of pyrethrum, which accumulates in the flower achenes [1,2]. Pyrethrum is non-persistent in the environment and possesses low mammalian toxicity. Pyrethrum extract contains three structurally closelyrelated insecticidal esters of chrysanthemic acid (pyrethrin I) and three corresponding esters of pyrethric acid (pyrethrins II). Pyrethrins are prototypes of pyrethroids, a large class of widely used synthetic insecticides [3]. Pyrethrins and pyrethroids target voltage-gated sodium channels for their insecticidal effects [4][5][6], which is critical for the initiation and propagation of action potentials in the nervous system. Pyrethrins and pyrethroids promote activation of sodium channels and inhibit deactivation and inactivation, which lead to the disruption of the function of the nervous system.
Besides the insecticidal activities, pyrethrum extract has also been used as an insect repellent against biting arthropods for thousands of years [7] and in mosquito coils for more than a century [8]. In addition, pyrethrum-producing Chrysanthemum spp. are recommended as companion plants to repel pest insects [9]. Recent studies experimentally demonstrated behavioral deterrence of pyrethrin-containing Chrysanthemum leaves against western flower thrips (Frankliniella occidentalis) [10] and spatial repellency of a pyrethrin precursor against cotton aphids (Aphis gossypii) [11]. Despite these studies, the mechanistic basis of pyrethrum repellency remains unknown until our recent study in Aedes aegypti [12] and this study in Drosophila melanogaster.
Drosophila melanogaster has been an excellent model for studying insect olfactory chemosensing, with distinct types of well-characterized olfactory receptor neurons (ORNs) [13][14][15][16][17][18][19][20]. ORNs are housed in hair-like olfactory sensilla on the antennae. With a few exceptions, each sensillum usually houses two (up to four) ORNs and each ORN expresses one specific odorant receptor (Or) protein. Activation of Ors by odorants excites ORNs which project axons to the antennal lobe in the brain, where signals are processed and transmitted to higher order centers, which triggers appropriate behavioral outcomes.
In this study, we discovered that pyrethrum activate antennal olfactory receptor neurons and elicit spatial repellency in D. melanogaster, a model insect, as well as D. suzukii, a serious global insect pest of economically valuable small fruit and tree fruit crops [21]. We then further investigated the underlying mechanism of pyrethrum repellency by taking advantage of D. melanogaster as a model for olfactory studies. We found that specific components of pyrethrum activate multiple odorant receptors (Ors) and that co-activation of these Ors are essential for pyrethrum repellency. Identification of pyrethrum-responsive Ors represents a major step forward in the understanding of the molecular basis of repellency of one of the most ancient and globally used insect repellents.

Pyrethrum repels D. melanogaster and D. suzukii
To evaluate whether pyrethrum repels D. melanogaster, we first used a two-choice assay ( Fig  1A) that is similar to that described previously [22]. We found that pyrethrum repelled D. melanogaster w 1118 adults at the 10 −2 dilution (v v -1 ) (Fig 1B). The avoidance behavior was also observed in a T-maze assay (S1 Fig) which was modified from a previously reported protocol [22,23]. Pyrethrum also repelled D. suzukii in these assays (S1 Fig). Furthermore, we performed the two-choice assay in the presence of an attractant, apple cider vinegar (ACV), i.e., a two-choice attraction assay [24] (S1 Fig) and found that both D. melanogaster and D. suzukii were repelled by pyrethrum in this assay as well (S1 Fig).

Electrophysiological responses of ORNs to pyrethrum in D. melanogaster and D. suzukii
To identify which ORNs respond to pyrethrum, we focused on ORNs housed in antennal basiconic (ab) sensilla, where most antennal Orco/Ors are expressed [30]. Except for ab1, which contains four neurons, all ab sensilla house two neurons. We conducted single sensillum recording (SSR) of the electrical activities (i.e., action potentials measured as spikes/second) of ORNs in ab sensilla, as described by de Bruyne et al. [13]. Neurons that generate larger spikes in response to odors are defined as A neurons, whereas neurons that produce smaller spikes are called B neurons. We first recorded SSR responses to a panel of standard discriminating odorants [31,32] to ensure accurate identification and normalcy of each sensillum. Using this method, we were able to locate ab1-5 and ab7-8 sensilla (S2 Fig). We then examined the response of ORNs in ab1-5 and ab7-8 sensilla to pyrethrum. Representative traces of SSR measurements from ab1-5 and ab7-8 are presented in Fig 2. Pyrethrum increased the firing frequency of five out of 16 neurons in ab1, ab2, ab3, ab4 and ab7 sensilla (Table 1). In contrast, pyrethrum did not activate any neurons in ab5 or ab8 sensilla.
Using the same panel of discriminating odorants, we identified ab1-5 and ab7-8 sensilla in the antennae of D. suzukii (S3 Fig). Interestingly, the response profiles to standard discriminating odorants were essentially identical to those in D. melanogaster with one exception. In D. suzukii, ab2B neurons displayed strong responses to 2-heptanone, which was not observed in D. melanogaster, as also reported by Keesey et al. [31]. As in D. melanogaster, pyrethrum activated neurons of ab1A, ab2A, ab3A, ab4A and ab7A of D. suzukii (S4 Fig and Table 1). In addition, ab2B neuron of D. suzukii responded to pyrethrum, which was not seen in D. melanogaster ( Table 1).

Identification of Ors activated by pyrethrum
Maps of Or gene expression in basiconic sensilla are well established in the D. melanogaster olfactory system [30]. To identify which Ors are activated by pyrethrum, we employed the ab3 "empty neuron" system [33] by genetically introducing Ors, individually, into the A neurons of the empty ab3 sensilla, in which its endogenous Or gene Or22a is deleted. SSR analysis of the recombinant ab3 sensilla expressing each of the heterologously introduced Ors confirmed that Or42b from ab1A, Or59b from ab2A, Or7a from ab4A, and Or98a from ab7A were activated by pyrethrum (Fig 3A and 3B). Since Or22a is expressed in ab3A, we cannot directly test the role of Or22a in sensing pyrethrum in the empty neuron system. Consistent with SSR results, Ors in pyrethrum-nonresponsive ab1D, ab2B, ab5A, ab5B, ab7B and ab8A/B neurons could not be activated by pyrethrum in the empty neuron system ( Table 2). In addition, we also examined Or49b from ab6B, and Or67a and Or85f from ab10A and ab10B sensilla, respectively, in the empty neuron system because we could not directly identify these two types of sensilla in SSR. We found that none of them were activated by pyrethrum ( Table 2). Taken together, our results showed that four Ors, Or7a, Or42b, Or59b and Or98a, are activated by pyrethrum in D. melanogaster.

Selective activation of pyrethrum-responsive Ors by different components in pyrethrum
Pyrethrum extract contains six structurally related esters: pyrethrin I, cinerin I and jasmolin I, which are three esters of chrysanthemic acid, and pyrethrin II, cinerin II and jasmolin II, which are esters of pyrethric acid (S5 Fig). The structures of these compounds differ only in the acid and alcohol termini. Pyrethrin I and pyrethrin II are predominant components (together constituting more than 50%) in pyrethrum extracts [1,3]. We tested the effects of the  Table 1. Response spectra of ORNs to pyrethrum in D. melanogaster and D. suzukii.

Discussion
In this study, we discovered that pyrethrum vapor evokes olfactory responses and elicits aversion in D. melanogaster and D. suzukii. Although the major components of pyrethrum, pyrethrins, are known to target voltage-gated sodium channels for their insecticidal activity [35], we show here that pyrethrins also activate three Ors, Or7a, Or42b and Or59b. In addition, we discovered that EBF, a minor component in pyrethrum, activates another Or, Or98a. The most intriguing discovery of this study is that three Ors, Or7a, Or59b and Or98a, that are activated by multiple components in pyrethrum are all essential for pyrethrum repellency. Our results provide insights into the molecular basis of repellency of one of the most ancient and globally used insect repellents. It appears that simultaneous activation of the Or98a-mediated   repellent pathway and pyrethrin-activating Or7a and Or59b pathways was exploited, unknowingly, by humans some thousands of years ago in the formulation of pyrethrum extract as a potent natural insect repellent. We speculate that similar mechanisms might exist for other natural repellents, which are often mixtures of multiple olfactory bioactive components. Insects respond to volatiles, which often exist as complex mixtures in their environments, by relying on their sophisticated olfactory input and central processing pathways in the peripheral and central nervous systems [36]. Among the pyrethrins-activating Ors, Or7a has previously been shown to be activated by aversive odorants [37,38]. Or7a-expressing ORNs project to the "aversive-specific" glomerulus DL5 in the antennal lobe, whereas Or42b-expressing ORNs activated by attractive odorants innervate the "attractive-specific" glomerulus DM1 [39][40][41][42]. Prior to our study, Or59b was shown to be exclusively activated by acetone [37], which elicits attraction in D. melanogaster [43]. Indeed, we also observed acetone attraction at the 10 −4 dilution (v v -1 ) (S7 Fig). Furthermore, we found that acetone attraction was abolished in Or59b -/lines, indicating that activation of Or59b mediates attraction (S7 Fig). However, our results also show that pyrethrum repellency is abolished in Or59b -/lines indicating that Or59b has a critical role in pyrethrum repellency. These seemingly contradictory findings may be explained by the differences in the olfactory stimuli: acetone is a single component activating one to a few Ors compared with pyrethrum which is a mixture activating multiple Ors with opposing valences.
EBF is part of herbivore-induced plant volatile blends in tobacco, bean, potato, corn, cotton, sorghum and pine [44][45][46][47][48][49][50][51], providing information on the presence of herbivores. In Helicoverpa assulta, EBF activates HassOr23 and one specific glomerulus in the AL and inhibits oviposition of female H. assulta in tobacco plants [52]. Aphids release EBF as an alarm pheromone when attacked by predators or parasites [53,54]. In Acyrthosiphon pisum, EBF activated ApisOr5 to signal alarm and trigger repellency; and knockdown of the ApisOr5 transcript by RNA interference abolished the repellency [55]. Repellency of EBF was abolished in the Or98 -/mutant flies (S8 Fig), demonstrating that EBF activates the Or98a-mediated repellent pathway in Drosophila. Recently, we reported that EBF activates Or31 from Aedes aegypti and Anopheles gambiae. Like Or98a in D. melanogaster and ApisOr5 in A. pisum [55], activation of AaOr31 mediates EBF repellency in Ae. aegypti [12]. However, there are less than 15% sequence similarities between AaOr31, ApisOr5, Or98a and also HassOr23 from H. assulta. So far, Or98, ApisOr5, HassOr23 and AaOr31 are the only Ors that have been reported to sense EBF.
EBF is a minor component of pyrethrum ranging from 1.25% to 1.97% based on our analysis of the pyrethrum extracts used in this study. At the 10 −4 dilution (v v -1 ), equivalent to the amount of EBF in our pyrethrum repellency assay in Fig 4, EBF did not elicit repellency (S8 Fig). Therefore, importantly, we have shown that activation of Or98a by EBF in pyrethrum is essential for pyrethrum repellency, even though EBF in pyrethrum by itself is not sufficient to evoke aversion. This suggests that EBF/Or98a contribution to pyrethrum repellency in Drosophila depends on pyrethrin-mediated activation of Or7a/Or59b repellency pathways. Notably, not only did the Or98 -/mutant flies lose avoidance response to pyrethrum (Fig 4A), but they also exhibited attraction to pyrethrum (but no attraction to EBF in S8 Fig), highlighting sophisticated interactions between various Or-mediated pathways in response to pyrethrum in determining an ultimate behavioral outcome. The Or98a-mediated repellent pathway could override pyrethrin-activated Or42b-mediated attractive pathway, similar to the geosmin-activated Or56a-mediated repellency, activation of which suppressed attraction by ethyl butyrate [23]. Our findings provide a foundation for further analysis of the neural circuitry that integrates these Or pathways into a potent avoidance response. Future analyses of combinations of double, triple, or quadruple mutants of Or7a, Or42b, Or59b and Or98a would be able to provide further insight into how these Ors interact. In our recent study on the mechanism of pyrethrum repellency in Ae. aegypti, we found that the low amount of EBF in pyrethrum also makes significant contribution to pyrethrum repellency [12]. Further functional analyses in both insect species could advance our understanding of inter-channel integration in the antennal lobe via lateral connections and/or further integration in the lateral horn [20,[56][57][58].
The two Drosophila species examined in this study have very distinct ecological niches. For example, D. suzukii exhibits stronger attraction to leaf odors than D. melanogaster in behavioral assays [31,59]. The fact that both D. melanogaster and D. suzukii respond similarly to pyrethrum, in electrophysiological and behavioral assays, suggests that the pyrethrum-sensing pathways are conserved between the two species although the mutant systems are not yet available to conduct experiments with the same detail in D. suzukii. Of note, ORNs activated by pyrethrum are identical between the two species except for ab2B. Pyrethrum activates ab2B neurons in D. suzukii, but not in D. melanogaster. Interestingly, Or85a, expressed in ab2B in D. melanogaster, is lost in D. suzukii [60]. Conversely, 2-heptanone activates ab2B in D. suzukii but not in D. melanogaster [31]. It seems likely that the loss of Or85a is responsible for the change in the response profiles of ab2B in the two species. It is also possible that the differential responsiveness of ab2B to pyrethrum in the two species could be due to the expression of a different (yet to be identified) Or in ab2B of D. suzukii. Future research should examine how differential activation of ab2B neurons in the two species might influence the integration of neural activities in the central processing of olfactory coding and whether such differential integration contributes to niche-adapted responses to natural odors as well as insect repellents, such as pyrethrum.
Chrysanthemum spp. are currently used as companion plants to repel pest insects [9]. Recent studies demonstrated behavioral deterrence of pyrethrin-containing Chrysanthemum leaves against western flower thrips (Frankliniella occidentalis) [10] and spatial repellency of a pyrethrin precursor against cotton aphids (Aphis gossypii) [11]. The Or98a/Or7a/Or59b triple receptors-mediated avoidance mechanism, discovered in this study, could represent an important general olfaction-based strategy for diverse insects to avoid natural insecticidal toxins from plants and for plants to avoid being consumed by insects in a dynamic plant-insect interactive natural world.

Fly stocks
Drosophila melanogaster w 1118 line was used as reference stock, and a D. suzukii (spotted-wing drosophila) line was field-collected in Michigan in 2016 and maintained in the laboratory since then. Two Orco mutants (herein called Orco -/-1; Orco -/-2), were obtained from the Bloomington Drosophila Stock Center (BDSC) (stock numbers: B23129 and B23130, respectively). The fly lines used in the empty neuron system were kindly provided by John Carlson (Yale University). All flies were raised on BDSC standard cornmeal food: 225 g agar, 2850 g yellow cornmeal, 675 g yeast, 390 g soy flour, 3 L light corn syrup, 39 L water, and 188 ml propionic acid; in an incubator with settings of 25˚C, 60% humidity and a 12 h /12 h day/night light cycle.
Pyrethrin I, cinerin I, jasmolin I, pyrethrin II, cinerin II and jasmolin II were purified by HPLC with a Shim-pack PREP-SIL silica gel column (20 x 250 mm; Shimadzu) at a flow rate of 10 mL min -1 by monitoring the absorbance at 230 nm. As the eluent, a hexane/ethyl acetate mixture (93/7) was used for the purification of pyrethrin I, cinerin I and jasmolin I purification, whereas an 85/15 mixture was used for the purification of pyrethrin II, cinerin II and jasmolin II.

Behavioral assays
A two-choice assay, as shown in Fig 1A, was modified from the previously described assay [22]. Briefly, to make an assay trap, the tapered end (0.2 cm) of a 1.7 mL microcentrifuge tube (Denville posi-click tubes, Natural color) was cut off; and a 1 mL pipette tip (Tips for Eppendorf Pipettes, Thomas Scientific Inc.) was cut at 2.5 and 0.5 cm from the narrow tip to produce a funnel-like small tip. A 1.6 cm × 1.6 cm filter paper was inserted through the open lid of the cut microcentrifuge tube and secured in by inserting the narrow end of the small tip into the cut microcentrifuge tube (Fig 1A). Fifty microliters of solvent or diluted test compound were applied onto the filter paper and the cut microcentrifuge tubes were then capped. The control and test traps were placed upside-down 2 cm apart in a 100 mL glass beaker and secured using small pieces of double-sided tape. Forty to fifty flies three-to six-day-old flies (both males and females) were gently tapped down from a food vial into the beaker which was already covered with cheese cloth secured with rubber bands. Individual beakers were then placed in individual plastic storage boxes (Snapware, Smart System; 40 cm x 30 cm x 15 cm) into a 25˚C incubator. The two-choice attraction assay setup was similar to that of the two-choice assay, except for the addition of 125 μL of 10% apple cider vinegar (ACV) to the upturned lids of cut microcentrifuge tubes as an attractant (S1 Fig). The T-maze assays was adapted from previously described assay [23] with some modifications (S1 Fig). Briefly, two 1 mL pipette tips (Tips for Eppendorf Pipettes, Thomas Scientific Inc.) and two 1.7 mL microcentrifuge tubes (Denville posi-click tubes, Natural color) were cut and assembled to form two traps. The traps were connected using a 4-cm length and 6.35 mm inner diameter Tygon tubing (Saint-Gobain, Tygon S3 E-3603). Before assembly, a piece of 0.8 cm × 3.2 cm filter paper was lined the wall of the microcentrifuge tube. A 50 μL solvent or test compound of 10 −2 dilution (v v -1 ) was applied to the filter paper. After the compound was applied, three-to six-day-old D. melanogaster or D. suzukii flies (10 males and 10 females) were gently introduced into the Tygon tubing via a third pipette tip which was connected to the tubing via a small hole made in the middle of the tubing. In both microcentrifuge tube lids, a small hole was made to let air flow through.
For all three assays, trials were run for 24 h at 25˚C and the number of flies entering each trap was counted. The Repellency Index (RI) was calculated as ((O-C)/(O+C)) � 100, where O is the number of flies in the test compound trap, C is the number of flies in the control (solvent) trap [22]. The RI ranges from −100% (complete attraction) to 100% (complete avoidance).

Electroantennography (EAG)
Flies (4-8 days old) were wedged into the narrow end of truncated 200 μL plastic pipette tip and mounted on a microscope slide. The tip of a glass micropipette was used to hold the antenna in a stable position. EAG recordings were conducted as described previously [61]. Reference and recording glass capillary electrodes (1.5 mm outer diameter) were filled with Drosophila Ringer's (in mM): NaCl 100, KCl 5, MgCl 2 20, CaCl 2 0.15, HEPES 5, sucrose 115, trehalose 5. The reference electrode was inserted into the contralateral eye. The recording electrode was capped onto the anterior distal region of the third antennal segment. The electrodes made electrical contact with a high impedance amplifier (World Precision Instruments, DAM 50) via silver/silver-chloride wires. The signals were digitized with a Digidata 1440A digitizer and Axoscope 10.4 software (Axon Instruments, Molecular Devices). Data were analyzed using Clampfit 10.4 software.

Single sensillum recording (SSR)
Single sensillum recording was conducted with electrolytically sharpened tungsten microelectrodes as previously described [13,24,31]. A 0.1 mm diameter tungsten wire was sharpened by repeatedly dipping its tip in a 10% KNO 2 solution electrified at 5-10 mV. Action potentials were recorded by inserting the recording microelectrode in the base of a sensillum, making contact with the lymph surrounding the dendrites of the ORNs. The reference electrode was inserted in the compound eye. The recording electrode was connected to an IDAC-4 signal acquisition system (Syntech, The Netherlands). Signals were fed into a computer and analyzed with Autospike software (Syntech). Signals were counted offline in a 500 ms period before stimulation and for 500 ms during stimulation. Stimulus was controlled using the CS-55 stimulus delivery system (Syntech). Thirty microliters of 10 −2 dilution (v v -1 ) of odorants or solvent was delivered on a filter paper strip (0.4 cm × 4 cm) which was placed in the shaft of a glass Pasteur pipet serving as an odorant cartridge.

Gene knockout
Knockout lines were constructed using the CRISPR/Cas9 technology following the method of Gratz et al. [62]. Two guide RNAs (gRNAs) for each Or were designed by searching the sense and antisense strands of the each ORs gene using the Chopchop (https://chopchop.cbu.uib.no/ ), CRISPR optical target finder (http://targetfinder.flycrispr.neuro.brown.edu/) and e-CRISPR (http://www.e-crisp.org/E-CRISP/). Sequences of gRNAs were selected based on recommendations by all the three websites for less likely off-target binding. For cloning of gRNA, sense and anti-sense oligos containing the overhang sequences (underlined in S1 Table) to anneal the vector (pU6-BbsI-chiRNA from Addgene), "G" (only for Or7a), and CRISPR target sequence were synthesized by Integrated DNA Technologies, Inc. The oligos are phosphorylated using T4 Kinase (Invitrogen) at 37˚C for 30min, followed by heating at 95˚C for 5 min, then ramp to 25˚C at a rate of -0.1˚C/sec. for annealing. Annealed oligo was then cloned into the BbsI site of pU6-BbsI-gRNA.
For donor construction for homology-directed repair, 5' arm and 3' arm regions of 1Kb upstream and downstream of the CRISPR target site were amplified using Platinum Taq DNA Polymerase, High Fidelity (Invitrogen). PCR reaction was heated to 94˚C for 2 minutes, followed by 35 cycles of 94˚C for 30 seconds, 55˚C for 30 seconds, and 68˚C for 70 seconds, then 68˚C for 7 minutes. PCR product was purified using Wizard SV Gel and PCR Clean-Up System (Promega), then 5' arm was digested by AarI and cloned into pDSRedattp (Addgene). PCR product of the 3' arm was then digested with SapI and cloned into the pDSRed-attp with the 5' arm.
Microinjection, generation, identification of transformants (with DsRed) and balancing were performed by BestGene Inc. Donor plasmid and gRNA plasmids are extracted by QIA-GEN Plasmid Midi kit (Qiagen). Each pair of gRNA plasmids and donor plasmid were co-injected into embryos of the transgenic line nanos-Cas9. Deletion of the Ors was confirmed by genomic PCR/sequencing. The sequences of the sgRNAs and details of the knockout lines obtained are summarized in S6 Fig. Primer sequences for PCR and sequencing are summarized in S1 Table. The knockout flies were then back-crossed for at least five generations with the wild-type strain to eliminate potential off-target events.

Statistical analysis
All statistical analysis was done using SigmaPlot 12.5 (Systat Software). Data are presented as mean ± SEM. Unpaired Student's t-test or Unpaired Mann-Whitney Rank Sum U-test (depending on whether assumptions for parametric tests were met) were used to compare results from two treatments. One-Way ANOVA (F-test) or One-Way ANOVA on Ranks (Kruskal-Wallis), depending on whether assumptions for parametric tests were met, were used, followed by Dunnett's test to compare multiple columns of data against a single control.  Table. List of primers used in this study. (PDF)