Surprisingly, D. suzukii did not show avoidance to CO2 at levels that elicit robust avoidance in D. melanogaster . These results suggest that while D. suzukii can detect CO2 other changes in processing this sensory information have occurred. Since D. suzukii are attracted to ripening fruits which emit CO2, we also tested whether CO2 can enhance attraction in the presence of food odors such as apple cider vinegar as has been reported in D. melanogaster6 . D. suzukii showed no significant increase in attraction to ACV in the presence of CO2 . While we do not know the behavioral significance of CO2 detection in this species yet, the ability to sense but not avoid CO2 may offer a distinct evolutionary advantage since D. suzukii feed on ripening fruits that respire and emit CO2. Avoidance of CO2 is robust in the D. melanogaster lab strain and also in two recently caught wild-type strains tested in the T-maze Assay which suggests that repellency is not due to artificial selection in our lab strains . These results pose an interesting question: how conserved is avoidance of CO2 in other related Drosophila species? In order to answer this question we performed a series of behavioral and electrophysiological experiments using 3 additional Drosophilid species . Using the T-maze Assay we found that the closely related D. yakuba showed avoidance to carbon dioxide,grow raspberries in a pot albeit to a lower degree than D. melanogaster .

However, like D. suzukii, the more distantly related D. virilis did not avoid carbon dioxide and D. pseudobscura was only mildly repelled at the highest concentration tested . The Gr21a and Gr63a amino acid sequences are highly conserved across all tested species: D. yakuba , D. pseudoobscuraand D .virilis . We find that a CO2-sensitive ab1C-like neuron is present in each of these species from single sensillum recordings . These results suggest that detection of CO2 is conserved; however, the behavioral changes could likely be due to other changes such as processing CO2-detection information in downstream circuitry in the brain. Since the CO2-pathway is the strongest known repellency pathway for some Drosophila species, we wondered whether other odorants that activate the CO2 receptors would serve as practical repellents that can be applied easily in areas to be protected unlike CO2, which is expensive and impractical for use. In a previous study we had identified pyridine, an animal skin odorant, as one of the strongest activators of the CO2 receptor14. At a 10−2 concentration, pyridine elicited avoidance behavior in D. melanogaster and D. yakuba as expected from the response to carbon dioxide observed, and also in D. pseudoobscura which avoids CO2 to a lesser degree . In D. pseudoobscura, it is conceivable that other olfactory receptors may also contribute to pyridine repellency. Also, as expected the D. suzukii and D. virilis showedvery little repellency to pyridine. A lower concentration of pyridine was also tested; however, none of the species showed behavioral response in the T-maze Assay . The T-maze assays the instantaneous behavioral response of walking flies offered a choice between an odor and control. In order to test behavioral response of flying Drosophila in a long-term behavior we utilized a Two-Choice Trap Assay.

Ten male and ten female starved flies are placed in a cylindrical chamber containing two entry traps lured with ACV, one of which also contained the CO2-neuron activator pyridine as a convenient substitute for CO2. Both D. suzukii and D. melanogaster showed no significant avoidance of the pyridine-treated trap . While behavioral responses of free flying Drosophila to CO2 have not been tested, tethered D. melanogaster that can beat their wings do not demonstrate clear anti-tracking behavior to CO2 15. This taken together with our results suggest that CO2-receptor activating odorants such as pyridine are unlikely to act as broad-spectrum repellents for Drosophila species, particularly for the agriculturally important pest D. suzukii.These findings prompted a systematic investigation of known repellent olfactory pathways in D. melanogaster and their conservation in related species. A second avoidance pathway in D. melanogaster is mediated by activation of Or85a, a member of another receptor gene family16. The strongest known activator of this receptor is the odorant ethyl 3-hydroxybutyrate17–19. We tested two concentrations of ethyl 3-hydroxybutyrate in a T-maze Assay. All species showed no preference at the lower concentration. D. melanogaster had some avoidance of ethyl 3-hydroxybutyrate at the higher concentration . This response was conserved in D. yakuba. The D. suzukii showed a small repellency; however, the distantly related species D. virilis and D. pseudoobscura showed no behavioral response to ethyl 3-hydroxybutryate. These behaviors are consistent with the observation that D. pseudoobscura and D. virilis lack a functional copy of the Or85a genes5,20–22. In order to test whether ethyl 3-hydroxybutryate can reduce attraction towards an attractive odor source and to act over a longer time period, we used a Two-Choice Trap Assay. D. melanogaster avoids the apple cider vinegar trap with 1% ethyl 3-hydroxybutyrate .

Participation in the Two-Choice Trap Assay for D. suzukii was very low and was not included in this analysis. The odorant was unable to repel the three other species from the lure. These results reinforce our view that for some odorants the avoidance is not conserved, because the receptor gene is not present in the genome. In addition, ethyl 3-hydroxybutryate may not activate other repellent receptors. A third known repellent is citronellal, a naturally occurring essential oil found in multiple plant species. Citronellal activates olfactory neurons in the antenna named ab11A and ab12A in D. melanogaster and a Trp channel is believed to play a role in citronellal’s activation of ab11A but not ab12A23. The odorant receptors in these neurons are unknown. In order to test the conservation in repellency to citronellal we used the previously described Direct Airborne Repellant Assay 23. For this non-contact assay, odorant is placed on filter paper at the bottom of a standard 15ml culture tube, with a mesh screen placed 0.5ml from the bottom to prevent flies from contacting the filter paper with odorant. Two tubes are placed together to form a long tunnel in which ~100 flies are given 30minutes to choose the odorant or solvent end of the tube. We found that 1% citronellal is avoided by D. melanogaster, D. pseudoobscura and D. virilis but not as strongly by D. yakuba and D. suzukii. In fact, a lower concentration of citronellal was slightly attractive for D. yakuba and D. suzukii . These results suggest that citronellal is unlikely to be useful as a strong repellent for D. suzukii. A recently identified repellent pathway is dependent on the olfactory ionotropic receptor, Ir40a,best grow pots which is expressed in the sacculus of the antenna and detects the commonly used insect repellent DEET24. DEET is also detected by bitter neurons in the gustatory system25. In Culex quinquefasciatus mosquitoes, DEET has been proposed to act via odorant receptor CquiOR136; however, none of the Drosophila species tested here have an ortholog of CqiOR13626. To examine the response to DEET in other Drosophilids, the Two-Choice Trap Assay in a plate is typically used27. Briefly, 10 female flies are placed inside a Petri dish containing two food odor-lured traps. In order to access the food, flies must crawl over a piece of filter paper impregnated with test compound. Four species strongly avoided DEET and D. pseudoobscura also avoided it, but to a lesser degree suggesting that this pathway is highly conserved, unlike the other repellent pathways tested . The Ir40a protein coding sequences are also well conserved compared to conservation in the Or gene family across the species D. yakuba , D. suzukii33 , D. pseudoobscura , and D. virilis but not as conserved as Gr21a and Gr63a. In fact, DEET repellency is conserved across most tested insects, as is Ir40a28.Recently, a number of new naturally occurring repellents were discovered that can substitute for DEET and are strongly repellent to D. melanogaster and mosquitoes24. Many of these repellent compounds are present naturally in fruits, have very mild and pleasant odors, are commonly used flavor and fragrance components belonging to a category called generally recognized as safe and are approved for human consumption through addition to food. We tested whether these compounds can be used to repel D. suzukii. We measured behavioral responses to three of these DEET-substitute compounds: butyl anthranilate , methyl N,N-dimethylanthranilate and ethyl anthranilate in the previously used Two-Choice Trap Assay in a plate. D. suzukii avoided the traps containing 10% of all three compounds ; however, at 1%, ethyl anthranilate did not repel D. suzukii . We then asked if DEET-like compounds would also act as oviposition deterrents by testing preference for Drosophila melanogaster egg-laying using a Two-Choice Oviposition Assay . Briefly, 15 male and 25 female flies are released into a 10 gallon closed glass chamber containing two Petri dishes with standard grape juice media.

After 24hours, eggs laid ongrape media with test odorant and solvent are counted. D. melanogaster did not oviposit on grape media infused with 0.4% DEET . At the lower concentration of 0.2% DEET, there is no avoidance and even initially a preference for ovipositing on the DEET containing media. D. melanogaster avoided ovipositing on MDA, BA or EA at both concentrations tested. Since these DEET-like compounds deter attraction and oviposition, we wondered if they would also deter D. suzukii oviposition on fruit. In order to test whether BA can protect fruit from D. suzukii we revised the previous assay by replacing the grape juice media with two bowls of fresh, ripe blueberries and extending the assay time to one week. This Two-Choice Assay in a glass chamber allowed us to infer egg-laying from a count of eggs, larvae and pupae emerging from each set of fruit. As expected from the time lapsed between the end of the experiment and dissection of exposed blueberries, few eggs were observed, with the exception of one of the six trials of 10% BA where 43 unhatched eggs out of 159 total D. suzukii that were counted. Of the unhatched eggs, 95% were laid on the control blueberries. More importantly, we found a clear dose-dependent decrease in numbers of larvae and pupae emerging from the BA treated blueberries. Remarkably, decreases were observed from the week-long experiment after only a single treatment, with substantial decreases at 2.5% and nearly complete protection at the 10% concentration . This proof of principle experiment indicates that insect repellents with different safety profiles can indeed be useful to reduce fruit damage during ripening.Each year D. suzukii damages hundreds of millions of dollars worth of fruits worldwide and there is a great need to find new ways to reduce this loss29,30. Toxic insecticides are often risky to use directly on fruits, and a safe affordable repellent could provide protection and reduce use of toxic chemicals. Although DEET is repellent to D. suzukii, it is a synthetic chemical that is unlikely to be useful in protecting crops given the human health concerns regarding food supply contamination, as well as the high production cost for the large volumes that would be required in agriculture. Other insect repellents we test here provide an opportunity to develop alternative effective strategies to reduce fruit damage. More generally, insects destroy a very large fraction of the global agricultural output and necessitate millions of tons of toxic insecticide use that is environmentally unfriendly and harmful to human populations. Further analysis of possibly environmentally safer and non-toxic repellents could decrease use of such insecticides. The analysis of conserved repellent pathways in the insect olfactory system offers an avenue to design behavioral control strategies of these dangerous pests and ultimately could form a foundation of novel and safe technologies that can improve both plant and human health.The T-maze Assay is ideally suited for highly volatile compounds such as CO2. Avoidance to carbon dioxide and pyridine trials were conducted as before7 . Briefly, approximately 40 flies are released from an elevator into the horizontal intersection of a T-shaped apparatus. A test odorant is applied to one arm of the T-maze and a control odorant to the opposite arm. For trials with other odorants, paraffin oil was used as the solvent and in the control arm. Flies are given one minute to choose an arm before the elevator closes. Orientation of arms for test and control were switched between trials. Preference index was calculated as = /. The Direct Airborne Repellent Testis suited for volatile compounds that could be confounded by responses from the taste system.