Plant essential oils and potassium metabisulfite as repellents for Drosophila suzukii (Diptera: Drosophilidae)

Spotted wing drosophila, Drosophila suzukii, is a globally invasive pest of soft-skinned fruit. Females oviposit into ripening fruit and larvae cause direct destruction of tissues. As many plant essential oils are permitted food additives, they may provide a safe means of protecting fruit from D. suzukii infestation in both conventional and organic production systems. Twelve oils and potassium metabisulfite (KMS) were screened in the laboratory as repellents for D. suzukii flies. Most essential oils deterred D. suzukii flies from cotton wicks containing attractive raspberry juice. Peppermint oil was particularly effective, preventing almost all flies from contacting treated wicks and remaining 100% repellent for 6 d post-application. Thyme oil was unique because it caused high male mortality and reduced the number of responding flies compared to other oils. KMS was not found to be repellent to D. suzukii, but may have fumigant properties, particularly at high concentrations. Peppermint oil appears to be the best candidate for field testing to determine the effectiveness and feasibility of using essential oils as part of a push-pull management strategy against D. suzukii. This is the first time that essential oils have been evaluated and proven effective in preventing fruit-infesting flies from contacting attractive stimuli.

nuisance pests 24,25 . Natural products, including essential oils, are perceived as posing a lower risk to the environment and humans compared to synthetic compounds 26 , although safety is dependent on biological properties of and exposure to chemicals that are not always consistent with their origin 27 . Second, the large diversity and redundancy of phytochemicals in a single essential oil, including many mono-and sesquiterpenes, can improve control efficacy and reduce selection pressure and resistance development in pests 25,28,29 . Finally, there is an increased availability of essential oils, as many are registered as flavouring or perfuming agents 24 Essential oils may have potential for use in organic small fruit production systems, as there are currently limited management options for D. suzukii available.
As crop protection agents, essential oils have been widely tested as repellents or fumigants for stored product pests 30 and as repellents for biting or nuisance pests, mainly flies, of livestock 31 . In the latter, essential oils act to prevent insects from flying to, landing on, or biting skin. House flies, Musca domestica L., are repelled by basil, lemongrass, lavender, peppermint, ginger, geranium, and elemol (osage orange) oils [32][33][34] , and eucalyptus oil was used as a fumigant against the horn fly (Haematobia irritans, L.) 35 . Essentials oils can also be effective repellents of herbivorous pests, including lavender and pennyroyal (mint) for codling moth larvae (Cydia pomonella (L.)) 36 , lavender, juniper and citronella against red bud borer midges (Resseliella oculiperda (Rübsaamen)) on apple saplings 37 , and common thyme and winter savory for Western flower thrips (Frankliniella occidentalis (Pergande)) 38 . Peppermint, thyme, lavender and rosemary oils caused mortality in Drosophila auraria flies 39 , but to our knowledge, essential oils have not been tested as repellents or deterrents to prevent fruit infesting flies from contacting host fruits or surfaces containing attractive stimuli.
In addition to essential oils, potassium metabisulfite (KMS) may be useful as a repellent for D. suzukii. KMS (K 2 S 2 O 5 ) produces sulfur dioxide when dissolved in water 40 and is a common additive in food and the wine-making process. Sulfur dioxide acts as an antimicrobial and antioxidant, preventing browning by inhibiting bacterial growth during wine-making 41 . In the only report on repellency towards insects, KMS repelled Harmonia axyridis (Pallas) in a Y-tube olfactometer and reduced numbers of beetles on grape vines 42 . KMS is not registered for field-use worldwide, but it may be useful for controlling Botyritis in vineyards 40 .
The objective of this study was to screen 12 essential oils and KMS as repellents for D. suzukii flies in the laboratory. A bioassay allowed male and female flies to choose between an attractant or the attractant plus an essential oil. Essential oils tested were selected from those which may have applicability in the field (eg., low-cost, high persistence, or from plants that could be grown in or near fruit fields in temperate regions). Due to its high volatility, KMS was assessed in a no-choice trapping assay. Our results are a first step towards selecting and further testing of repellents for developing a push-pull management strategy against D. suzukii.

Results
In Exp. 1, all essential oils, except white pine, showed a significant level of repellency to male D. suzukii (Fig. 1). Few male flies chose geranium, peppermint, citronella, or thyme oil-treated wicks throughout the duration of the experiment, whereas rosemary and eucalyptus oils did not repel male flies after 24 h. Eastern white cedar, balsam fir, white spruce, ginger, and lavender oils were significantly repellent after 24 h, but approximately 20% of flies settled on wicks treated with these oils.
There were few differences among essential oils in the percent of flies responding or dying during this experiment (Table 1). There were fewer responding female flies in the peppermint oil than geranium oil treatments in group one, and thyme oil caused higher male fly mortality than the other three oils in group three (Table 1).
In Exp. 2, peppermint oil remained 100% repellent to female flies for up to 6 d after application on the wick (Fig. 3). Repellency of geranium oil to female flies decreased as time since application increased (F 1,14 = 27.9, P < 0.001), decreasing to about 35% by 4 d post application (Fig. 3).
In Exp. 3, repellency of peppermint and geranium oils to male and female flies was dependent on concentration (peppermint, males: χ 2 = 48.9, P < 0.001, females: χ 2 = 26.6, P < 0.001; geranium, males: χ 2 = 29.9, P < 0.001, females: χ 2 = 65.6, P < 0.001) (Fig. 4). Concentrations required to repel 95% of flies were 1.1 times greater for males than females for both peppermint and geranium oils, and 1.3 times more peppermint than geranium oil was required to repel both male and female flies ( Table 2). Essential oil release rates. Volatility, measured as decrease in weight over 96 h, varied by essential oil (F 11,33 = 66.9, P < 0.001). Volatility was highest in tree-derived and rosemary essential oils, although volatility of Eastern white cedar oil was lower than for other tree oils (Table 3). Peppermint oil was least volatile, followed   closely by geranium, citronella, and lavender oils. Release rates of all oils tended to decline over time, although differences in release rates between the first 6 h and the last 72 h were much greater for some oils (e.g., thyme, lavender) than others (e.g., ginger, peppermint).    Table 2. Response of Drosophila suzukii flies when given a choice of a cotton wick with raspberry juice or a wick with juice and an essential oil. EC95 is the effective concentration required to achieve 95% repellency.

Discussion
In these experiments most essential oils deterred D. suzukii male and female flies from wicks containing attractive, fresh raspberry juice. Peppermint oil was particularly effective, preventing all but a few flies from contacting treated wicks and remaining 100% repellent for up to 6 d after it was applied to wicks. Peppermint oil consists mainly of two monoterpenoids: menthol and menthone 43 . Mint oils and their constituents were the best of 86 essential oil repellents tested against German cockroaches (Blattella germanica) 44 , and peppermint oils have shown strong and prolonged repellent action against flies, including mosquitoes 45 , house flies 30,31 and other nuisance flies 46 . Prolonged repellency of peppermint oil may be due to its consistently low release rate. It was the least  Table 3. Volatility as measured by total weight lost and release rates for 12 essential oils in 2 mL vials with 3 mm holes in lids held in a fume hood at 22 °C ± 1 °C. a Means followed by the same letter are not significantly different, Tukey's HSD, P > 0.05. volatile of all tested essential oils, losing only 1.2% of its weight over 96 h. However, in at least one study, after aging for 72 h peppermint oil at multiple concentrations was not more repellent than other essential oils to the housefly, Musca domestica L. 31 . Mint oils have also been reported as good adulticides and larvicides against insect pests 30 , and pennyroyal oil, Mentha pulegium L., caused high D. auraria mortality 39 . We did not find greater mortality of D. suzukii adults due to peppermint oil. There were fewer responding female flies in assays with peppermint oil than with geranium oil, suggesting that peppermint oil may negatively affect flies at a greater distance from its release point than geranium oil. Peppermint oil appears to be a good candidate for testing under field conditions, where persistence of repellency and distance of repellency from ripe fruit will be important factors determining the effectiveness and feasibility of using essential oils as part of a management strategy for D. suzukii.
Amounts of peppermint and geranium oils estimated to achieve 95% repellency of female and male flies were similar, ranging from 6.8-9.9 mg of oil per wick or 0.39-0.57 mg cm 2 −1 based on the wick surface area. Studies on third-instar larvae of small dipterans have found an LC 50 of 0.70 mg cm 2 −1 of peppermint oil for Camptomyia cortacalis (Diptera: Cecidomyiidae) 47 and an LD 50 of 2.1 and 1.1 μ L of pennyroyal and spearmint oils, respectively, for Drosophila melanogaster 48 . Konstantopoulou 43 found 100% mortality of D. auraria when eggs and adults were exposed to M. pulegium but stated only that 1-20 μ L of an essential oil was used. The estimated amount of peppermint oil on the exposed part of the wick required to achieve 95% repellency in our study was 4.5-6.7 μ L. It is difficult to estimate the actual dose experienced by flies as we do not know whether the wicking action and evaporation rate of raspberry juice changed and affected the oil concentration during the experiment. Distances at which oils are repellent and concentrations required to achieve repellency and deter oviposition in the field will need to be determined.
Thyme oil was unique among oils tested because it caused higher mortality of male flies than other oils, and it reduced the number of responding males and females to 56-65% and 75-81%, respectively, of that of flies responding in arenas with citronella, lavender, and rosemary (Trial 3). Thyme essential oil constituents, particularly the monoterpene thymol from the chemotype of Thymus vulgaris used in this experiment, have strong toxic effects on other insect pests, including southern green stink bug, Nezara viridula (L.), nymphs and adults 49 , larval tobacco cutworm, Spodoptera litura Fab. 50 , and third-instar gall midge C. cortacalis (Diptera: Cecidomyiidae) 47 . Toxicity of thymol has been shown in D. melanogaster to be related to interference in tyramine receptor cascades that are involved with cAMP and calcium at the molecular level 51 . It also binds to GABA receptors associated with chloride channels, disrupting the function of GABA synapses 52 . Excellent levels of repellency or deterrence using thyme have also been recorded against pest insects, including Western flower thrips adults from leaf discs 36 and Culex pipiens pallens Coquillet mosquitoes 53 . As with peppermint oil, thyme oil appears to be a good candidate for further testing, and thyme oil may soon be available as an insecticide in the European Union 24 .
Our results show that KMS is not repellent to D. suzukii flies but may have fumigant properties, particularly at high concentrations. At lower concentrations (Exp. 4), there was a small but significant increase in fly mortality from less than 25% in controls to about 40% at 5 g L −1 . At higher concentrations (Exp. 5), 10-15% of flies died outside traps at all concentrations, but mortality in traps reached 60% at 30 g L −1 , a significant increase from no mortality in traps at 0 and 5 g L −1 . KMS is highly volatile in solution, releasing half its weight in SO 2 and repelling H. axyridis within 1 min after the start of assays 42,54 . Sulfur dioxide has been shown to be acutely toxic to omnivorous leafroller (Platynota stultana Walshingham), providing protection to packed table grapes at low doses in combination with low temperature over long periods 55 . Although no concentration of KMS repelled flies from traps after 24 h, likely due to rapidly declining concentration, sufficient SO 2 was produced to kill flies, particularly within traps where SO 2 would have been more concentrated. It is also possible that once inside traps, flies died by contact with or ingestion of KMS on wicks. KMS may be more effective as a fumigant in enclosed spaces than as a repellent in open fields and could be investigated against D. suzukii larvae as a post-harvest alternative to methyl bromide 56 . This is the first time that essential oils have been shown to be effective in preventing fruit-infesting flies from contacting attractive host stimuli. Essential oils from the same plant species may vary in their constituent chemical composition due to production practices or plant material sources and thus affect their bioactivity. Using essential oils or their active components (see gas chromatograph results at www.aliksir.com for most of the essential oils used in these experiments) as deterrents for D. suzukii and alternatives to synthetic insecticides is desirable because they pose little or no risk to mammalian health or beneficial insects 23 . As many essential oils are permitted food additives, repellent essential oils may provide a safe means of protecting fruit from D. suzukii infestation applicable in both conventional and organic production systems.
Other compounds tested as D. suzukii deterrents include geosmin, 1-octen-3-ol, and butyl anthranilate in the laboratory; 1-octen-3-ol reduced infestation in fruit in the field by 40-50% 57,58 . Further research will be needed to compare these molecules with those in essential oils and to test blends of active compounds. Factors such as compound dose, persistence, volatility, availability and cost will be important, but determining effective dispersal mechanisms for field application of repellents will be crucial to success. Products that are phytotoxic or compromise fruit quality cannot be applied directly to ripening fruit, and small dispensers required near each bunch of fruit 58 may not be feasible on a large scale. Methods for dispensing insect pheromones (e.g., biopolymer flakes, rubber septa, aerosols) may be appropriate and adaptable for repellent dispersal. If repellents are to be used in the field, then they will likely need to be part of a push-pull management strategy to provide sufficient control. Mass trapping, attract-and-kill, or other 'pull' strategies also need to be further developed and integrated with other cost-effective management practices in order to achieve the high-level of control needed for this important pest 59 . Halton Hills, ON, Canada (43 N 34′ 43; 79 W 57′ 38″ ). Flies were kept in Plexiglas ® cages (26 × 26 × 26 cm), with mesh backings and sleeves for access, at 22-23 o C, 20-30% RH and 16:8 h L:D. Flies were provided with moist cotton batting and fed and reared on diet prepared by combining: water (4 L), agar (45 g), cornmeal (125 g), white sugar (200 g) and nutritional yeast (70 g) with propionic acid (17.7 mL) and methyl paraben (3.3 g) dissolved in 95% ethanol (33.3 mL). The mixture was boiled and then cooled before pouring into Petri dishes (9 mm diameter). Flies that were 5-8 d old were used for experiments. Flies were separated by sex and held without food for 20 h and without water for the final 2 h before the start of the experiments. New York) and 10 mL of raspberry juice was added to each vial. A second dental wick was treated with either an essential oil (2 mL acetone solution containing 30 mg of essential oil) or acetone only (2 mL) and placed in a fumehood for one hour to allow acetone to volatilize. Raspberry juice (2 mL) was applied to each wick before it was put in a vial; the wick rested on the wick already in the vial and protruded 2 cm above the rim of the vial. Raspberry juice was prepared by heating fresh, store-purchased red raspberries (400 g) mixed with distilled water (250 mL) and pectinase (1.5 g; Vineco International Products, St. Catharines, Ontario) at 45 o C for 1 h. The mixture was strained under vacuum through a paper coffee filter, and juice was stored overnight in a refrigerator. After each experiment, arenas and vials were washed with water and soap (Sparkleen, Fisher Scientific Co., Pittsburgh, Pennsylvania), air-dried and wiped or rinsed with hexane.

Essential oil choice bioassays. Essential oils tested in repellency bioassays
Arenas were arranged in a completely randomized design on a table covered with white paper. Three 60 W incandescent bulbs were hung 1 m above the table at 16:8 h L:D with photophase beginning at 0600 h. Trials began at 1100 h by placing plastic vials (70 mL) containing approximately 20 flies upright in the center of the container and removing lids so that flies could walk or fly from vials. The number of responding flies (those on treated or control wicks) and the number alive or dead in the arena and release vial were counted at 1, 6, and 24 h after flies were introduced. Experiment 1. Essential oils were tested in three trial periods. The first trial tested peppermint, geranium, ginger, eucalyptus, the second trial tested eastern white cedar, balsam fir, white spruce, white pine, and the third trial tested citronella, lavender, rosemary, thyme with female and male flies separately. Each oil was replicated six times for each sex, except peppermint, geranium and group 2 oils that were replicated five times for males due to lower than expected numbers of available flies.

Experiment 2.
Persistence of repellency to female flies was assessed by treating wicks 0, 1, 2, 4 d or 0, 1, 2, 4, and 6 d before the start of the experiment with geranium and peppermint oil, respectively. Wicks were held on aluminum foil in a fumehood at 22 °C ± 1 °C for aging. There were four geranium and five peppermint oil replicates. Experiment 3. The effect of geranium and peppermint oil concentrations on repellency to male and female flies was evaluated by treating wicks with 0, 0.6, 1.2, 6, and 30 mg of peppermint and geranium oil; 18 mg of geranium oil was also included in assays with females. Each peppermint and geranium concentration was replicated five times, except geranium concentrations were replicated four times for females because of the extra concentration and a shortage of female flies. Essential oil release rates. Release rates of essential oils were determined by measuring weight loss from 2 mL polypropylene vials (Fisher Scientific, Toronto, Ontario), each with a 3 mm hole in the lid. Vials were loaded with 400 μ L of an undiluted essential oil and placed in a fume hood at 22 °C ± 1 °C. Vials (three per oil) were weighed 0, 2, 6, 12, 24, 48, 72, and 96 h after loading. A no-choice trapping bioassay was conducted in arenas consisting of upright clear Plexiglas ® cylinders (31 × 8.8 cm) with ends covered by fine white mesh held in place at the top of the cylinder by an elastic band and glued to the base of the cylinder. There were two ventilation holes (5 cm) covered with fine white mesh 20.3 cm from the bottom of each cylinder. Traps were plastic jars (7.3 × 4.8 cm) with aluminum foil secured by an elastic band over the jar opening. A cut microcentrifuge tube (1.5 mL) was inserted through the middle of the foil. Two store-purchased red raspberries (6.9 ± 1.4 g) and a cotton wick that was treated with KMS solution (1.5 mL) or Scientific RepoRts | 6:21432 | DOI: 10.1038/srep21432 untreated (1.5 mL water) were placed in each trap. One trap was placed in each cylinder. Cylinders were arranged in a completely randomized design on a table under the same conditions as described above for choice bioassays.
Experiments 4 and 5. KMS was tested at four concentrations between 1-5 g L −1 with five replicates (Exp. 4). Because there was no repellent effect of KMS at these concentrations, Exp. 5 was conducted using concentrations of 5, 15, and 30 g L −1 with six replicates. For each experiment, 25-35 flies (approximately 1:1 sex ratio) were released into each cylinder. After 24 h, the number of trapped flies and the number of dead flies inside and outside traps were counted. Data analysis. For Exp. 1, numbers of D. suzukii on treated and untreated wicks were compared using goodness-of-fit G-tests. The pooled G-test statistic is presented as results of each essential oil for either male or female flies were consistent between replicates (heterogeneity G-test, P > 0.05) 60 . Analysis of variance (ANOVA) was used to test effects of essential oils by groups on the percent responding flies at 6 h (percent on treated wicks + percent on untreated wicks) and percent fly mortality at 24 h. Percentages were square-root transformed to normalize residuals. ANOVA was also used to evaluate differences between release rates of essential oils over 96 h. All means were separated using Tukey's HSD test.
For Exps. 2 and 3, a repellency index (RI) 31 was calculated at 6 h after the start of the experiment: (1) RI = (flies on treated wicks − flies on untreated wicks)/(flies on both wicks) Linear regression was used to evaluate geranium oil persistence, where Y was the repellency index and X was the age of the treated wicks in days. For male and female flies, probit analysis was used to generate concentrationrepellency regressions for peppermint and geranium oils. Concentrations that achieved 95% repellency (EC 95), confidence limits, slopes, and goodness-of-fit χ 2 vales were determined.
For Exps. 4 and 5, linear regression was used to assess repellency, where Y was the percent trapped flies or percent dead flies inside and outside traps and X was the KMS concentration (g L −1 ). A quadratic regression was also fit to the percent dead flies inside traps in experiment five. Significance of slopes (deviation from zero) and coefficients of determination (R 2 ) are reported for regressions.
JMP software 61 was used for ANOVAs, regressions, and PROBIT analyses. In all cases, α = 0.05.