Journal list menu

Volume 54, Issue 1 p. 170-177
Standard Paper
Free Access

Enhanced yeast feeding following mating facilitates control of the invasive fruit pest Drosophila suzukii

Boyd A. Mori

Boyd A. Mori

Department of Plant Protection Biology, Swedish University of Agricultural Sciences, Box 102, 23053 Alnarp, Sweden

Search for more papers by this author
Alix B. Whitener

Alix B. Whitener

Department of Entomology, WSU Tree Fruit Research and Extension Center, 1100 N. Western Avenue, Wenatchee, WA 98801, USA

Search for more papers by this author
Yannick Leinweber

Yannick Leinweber

Department of Plant Protection Biology, Swedish University of Agricultural Sciences, Box 102, 23053 Alnarp, Sweden

Search for more papers by this author
Santosh Revadi

Santosh Revadi

Department of Plant Protection Biology, Swedish University of Agricultural Sciences, Box 102, 23053 Alnarp, Sweden

Search for more papers by this author
Elizabeth H. Beers

Elizabeth H. Beers

Department of Entomology, WSU Tree Fruit Research and Extension Center, 1100 N. Western Avenue, Wenatchee, WA 98801, USA

Search for more papers by this author
Peter Witzgall

Corresponding Author

Peter Witzgall

Department of Plant Protection Biology, Swedish University of Agricultural Sciences, Box 102, 23053 Alnarp, Sweden

Correspondence author. E-mail: [email protected]Search for more papers by this author
Paul G. Becher

Paul G. Becher

Department of Plant Protection Biology, Swedish University of Agricultural Sciences, Box 102, 23053 Alnarp, Sweden

Search for more papers by this author
First published: 10 May 2016
Citations: 70

Summary

  1. The highly invasive spotted wing Drosophila Drosophila suzukii is a key pest of soft fruit and berries in Europe and North America, and development of control techniques is an urgent research challenge. Drosophila suzukii is widely associated with the yeast Hanseniaspora uvarum. Yeasts are symbionts of drosophilid flies and communicate with insects through volatile metabolites for spore dispersal. Accordingly, yeasts and behaviour-modifying chemicals produced by yeasts are prospective tools for environmentally sound insect management.
  2. We first bioassayed flight attraction, feeding and oviposition of D. suzukii females in response to H. uvarum yeast and blueberries, which are a preferred host fruit. We then investigated the combined effect of yeast and insecticide on adult female oviposition behaviour and mortality towards the development of a yeast-based control method.
  3. Following mating, attraction of female flies to blueberry and yeast odour cues was strongly enhanced. Yeast feeding significantly increased in mated females, while yeast did not increase oviposition on blueberries. This observation suggests that mated flies become attracted to yeast for feeding and to fruit for egg laying. A combined feeding–oviposition assay demonstrated different roles and interference between yeast and fruit stimuli: during the day after mating, females laid fewer eggs when yeast was available.
  4. The post-mating yeast-feeding response is an opportunity for the development of an attract-and-kill technique for population control of D. suzukii. Exposing flies to a blend of yeast and insecticide reduced oviposition and greatly enhanced adult fly mortality compared with an insecticide treatment alone.
  5. Synthesis and applications. Mated females are the key life stage for Drosophila suzukii population control. Egg-laying females perforate fruit skin and fungal infestations ensue, even when eggs and larvae are killed off by insecticide sprays. Behaviour-modifying chemicals, including yeast metabolites, enable environmentally safe insect management via manipulation of olfactory-mediated reproductive behaviour. Our results highlight that yeast and yeast semiochemicals hold potential for D. suzukii management and that response modulation to olfactory stimuli following mating is a vital element for the development of D. suzukii control methods. Yeast feeding is enhanced in mated D. suzukii females, and this change in post-mating behaviour can be exploited by an attract-and-kill strategy, combining a fly-associated yeast with an insecticide. Furthermore, using the D. suzukii yeast mutualist, H. uvarum, may reduce non-target effects and increase species specificity, which further contributes to the development of an efficient and safe control method.

Introduction

The search for alternative insect management strategies in food crops is urgent (Chandler et al. 2011). Only a few insecticides are available to cope with native and an increasing number of alien pests. Moreover, insecticide impact on biodiversity and ecosystem services, in addition to residues in food and water, is of concern (e.g. Goulson 2013; Hallmann et al. 2014; Rundlof et al. 2015; Stanley et al. 2015).

Sustainable control techniques have been established with behaviour-modifying olfactory signals (semiochemicals), used by insect herbivores to locate food plants and mates (Ridgway, Silverstein & Inscoe 1990; El-Sayed et al. 2006; Witzgall, Kirsch & Cork 2010; Bruce & Pickett 2011). Olfactory responses, however, change with physiological state, especially after feeding or mating (Carvalho et al. 2006; Anton, Dufour & Gadenne 2007; Vargas et al. 2010; Saveer et al. 2012; Lebreton et al. 2015). The consequences of olfactory response modulation for population control with semiochemicals, which depends on behavioural manipulation of individual insects, have received little attention.

Areawide applications of semiochemicals rely mainly on air permeation with sex pheromones, which leads to disruption of the mate-finding behaviour of male insects (El-Sayed et al. 2006; Witzgall, Kirsch & Cork 2010). In order to supplement pheromone-based control methods, attempts have been made to manipulate the behaviour of mated, egg-laying females. Research has focused on plant volatiles mediating host-plant finding and oviposition, but direct use of plant compounds has not been brought to wide practical application.

Micro-organisms link insects to their food plants (Janson et al. 2008; Biere & Bennett 2013; Hansen & Moran 2014), and microbial volatile cues contribute to host-plant finding, in addition to plant compounds (Ganter 2006; Witzgall et al. 2012; Davis et al. 2013). Consequently, semiochemicals of microbial origin are emerging as promising tools for insect control, especially in combination with insect pathogens or insecticides in attract-and-kill formulations (Knight & Witzgall 2013; Knight, Basoalto & Witzgall 2015; Knight et al. 2015).

Insect–microbe interactions have been widely studied in drosophilid flies, with particular focus on yeasts (Dobzhansky et al. 1956; Heed et al. 1976; Becher et al. 2012; Buser et al. 2014; Christiaens et al. 2014; Günther et al. 2015). This knowledge can now be employed to design control strategies against spotted wing Drosophila Drosophila suzukii (Matsumura) (Diptera: Drosophilidae), which emerged as a major pest of soft fruit and berries soon after its invasion of Europe and North America in 2008. In contrast to the taxonomically close Drosophila melanogaster Meigen (Diptera: Drosophilidae), D. suzukii females possess a sclerotized and serrated ovipositor to penetrate ripening fruit for egg laying (Hauser 2011; Lee et al. 2011a,b; Walsh et al. 2011; Calabria et al. 2012; Atallah et al. 2014).

The abundant occurrence of the yeast Hanseniaspora uvarum (Niehaus) in D. suzukii-infested fruit, together with strong olfactory attraction, suggests a specific association (Hamby et al. 2012; Scheidler et al. 2015). Field studies have confirmed D. suzukii attraction to fermentation odours, and yeast volatiles are currently employed in trap lure and attract-and-kill formulations (Cha et al. 2012; Landolt, Adams & Rogg 2012; Iglesias, Nyoike & Liburd 2014; Knight et al. 2015). Notwithstanding, D. suzukii lays eggs in ripening fruit, preceding yeast colonization and fermentation. In view of the apparent discrepancy between attraction to yeast and oviposition on fresh fruit, it has been hypothesized that D. suzukii females use yeast volatiles to locate adult food sources and fresh fruit volatiles to locate oviposition sites (Cha et al. 2012; Keesey, Knaden & Hansson 2015; Revadi et al. 2015a).

We therefore investigated whether mating modulates upwind attraction, feeding and oviposition response of D. suzukii to fruit and yeast volatiles. The results show a mating-modulated yeast attraction and feeding response, which is fundamental for the development of attract-and-kill formulations against spotted wing Drosophila.

Materials and methods

Flies

A laboratory colony of D. suzukii was established from insects collected in Northern Italy (San Michele all'Adige, Italy) in 2011. Flies were reared on a standard Bloomington diet at 22–24 °C, 35–60% RH and under a 12 : 12 h L : D (light : dark) photoperiod. Newly eclosed, unmated flies were collected from rearing vials twice daily, anaesthetized with CO2 and sorted by sex. Flies were 3–7 days old when used in experiments. In all bioassays, mated females were obtained via controlled matings in which unmated males and females of the same age were placed in a vial for 1 h at the start of the photophase. Mating pairs were transferred in copula to another vial until mating ceased (Revadi et al. 2015b).

For insecticide tests, 3- to 7-day-old flies were obtained from a laboratory colony of D. suzukii established from insects collected in central Washington (Orondo, Washington, USA) in 2014. The colony was maintained on a standard Drosophila diet (formula 4–24; Carolina Biological Supply Company, Burlington, NC, USA) in a climate chamber at 22 °C, 50–60% RH, under a 16:8 h L : D photoperiod.

Yeast and Fruit

Hanseniaspora uvarum (CBS 2570) was obtained from Centraalbureau voor Schimmelcultures (Utrecht, the Netherlands). Suspensions of H. uvarum were grown in 125-mL culture flasks in 50 mL of liquid minimal medium (Merico et al. 2007) during 24 h in a shaking incubator (25 °C, 260 RPM). After 24 h, optical density was ~1·5–1·8 at λ = 595 nm and cell counts were adjusted to ~1·5 × 108 cells mL−1.

Frozen organic blueberries (Vaccinium myrtillus L., Coop, Malmö, Sweden) were thawed and used for wind tunnel bioassays, and locally grown blueberries (V. corymbosum L., Alnarp, Sweden) with the stem attached were used in oviposition experiments. Berries were washed twice with deionized water before experiments.

Wind Tunnel Bioassay

A wind tunnel bioassay (Becher et al. 2010) was conducted to determine the effect of mating status on the behavioural response of D. suzukii females to volatiles from a suspension of H. uvarum (50 mL) or thawed, whole blueberries (100 g) (= 40–45). Females were placed individually in open-ended cylindrical glass release vials (2·5 cm × 15 cm). One vial end was plugged with a dry cotton ball and one with a moist cotton ball to provide females with a water source. Females were tested during the photophase, 4 to 5 h after mating. Each fly was exposed to a single odour source for 5 min; attraction was scored as landing at the odour source. Each day, glass vials were washed with soap and water, rinsed with 96% ethanol and heated to 200 °C for 8 h.

A generalized linear model (GLM) with a binomial error distribution was used to determine the effect of mating on the response of females to H. uvarum yeast and blueberry odour volatiles. The number of flies landing at the odour source was the response variable, and mating status, odour source and their interaction were explanatory variables. The interaction was not significant and removed from the final model. All analyses were calculated with R 3.1.2 (R Core Team 2014).

Feeding Bioassay

A modified capillary feeding (CAFE) assay (Ja et al. 2007) was used. The chamber was made from a standard Drosophila vial (polystyrene, 25 mm × 95 mm; VWR International, Vienna, Austria) filled with 10 mL of deionized water to maintain humidity. A foam platform (30 mm × 5 mm) 0·5 cm from the bottom of the vial prevented direct access to the water. A foam plug (32 mm × 15 mm) capped the tube, and a calibrated glass capillary (5 μL; Karl Hecht KG, Sondheim/Rhön, Germany) was inserted through the foam plug. Capillaries were filled with a H. uvarum suspension. Identical chambers without flies were used to measure evaporation, which was subtracted from the experimental readings. Females were transferred into individual CAFE chambers 4 h after mating (= 20) or the equivalent time period for unmated flies (= 19). The amount of yeast suspension consumed after 2 h was compared with a Wilcoxon rank-sum test.

Oviposition Bioassay

A two-choice bioassay was conducted to determine the effect of yeast on oviposition preference of D. suzukii females (= 24). Two blueberries were placed on a moistened filter paper at the bottom of a plastic cylindrical cage (9·5 cm × 4 cm). One hour prior to the experiment, a berry was inoculated by placing a 3-μL droplet of H. uvarum suspension directly on top of the skin adjacent to the stem. As a control, a 3-μL droplet of minimal media was added to the other berry. Four hours after mating, females were placed into each cage and enclosed with a fine mesh lid. Females were given 23 h to oviposit, after which eggs were counted. The number of eggs laid per berry was compared with a Wilcoxon rank-sum test.

Combined Feeding and Oviposition Bioassay

To determine if yeast influences oviposition on nearby blueberries, we modified the CAFE bioassay to include an oviposition substrate (OviCAFE). The CAFE chamber was set up as previously described with the addition of a blueberry on the platform. Capillaries were filled with either 5 μL of H. uvarum suspension or minimal medium. Four hours after mating, individual females were transferred into the OviCAFE chambers (= 19). Twenty-three hours after mating, consumption was measured and the number of eggs laid counted. Identical chambers without flies were used to account for evaporation from the capillaries, and these measurements were subtracted from the experimental readings. A Wilcoxon rank-sum test evaluated the differences in consumption of minimal medium and yeast, as well as the total number of eggs laid per female over 23 h.

Insecticide Bioassay

To determine the combined effect of yeast and insecticides, an oviposition and mortality bioassay was conducted using cherries Prunus avium L., va. ‘Sweetheart’ at the Tree Fruit Research and Extension Center, Washington State University, Wenatchee, WA, USA. Cherries and leaves were picked from an untreated orchard in Douglas Country (WA, USA) and washed twice with deionized water before use.

Spinosad (22·5% active ingredient (AI), Entrust SC; Dow AgroSciences, Indianapolis, IN, USA), an insecticide based on a bacterial metabolite and approved for organic growing, was used in all trials. Treatments included spinosad alone (5 mg AI L−1), spinosad combined with H. uvarum (5 mg AI L−1), H. uvarum alone and a distilled water control. The insecticide dilution (0·0005% AI, 5 mg AI L−1) made in distilled water or yeast suspension was based on a previously determined LD50 (results not shown). H. uvarum was grown as described above.

The bioassay was conducted in 946-mL plastic bowls (Sabert, Sayreville, NJ, USA) (= 5 per treatment), each lined with three cherry leaves. In each bowl, three 20-μL droplets of a single treatment (water control, yeast, spinosad or yeast–spinosad blend) were added to each leaf (9 droplets bowl−1). As an alternative food source, 2 mL of 10% honey agar was placed on the lid of each arena. Five cherries were suspended in a circular pattern by threading the stems through holes in the bowl lids and secured with hot-melt glue. For ventilation and introduction of mated females, a 1-cm hole in the centre of the lid was covered with tape (Micropore; 3M Healthcare, St. Paul, MN, USA). Mated females were briefly anaesthetized with CO2 and 10 flies were added to each arena. To minimize possible differences in exposure times, flies were added simultaneously to four bowls (blank, yeast, spinosad, yeast–spinosad blend). The four bowls were then treated as an experimental block. After fly introduction, bowls were held in a growth chamber at ca. 21 °C, 16:8 L : D photoperiod for 24 h. After 16 h, the number of eggs laid was evaluated by briefly anaesthetizing flies, removing the lids with cherries and placing a new lid (without fruit) on the bowl. Mortality was evaluated 24 h after fly introduction.

A generalized linear mixed-effects model (GLMM) with a Poisson distribution was used to evaluate the differences in the number of eggs laid per treatment, with block specified as a random factor. A GLMM with a binomial error distribution was used to evaluate the differences in mortality of flies in each treatment, with block specified as a random factor (lme4 package) (Bates et al. 2014). Multiple comparisons were calculated with a Tukey's honest significant difference (HSD) test.

Results

Wind Tunnel Bioassay

In view of the attraction of D. suzukii to fermentation volatiles in laboratory and field studies (Cha et al. 2012; Landolt, Adams & Rogg 2012; Iglesias, Nyoike & Liburd 2014) and the association between D. suzukii and H. uvarum (Hamby et al. 2012; Scheidler et al. 2015), we investigated attraction of D. suzukii to H. uvarum and blueberry volatiles in a wind tunnel bioassay. More mated than unmated females responded to yeast and blueberry odours (χ2 = 7·06, d.f. = 1, < 0·01), while attraction to yeast and berry odours did not differ significantly (χ2 = 1·95, d.f. = 1, > 0·05) (Fig. 1a).

Details are in the caption following the image
Effect of mating status on attraction (a) and feeding (b) of D. suzukii females. (a) Significantly more mated (+) than unmated (−) females were attracted by upwind flight and landed at blueberry V. myrtillus and yeast H. uvarum odour sources. Different letters indicate significant differences based on mating status (GLM < 0·01). (b) Mated (+) females consumed significantly more yeast H. uvarum suspension than unmated (−) females during 2 h (< 0·01). [Colour figure can be viewed at wileyonlinelibrary.com

Feeding and Oviposition Bioassays

We next asked whether mated females were attracted to yeast for feeding or oviposition. Mated females consumed significantly more H. uvarum yeast suspension than unmated females (W = 300, < 0·01) (Fig. 1b). In contrast, yeast did not influence oviposition rate (W = 270·5, = 0·73) (Fig. 2a). A slight increase in the number of eggs laid on yeast-inoculated berries may be due to initial attraction to yeast for feeding, followed by oviposition.

Details are in the caption following the image
Effect of yeast, H. uvarum, on oviposition and feeding in D. suzukii females. (a) Mean number of eggs (+SE) laid on blueberries without (−) and [berries inoculated] with yeast (+) (= 0·73). (b) Mean amount of H. uvarum suspension (+SE) consumed and mean number of eggs laid (+SE) by individual females in the vicinity of a blueberry after 23 h, when yeast was added medium (+) or absent (−). Different letters indicate significant differences in feeding (< 0·0001) and oviposition (< 0·05). [Colour figure can be viewed at wileyonlinelibrary.com

We further investigated the interaction between yeast feeding and oviposition, with yeast available near blueberries. Females fed significantly more on a H. uvarum yeast suspension than on minimal medium (W = 37, < 0·0001); females with access to yeast suspension laid significantly fewer eggs on blueberries than females with access to minimal medium over 23 h (W = 263·5, < 0·05) (Fig. 2b).

Insecticide Bioassay

The combined results indicate a trade-off between yeast feeding and oviposition after mating. We therefore examined the effect of yeast for control of D. suzukii with or without spinosad insecticide. Hanseniaspora uvarum yeast treatment alone reduced oviposition, without significantly increasing mortality (Fig. 3). Adding spinosad to yeast further reduced the number of eggs laid by 36·8% (χ2 = 10·72, d.f. = 3, < 0·05) (Fig. 3). Moreover, the combination of H. uvarum yeast and spinosad significantly increased mortality of D. suzukii females over spinosad alone by 26% (χ2 = 62·49, d.f. = 3, < 0·0001) (Fig. 3).

Details are in the caption following the image
Effect of insecticide (spinosad 5 mg AI L−1) with (+) or without (−) the addition of yeast H. uvarum on oviposition and mortality in D. suzukii females. (a) Mean number of eggs laid per female (+SE) on 5 cherries during 16 h, and mean adult mortality (+SE) during 24 h, in presence (+) and absence (−) of yeast and insecticide. Different letters indicate significant differences between the treatments (GLMM followed by Tukey's HSD test < 0·05). [Colour figure can be viewed at wileyonlinelibrary.com

Discussion

Co-evolution between higher plants and phytophagous insects is linked to associated micro-organisms and their effect on host plant chemistry and physiology (Farrell et al. 2001; Janson et al. 2008; Gibson & Hunter 2010; Boone et al. 2013; Douglas 2013; Casteel & Hansen 2014; Hansen & Moran 2014). The dialogue between insects and microbes, providing an interface with food plants, is therefore ecologically fundamental. Microbial volatile signatures, in addition to plant volatiles, contribute to host plant choice for adult and larval feeding and oviposition (Becher et al. 2012; Davis & Landolt 2013; Davis et al. 2013; Andreadis, Witzgall & Becher 2015).

We investigated how mating affects olfactory-mediated behavioural responses of female D. suzukii to host fruit and yeast. Mating increased attraction to fruit and yeast volatiles and triggered yeast feeding (Figs 1 and 2), supporting the hypothesis that adult females use damaged fermenting fruit, and the yeast contained therein, for feeding (Cha et al. 2012, 2014). Mating did not modulate or differentially affect D. suzukii attraction to fruit and yeast odours (Fig. 1). In contrast, mating induced a behavioural switch from feeding to oviposition in females of the cotton leafworm moth. This post-mating behavioural change was paralleled by an attraction modulation from floral to green leaf odours, signalling food and oviposition sites, respectively (Saveer et al. 2012).

Mating-induced yeast feeding in Dsuzukii is comparable to Dmelanogaster, where consumption of protein-rich and salt-containing food is caused by the transfer of a male sex peptide (SP) during copulation (Carvalho et al. 2006; Vargas et al. 2010; Lee, Kim & Min 2013; Walker, Corrales-Carvajal & Ribeiro 2015). When D. melanogaster SP was applied exogenously to unmated D. suzukii females, it elicited rejection in the presence of courting males and increased egg laying. D. suzukii SP had a similar effect when applied to D. melanogaster females, indicating a similar function (Schmidt et al. 1993).

Yeast intake after mating is likely to supply nutrients for egg development (Drummond-Barbosa & Spardling 2001; Ribeiro & Dickson 2010), which explains the trade-off between oviposition and feeding noted in our combined feeding–oviposition assay. Although egg laying during 24 h was reduced in females with access to yeast (Fig. 2), nutrient intake through yeast feeding could increase long-term egg production and fitness. In D. melanogaster, the amount of yeast in the diet is positively correlated with the number of eggs produced (Chippindale et al. 1993); increased egg laying results mainly from nutrient intake, since very little stored reserves are redirected for egg production (Simmons & Bradley 1997).

Surprisingly, yeast on the berry surface did not significantly influence D. suzukii oviposition. Fruit is a poor protein source and development of drosophilid flies is constrained when reared on yeast-free diets (Loeb & Northrop 1916; Becher et al. 2012; Hardin, Kraus & Burrack 2015), which suggests that D. suzukii larvae likely encounter naturally occurring yeast in host fruit (Jaramillo, Mehlferber & Moore 2015). Yeast is present in the alimentary canals of adult flies (Hamby et al. 2012) and known to survive the gut passage (Starmer & Fogleman 1986; Ganter 1988; Starmer, Peris & Fontdevila 1988). Fruit is therefore likely inoculated by ovipositing females through faecal deposition, regurgitation or by mere physical contact (Bakula 1969; Gilbert 1980).

Sustainable control strategies for D. suzukii are not yet available. The fly is among the most economically important plant-feeding insects world-wide, and inflicts extensive damage in soft fruit and berry crops. In a 400-ha fruit-growing area in Northern Italy, the damage was in the order of 3 million Euros in 2011 (Cini, Ioriatti & Anfora 2012). Growers are therefore compelled to abandon integrated management programmes in favour of calendar-based spraying of all available insecticides (Beers et al. 2011; Dreves 2011; Goodhue et al. 2011; Lee et al. 2011a; Cini, Ioriatti & Anfora 2012; Knight et al. 2015).

An increased post-mating olfactory attraction response, which aligns with increased yeast feeding (Figs 1 and 2), has consequences for population monitoring employing fermentation or fruit volatiles (Cha et al. 2012; Landolt, Adams & Rogg 2012; Abraham et al. 2015). The mating status of trapped females, which is decisive for timing of control measures, has not been established. Our results suggest that both yeast and fruit volatiles attract more mated than unmated females. Lure composition, and particularly the use of synthetic fruit vs. yeast volatiles, may have an effect on the ratio of mated to unmated trapped females. Our flight tunnel tests showed a tendency for unmated flies to preferentially respond to yeast instead of fruit volatiles (Figs 1 and 2).

Another urgent research goal is the design of a more species-specific lure, since current lures attract a wide range of flies and taxonomic determination of D. suzukii among other trapped drosophilid flies is tedious (Burrack et al. 2015). The association with H. uvarum (Hamby et al. 2012) is not species-specific, and many other drosophilid flies are also attracted (Palanca et al. 2013; Lam & Howell 2015; Scheidler et al. 2015); however, D. suzukii is attracted to fruit at an earlier phenological stage than most other drosophilids, and chemical and behavioural analysis of volatiles signalling fruit in an unripe or ripening stage may lead to the design of more specific lures.

A principal finding of this work is that D. suzukii females increase feeding at the expense of oviposition shortly after mating (Fig. 2). This behavioural response modulation can be exploited to augment the efficacy of insecticide treatments by combining insecticide and yeast into attract-and-kill formulations. Mortality increased and females laid fewer eggs when they were exposed to the bacteria-derived insecticide spinosad together with H. uvarum yeast (Fig. 3). Attract-and-kill technology has previously been investigated for control of D. suzukii, by blending sugar and yeast feeding stimulants with insecticide (Knight, Yee & Hilton 2013; Cowles et al. 2015; Knight et al. 2015). The addition of sugar and Saccharomyces cerevisiae or the yeastlike fungus Aureobasidium pullulans enhanced the effect of two insecticides, spinosad and cyantraniliprole, on adult mortality and survival of eggs and larvae (Knight et al. 2015), and is in line with our findings. Interactions of the orchard microbiome with spray applications of yeasts, or other insecticide and fungicide sprays, should be the subject of future investigations.

Attract-and-kill formulations reduce crop contamination through a reduction in insecticide application rates, compared with cover sprays. In addition, species-specific attractants avoid non-target effects on natural enemies or pollinators. The addition of yeast adjuvants may also enhance less effective insecticides (Beers et al. 2011) and mitigate insecticide resistance by making different classes of insecticides available.

We conclude that mating in Dsuzukii females increases host fruit attraction and yeast feeding, which translates into an opportunity for the development of an attract-and-kill control technique. Our results underscore that the physiological state of individual insects and resulting behavioural response modulation is a decisive element of olfactory-mediated population management strategies. This concept is substantiated by another fundamental aspect of yeast-induced behaviour, which concerns unmated insects, wherein the presence of yeast enhances mating (Gorter et al. 2016). The interaction between sex and food olfactory cues adds yet another degree of freedom to future work on behavioural manipulation for population control.

Acknowledgements

This work was funded by Carl Tryggers Stiftelse för Vetenskaplig Forskning (Stockholm), the Swedish University of Agricultural Sciences (LTV Faculty), the Linnaeus grant ‘Insect Chemical Ecology, Ethology, and Evolution’ IC-E3 (Formas, SLU) to PW, Formas (grants 2011-390 and 2015-1221) to PB, the Washington State Department of Agriculture (grant K1276), Washington Tree Fruit Research Commission (grant CH-14-106), and the Washington Commission on Pesticide Registration (grant PN1537) to EHB.

    Data accessibility

    Data available from the Dryad Digital Repository http://dx.doi.org/10.5061/dryad.8qj84 (Mori et al. 2016).