Predation on Drosophila suzukii within Hedges in the Agricultural Landscape

Simple Summary Drosophila suzukii is an invasive species that feeds and reproduces on various cultivated and wild fruits and moves between agricultural and semi-natural habitats, such as hedges and forest patches. These semi-natural habitats are known to harbor a diverse community of natural enemies of pests. When we exposed D. suzukii pupae in dry and humid hedges, we found that on average 44% of them were predated within four days. The most common predators in the hedges were earwigs, spiders, and ants. Using a molecular assay that detects the DNA of D. suzukii in the gut of predators, we could show that 3.4% of the sampled earwigs, 1.8% of the spiders, and one predatory bug had fed on D. suzukii. This small proportion may be due to methodological constraints. However, the overall predation rate helps to reduce D. suzukii populations, in particular in hedges that are scarce of host fruits. Abstract The invasive Drosophila suzukii feeds and reproduces on various cultivated and wild fruits and moves between agricultural and semi-natural habitats. Hedges in agricultural landscapes play a vital role in the population development of D. suzukii, but also harbor a diverse community of natural enemies. We investigated predation by repeatedly exposing cohorts of D. suzukii pupae between June and October in dry and humid hedges at five different locations in Switzerland. We sampled predator communities and analyzed their gut content for the presence of D. suzukii DNA based on the COI marker. On average, 44% of the exposed pupae were predated. Predation was higher in dry than humid hedges, but did not differ significantly between pupae exposed on the ground or on branches and among sampling periods. Earwigs, spiders, and ants were the dominant predators. Predator communities did not vary significantly between hedge types or sampling periods. DNA of D. suzukii was detected in 3.4% of the earwigs, 1.8% of the spiders, and in one predatory bug (1.6%). While the molecular gut content analysis detected only a small proportion of predators that had fed on D. suzukii, overall predation seemed sufficient to reduce D. suzukii populations, in particular in hedges that provide few host fruit resources.


Introduction
The invasive, frugivorous Drosophila suzukii Matsumura (Diptera: Drosophilidae) oviposits into ripening cherries, berries, and grapes, causing large economic damage [1,2]. Besides these agricultural crops, D. suzukii infests more than 100 species of wild fruits, many of them commonly found in agricultural hedges [3][4][5]. Therefore, this highly mobile species may use fruits in hedges for reproduction when host crops are not available. Hedges in the agricultural landscape may consequently provide a source for the re-infestation of crops after pest management interventions. Drosophila suzukii further uses hedges and woodland as a refuge during unfavorably hot and dry weather in summer [6] and as an overwintering habitat [7]. In addition, other polyphagous pests, for example aphids, are known to utilize the diverse resources and shelter afforded by semi-natural habitats [8,9]. Several studies have shown that nearby hedges and forest promote D. suzukii in crops [10][11][12]. In a recent study, it was observed that a greater proportion of non-cropping habitat promoted larger populations of D. suzukii, whereas fewer generalist predators were captured in those landscapes [13]. However, many natural enemies that provide biological control of agricultural pests also respond positively to landscape complexity, such as semi-natural habitats between agricultural plots [14,15]. Hedges, for example, not only connect natural and semi-natural habitats with crops [16], but they can also be the source of biocontrol functions in agricultural landscapes [17] by satisfying natural enemy needs not covered by crops [18]. For instance, the microclimate within hedges is buffered compared to open fields and therefore offers shelter during adverse climatic conditions [17,19]. Hedges may also form refuges for natural enemies, as they are usually not sprayed with insecticides [20]. Furthermore, diverse hedges can provide food and alternative hosts when target pests are scarce in the crop [21].
Thus, hedges support a diverse fauna of natural enemies, among them generalist predators such as carabids, staphylinids, earwigs, spiders, and predatory bugs [17,22]. It has been shown that these generalist predators can provide baseline control of agricultural pests [23]. Some of them are known to prey on D. suzukii and may thus reduce its abundance [22,24,25]. We aimed to gain further insight into the role of generalist predators in hedges in agricultural landscapes for the control of D. suzukii. We exposed pupae of D. suzukii in dry and humid hedges at different locations in Switzerland during the growing season (June-October) and determined predation rates, assessed the community of arthropod predators within these hedges, and analyzed the gut content of captured predators with D. suzukii-specific primers in order to confirm their feeding on this prey.

Field Sites
Ten hedges embedded in the agricultural landscape were chosen at five different locations in Switzerland. Hedges were selected according to the known presence of D. suzukii from surveys conducted in previous years and all hedges contained a variety of potential host plants of this vinegar fly. One humid and one dry hedge were assessed at each location to cover a wide diversity of predator species. Dry hedges were sun exposed, whereas humid hedges were at shady locations, often close to a water course. Locations and plant species of each hedge are shown in Table 1. Cornus spp., Hedera helix, Ligustrum vulgare, Rubus spp., Sambucus nigra * Only 15 instead of 30 pupae were exposed due to rearing limitations.

Exposure of D. suzukii Pupae in the Field
Laboratory rearing of D. suzukii were established in the infrastructures of the authors' affiliations, at Agroscope and FiBL. Adult flies were kept on an artificial banana-based diet (for a detailed recipe, see Boycheva et al. [26]) for oviposition. The diet containing eggs was replaced every 2-3 days, stored in ventilated jars, and kept for preparation of sample pupae or emergence of adult flies. For sample preparation, tissue paper was placed into the jars with last instar larvae for 24 h. During this time, larvae crawled onto the paper and pupated. The paper was then cut to contain the desired number of pupae for experiments. Rearing was conducted in climate chambers at 22 ± 1 • C, 70-75% RH and 16:8 L:D.
Exposure of pupae in the hedges took place at each location simultaneously in both hedge types at intervals of about one month between June and October 2019, accounting for four sampling periods: First: 17 June-4 July; Second: 5 July-26 July; Third: 12 August-2 September; Fourth: 17 September-18 October (Table 1). Three sets of thirty pupae per exposure period and per hedge were prepared. The samples were placed on the bottom of small plastic Petri dishes (5 cm diameter; Conthey, Reckenholz, Wädenswil) or plastic cylinders (10 cm diameter; Changins, Frick) and enclosed in a metal grid (1.5 cm diameter mesh size) to exclude larger predators such as mice. One sample was clamped between two branches at a height of 1.5 to 1.8 m and the two others were installed on the ground. One of the latter two served as a control and was protected from predation by a ventilated plastic jar (0.8 L). Care was taken to place the samples in locations protected from rain and direct sunlight. The samples were exposed in the hedges for four days, but the period was prolonged or shortened on a few occasions to two or five days in the case of cold weather, and thus low predator activity, or to avoid the emergence of adult flies when high temperatures prevailed. After field exposure, the samples were collected and the pupae were counted and classified into: Intact pupae, damaged pupae and missing pupae. The samples were then put into plastic jars and transferred into climate chambers to check for the capacity of collected pupae to complete their development and allow adults to emerge.

Collection of Arthropod Predators
Predators were captured at each location in both hedges at several meters distance from the exposed pupae at intervals of about one month between June and October 2019 ( Table 1). Predators were collected by three different trap types: pitfall traps, i.e., vertically buried soil traps filled with vegetal material (a 10 cm diameter PVC tube completed with a plastic bottle (200 mL) and a funnel connecting the top of the soil with the top of the bottle, two traps per hedge), bamboo pipes (0.5-1.0 cm diameter and 10-15 cm long, five pipes per hedge), and corrugated cardboard strips (band of 2 × 30 cm wrapped around a branch, five strips per hedge). Traps were inspected 12 and 24 h after deployment and then removed until the following scheduled collection. Additionally, predators were collected by beating the shrubs and gathering the fallen arthropods from a net placed below. Approximately 15-30 min of beating were invested per hedge and sampling date. Each predator was directly trapped into an Eppendorf tube (1.5 mL) and put on ice until reaching the laboratory where the samples were stored at −20 • C. Predators were morphologically determined to family or order level while being kept on crushed ice. Spiders were imaged with a high-resolution 4K numeric microscope camera, series VHX-7000 (Keyence, Osaka, Japan) and determined to family level using the photographs.

Molecular Gut Content Analysis
Field sampled predators were subjected to analyses using a PCR-based approach based on the Dro-suz-S390 and Dro-suz-A380 primers (Eurofins Genomics, Ebersberg, Germany), and a modified protocol based on Wolf et al. [22]. Although the whole body of the predator was used in the analyses, we refer to common terminology and use the term molecular gut content analyses for this approach. The used primers specifically target a 171 bp sequence of the COI gene from D. suzukii. To avoid any contamination of the predators' body surface with D. suzukii DNA, samples were washed with bleach and water prior to analysis [27,28]. To each 1.5 mL tube containing a single predator, 1 mL bleach 1-1.5% solution (1 part 10-15 % sodium hypochlorite (VWR Chemicals, Radnor, PA, USA) 9 parts Milli-Q water, and Tween 20 (0.1%) was added and the tubes were then shaken for 30 s. They were then rinsed twice with 1 mL Milli-Q water and shaken for another 15 s.

DNA-Extraction
The whole predator body was added to 180 µL TES buffer (0.1 M TRIS, 10 mM EDTA, 2% SDS, pH 8) and 20 µL Proteinase K (20 mg/mL) and crushed manually with sterilized pipette tips. The samples were then centrifuged for 5 min at 17,000× g and incubated overnight on a rocking platform (500 rpm; Hettlich Benelux, Geldermalsen, The Netherlands) at 56 • C. DNA was extracted with the DNeasy blood & tissue Kit (Qiagen, Hilden, Germany) following the manufacturer's instructions, except for the lysis step described above and the elution with 1× TE buffer. To determine the samples' DNA concentrations, the Nanodrop 1000 Spectrophotometer (Thermo Fischer Scientific, Waltham, MA, USA) was used. Each sample was adjusted to the same concentration of DNA (5 ng/µL) using the PIPETMAX ® device (Gilson, Lewis Center, OH, USA).

Gel Electrophoresis
To determine presence of D. suzukii in a sample, 3 µL of the PCR product were analyzed on a 2% agarose gel. Visualization of the PCR product was performed with a UV analyser Quantum (Vilber, Collégien, France). A sample was defined as containing D. suzukii DNA if the PCR product of 171 bp was visible. To obtain a positive control, 5 earwigs were collected in the field and then kept individually in petri dishes (5 cm) for 24 h with a humid cotton pad only. The next day, they were fed two D. suzukii pupae and were observed during 3 h. Finally, their gut content was analyzed as described above. The DNA isolated from an earwig that had eaten both D. suzukii pupae and thus produced a strong band on the gel was kept and used as a positive control. Water instead of DNA was added for the negative control. To exclude false-positives, all positive samples were re-analyzed. Only samples that tested positive twice were considered to have fed D. suzukii.

Statistical Analyses
The proportion of predated pupae and the number of predators per group were analyzed with generalized linear mixed models using the R-package lme4 [29] in R version 4.0.1. [30] The model for the analysis of the proportion of predated pupae used binomialdistributed errors, the model for the number of predators used Poisson-distributed errors. The models used the fixed effects hedge type (dry and humid) and exposure period (1)(2)(3)(4). The model analyzing predation further included the fixed effects exposure place (control, ground and branch) and exposure time (2-5 days). The model analyzing the number of predators additionally included the fixed effect predator group. Non-significant interactions and the non-significant variable exposure time and its quadratic term were removed from the models in a stepwise process. The random effect location explained a significant proportion of the variance only in the model for the analysis of the proportion of predated pupae. A random effect observational level was included in both models because of overdispersion. For better convergence, the models used the optimizer "bobyqa" instead of "Nelder-Mead" for the second phase. Tukey post-hoc multiple comparisons were performed using the R-package lsmeans [31]. Model assumptions were validated.
The composition of predator groups and Araneae families was analyzed with transformation-based redundancy analyses with the R package vegan [32]. Data was Hellinger transformed. The models used the describing variables hedge type (dry and humid) and sampling period (1-4) and the random effect location. Because of unbalanced data from different dates and sites, permutation tests used type III sum of squares. The number of permutations was 999.

Exposure of D. suzukii Pupae in the Field
Overall 2113 pupae of D. suzukii were exposed to predation in the hedges. Adults emerged from 33.1 ± 4.0% of them after taking them back to the laboratory, and 43.7 ± 4.1% were considered to be predated (pupae were missing or damaged) while the fate of 23.2 ± 3.1% remained unaccounted for (pupae were returned visibly undamaged but no adult flies emerged; Table 2). From the 1090 control pupae, 67.4 ± 4.9% yielded adult flies, while 1.9 ± 1.1% were lost and 29.6 ± 4.7% remained unaccounted for. Mean predation over the whole assessment period varied largely among hedges and ranged from 6.7 ± 3.8% on the ground at the humid site in Wädenswil to 70.0 ± 21.3% on branches at the dry site in Changins. The proportion of predated pupae was significantly higher at dry than at humid sites (χ 2 1,103 = 4.206, P = 0.040; Figure 1A). Furthermore, predation significantly differed between exposure sites (χ 2 2,103 = 67.295, P < 0.001): predation on pupae exposed on the ground (z = 6.965, P < 0.001) or on branches (z = 7.871, P < 0.001) was significantly higher as compared to the control pupae kept in enclosed plastic jars ( Figure 1B). However, predation on pupae exposed on branches or the ground did not differ significantly (z = 1.215, P = 0.444). The proportion of predated pupae did not significantly vary between sampling periods (χ 2 1,103 = 1.915, P = 0.590).

Exposure of D. suzukii Pupae in the Field
Overall 2113 pupae of D. suzukii were exposed to predation in the hedges. Adults emerged from 33.1 ± 4.0% of them after taking them back to the laboratory, and 43.7 ± 4.1% were considered to be predated (pupae were missing or damaged) while the fate of 23.2 ± 3.1% remained unaccounted for (pupae were returned visibly undamaged but no adult flies emerged; Table 2). From the 1090 control pupae, 67.4 ± 4.9% yielded adult flies, while 1.9 ± 1.1% were lost and 29.6 ± 4.7% remained unaccounted for. Mean predation over the whole assessment period varied largely among hedges and ranged from 6.7 ± 3.8% on the ground at the humid site in Wädenswil to 70.0 ± 21.3% on branches at the dry site in Changins. The proportion of predated pupae was significantly higher at dry than at humid sites (χ 2 1,103 = 4.206, P = 0.040; Figure 1A). Furthermore, predation significantly differed between exposure sites (χ 2 2,103 = 67.295, P < 0.001): predation on pupae exposed on the ground (z = 6.965, P < 0.001) or on branches (z = 7.871, P < 0.001) was significantly higher as compared to the control pupae kept in enclosed plastic jars ( Figure 1B). However, predation on pupae exposed on branches or the ground did not differ significantly (z = 1.215, P = 0.444). The proportion of predated pupae did not significantly vary between sampling periods (χ 2 1,103 = 1.915, P = 0.590).

Discussion and Conclusions
Our field survey at five different locations in two types of hedges showed that, on average, 44% of the exposed D. suzukii pupae were predated with few incidences of up to 95% of the provided pupae unrecovered. Predation took place on the ground as well as in the branches and across all sampling periods. However, although several families of predatory arthropods were present in the hedges and more than thousand individuals were collected and analyzed, less than 2% of the captured predators tested positive for the D. suzukii COI marker and only earwigs, spiders, and one predatory bug were identified to have recently fed on this invasive pest species.
Although web spiders may mainly capture adult flies and predatory bugs mainly feed on eggs, immobile pupae are probably the life stage of D. suzukii most vulnerable to predation [22,24,33]. Developing eggs and larvae are well hidden within the fruits, whereas pupae often protrude through the skin of the fruit or are found on the ground. Even though pupae might have been more exposed and aggregated in our study than under natural conditions, the observed predation rate of 44% of exposed pupae is considerable and likely affects the fly's population dynamic. The calculated predation rate largely exceeds OECD requirements of a 10% reduction in the pest population for the registration of an invertebrate biocontrol agent [34]. Moreover, the calculated rates of predation are in accordance with previous studies. Woltz and Lee [35] observed a decrease of 61-91% in the number of recovered pupae exposed on the soil in strawberry, blueberry, and blackberry fields. Similarly, Ballman et al. [36] found predation rates on exposed pupae from 34 to up to 100% in wild blueberry fields. Considering all three studies, it seems that generalist predators might have a higher impact on population dynamics of D. suzukii in Europe and North America than the more specialized parasitoids, whose natural parasitism rates rarely exceed 10 % [33,37,38]. Since parasitoids in the invaded areas mainly attack pupae, an additive effect of the two guilds of natural enemies is plausible [39,40].
We observed a high spatial variability in predation between hedges at the five locations, with calculated mean predation rates ranging from 6.7 ± 3.8% on the ground of a humid hedge to 70.0 ± 21.3% on the branches of a dry hedge. Significantly more D. suzukii pupae were predated in dry than humid hedges, but there was no significant difference in predation between pupae that were exposed on the ground and on branches. Likewise, predation did not significantly vary among the four sampling periods. These findings indicate that dry and sunny hedges harbor more voracious or higher densities of pupal predators than wet and shady hedges. Overall, dry hedges also hosted a higher plant diversity (see Table 1), which might thereby provide more food and alternative prey to generalist predators [21]. Accordingly, a general relationship between plant diversity and predation also became visible. In Conthey and Frick, where humid hedges showed a high plant diversity, higher predation in humid hedges was also observed, whereas in Changins and Wädenswil, the lower plant diversity in humid hedges was accompanied by less predation. In addition, pupal predators on the ground and on branches were of comparable voracity or even consisted of similar complexes of species. The latter might be the case, since in particular ants and earwigs are assumed to be important predators of D. suzukii pupae [22,24,25,35,41]. Typically, they live and forage on the ground as well as in the canopy and both are probably able to remove pupae entirely [35], which could explain the missing 34.4 ± 3.6% of the exposed pupae. Furthermore, ants and earwigs are present and active over the whole vegetative season, which might further explain the lack of significant differences between exposure periods. Yet, the presence of the D. suzukii COI marker could only be confirmed in the gut of earwigs but not in any one of the 160 collected ants. Since ants usually carry captured prey to their colonies where it is fed to larvae and shared among its members, it might, however, be difficult to actually detect any D. suzukii DNA in ants.
Earwigs, ants, spiders, and ground beetles were the most collected predator taxa. The composition of the predator communities and their abundance did not significantly differ between hedge types. Similarly, hedge type did not significantly influence the taxonomic composition of spiders at the family level. It can be assumed that the rather coarse identification up to the genus or family level, as was limited by the sample preservation for molecular analyses, does not permit to find more subtle habitat-specific differences at the species level. Although predator abundance was significantly higher in June to early July, as compared to late September to October; species composition of predator groups and Araneae families did not vary between sampling periods. The former could be due to warmer temperatures in summer than in fall, which is similar to the higher predation rates found in dry and warm hedgerows compared to humid hedges.
In the 1,101 captured predators, the presence of the D. suzukii COI marker could be confirmed in fewer than 2% of the tested individuals and only in the three orders Forficulidae, Araneae, and Heteroptera. The proportion of individuals that had fed on D. suzukii was highest in earwigs, whereas fewer proportions tested positive among spiders and bugs. The same predator orders, and furthermore rove beetles (Staphylinidae), have also been identified to feed on D. suzukii in the study by Wolf et al. [22]. The proportion of predators that tested positive for the D. suzukii COI marker was, however, considerably higher in the study of Wolf et al. [22] with 43.4% for earwigs and around 15% for spiders and bugs. This might in part be explained by higher D. suzukii density in the study of Wolf et al. [22], in which arthropod predators have been captured in the immediate vicinity of heavily infested cherries, blackberries, and raspberries, whereas we also collected many specimens on non-fruiting plant species. In order to prevent effects of volatiles on the predators and to avoid reducing prey availability, we deliberately refrained from setting up D. suzukii traps in the monitored hedges. We therefore have no observational data on the effective abundance of D. suzukii in the studied hedges. Furthermore, semi-natural habitats are known to harbor alternative prey to generalist predators [21]. Thus, D. suzukii density might have been lower in our study, while alternative prey was more common, which probably reflects the natural pattern of predation in agricultural hedges.
The proportion of predators that was tested positive for the D. suzukii COI marker was surprisingly low in comparison to the observed predation rate. It is likely that we did not cover the full range of predators in our molecular analyses, either because they carried away the prey such as ants or because they were not analyzed, such as predatory flies or mollusks. Furthermore, the relatively small proportion of positive samples can be explained by the detection limits of the analysis. Molecular gut content analyses allow the detection of prey DNA before it is digested to below the detection limit, which usually takes between several hours to a few days [42]. Under laboratory conditions, the DNA of D. suzukii was detectable in the gut of the common earwig for up to 48 h [43]. Moreover, we also decided to use a conservative approach to effectively validate the presence of D. suzukii DNA in the gut of predators, since individuals had to be tested positive twice in independent PCR runs. Hence, we might have dismissed samples where a very small amount of D. suzukii DNA failed to be confirmed in the repetition. It can therefore be assumed that our molecular assay underestimated the proportion of predators, which had fed on D. suzukii pupae.
Overall, a rather high proportion of exposed pupae was preyed by generalist predators in the studied hedgerows. Due to their broad prey range, the identified predators might however be less suited for targeted biological control attempts compared to more specialist parasitoid species from D. suzukii's native range or even from invaded regions [44]. Furthermore, high densities of some predators like the omnivorous earwigs can also inflict damage to thin-skinned fruits such as cherries or grapes [45][46][47]. Nonetheless, these generalist predators might provide a non-negligible background regulation of D. suzukii and their presence in hedgerows in the agricultural landscape may contribute to the regulation of pest populations [33,48]. Differences in composition of the predator community and the predation of pupae between sites were high, indicating that landscape context and hedge quality affect the extent of biological control provided by natural enemies, and suggesting a still underexploited potential to improve natural pest regulation. However, whether an individual hedge functions as a sink or a source for D. suzukii may also depend on the community of host plant species. For example, plants from the genus Cornus, Prunus, Rubus, or Sambucus are attractive and favorable to D. suzukii [4] and probably enhance its spillover from hedges to nearby crops (e.g.,: [10,11,49,50]). Hedges with less suited plant species or even harboring dead-end hosts (i.e., fruits that are used for oviposition but do not support larval development) might however function as sinks [5,51,52]. Moreover, hedgerows might provide other ecosystem services such as wind breaks, the conservation of biodiversity, the control of other pests, the supply of edible fruits, and attractiveness for recreation, thereby enhancing the overall quality of agricultural landscapes.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/ 10.3390/insects12040305/s1, Figure S1: Transformation-based redundancy analyses of the composition of predator groups in dry and humid hedges and over four sampling periods, Figure S2. Transformation-based redundancy analyses of the composition of Araneae families in dry and humid hedges and over four sampling periods.