The Wide Potential Trophic Niche of the Asiatic Fruit Fly Drosophila suzukii: The Key of Its Invasion Success in Temperate Europe?

The Asiatic fruit fly Drosophila suzukii has recently invaded Europe and North and South America, causing severe damage to fruit production systems. Although agronomic host plants of that fly are now well documented, little is known about the suitability of wild and ornamental hosts in its exotic area. In order to study the potential trophic niche of D. suzukii with relation to fruit characteristics, fleshy fruits from 67 plant species were sampled in natural and anthropic ecosystems (forests, hedgerows, grasslands, coastal areas, gardens and urban areas) of the north of France and submitted to experimental infestations. A set of fruit traits (structure, colour, shape, skin texture, diameter and weight, phenology) potentially interacting with oviposition choices and development success of D. suzukii was measured. Almost half of the tested plant species belonging to 17 plant families allowed the full development of D. suzukii. This suggests that the extreme polyphagy of the fly and the very large reservoir of hosts producing fruits all year round ensure temporal continuity in resource availability and contribute to the persistence and the exceptional invasion success of D. suzukii in natural habitats and neighbouring cultivated systems. Nevertheless, this very plastic trophic niche is not systematically beneficial to the fly. Some of the tested plants attractive to D. suzukii gravid females stimulate oviposition but do not allow full larval development. Planted near sensitive crops, these “trap plants” may attract and lure D. suzukii, therefore contributing to the control of the invasive fly.


Introduction
Biological invasions are considered as one of the major causes of biodiversity loss on the planet [1,2]. Alien species may have serious impact on native communities, habitats, and ecosystem processes [3], therefore altering ecosystem services that might be key elements for human wellbeing [4][5][6]. In addition, if the life cycle of invasive species coincides with that of production The fruits of different plant species are not equally suitable to the different life stages of D. suzukii (egg, larva, adult), as suitability depends on a large set of traits (volatile compounds, pH, shape, structure, firmness, quantity and quality of resources, colour. . .) either preventing, limiting or favouring the development of the fly [30,31,39]. Hence, analyzing relationships between the functional traits of the fruit and D. suzukii oviposition behaviour or larval development would help to understand the mechanisms underlying successful invasion [40] that are not explained by a taxonomic approach (i.e. species of different families can show similar fruit traits, while a great diversity of fruit shapes or colours can be found in a single family). In functional ecology, plant species and traits are grouped according to common responses to the environment ('response traits') or common effects on ecosystem processes ('effect traits') [41]. Fruit colour and shape are commonly considered as responses ('response traits') to fruit consumer behaviour-as with D. suzukii's oviposition behaviour-, while fruit structure (including the presence of septa within complex fruits and internal partitions, as in Rubus polydrupes) and size/diameter will rather influence larval development by limiting resources ('effect traits'). Therefore, fruit nutrients and weight may influence ecosystem processes through the quantity and quality of organic matter entering the trophic networks in the presence of D. suzukii populations.
In this study, we examine the relationships between the community of fleshy-fruited plants and the invader D. suzukii in a temperate region of Europe. More specifically, we addressed the following research questions: (1) how many plant species may host D. suzukii larvae and lead to the full development of imagos? (2) Do fruit traits (structure, colour, shape, skin texture, diameter and weight) affect D. suzukii oviposition choices and development success? As D. suzukii is able to migrate across regions, along altitudinal and climatic gradients and within and between ecosystems according to the season [18,27,42], fruits of 67 wild and ornamental plants were collected over the course of an entire year in various ecosystems (forest, hedgerow, grassland, garden, coastal area. . .) of the north of France and exposed to infestation by D. suzukii under experimental conditions using no-choice tests.

Study area and field sampling
The study was carried out in the Picardy region, in the north of France (N 48°50'19''-50°21'59''; E 1°22'50''-4°15'23''; alt. 0-296 m). The climate is of the oceanic type, with a mean annual temperature of 10°C and an average annual rainfall of 700 mm. The geological substrate is mainly composed of Cretaceous chalk covered by clay and/or Quaternary loess. Sandy substrates and salt alluvium are locally found along the coast. Landscapes are diversified and consist of mosaics of openfields, bocages, forests, urban areas and coastal vegetation. All these types of landscape have been visited to identify the pool of fleshy-fruited plants that could be potentially used as host plants by D. suzukii over the four seasons of one year.
Between October 2011 and November 2012, fruits from 67 fleshy-fruited plant species (see details in Appendix A) were collected in forests, hedgerows, grasslands, coastal areas, gardens and urban areas in the Picardy region. This exhaustive species sampling covered the majority of fleshy-fruited plant species present in the north of France. Only 16 fleshy-fruited wild plant taxa of the north of France were not tested in the study: Arum italicum (exotic), Convallaria maialis (irregular fruiting in the forests of the north of France), Cornus mas, Crataegus laevigata, Daphne  separated by a minimum distance of 200 m were collected randomly and stored in individual paper bags. Fruits from each species of plant were collected only once at maturity during the season from all five individual plants. Every month, the presence of fruit of each species was recorded in the field. The fruits of 10 additional species (Appendix B) were collected but not included in the analyses because the number of fruit per individual or per species was too low to be tested. The colour of the fruit skin is commonly used to characterize their maturity in D. suzukii studies [24,43,44]. For instance, Atropa belladonna fruits turn from green (unripe) to black (ripe) before falling, while Arum maculatum fruits change from green to red at maturity [45]. All the fruits collected in this study were ripe, i.e. their skin had the colour typically associated with their maturity (e.g. entirely black for Atropa belladonna and entirely red for Arum maculatum), according to the indications of floras [45][46][47][48]. Plants were sampled on sites receiving no chemical treatments, excepted Solanum tuberosum (listed in the additional set of plant species of Appendix B, not included in the analyses) which was collected on a conventional agricultural field at the end of the production period. The sampling was conducted with the permission of the owners of private land and gardens. In public areas, no specific permissions were required (e.g. Rubus fruticosus is traditionally harvested in public forests for preparing jam). Field studies did not involve species currently known to be neither endangered nor protected, and many sampled species were ornamental plants commonly marketed in plant nurseries.

Allotments of fruit and laboratory tests
For each plant species, a total of 150 fruits (1 species x 5 individuals x 3 tests x 10 fruits) was used in laboratory tests. The fruits were tested immediately after sampling. For each individual of each species, 30 fruits were randomly selected. The fruits were carefully examined with a stereomicroscope (Leica M 165C) and those already damaged or attacked by animals or pathogens were excluded. The 30 fruits were split into three sets of ten fruits and placed in ventilated transparent plastic boxes (15 cm x 10 cm x 5cm) to perform three types of tests. In a first test, subsequently termed 'adult emergence test', 10 fruits were exposed to 3 D. suzukii mated females for 24 hours following the protocol set up by Poyet et al. (2014) [24]. After 24 hours, the number of eggs laid in each fruit was counted under a Leica M 165C stereomicroscope. The presence of eggs oviposited in fruits was identified by the holes drilled by the females' ovipositor and by the presence of egg filaments [24]. During those experiments, every hole observed contained one single D. suzukii egg. The number of D. suzukii flies emerging from each test sample was checked daily for two months, the flies were counted then removed from the experimental boxes to avoid new oviposition. In the second test ('larvae development test'), 10 fruits were exposed to 3 D. suzukii females for 24 hours. After 24 hours, the number of eggs laid in each fruit was counted. After one week, the fruits were dissected and the total number of larvae that were present in and on the fruits and in the test plastic box was recorded. In the third test, a set of 10 fruits was used as a control to monitor the potential fruit contaminations by other insects.
In all the tests, the strain of D. suzukii collected in 2011 in the forest of Compiègne, in Picardy [24,49], was used and maintained under an LD 13: 11 h photocycle at 20°C. The D. suzukii strain was mass reared and fed with a regular banana Drosophila diet [50]. Five-days-old Appendix B. Emergence of Drosophila suzukii imagos from the fruits of an additional set of plant species. These additional species were collected but not included in the analyses because the number of fruit collected per individual or per species was too low to be tested.

Plant species
Number of sampled individuals

Fruit traits
We targeted fruit characteristics associated with three fundamental stages of the D. suzukii life cycle, i.e. egg laying, larval development and adult emergence. A set of five biological traits (including a total of 16 trait categories described hereafter) commonly related to oviposition choice and larva survival [51][52][53][54][55] was retained: maximum fruit diameter (fruit weight was also measured, but was redundant with fruit diameter and not retained in the final analyses; see explanations in the next paragraph), type (i.e. structure, defined hereafter), colour, shape and skin texture. Information was collected or extracted from flora [45][46][47][48] and plant databases [56,57]. Fruit weight and maximum diameter were individually measured before laboratory tests on the 6694 fruits used for larval development and adult emergence experiments, and treated as continuous variables. Fruit type, colour, shape and skin were treated as categorical variables and the optimal number of categories was defined so as to be ecologically meaningful and have balanced sizes [58,59].  [48] and plant families follow the APG III phylogeny [60].

Data analyses
The effects of fruit traits on the number of D. suzukii eggs, larvae and adults emerging from fruits were examined using generalized linear models (GLM) with a Poisson distribution and a log-link term [61]. In the egg-model (n = 67 plant species tested), the number of D. suzukii eggs was the response variable, the fruit diameter was used as a fixed covariate, and the fruit type, colour, shape, surface type and skin thickness were used as fixed factors. In the larvamodel, the number of D. suzukii larvae (n = 60 plant species tested) was the response variable, the fruit diameter was used as a fixed covariate, and the fruit type was used as a fixed factor. Seven species were excluded from this analysis (Berberis julianae, Gaultheria procumbens, Ligustrum vulgare, Lonicera caprifolium, Mespilus germanica, Pyrus calleryana 'Chanticleer', Sorbus aria) because too few undamaged fruits per individual plant were available to be tested.
In the adult-model (n = 67 plant species tested), the number of D. suzukii adults emerging from fruits was the response variable, the fruit diameter was used as a fixed covariate, and the fruit type was used as a fixed factor. To fulfil the normality assumption, fruit diameter and weight were log 10 -transformed. As fruit diameter and weight were highly correlated (R = +0.813; p<0.0001), fruit weight (less correlated with the number of eggs laid in the fruits by the flies than the diameter was) was excluded from the analyses to avoid redundant explanatory variables in the models. SPSS v. 17.0 (IBM Corp., Somers, NY, USA) was used in all the analyses.

Results
A very wide range of plant families was used by D. suzukii females and allowed adult emergence:  (Fig 2). The highest mean number of eggs per fruit was 10.78 (± 1.68) for Phytolacca americana, the highest mean number of larvae per fruit was 6.84 (± 1.37) for Rubus fruticosus agg., and the highest mean number of adult emergences per fruit was 5.20 (± 1.83) for Rubus fruticosus agg. The developmental times of D. suzukii from eggs to adults among the tested fruits were reported in Appendix C. Among the 67 plant species tested, 33 (49.25%) allowed the emergence of D. suzukii imagos, 6 (8.96%) hosted larvae that did not reach the adult stage, 11 (16.42%) hosted eggs that did not hatch, and 17 (25.37%) did not host any eggs (Figs 1 and 2).

Effects of fruit traits
The GLM showed significant effects of fruit traits on the number of D. suzukii eggs, larvae, and adults emerging from fruits ( Table 1). The number of eggs, larvae and adults increased significantly with the fruit diameter. The number of eggs was higher in berries and drupes than in the other fruit types, higher in white or black fruits than in fruits of other colours, higher in oval fruits than in spherical ones and higher in rough fruits than in smooth ones, especially in those with a pruinose coating. The number of larvae and emerging adults was higher in polydrupes than in the other types of fruit and was also higher in drupes and berries than in fruits with a complex structure.

Range of plant species and fruit phenology
Among the 33 plant species that allowed the emergence of D. suzukii adults, 21 (63.6%) were ornamental or cultivated species, 14 (42.4%) were exotic, and 4 species (12.1%) were invasive in the region. Several naturalized plant species (i.e. observed in nature but not considered as invasive at present: Lonicera nitida, Ribes sanguineum and Symphoricarpos albus) also allowed the emergence of D. suzukii adults (Appendix A and Fig 2). Among the 67 plants tested, the number of larvae per fruit was lower in the ornamental plant species (0.49 ± 0.12 larvae.fruit -1 ) than in the others (1.57 ± 0.50 larvae.fruit -1 ; t = -2.799, p = 0.007). The number of emerging adults per fruit was lower too, in the ornamental plant species (0.36 ± 0.09 larvae.fruit -1 ) than in the others (1.06 ± 0.35 larvae.fruit -1 ; t = -2.548, p = 0.013).
The number of plant species with fruits potentially suitable for the development of D. suzukii offspring was high between September and December with a peak in October (Fig 3). Only a few host plants suitable for D. suzukii were fruiting in early spring. However, the fruits of several potential native and exotic hosts remained hanging on plants during the winter (Viscum album, Aucuba japonica, Hippophae rhamnoides subsp. rhamnoides, Symphoricarpos albus).  Firstly, the extremely high number of host plant species suitable for the development of D. suzukii indicates that the fly benefits from a large amount and diversity of resources that may have efficiently contributed to the success of its invasion. This wide polyphagy was mainly reported on agronomic varieties of cultivated fruits [31,39]  The Polyphagy of Drosophila suzukii advantages. A multi-fruit diet reduces the time spent searching for food, limits the effect of resource stochasticity within and between years, increases enemy-free space, provides the fly with nutrient complementation, or attenuates the possible toxic effects of some fruits [62][63][64]. Moreover, a mixed diet may also increase the survival rate and the fecundity of adult females [64]. The large pool of host plants potentially used by D. suzukii in Picardy is widely distributed among both natural ecosystems and gardens throughout Europe. Among the suitable hosts detected in the present study are plants that D. suzukii probably encountered on its historical  [18,27,66]. Secondly, the host's phylogeny is not a barrier to infestation by D. suzukii. Indeed, host families are phylogenetically very different and extremely distant on the APG III classification tree [60] as they belong to different Orders, Classes and Divisions: for example, D. suzukii successfully develops in the fruits of Taxus baccata (Family: Taxacea suzukii. This suggests that the fruit's traits matter more than the plants' evolutionary history. A broader analysis of the hosts' phylogeny across the different continents invaded by the fly could help understand the influence of preference-and performance-related traits of D. suzukii on host range [67]. Among the plant families infested by D. suzukii, many are known to produce toxic secondary compounds [68]. For example, alkaloids that are specific to each taxonomic group of plants and neurotoxic to mammals [68], have been used as deterrents to larvae or biocides to control insect pests [69][70][71][72][73][74][75][76]. Solanaceae are one of the plant families that produces the largest variety of toxic molecules, including alkaloids, concentrated in the fruits [68,76]. Surprisingly, among the Solanacea family, Atropa belladonna is one of the best hosts for D. suzukii (Figs 1 and 2) while it predominantly contains tropane alkaloids [77][78][79], in addition to cuscohygrine, apotropine, belladonine and scopoline [80][81][82][83]. Solanum dulcamara and S. nigrum also allow the emergence of imagos (Fig 2) though harbouring numerous alkaloids [76,[84][85][86][87] that may act as feeding deterrents for insect larvae [70]. D. suzukii also successfully develops in a large set of other tested fruits (Fig 2) characterized by the presence of alkaloids or other toxic compounds such as glycosides, terpenoids and phenylpropanoids [88]. The high number of toxic plants permitting D. suzukii development suggests that larvae may possess a substantial set of enzymes enabling them to process these secondary compounds. Enzymatic detoxification ability has already been reported in polyphagous insects [89] and in particular in Drosophila melanogaster which shows a resistance to the alkaloids produced by the cacti on which larvae and adults feed [90]. D. melanogaster thus evades competition with other Drosophila species unable to use this toxic resource [91]. Feeding on toxic plants may confer several other advantages to the invasive fly. Numerous insects, especially at larval stages, can store in their tissues high quantities of alkaloids and other toxins present in their diet, thus increasing their resistance to pathogens [92] and parasites [93], or avoiding attacks by predators such as birds [94,95]. The consumption of toxins by the Drosophila species can also be a strategy of medication against parasites [93,96], which lays down a valuable hypothesis explaining why toxic fruit can be beneficial to D. suzukii.
The wide polyphagy of D. suzukii contrasts with the diet of other invasive insect pests in France and Europe-monophagous or oligophagous species-like the grapevine phylloxera (Daktulosphaira vitifoliae) and the Colorado potato beetle (Leptinotarsa decemlineata) [97, 98], two historical and emblematic pests on the continent. The oligophagy among insect pests facilitates their control, as they do not benefit by alternative food resources when their main host plant has become resistant to infestation. For example, the damages caused by Daktulosphaira vitifoliae and its survival rate are dramatically reduced by using plants naturally or artificially resistant to phylloxera as rootstocks [99]. This type of pest control by host resistance would be difficult to use against the recent invasions of polyphagous pests like D. suzukii, Halyomorpha halys [100,101], Bemisia tabaci [102] or Popillia japonica [103], since they all have a large set of host plants in their exotic range. For these new multi-host pests, other integrated and ecological control strategies need to be developed, including the management of their natural reservoirs in the vicinity of crops and that of crop diversity. Although host diversity is generally beneficial to polyphagous insects, in the presence of a mixture of host plants these insects may experience difficulties in matters of decision-making when selecting food and oviposition sites [102]. This behavioural disturbance may reduce their performance. For example, increasing the host plant diversity in the environment of multi-host pests may lead individuals to move more, to switch between plants more frequently, and to lose energy by feeding in each place for short periods of time [102,104]. Unlike D. suzukii, other invasive Drosophila species (for example, D. subobscura in America [105], D. melanogaster in Australia [106], or D. simulans in Europe [107]) are mainly used as model species for genetic studies but are not considered to be important pests because they mainly feed on diversified but rotten substrates and do not damage fruit and vegetable productions.
The controlled environment in our experiments provided the homogeneous conditions necessary to perform sample comparisons and to avoid the biases caused by the environmental variability commonly observed in natural infestations (depending on climate or other stochastic factors in the field). The findings recorded here do not necessarily describe overall fruit infestation in natural environments but they do point to a potential trophic niche that can be partially occupied by the fly according to local factors. Moreover, fruit maturity may also influence the rate of infestation in the field as well as in laboratory conditions [24,43]. Indeed, D. suzukii lays more eggs on ripe fruits for some plant species [43], while for other host plants more eggs are found on ripening fruits [24]. The influence of fruit maturation on the fly's behavioural preferences still remains debated as a recent study showed that the developmental stage of fruit alone does not explain the ecological niche observed for D. suzukii [108], and that other plant traits need to be examined to understand fly-fruit interactions.
The role of fruit traits in D. suzukii infestation All along their evolutionary history, Angiosperms developed a wide range of fruit traits that protect them against predators or facilitate the activity of frugivorous insects [89,109,110]. To continue to exist in the environment, D. suzukii needs to ensure the success of the fundamental steps of its relationship with fruits: oviposition, development, and adult emergence.
Fruit size. The first important fruit trait associated with the success of the different D. suzukii life stages is fruit size, which is measured by the fruit's diameter and weight (Table 1). Larger fruits increase the number of D. suzukii adults emerging from the fruits, as previously reported with Prunus serotina [24]. The correlation between fruit size and egg-clutch size is proved for many plant-insect associations [53,111] but the fruit size can also be even more important than the fruit type itself [53] as it may indicate to gravid females what amount of resources is available for their progeny. Fruit diameter is also correlated to the number of developing larvae and emerging imagos (Table 1), but the developmental time was negatively correlated with the number of imago emergences. Indeed, multiple infestations of small fruits may prove lethal to competing larvae whilst pupal mass is known to increase with fruit size [112].
Our results corroborate those of Dukas et al. (2001), showing that, in small fruit, competition between larvae affects their fitness, therefore reducing possible benefits from social facilitation [54].
Fruit type. The fruit type strongly influences the egg-clutch size and the outcome of larval development (Table 1). D. suzukii females lay more eggs on simply-shaped fruits (berries and drupes) than on fruits with a complex structure such as polydrupes (an assemblage of a high number of drupeoles), capsules (with internal physical partitioning as in Euonymus europaeus), pseudo-or composed fruits (like the rosehip of Rosa canina and pome fruits of Pyrus calleryana, Mespilus germanica and Crataegus monogyna). These complex fruits, more fibrous than the others, may hamper the migration of larvae within their flesh and reduce the efficiency of food intake by the larvae. Although the polydrupes of the Rubus species force the larvae to exit and then migrate to access other parts of the fruit, this physical barrier does not decrease the success of larval development (Table 1 and Figs 1 and 2). Many studies have shown the impact of fruit type on fruit attractiveness to frugivorous insects [53], including Drosophila species [113][114][115], although fruit discrimination is complex and depends on both the physical and chemical traits of the fruit [116]. The recent works on D. suzukii attractants focus almost exclusively on chemical components because they significantly increase our fundamental knowledge of the biology of the fly [108] and, predominantly, because the optimal chemical composition of bait is actively researched for the elaboration of drosophila traps used for fly monitoring and crop protection [117,118]. Although the shape of bait is neglected in fly-trap design, our results show that the simply shaped structures of berries and drupes (oval or sphere) are preferred by D. suzukii females and could be drawn at the surface of plastic traps or used to pattern a new generation of solid baits containing chemical attractants.
Fruit colour. Fruit colour is an important visual trait for fruit flies' oviposition behaviour [119,120]. D. suzukii can lay eggs in fruits of various colours: black (Atropa belladonna, Sambucus nigra), blue (Ribes sanguineum), red (Aucuba japonica, Arum maculatum, Fragaria vesca, Ribes rubrum), pink and purple (Gaultheria procumbens, Lonicera nitida), orange (Hippophae rhamnoides), brown (Pyrus calleryana 'Chanticleer') or white (Symphoricarpos albus, Viscum album; Fig 2 and Appendix B). Therefore, the way D. suzukii uses colour to select elicited fruits remains a moot point and experimentation using coloured objects or host fruit mimics [51,121] is still needed. The coloured area of the fruit could also be used by sexually mature flies as a rendezvous site for courtship and mating [121,122]. Whatever the mechanisms governing fruit choice, D. suzukii tolerance to various fruit colours may be one key-factor explaining its polyphagy. Without this plasticity in the visual selection of the fruit, the fly would not have been able to infest such a large range of fruit. By using a great diversity of fruit colours, D. suzukii might also disorientate some potential parasitoids which could be attracted to a specific range of colours [123], even if kairomones also play an important role in host selection by parasitoids [124]. This plasticity in visual attractiveness could be a behavioural innovation accompanying the evolutionary changes in the fly's morphology [29,66].
The evolution of the fruit-penetrating ovipositor is a major morphological innovation that differentiates D. suzukii from its close relatives. The ability to drill holes into the skin of fresh fruits allows access to a new ecological niche [29]. This mutation in the fly's ecology necessarily includes additional neurological, lifecycle, and physiological adaptations to find non-rotted fruits [66] and to identify their various colours.
Fruit skin. The first physical barrier developed by fruits against insects is their skin [39,125]. D. suzukii females lay more eggs in rough-skinned fruits than in those showing a smooth, waxy, shiny or pruinose coating (Table 1). A waxy texture may indicate the presence of hydrophobic coatings and the presence of specific molecules [126] that may act as deterrents to insects. Pruinose surfaces contribute to the mechanical strength of plant tissues, to the cueing of host-pathogens/insects recognition, to the reduction of contamination and pathogen attacks, and to promoting or preventing insect attachment and locomotion [127][128][129]. This may contribute to explain why some pruinose fruits (Viburnum tinus, Polygonatum multiflorum, Paris quadrifolia, Mahonia x media, Berberis julianae) or shiny waxy fruits (Ruscus aculeatus, Rosa canina, Berberis thunbergii) have been less attacked by D. suzukii. Nevertheless, host attractiveness may also depend upon many additional factors [30,31,39] like, for example, soluble sugar content, pH or volatile compounds. Therefore, complementary experiments must be conducted for an accurate understanding of fruit attractiveness and suitability to D. suzukii.
The capacity of the fly to lay eggs in fruits protected by various skin textures (waxy, pruinose skins) partly relies on its saw-tooth ovipositor which bypasses the textural defence of the skin, and represents an evolutionary innovation [18] and a new weapon [130,131] in the introduction areas as well as an advantage over other Drosophila species (see Atallah et al. 2014 [29] for the rapid evolution of that weapon). The evolution of that organ, together with the associated behaviour of fresh fruit recognition [18] provided D. suzukii with the ability to use the flesh of a large diversity of fresh fruits and to acquire these resources before other Drosophila species which mostly lay eggs on rotten fruits. As fresh fruits hanging on plants are neglected by other local Drosophila species (see D. subobscura feeding only on fruits fallen on the ground in Poyet et al. 2014 [24]), D. suzukii benefits from an empty niche [132], avoids major resident competitors and natural enemies [133], and may consequently increase its performance in the interactions network of its novel ecosystem [134]. Increased invasiveness coupled with the absence of Drosophila competitors is likely to have helped the fly to colonize a great diversity of habitats and reproduction/feeding substrates.

Temporal continuity of host availability along the four seasons
Two recent studies [135,136] examined the seasonal variations of D. suzukii populations and their relationship with the phenology of cultivated fruits, but the phenology of wild and ornamental fruits has not been studied satisfactorily yet. With its polyphagous behaviour, D. suzukii is likely to find alternative host plants in natural and urbanized systems throughout the year (see Fig 3). The number of host plant species fruiting between April and May is lower but many flowers (Prunus spinosa, P. avium, R. rubrum, R. nigrum, Cornus sanguinea, Crataegus monogyna. . .) producing nectar may compensate for the absence of fruit and help D. suzukii adults to survive until the summer [27]. Other plant species, which are resistant to the infestation by D. suzukii, produce a large amount of nectar in autumn (Hedera helix) and may contribute to D. suzukii overwintering. In winter, the fruits of several native and exotic host plants (Viscum album, Symphoricarpos albus, Hippophae rhamnoides) remain hanging on parent plants, especially along coastal areas where the oceanic climate protects fruits from frost. Therefore, one issue is still at stake: do D. suzukii populations migrate between ecosystems (from forests in autumn and winter to gardens in spring and summer, for example) and between regional areas, in order to find fruiting or flowering plants? Coastal areas with a warm climate, small temperature variations and many host plant varieties (such as Hippophae rhamnoides subsp. rhamnoides) may represent a continental climatic corridor for the dispersal of the fly.
Drosophila suzukii, a new environmental filter among plant communities?
Drosophila suzukii may be a time bomb in natural plant communities. As numerous plants with thin-skinned fleshy fruit reproduce and disperse in temperate European landscapes and given the damages caused on fruits [137], significant impacts of D. suzukii invasion could be expected on the plant communities. However, according to recent advances in the role of frugivorous insects on the regeneration of host plant species [110], the magnitude and direction of the impact of D. suzukii on plant species fitness remains uncertain. Indeed, frugivorous insects are not always pests [110] and they may produce negative, neutral or beneficial effects on plants according to their hosts' autecology.
Phytophagous, and especially frugivorous, insects were first considered to be particularly damaging to hosts because they pre-empt the plants' reproductive tissues [138], accelerate fruit decay, cause premature abscission of infested fruits [139] containing immature seeds, reduce fruit attractiveness for vertebrate endozoochorous dispersers (animals that disperse seeds via the ingestion of fruits, such as birds or mammals) [140,141], which consequently increases competition between parent plants and offspring growing under their canopy. They are also major drivers for the evolution of host plant defensive traits [142]. Seed transport by animal dispersers is the main way for many plants to reproduce, colonize new habitats, and survive in the landscape matrix [143,144]. Fruit removal and seed arrival in different habitats are processes related to frugivorous species [144,145]. Among frugivorous vertebrates, birds are major vectors for fleshy fruits in Europe. They are highly sensitive to fruit quality and they discriminate and significantly reject fruits that have been attacked by insects [140,141,146,147]. The alteration of the dispersal service provided by avian seed dispersers is known to cause plant regeneration collapses [144] and to produce cascading effects on ecosystem services [148]. As the fruits of 50% of the tested plant species can be damaged by D. suzukii, a high number of species could lose their ability to disperse and regenerate successfully. Even if the species damaged by D. suzukii are not economically important, they contribute to the equilibrium and services of the ecosystem. For instance, Prunus avium is used for wood production and many fruits from wild plants (Rubus fruticosus, R. idaeus, Sambucus nigra, Fragaria vesca, Prunus spinosa or Hippophae rhamnoides) are harvested for cooking and jam preparation or medicinal use, and marketed. By damaging ripening fruits, D. suzukii may modify the fertility and the dispersal performances of plant species, and may consequently change their frequency and place in the communities. A shift in the functional composition of forest communities could then occur: fleshy-fruited species infested by the fly would be disadvantaged to the benefit of plants reproducing through vegetative organs (stolons) or dry fruits/seeds (achenes, wing-bearing samaras, for example) not attacked by D. suzukii larvae. Consequently, anemochorous (seed dispersal by wind) and ectozoochorous (external transport of seeds by animals, in their fur for example) processes would be more likely to determine the patterns of species distribution than endozoochorous processes in the future ecosystems invaded by D. suzukii.
Several studies showed the neutral or positive effects of frugivorous insects on plant fitness [110,142]. Indeed, the viability of embryos and the production of seeds are not necessarily affected by insect activity. By perforating the seed coat, removing the fruit pulp or accelerating fruit decay, insects may even stimulate germination, increase seed viability or attract frugivorous birds and mammals responsible for seed dispersal [110,149,150]. Regarding these contradictory plant-insect interactions and the complexity of the triad fruits-insects-frugivorous vertebrate dispersers [151], it seems difficult to predict the effect of D. suzukii on plant regeneration accurately. However, by pre-empting the organic matter contained in the fruit flesh of many plant species, D. suzukii will change the fluxes of matter and energy and the interactions in trophic networks and will become a new, non-negligible component of the ecosystem.
A perspective in the ecological control of Drosophila suzukii Although half the wild and ornamental plants with fleshy fruits may be considered to be reservoirs for the fly, some of the tested plants (Pyracantha, Cotoneaster. . .) attractive to D. suzukii gravid females are characterized by low success or absence of imago emergence (see Figs 1 and 2). These "trap plants" may lure the fly as they induce oviposition but do not allow the full development of larvae. Further investigations are needed to understand why larvae do not reach the adult stage in these fruits (presence of toxic compounds, low water and nutrient availability, fibrous structure. . .). These potential trap plants could be planted near sensitive crops (strawberries, grapes) to reduce the amount of pesticides. Choice tests between potential wild (and non invasive) trap-plants and agronomic ones will be the next step towards the ecological control of the invasive fly.