Introduction

The spotted wing Drosophila, Drosophila suzukii Matsumura (Diptera, Drosophilidae), is native to Asia and has recently invaded several regions and continents including Europe, North and South America, Reunion Island and Central Asia1,2. The economic impact of this invasive fly is increasing proportionally to its geographic range. Unlike most other Drosophilidae, D. suzukii is able to lay eggs in fresh fruits, due to the possession of a serrated ovipositor. With this feature, D. suzukii has become a major pest of small and stone fruits in most invaded regions3,4,5. It also has a very wide host range comprising many cultivated fruits as well as fruits from ornamental and wild plants6,7,8, and a short development time, which allows the completion of several generations per year1. As a result, crops are constantly reinvaded from neighbouring habitats, which complicates management strategies. Current control methods include chemical treatments and good cultural practices (e.g. sanitation, bait traps and insect proof nets)9. In addition, in invaded regions, D. suzukii encounters very few competitors. It is attacked by generalist predators10 and, to a much lower extent, by generalist pupal parasitoids11,12,13,14. In contrast, larval parasitoids, which are often considered as major mortality factors in Drosophilidae15,16, are so far absent from the natural enemy complex of D. suzukii in invaded regions8,9,17. Indeed, larval parasitoids of local Drosophilidae either do not show interest for D. suzukii or are not able to develop successfully in D. suzukii larvae, partly because of the strong host immune response of the fly larvae18,19,20,21,22.

Classical biological control, i.e. the introduction of larval parasitoids from the region of origin of the pest that are specialised in parasitizing D. suzukii, could help reduce populations at the landscape level and, consequently, decrease the need for management. However, little is known of the larval parasitoids of D. suzukii in Asia. The most comprehensive studies have recently been carried out in Japan but these often focused on specific parasitoid species or genera and rarely provided quantitative data on the role of parasitoids in the natural control of D. suzukii23,24,25,26,27. They concluded that the most promising biological control agent would be the figitid Ganaspis brasiliensis (Ihering) (Hymenoptera, Figitidae), a species that has been of unclear taxonomic status (first named G. xanthopoda24, and then G. brasiliensis26,27, following Buffington and Forshage28). This species is composed of different host races, some of which may warrant species status, and a specific host race appears to be specialised on D. suzukii26. In Asia, data on larval parasitism are restricted to a parasitoid survey in South Korea29 and a survey of the braconid genus Asobara in the Yunnan Province of China and South Korea30. Both studies used fruit collection and traps baited with uninfested fruits and suggest that field collection of suitable fresh fruits is a more reliable method to collect parasitoids. The most abundant larval parasitoids collected in South Korea were Asobara japonica Belokobylskij (Hymenoptera, Braconidae), Ganaspis brasiliensis and Leptopilina japonica Novković & Kimura (Hymenoptera, Figitidae)29.

In this publication, we report on surveys made from 2015 to 2017 in 12 Chinese Provinces and five Japanese prefectures to gather quantitative data on larval parasitism of D. suzukii. These surveys were made by sampling potentially suitable fresh fruits and, thus, also allowed us to collect two other Drosophila spp. living in the same habitat, D. pulchrella and D. subpulchrella Takamori and Watabe. The parasitoid complex of these two latter species is presently unknown.

Results and Discussion

Parasitism of Drosophila species

Details of all collections are provided in Table S1 (Supplementary information S1) and the results of the larval parasitoid surveys, with parasitism rates, are summarised in Table 1. For 2015 and 2016, only parasitism rates from samples that produced at least one larval parasitoid are shown because, in some cases, a high mortality occurred in the host pupal stage and the lack of parasitism may be due to sample deterioration in the period between fly and parasitoid emergence. For 2017, samples without parasitoid emergence were added because the rearing techniques had improved and we were more confident that all samples could be kept clean and healthy until parasitoid emergence. In any case, parasitism rates provided in Table 1 should be considered with caution. These rates are based on fly and parasitoid adult emergence and the rearing and transport conditions experienced during these surveys may have affected D. suzukii and its parasitoids very differently. On the one hand, most parasitoids emerge about two weeks after the D. suzukii adults. During this period, fruits became frequently covered by fungi, which surely prevented some parasitoids from emerging. On the other hand, D. suzukii pupae are extremely sensitive to high temperatures and low humidity. Unpublished observations by the authors showed that no fly emerges when pupae are exposed to temperature above 30 °C or humidity below 50% RH. It is possible that parasitoids in the host pupa are less sensitive than their host, however, nothing is known yet about the developmental requirements of the parasitoid species.

Table 1 Larval parasitoids and parasitism rates observed from fruit collections in China and Japan in 2015, 2016 and 2017.

Larval parasitism rates were highly variable, from 0 to 75.6%. The highest rates of parasitism were observed in Yunnan Province (China) and Nara prefecture (Japan). In contrast, parasitism seems to be lower in northern and eastern China, as shown by collections in Beijing, Jilin, Inner Mongolia, Jiangsu and Zhejiang Provinces, which did not yield parasitoids (Tables 1 and S1 and unpublished data). A possible explanation could be that D. suzukii is likely non-native in these areas and parasitoids may be less well adapted to cold winter and hot summer conditions experienced in these regions. Strong variations in parasitism were observed between nearby sites but also from year to year at the same sites. For example, parasitism in Japan on the same Prunus serrulata Lindley (Rosaceae) trees in Tokyo climbed from 9% in 2015 to 28% in 2016 at the same period of the year. These strong variations in parasitism are typical for insects that have short development times and many annual generations. Indeed, abiotic factors may affect hosts and parasitoids differently. Another factor that may affect the estimation of parasitism rates is the fact that parasitoids attack young larvae whereas we probably collected Drosophila spp. as eggs as part of the sample, i.e. before the attack of the parasitoids, which may result in an underestimation of the parasitism rate. On the other hand, parasitoids emerge later than D. suzukii and it cannot be ruled out that some samples contained empty fly pupae and immature parasitoids in host pupae. This, however, is rather unlikely because most larvae leave the fruits to pupate10 and we specifically sampled fruits on the plant that looked undamaged, to avoid collecting other Drosophila species.

Other data on parasitism rates of D. suzukii in Asia are scarce. Data on parasitism in South Korea have been recently published29 and parasitism rates of 0–17% have been reported, i.e. in line with our results from Northern/Eastern China. Another survey focused on the genus Asobara (Braconidae) in South Korea and Yunnan30, and found two species, A. japonica and A. leveri (Nixon), associated with D. suzukii, but it did not provide parasitism rates. Parasitism of D. suzukii has been studied more extensively in Japan but the information provided in the literature is mostly qualitative. Quantitative data were published from Prunus serrulata in Tokyo in 2015 and 201627. These results are very similar to our findings because fruits were collected at the same time. Other quantitative data from the same site in Japan have been published24.

In the majority of samples, D. suzukii was accompanied by two congeneric species that are also able to attack fresh fruits, D. pulchrella at high altitudes in Yunnan and Sichuan Provinces (China) and D. subpulchrella in Japan and at lower altitudes in China (Beijing and Hubei Provinces). This confirms the geographic range of the two species described31.

To our knowledge, this is the first time that D. pulchrella and D. subpulchrella have been sampled for parasitism. However, these two species occur nearly always together with D. suzukii and pupae of the three Drosophila species are morphologically indistinguishable, hence it was impossible to determine from which host the parasitoids emerged. Several samples, including very large ones, contained only D. suzukii, from which we deduced that the parasitoids emerged from this host. In contrast, only one small sample gave rise to D. subpulchrella only and no sample provided exclusively D. pulchrella. Thus, the association of the parasitoids with these two species could not be ascertained. The only solution would be to keep all pupae singly and to identify pupae from which parasitoids have emerged using molecular tools32.

Larval parasitoid species

Ganaspis cf. brasiliensis (Hymenoptera, Figitidae)

A figitid wasp of the genus Ganaspis was the most frequently reared parasitoid of D. suzukii in China and Japan, being present in all samples from which parasitoids emerged. It was also the species that reached the highest levels of parasitism in both countries. The same parasitoid was also reared from the sample in Hubei province, from which D. subpulchrella emerged without D. suzukii, suggesting that G. cf. brasiliensis also parasitizes this host (Table 1).

All specimens were morphologically similar to specimens from D. suzukii collected in South Korea and Japan and were identified as G. brasiliensis by one of the co-authors (MB) who had examined the specimens of all studies26,28. Ganaspis brasiliensis was also the most abundant parasitoid collected in South Korea29 and Japan27. By using molecular tools and behavioural studies, it has been suggested that G. brasiliensis may be a complex of cryptic species and/or host races with unique distribution patterns and host ranges26. Sufficient evidence for taxonomic assignment have not been provided, and hence, no new names were proposed for these lineages. However, it has been shown that the host race of G. brasiliensis, attacking D. suzukii is the ‘suzukii-specialised’ type of Ganaspis xanthopoda24. It has been suggested that this host race only attacks and develops on D. suzukii24,26 and possibly other species inhabiting fresh fruits such as D. pulchrella and D. subpulchrella. Its specificity is largely confirmed by our studies in quarantine conditions in Switzerland33,34. The same species is also being considered for release in USA, where the biology of a Korean population has recently been studied in quarantine conditions35. Future research, combining data from multiple gene regions in specimens from across a wide geographic area and cross-mating experiments are required to further elucidate the taxonomic status of Ganaspis brasiliensis s.l. In the meantime, the specimens collected in this study from Drosophila spp. in fresh fruits are referred to as Ganaspis cf. brasiliensis.

Leptopilina japonica (Hymenoptera, Figitidae)

Leptopilina japonica has been found in all regions of China and at two sites in Japan, but rarely reached high parasitism rates. Among the 20 sites where this species was collected, we observed parasitism above 10% at only six sites, and with a maximum rate of 34.5%. However, it was at least as abundant as Ganaspis cf. brasiliensis in Beijing, and more abundant in the single sample from Sichuan (Table 1). This parasitoid was already known from D. suzukii in Japan and Taiwan23,27 and was also reared frequently from D. suzukii in South Korea29. To our knowledge this is the first record of L. japonica in the People’s Republic of China. Two sub-species23, L. japonica japonica Novković & Kimura and L. japonica formosana Novković & Kimura, have been reported from Japan and Taiwan, respectively. Both sub-species were also found in South Korea29. Morphological observations suggest that most species collected during this study belong to the subspecies L. japonica japonica. However, at least one sample collected in Hasuike (Japan) in 2017 provided specimens of L. japonica formosana and not all specimens reared during this study were identified to subspecies level. Leptopilina japonica was successfully reared on D. suzukii in the laboratory34 and is also known to successfully develop in Drosophila biauria Bock & Wheeler and Drosophila rufa Kikkawa & Peng in Japan under natural conditions and in Drosophila simulans Sturtevant in the laboratory23. Its biology has been recently studied in quarantine conditions in USA35.

Leptopilina sp. (Hymenoptera, Figitidae)

Another Figitidae was reared in high numbers from bayberry fruits collected at two sites in Yunnan (China). While one sample yielded a mixture of D. suzukii and D. pulchrella, only D. suzukii emerged from the other. This parasitoid has also been successfully reared on D. suzukii in the CABI laboratory in Beijing (J. Zhang, unpublished data). It is morphologically very different from L. japonica, and a description is presently being prepared (Buffington et al., in prep).

Asobara spp. (Hymenoptera, Braconidae, Alysiinae)

Asobara is the third important genus of larval parasitoids of Drosophilidae worldwide15. At least three species of Asobara spp. have been collected, but usually in very low numbers. Asobara sp. TK1 has been found in Tokyo. This undescribed species could be specific to D. suzukii, based on a study of the capacity of eight Asobara species associated with Drosophila spp. in Japan to parasitize D. suzukii25. So far, Asobara sp. TK1 has only been collected in Tokyo25,27, but it has been suggested that Asobara sp. TK1 could be the recently described species, A. triangulata van Achterberg & Guerrieri, based on the molecular analysis of one specimen from Yunnan, China30.

Other species found in this study include Asobara pleuralis (Ashmead) and A. mesocauda van Achterberg and Guerrieri, reared from pupae of D. suzukii and possibly D. pulchrella collected in Sichuan and Yunnan. Asobara pleuralis has also been reared from D. suzukii and/or D. subpulchrella in Beijing. These two species are known to be associated with Drosophila spp.30, but not yet with D. suzukii, D. pulchrella or D. subpulchrella. Other surveys for parasitoids of Drosophilidae in Yunnan (China) and South Korea found Asobara japonica and possibly A. brevicauda van Achterberg & Guerrieri and A. leveri (Nixon) associated with D. suzukii29,30.

Tanycarpa chors (Hymenoptera, Braconidae, Alysiinae)

Tanycarpa species are also known as parasitoids of Drosophila flies15. However, no species have ever been recorded from D. suzukii, D. pulchrella or D. subpulchrella36. In this study, Tanycarpa chors Belokobylskij was obtained from two sites in China and one site in Japan (Table 1).

Areotetes striatiferus (Hymenoptera, Braconidae, Opiinae)

Opiinae are very common larval parasitoids of Diptera, but more frequently attack fruit-infesting Tephritidae and mining Agromyzidae37. However, parasitism of Drosophilidae has been occasionally reported38. A few specimens of Areotetes striatiferus Li, van Achterberg and Tan emerged from pupae of Drosophila suzukii obtained from two sites in Yunnan (China). At one site D. suzukii was the only host species to emerge. Areotetes striatiferus was previously known only from Hunan province (China) and its biology was unknown37. It is also the first time that a species of Areotetes is recorded as a parasitoid of a Drosophila species.

Conclusions - Prospects for biological control

For the first time, a large survey for larval parasitoids of D. suzukii was carried out in several Chinese Provinces. These surveys, and those made in Japan, revealed that most populations are parasitized and a complex of at least eight parasitoid species has been identified. The most abundant species collected during these surveys, i.e. Ganaspis cf. brasiliensis and L. japonica, are similar to those observed in previous surveys in Japan24,27 and South Korea29. Except Asobara sp. TK1 that was already known from Japan25,27, all other species found in this survey have been recorded for the first time from D. suzukii and should be further investigated. However, current studies presently on the biology of the parasitoids33, in accordance with previous research24,27, suggest that our Ganaspis cf. brasiliensis, or the suzukii-specialised-race of Ganaspis brasiliensis, is the most specific parasitoid among those associated with D. suzukii. Since it is also the most abundant parasitoid of D. suzukii in Asia, it is clearly the first candidate for introduction into Europe, North America and other regions invaded by D. suzukii. The fact that it also probably attacks two other species also found in fresh fruits in Asia, D. pulchrella and D. subpulchrella, suggests that it may be specific to fresh fruits rather than purely host specific. This would not prevent its introduction in Europe and North America since native Drosophilidae in these regions are not able to attack fresh, undamaged fruits. These surveys also showed that, in some East Asian regions, D. pulchrella and D. subpulchrella are nearly as common in fresh fruits as D. suzukii, and their introduction to other continents should be avoided at all costs. It remains to be discovered whether Ganaspis cf. brasiliensis would be able to colonise all invaded regions or whether it would be limited by climatic constraints. The highest parasitism rates were observed in sub-tropical and warm-temperate climates in Yunnan (China) and Japan. However, the fact that Ganaspis cf. brasiliensis was also rather abundant at a ski resort in Nagano Prefecture (Central Japan: Hasuike, 1490 m average T° in January: ca. −13 °C) suggests that it may establish in temperate regions in Europe and North America. Since larval parasitism of D. suzukii in invaded regions has not been reported, introduction of a parasitoid may help lower populations and reduce the need for other management methods. Nevertheless, before its introduction, completion of host specificity tests and resolution of the taxonomic status of Ganaspis cf. brasiliensis are needed.

Methods

Collection sites and methods

Surveys for D. suzukii and associated parasitoids were carried out in China and Japan from 2015 to 2017. Fruits that could potentially host D. suzukii were collected at more than 100 sites in five prefectures in Japan and 12 provinces in China. Among these, seven sites in Japan (5 prefectures) and 29 sites in China (7 provinces) provided a sufficient number of D. suzukii and parasitoids (arbitrarily set at 10 individuals) to assess parasitism (Figs 1 and 2 and Table S1, Supplementary Information). Only fresh fruits that were still on the plant were sampled to avoid the collection of Drosophila spp. that prefer rotten or damaged fruits on the ground. The protocol to obtain Drosophila spp. and parasitoids varied slightly among years and regions but, in general, was as follows. Collected fruits were placed in a cooler box during their transport to the laboratory. Records of the location (GPS coordinates) and names of the collected fruit were annotated for each sample. In most cases, fruits were counted and then placed on a layer of slightly moist cellulose paper in plastic containers of various sizes, with ventilated lids. The boxes were inspected daily and Drosophila spp. and parasitoids that emerged were reared in a cage or placed in alcohol (96%). The cellulose paper was checked and moistened if necessary. After about a week, the paper was inspected and each fruit dissected to collect remaining drosophilid pupae as some larvae may have pupated inside the fruits. All pupae were placed in Petri-dishes on slightly moist cellulose paper. The Petri-dishes were then inspected daily. Emerged drosophilids and parasitoids were either put directly in alcohol or placed in cages for laboratory rearing.

Figure 1
figure 1

Geographic distribution of the sampling sites in China. A: Kunming - Fumin, B: Shiping, C: Dali, D: Panzihua, E: Wenshan, F: Qujing, G: Jiu Mountain - Lija Farm - Yiangtai Mountain, H: Dazhou, I: Xiaoguan and J: Jilin.

Figure 2
figure 2

Geographic distribution of the sampling sites in Japan. K: Tokyo, L: Nara, M: Hasuike and N: Tsukuba.

Identification of Drosophila spp. and parasitoids

Drosophila suzukii, D. pulchrella and D. subpulchrella were identified using keys31,39. Other Drosophila spp. were not determined to species level. Parasitoids were identified by MB, FRPF, MTK and MK using morphological characters and reference collections housed at the National Insect Collection (NIC), NMNH, Washington DC, and the Zoological Institute RAS of St Petersburg, Russia (ZISP). Vouchers for this study are housed at the NIC, the CABI collection in Delémont, the entomological collection of Natural History Museum Basel (NMBA) and the ZISP.

Calculation of parasitism rates

Parasitism rates were calculated, for each sample, by dividing the number of individuals of one or all parasitoid species by the total number of parasitoid and Drosophila spp. adults emerged from the sample. Samples that contained more than 1% of Drosophila spp. unable to attack fresh, undamaged fruits, i.e. other than D. suzukii, D. pulchrella and D. subpulchrella, were discarded because it is supposed that these species have their own assemblage of parasitoids, specialised in rotten habitats, that may be different from those attacking Drosophila spp. specialised in fresh fruits”.

Data Availability

The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.