Non-target effects of neurotoxic insecticides on Ganaspis cf. brasiliensis, a classical biological control agent of the spotted wing Drosophila

Background The spotted wing drosophila, Drosophila suzukii , is an invasive pest causing significant economic losses worldwide. Current pest control strategies mainly rely on insecticides, which negatively impact fruit marketability and the sustainability of integrated pest management (IPM) programs. In addition, pesticides can have dramatic con‑ sequences on non‑target species when persisting in the environment at low concentrations after field applications. In this context, chemical control can strongly interfere with the releases of the G1 strain of the Asian larval parasitoid Ganaspis cf . brasiliensis , which is currently the adopted classical biological control agent to manage D. suzukii infesta‑ tions worldwide. Methods Probit analysis was used to assess the baseline toxicity of acetamiprid, cyazypyr, lambda‑cyhalothrin, phosmet, and spinosad on G1 G. cf . brasiliensis adults through residual contact exposure in the laboratory. Then, adult parasitoids were exposed to insecticide low Lethal Concentrations (LC 5 and LC 30 ) and their mortality was checked daily to assess the survival of treated wasps. Results Lambda‑cyhalothrin showed the highest toxicity on the parasitoid with a LC 50 of 1.38 × 10 –3 g active ingredi‑ ent (a.i.) /L, while cyazypyr seemed the safer active ingredient with an estimated LC 50 of 0.20 g a.i./L without affect‑ ing parasitoids at sublethal doses. Spinosad and phosmet significantly reduced wasp survival at both LC 30 and LC 5 , while lambda‑cyhalothrin and acetamiprid affected parasitoid lifespan only at LC 30 . Spinosad, lambda‑cyhalothrin and phosmet LC 30 caused the major survival reductions, followed by acetamiprid LC 30 . The least significant reduction in parasitoid survival was 21.6% by spinosad LC 5 . Conclusions Overall, this study highlighted the importance of carefully selecting insecticides to minimize adverse effects on non‑target organisms. In particular, cyazypyr was the most promising candidate to integrate inocula‑ tive biological control with chemical treatments. By contrast, the application of phosmet, spinosad and lambda‑ cyhalothrin should be avoided alongside parasitoid field releases. Although acetamiprid is less used against D. suzukii in the field than the other tested molecules, it should be used with caution due to its sublethal toxicity on the para‑ sitoid. These results provide the first evidence of G . cf. brasiliensis susceptibility to insecticides in order to promote sustainable and efficient pest management strategies.


Introduction
Classical biological control (CBC) is considered the most promising strategy to control the spotted wing drosophila, Drosophila suzukii (Matsumura) (Diptera: Drosophilidae) (Lisi et al. 2022), which is an Asian pest of soft-skinned fruits causing severe economic losses worldwide (De Ros et al. 2020;Boughdad et al. 2021).The CBC approach consists of the importation of specialist natural enemies from the pest's native range (Heimpel and Mills 2017).This is because they are more efficient in targeting the prey/host and providing a stable control of D. suzukii populations than natural enemies resident in the pest infested areas (Wang et al. 2020a, b).Indeed, some pupal parasitoid species showed promising effectiveness under laboratory conditions in invaded countries (Kaçar et al. 2017), but not in the field because too generalist (Lee et al. 2019), while larval parasitoid success is prevented by the lack of co-evolution with D. suzukii due to a strong immune response of the pest against the wasp eggs and larvae (Kacsoh and Schlenke 2012).In this context, shortage of effective natural enemies prompted the need for foreign explorations in South Korea, China, and Japan targeting D. suzukii co-evolved parasitoids (Daane et al. 2016;Girod et al. 2018a;Giorgini et al. 2019).
Field surveys and laboratory quarantine investigations identified the larval parasitoid Ganaspis cf.brasiliensis (Ihering) (Hymenoptera: Figitidae) as the most suitable candidate for the CBC program (Wang et al. 2020a, b;Biondi et al. 2021;Daane et al. 2021), although genetic and molecular studies raised uncertainty on the parasitoid's taxonomic status.Indeed, four to five genetic groups (G1-G5), mainly differing in the host range, were identified in the G. cf.brasiliensis complex (Nomano et al. 2017;Giorgini et al. 2019;Seehausen et al. 2020).Among these, the G1 strain resulted as the most hostspecific lineage mostly parasitizing D. suzukii larvae within fresh and ripe fruits (Girod et al. 2018a;Seehausen et al. 2022).
These features prompted various governments, such as the Italian and the American one, to grant approvals for an area-wide release program of the parasitoid G1 strain (Beers et al. 2022;Lisi et al. 2022), and first release efforts confirmed the parasitoid's ability to disperse, overwinter and parasitize D. suzukii in the field (Fellin et al. 2023).Despite these encouraging results, assessing the parasitoid's compatibility within the integrated pest management (IPM) is crucial for ensuring greater synergy (Tait et al. 2021;Kenis et al. 2023).In this context, the abundant presence of D. suzukii close to harvest season has induced growers to rely on chemical control, which is mainly based on conventional and broad-spectrum insecticides sprayed according to calendar schedules (Tait et al. 2021).Among these, pyrethroids, carbamates, organophosphates and diamides are the most common insecticides used by conventional farmers (Shawer et al. 2018;Shaw et al. 2019;Tait et al. 2021), while spinosad, azadirachtin and pyrethrins are among the best options for the organic management of D. suzukii (Gress and Zalom 2019;Noble et al. 2023).To date, most of the toxicological studies have focused on application methods, insecticide resistance and lethal toxicity towards D. suzukii (Van Timmeren et al. 2018;Mermer et al. 2021Mermer et al. , 2023;;Noble et al. 2023).Toxicity and sublethal effects of insecticides have been evaluated for several non-target biocontrol arthropods species (Desneux et al. 2007), including predators (Biondi et al. 2012;Ricupero et al. 2020) and parasitoids (Biondi et al. 2013;Teder and Knapp 2019), but the agrochemical impact on D. suzukii parasitoids is still neglected.In particular, only few studies showed that synthetic neurotoxicants and spinosad can affect viability of pupal parasitoids (Cossentine and Ayyanath 2017;Schlesener et al. 2019;Morais et al. 2022), and strongly compromise their effectiveness as biological control agents at sublethal concentrations (Lisi et al. 2023).However, no toxicological reports are available for G. cf.brasiliensis and other D. suzukii larval parasitoids.
In this scenario, we first aimed at evaluating the baseline toxicity of five insecticides on adults of G1 G. cf.brasiliensis exposed by residual contact to chemicals.Then, we assessed sublethal effects of low insecticide concentrations (Lethal Concentrations 5 and 30%, LC 5 and LC 30 ) on parasitoid survival.These results provide a first screening on insecticide toxicity on G. cf.brasiliensis and can support ongoing biological control efforts in Europe and the US.

Insect colonies
An isofemale colony of D. suzukii was established in September 2015 from a field sampling on infested wild blackberries (Rubus sp.) in the Catania province (Sicily, Italy).Adult flies were fed with a nutrient honey-water solution (1:1) and rearing conditions were kept constant with a photoperiod of 16: 8 (L:D) at 24 ± 2 °C and 60 ± 10% R.H. inside insect cages (BugDorm ® , MegaView, Taiwan, 32.5 × 32.5 × 32.5 cm).A cornmeal artificial diet within Dutscher rearing tubes (Ø x h: 25 × 95 mm) was provided to adult females as oviposition substrate and food source to the larvae after the egg hatching.Artificial diet was prepared according to the following recipe.Briefly, 45 g Agar, 125 g cornmeal, 200 g sugar, 70 g yeast in granules were added in 4.8 L of boiling water, and 3.3 g of methyl paraben (Methyl-4-hydroxybenzoate) dissolved in 33.3 mL of 95% ethanol and 25 g of 1 M propionic acid were added to avoid the formation of mould and bacteria.
A colony of G1 G. cf.brasiliensis was initiated with specimens provided by the Sicilian Phytosanitary Service (Regione Siciliana), in the context of the classical biological control program (Lisi et al. 2022) in collaboration with the Edmund Mach Foundation (FEM) (San Michele all' Adige, Italy), in July 2021.Wasps were originally sampled, in 2017, in Japan during foreign explorations targeting D. suzukii co-evolved parasitoids (Girod et al. 2018b) and then reared in quarantine laboratories at CABI (Delemont, Switzerland) and later at FEM until government approval for field releases (Lisi et al. 2022;Fellin et al. 2023).Parasitoids were then reared at the Department of Agriculture, Food and Environment of the University of Catania (Catania, Italy), according to Rossi-Stacconi et al. (2022), on fresh blueberries previously infested by D. suzukii.
Spinosad is a naturally derived insecticide compatible with conventional and organic pest management programs, while the other molecules are synthetic insecticides currently employed in D. suzukii agroecosystems (Table 1).Phosmet was recently banned for field application in European Union (EU pesticide database 2022).Label information of each pesticide is shown in Table 1.

Chemical residual exposure and insecticide baseline toxicity on Ganaspis cf. brasiliensis
The experimental arenas were composed by a PVC cylindric-shaped section (88 cm 3 ) interposed between two glass plates (9 × 9 cm) held together by two elastic strands.To prevent fumigant effect of insecticide residues inside the arena, forced ventilation was triggered by a ventilator (100 L/h) (Air fizz 100 ® , Ferplast Spa) able to recycle air approximately once an hour in five experimental arenas, through PVC tubes connections.The airflow movement in and out was ensured by two specular holes (diameter 0.3 cm) covered with a fine mesh net integrated into the plastic section.A third hole in the PVC section allowed for the introduction of wasps with a mouth aspirator and the placement of a cotton dispenser filled with a honey-water solution (1:1) to feed adult wasps during the chemical exposure (Fig. 1).
For each tested insecticide concentration, glass plates were sprayed with 1.0-1.2mL of insecticide solution under a constant pressure of 34, 5 kPa using the Potter spray tower (Burkard Manufacturing Co. Ltd) and ensuring a standard deposit of 1.5-1.8mg/cm 2 of insecticide solution on each surface area (Suma et al. 2009).Control arenas were sprayed with distilled water.The treated glass plates were allowed to dry for 1 h inside a laminar flow hood and were then assembled to expose wasps to the inner and contaminated surfaces inside the arena.At the end of the laboratory trials, glass sections were cleaned with a 10% KOH solution before reuse.
Insecticide concentration-mortality response was assessed by exposing five G. cf.brasiliensis adult couples, three days old, in each arena for 48 h.For each insecticide, concentration-mortality regression lines were assessed by testing a range of six to eight decreasing concentrations starting from the label field rate (Table 1).Following preliminary observations, acetamiprid, cyazypyr and lambda-cyhalothrin label field doses were decreased with a geometrical ratio.While spinosad and phosmet were tested by decreasing the label rate with a logarithmic ratio.Each concentration was replicated three to six times and consisted of five G. cf.brasiliensis adult couples tested.Insecticide dilutions were prepared individually in a laminar flow hood using distilled water, starting from the recommended label dose for each chemical compound (Table 1).Mortality of treated wasps was assessed after 48 h of exposure by stimulating wasps with a fine paintbrush and considering them dead in absence of parasitoid movements after being touched.

Effects of insecticide sublethal concentrations on parasitoid survival
The survival of 17-20 male and female parasitoids following sublethal chemical exposure was tested for each insecticide and concentration.Ganaspis cf.brasiliensis adults, 3-4 days old, were exposed to the estimated LC 30 or LC 5 of each insecticide, as described in the baseline toxicity trials, and control arenas were treated with distilled water.After chemical exposure, survived wasps were individually moved into Dutscher rearing tubes (Ø x h: 28.5 × 95 mm).A moistened cottony ball was placed on the bottom of the tube to ensure proper humidity and honey drops were applied once a week on the inner part of the tube lid as food source.Parasitoid mortality was monitored daily to calculate survival of treated male and female wasps following the sublethal chemical exposure.

Data analysis
The baseline toxicity of tested insecticides on G. cf.brasiliensis adults, exposed via residual contact, was assessed using a probit regression model through a logarithmic transformation of the data (Finney, 1971).The dosemortality relationships were considered valid when there was no significant deviation between the observed and expected data (P > 0.05).Survival raw data were analysed for normality and homogeneity of variance using the Kolmogorov-Smirnov and Levene tests.A general linear model (GLM) analysis was performed to assess the effects of the "insecticide" (five insecticides and untreated control), "concentration" (LC 30 and LC 5 ), and "sex" (male and female wasps), as well as their interactions, on the survival of exposed parasitoids.Additional one-way ANOVA followed by Tukey post-hoc test (P < 0.05) was used for multiple mean comparisons among insecticides within each tested concentration.The unpaired student t test (P < 0.05) was employed to reveal significant differences between the two concentrations of each tested insecticide.All statistical analyses were performed at 95% level of significance using IBM SPSS Statistics for Windows version 20.00 (IBM Corp. 2011.Armonk, NY: IBM Corp.), and Excel ® (Microsoft) was used to generate means, standard errors (SE) and graphs.

Insecticide baseline toxicity on Ganaspis cf. brasiliensis
The probit model was fitted to the observed data for all the treatments (i.e., there was no significant differences between the observed data and the expected data withall treatments being at P > 0.05) and the estimation of lethal concentrations was considered valid (Table 2).Parasitoids treated only with distilled water survived throughout the period assessment.
Following 48 h of exposure to insecticide residues, the highest toxicity towards G. cf.brasiliensis was estimated for lambda-cyhalothrin, as it exhibited the lowest LC 50 values, followed by spinosad, phosmet, and acetamiprid, respectively.Cyazypyr was the least toxic molecule to G. cf.brasiliensis showing the highest concentration required to kill 50% of treated wasps.The same toxicological trend was observed for the two sublethal concentrations, although spinosad LC 5 was more toxic than lambda-cyhalothrin LC 5 with values of 1.20 × 10 -5 and 1.4 × 10 -4 g a.i./L, respectively (Table 2).
Except for cyazypyr, all chemical molecules had lower estimated LCs values than their label field dose.The ratio between the recommended field doses and LC 30 ranged (3) specular hole covered with a fine mesh net to ensure the air movement outside the arena from 0.53 to 274.7 for cyazypyr and fosmet, respectively, and from 2.81 to 10169 for LC 5 of cyazypyr and spinosad.
In the present study, cyazypyr resulted as the least toxic insecticide toward G. cf.brasiliensis, showing an estimated LC 50 equal to 0.20 g a.i./L, which is 0.375fold higher than the recommended field dose against D. suzukii (Table 2).These results align with those of other studies testing cyazypyr toxicity towards biological control agents of insect pests.For example, Zhang et al. ( 2021) estimated a cyazypyr LC 50 of 0.22 g a.i./L on the parasitoid Encarsia formosa (Gahan) (Hymenoptera: Aphelinidae), following insecticide residual exposure.Interestingly, Ahumada and Chorbadjian (2019) reported a cyazypyr LC 50 on Chrysoperla defreitasi (Brooks) (Neuroptera: Chrysopidae) three times higher than that estimated for G. cf.brasiliensis in this study, thus elucidating potential differences in cyazypyr susceptibility between parasitoids and predators.On the other hand, our results showed lambda-cyhalothrin as the most toxic compound toward the parasitoid, with a LC 50 of 1.38 × 10 -3 g a.i./L (Table 2), followed by spinosad, phosmet and acetamiprid.Similar findings were also reported on the larval parasitoid Cotesia flavipes (Cameron) (Hymenoptera: Braconidae) exposed to the dry residues of 12 insecticides commonly used against its host Chilo partellus (Swinhoe) (Lepidoptera: Pyralidae) (Akhtar et al. 2021).In particular, pyrethroids, organophosphates and carbamates exhibited high lethal toxicity towards the parasitic wasp, and lambda-cyhalothrin showed a LC 50 equal to 1.85 × 10 -3 g a.i./L, which is comparable to that obtained on G. cf.brasiliensis.At the same time, acetamiprid and other neonicotinoids were considered safer to C. flavipes than the aforementioned pesticides (Akhtar et al. 2021).
Spinosad is gaining increased relevance in IPM strategies against D. suzukii because it stands out as one of the most effective options in both organic and conventional farming (Gress and Zalom 2019;Tait et al. 2021).The LC 50 estimated for spinosad against G. cf.brasiliensis was equal to 3.94 × 10 -3 g a.i./L (Table 2) , which is very close to the 4.94 × 10 -3 g a.i./L estimated on the larval wasp Oomyzus sokolowskii (Kurdjumov) (Hymenoptera: Eulophidae), following exposure to spinosad residues (Cordero et al. 2007).Moreover, Cordero et al. (2007) showed that spinosad and acetamiprid exhibited higher toxicity on Diadegma insulare (Cresson) (Hymenoptera: Ichneumonidae) than O. sokolowskii, with LC 50 values of 3.4 × 10 -4 and 23.9 × 10 -3 g a.i./L for spinosad and 4.94 × 10 -3 and 35.2 × 10 -3 g a.i./L for acetamiprid, respectively (Cordero et al. 2007).Despite several reports confirming the toxicity results obtained here on G. cf.brasiliensis, it is worth noting that insecticide impact on parasitoids could depend on multiple biotic and abiotic factors, even within the same experimental conditions (Prashar and Shah 2016), as shown by Cordero et al. (2007).Unsurprisingly, lambda-cyhalothrin LC 30 and both concentrations of spinosad and phosmet were the most impactful insecticides on G. cf.brasiliensis, along with a moderate toxicity of acetamiprid LC 30 in reducing survival of treated parasitoids.Several studies reported reduction of wasp lifespan following residual exposures to pyrethroids at non-lethal concentrations (Bayram et al. 2010;Garcia 2011;Desneux et al. 2006).Lambda-cyhalothrin significantly reduced Aphidius colemani Viereck (Hymenoptera: Braconidae) survival at sublethal concentrations estimated for its host (Alfaro-Tapia et al. 2021) and for the parasitoid itself (D' Ávila et al. 2018), following wasp exposure on pesticide dry residues.In the latter case, D' Ávila et al. ( 2018) also showed that A. colemani was about 20-fold more susceptible to spinosad than lambda-cyhalothrin and imidacloprid sublethal doses, with very low concentrations of spinosad residues (i.e., 200 ng a.i./cm 2 ) reducing wasp survival by half.
Interestingly, this neonicotinoid was reported to reduce Peristenus spretus (Chen et van Achterberg) and P. relictus (Ruthe) (Hymenoptera: Braconidae) survival in a range of concentrations (Yang and Lu 2023) similar to the LCs estimated in this study for G. cf.brasiliensis.However, few studies are available for cyazypyr sublethal toxicity on wasp survival.Current literature showed that residual exposures of Trichogramma atopovirilia (Oatman & Platner) (Hymenoptera: Trichogrammatidae) to cyazypyr increased, even not significantly, survival of treated wasps (Cantori et al. 2023), while Tamarixia triozae (Burks) (Hymenoptera: Eulophidae) adults fed on sucrose solution contaminated with cyazypyr at 0.12 g a.i./L lived half than untreated wasps (Liu et al. 2012).In this context, it is well known that pesticide sublethal impact in insects is influenced by a plethora of biotic and abiotic variables and interacting factors.In particular, the pesticide penetration route into the insect body, particularly at sublethal concentrations, could play a key role in determining the amount of insecticide molecules affecting the organism, and therefore the toxicity degree.For example, ingestion is considered the primary route of penetration in insect body both for acetamiprid and cyazypyr (Sparks et al. 2001;Balabanidou et al. 2018), therefore higher oral than contact penetration rate could explain their low and absent toxicity on G. cf.brasiliensis at sublethal doses, as also supported by previous reports (Cantori et al. 2023;Liu et al. 2012).On the other hand, spinosad is highly toxic when ingested and by residual contact, causing quick death of several organisms under both routes of exposure (Biondi et al. 2012;Bacci et al., 2016;Eger & Lindenberg, 1998).
Indeed, in our study spinosad reduced parasitoid survival even when tested at a concentration 10169 times lower than the label rate (Fig. 2).Phosmet and lambdacyhalothrin act as contact insecticides, penetrating insect cuticle and crossing biological membranes and tissues due to their biochemical structure (He et al. 2008).Therefore, their strong impairment on G. cf.brasiliensis lifespan could be related to a higher penetration rate into the parasitoid body than the other tested molecules.However, many other features would influence the non-target effects of these insecticides at very low concentrations (Pazini et al. 2019).Special emphasis should be placed on their secondary mechanisms of action under a sublethal scenario, which can strongly interfere with the behaviour and physiology of surviving insects (Guedes et al. 2016).Indeed, several studies revealed that pesticide sublethal doses may interfere with the feeding and mating behaviour, as well as many other physiological processes influencing parasitoid longevity.This suggests that alterations of a specific key life-history trait of the parasitoid could indirectly affect other behavioural and/or physiological features and directly impact on the entire biocontrol service provided by natural enemies (Desneux et al. 2007).
According to these results, the significance of evaluating insecticide sublethal effects on parasitoid survival is closely related to the general assumption that longevity is a main component of individual fitness and population dynamic in hymenopteran parasitoids (Jervis 2007).Indeed, the longer parasitoids can survive, the longer they can mate and lay eggs, ultimately leading to increased parasitism and host mortality (Jervis 2007).In this context, decreased wasp survival by pesticides can result in cascading effects at the population level, which mainly depend on the parasitoid biological features (Desneux et al. 2007).Ganaspis cf.brasiliensis is a weakly pro-ovigenic species (Wang et al. 2018), which means that females are characterized by a short oviposition period, short lifespan and by decreased fecundity with increased female age (Jervis et al. 2001).From a practical perspective, pesticides persisting at sublethal concentrations in the agroecosystem decrease parasitoid survival, therefore they would cause an additional decrease of the available oviposition timeframe and the effectiveness of parasitoids.Beside our hypothesis, it is worth noting that life table analysis should be used to better understand the implications of reduced survival at the population level (Desneux et al. 2007).To the best of our knowledge, there are no reports of successful programs integrating parasitoid release with chemical control for the integrayed management of D. suzukii, although different reports against other pests showed the feasibility of combining biological and chemical control (Mansour and Biondi 2021;Wright and Verkerk 1995).As classical biological control programs continue to be developed against this invasive pest worldwide, these results could be used as a proxy for future ecotoxicology research focusing on other key traits of G. cf.brasiliensis, such as fertility, fecundity and parasitism behaviour.Moreover, further routes by which the parasitoid could be exposed to insecticides (e.g., topical exposure of adult parasitoid, juvenile parasitoid intoxication while developing on treated hosts) need to be investigated, as well as semi-field and field evaluations are required to achieve a comprehensive understanding of parasitoid susceptibility to insecticides, therefore on its potential compatibility with chemical control within the current D. suzukii IPM programs.

Conclusions
This study sheds light on the susceptibility of G1 G. cf.brasiliensis to five insecticides commonly used for managing D. suzukii infestations.Cyazypyr would be the most recommended insecticide to integrate parasitoid releases with the current IPM strategies, as a result of its low toxicity and lack of sublethal effects on the parasitoid lifespan.Acetamiprid was considered as a moderate risk molecule because it was the least impactful insecticides among the toxic ones.Moreover, it is not frequently employed in the agroecosystem affected by D. suzukii.The application of lambda-cyhalothrin, phosmet and spinosad would not be recommended due to their strong non-target effects at sublethal and low lethal concentrations.Further research on other key traits of G. cf.brasiliensis and field trials would allow to better evaluate the toxicological profile of these insecticides and integrate this biological control agent within the current IPM strategies.

Fig. 1
Fig. 1 Lateral (left) and ventral (right) view of the experimental arena used to expose Ganaspis cf.brasiliensis adults to dry insecticide residues.(1) ventilator triggering forced ventilation inside the arena to prevent fumigation (each ventilator was connected to five experimental units, despite the figure shows only one); (2) hole to introduce wasps in the arena and place cotton dispenser filled with a honey-water solution; (3) specular hole covered with a fine mesh net to ensure the air movement outside the arena

Fig. 2
Fig. 2 Mean (± SE) Ganaspis cf.brasiliensis adult survival following LC 30 and LC 5 chemical exposure by residual contact.Different capital letters indicate significant differences among insecticides within the LC 30 exposure, while different lower-case letters indicate significant differences in the LC 5 treatments (Tukey post hoc tests, P ≤ 0.05).Asterisks show significant differences between the LC 30 and LC 5 for each tested insecticide according to the unpaired student t test (P ≤ 0.05)

Table 1
Tested insecticides and label information

Table 2
Results of the baseline toxicity assessment for five insecticides on Ganaspis cf.brasiliensis adults by residual contact FR field rate, SE standard error, χ2 chi-square testing goodness of fit of concentration-mortality response, df degree of freedom, FR/LC ratio between Field Rate and Lethal Concentration