Insecticidal and oviposition deterrent effects of essential oils of Baccharis spp. and histological assessment against Drosophila suzukii (Diptera: Drosophilidae)

The diverse flora of the Atlantic Forest is fertile ground for discovering new chemical structures with insecticidal activity. The presence of species belonging to the genus Baccharis is of particular interest, as these species have shown promise in pest management applications. The objective of this study is to chemically identify the constituents expressed in the leaves of seven species of Baccharis (B. anomala DC., B. calvescens DC., B. mesoneura DC., B. milleflora DC., B. oblongifolia Pers., B. trimera (Less) DC. and B. uncinella DC.) and to evaluate the toxicological and morphological effects caused by essential oils (EOs) on the larvae and adults of Drosophila suzukii (Diptera: Drosophilidae). Chemical analysis using gas chromatography-mass spectrometry (GC–MS) indicated that limonene was the main common constituent in all Baccharis species. This constituent in isolation, as well as the EOs of B. calvescens, B. mesoneura, and B. oblongifolia, caused mortality in over 80% of adults of D. suzukii at a discriminatory concentration of 80 mg L−1 in bioassays of ingestion and topical application. These results are similar to the effect of spinosyn-based synthetic insecticides (spinetoram 75 mg L−1) 120 h after exposure. Limonene and EOs from all species had the lowest LC50 and LC90 values relative to spinosyn and azadirachtin (12 g L−1) in both bioassays. However, they showed the same time toxicity over time as spinetoram when applied to adults of D. suzukii (LT50 ranging from 4.6 to 8.7 h) in a topical application bioassay. In olfactometry tests, 92% of D. suzukii females showed repellent behavior when exposed to the EOs and limonene. Likewise, the EOs of B. calvescens, B. mesoneura, and B. oblongifolia significantly reduced the number of eggs in artificial fruits (≅ 7.6 eggs fruit−1), differing from the control treatment with water (17.2 eggs fruit−1) and acetone (17.6 eggs fruit−1). According to histological analyses, the L3 larvae of D. suzukii had morphological and physiological alterations and deformations after exposure to treatments containing EOs and limonene, which resulted in high larval, pupal, and adult mortality. In view of the results, Baccharis EOs and their isolated constituent, limonene, proved to be promising alternatives for developing bioinsecticides to manage of D. suzukii.

Based on the concentration-response curves and the overlapping confidence intervals of the LC 50 and LC 90 values for the ingestion and topical application bioassays, we found that these values were lower for all Baccharis EOs and limonene than for the spinosyn-and azadirachtin-based insecticides after 120 h of exposure (Table 2). Topical application of the Baccharis EOs and spinosyn showed no difference in LT 50 values, which ranged from 4.55 to 8.71 h (Table 3). Meanwhile, the spinosyn-based insecticide had the lowest LT 50 value in the ingestion bioassay (17.95 h; Table 3).

Discussion
This study provides the first verification that EOs extracted by hydrodistillation from the leaves of seven species of Baccharis and limonene, a constituent of these EOs, exhibit high toxicity against adults and larvae of D. suzukii. The EOs of Baccharis species are known for the predominance of monoterpenoids and sesquiterpenoids 1,4,5 , which have been reported to have the potential to cause mortality in different larval stages 31 , malformations in adults 31 , and to repel insects 7 . Of the different Baccharis species examined in this study, the only one whose oil has been reported in the literature as having insecticidal properties is B. trimera, which has been shown to be effective against pests of stored products 32 . www.nature.com/scientificreports/ www.nature.com/scientificreports/ The gas chromatography-mass spectrometry (GC-MS) analysis showed that limonene was the only major constituent found in all Baccharis species, the content of which varied between 12.5 and 88.8% in the studied species. In Brazil, limonene is a product marketed for use in treatments against fleas in domestic animals in the form of shampoos, sprays, and aerosols 8 . However, previous studies have found that the compound exhibits toxic activity against several arthropods, such as Leptinotarsa decemlineata Say (Coleoptera: Chrysomelidae )33 , Sitophilus zeamais Motschulsky, (Coleoptera: Curculionidae) 34 , Tribolium confusum du Val (Coleoptera: Tenebrionidae) 35 , Figure 2. Drosophila suzukii mortality when submitted to various treatments in topical application and ingestion bioassays. Means followed by different letters on the columns (within each exposition bioassay) indicate significant differences between treatments (GLM with quasi-binomial distribution followed by post hoc Tukey test, P < 0.05). Table 2. Estimation of LC 50 and LC 90 (in mg L −1 ) and confidence interval of Baccharis spp., limonene, spinosyn-based synthetic insecticide and azadirachtin on adults of Drosophila suzukii at 120 HAE in topical bioassays and ingestion. df degrees of freedom. a LC 50 and LC 90 : Insecticide concentrations (mg L −1 ) required to kill 50% or 90% of D. suzukii adults, respectively (CI 95% confidence interval). b LC 50 34 , and Tyrophagus putrescentiae Schrank (Acari: Acaridae) 36 . Besides, in vitro bioassays reduced feeding by larvae of Thaumetopoea pityocampa Schiff (Lepidoptera: Thaumetopoeidae) on leaves of Pinus spp. 37 and feeding by Bemisia tabaci Gennadius (Hemiptera: Aleyrodidae) on Solanum esculentum L. 38 ; they also had a repellent effect on Coptotermes formosanus Shirak (Isoptera: Rhinotermitidae) in wood 39 . Similar effects were observed in the larvae and adults of D. suzukii in this study. We also found high concentrations of other constituents, including α-pinene, β-pinene, spatulenol, and thujopsan-2-α-ol (B. calvescens); carquejyl acetate and palustrol (B. trimera); β-pinene(B. milleflora); α-pinene, β-pinene www.nature.com/scientificreports/ (B. mesoneura); α-thujene, α-pinene, β-pinene (B. oblongifolia); and β-pinene, thujopsan-2-α-ol, globulol (B. anomala). These findings are corroborated by previous studies that found that Baccharis species contained large amounts of monoterpene hydrocarbons (α-thujene, α-pinene, β-pinene, and limonene), oxygenated monoterpenes (carquejyl acetate), and oxygenated sesquiterpenes (palustrol, spathulenol, and thujopsan-2-α-ol) 5 . However, for both bioassays performed in this study, we found that the substances contained in the EOs of B. calvescens, B. mesoneura, and B. oblongifolia had the greatest effect on adults of D. suzukii, with similar mortality rates (over 90%) to synthetic spinosyn-based insecticides. These species showed efficacy comparable to the organophosphates, pyrethroids and spinosyns used to management adults of D. suzukii [40][41][42][43] . It is known that the potential of EOs depends on the chemical constituents and their proportions found in the samples 44 . Likewise, the interactions of constituents contained in EOs have been reported to have synergistic action, providing a significant increase in the effectiveness of formulations 44,45 . The insecticide azadirachtin, meanwhile, showed the lowest toxicity on adults of D. suzukii. However, even though this product exhibits low toxicity for this pest, it may favor pest suppression by repelling the insects or reducing oviposition capacity, as verified in a previous study 40 .
In the topical application bioassays, we observed that adults of D. suzukii died more quickly (LT 50 of 4.55-11.78 h) than during the ingestion bioassays (LT 50 of 42.76-66.41 h). This difference in the toxicity of Baccharis spp. oils evaluated using the two bioassay methods can be attributed to the fact that topically applied EOs directly penetrate the insect hemolymph in a single dose. In contrast, ingested EOs are administered gradually and in small amounts over the feeding period (24 h). This also suggests that the higher toxicity by topical application results from damage to the nervous and/or respiratory systems of insects since these are the main routes of intoxication by substances absorbed by the cuticle 46 . Furthermore, during the ingestion period, treatments remain in the intestine of the insects for longer, requiring a longer time for metabolization and/or excretion of the chemical 29 . These results too may be related to the lipophilic constitution and the low molecular weight of  Table 4. Larval mortality (LM), pupation rate (PR), pupal mortality (PM), and deformity of Drosophila suzukii adults exposed to different treatments. Columns followed by the same letter are not significantly different from one another (GLM with an almost binomial distribution followed by Tukey's test: P > 0.05).  www.nature.com/scientificreports/ the chemical constituents of these EOs 47 . These characteristics may enable diffusion through the cellular membrane, causing physiological disruptions in the insect membrane and leading to mortality 48,49 . Likewise, they can trigger the inhibition of acetylcholinesterase (AChE) activity, which has been verified in adults of S. zeamais, T. castaneum 50 and the spider mite, Tetranychus urticae Koch (Acari ) 51 .
In addition to their high toxicity, the EOs of B. calvescens, B. mesoneura, and B. oblongifolia reduced the oviposition capacity of D. suzukii by up to 43%. This fact corroborates the observations in the double-choice olfactometry repellency tests, in which females of D. suzukii avoided the olfactometer arm that contained a piece of filter paper containing 5 µL of EOs, preferring instead to move into arm containing the negative treatment (acetone). Products that reduce oviposition or repel D. suzukii females can reduce the incidence of epidermal rupture by oviposition, which consequently reduces phytopathogen infestation 43 , while also avoiding the attraction of other drosophilids such as Zaprionus indianus Gupta (Diptera: Drosophilidae), which can accelerate damage to crops, as seen in strawberry 13 and persimmons 15 . In addition, it helps decrease pest population density in crops 13,29 .
In addition to their repellent effects and, consequently, their ability to reduce oviposition, B. anomala, B. calvescens, B. mesoneura, B. milleflora, B. oblongifolia, and limonene had a major impact on the L3 larvae of D. suzukii. Specifically, they were able to affect the species' pupation rate and pupal mortality negatively. The larvicidal effect of these materials may be related to the polarity of the EOs (lipophilic substances), which allows oils to penetrate the cuticle of the larvae, interfering in their physiological functions 52  Most insecticides used to control D. suzukii act on the AChE receptors or the sodium [23][24][25]41,57,58 . Products that use a different mode of action can thus help avoid the emergence of resistance to such compounds 27,28 . Studies of these morphological markers, including the histopathological evaluation of larvae, is of the utmost importance when seeking to understand how exposure to EOs and their individual constituents can damage target cells. In this study, we observed morphological damage to organs such as the brain, fat body, and Malpighian tubules of D. suzukii larvae subjected to the EOs of Baccharis spp. and limonene. We examined these organs in particular because the brain is the organ that transmits the stimuli received through physical and chemical impulses, while the fat body and Malpighian tubules are the main sites of metabolization and excretion of substances, analogous to the liver and kidney of vertebrates, respectively 59 . www.nature.com/scientificreports/ In this study, B. calvescens, B. mesoneura, and B. oblongifolia were shown to have neurotoxic mechanisms, including the neurodegeneration and alteration of the morphology of the cortical layer and neuropils. Similar observations were reported with larvae of Cochliomyia macellaria (Diptera: Calliphoridae) after being treated with the oil of Curcuma longa L. 52 . In that study, the authors demonstrated the occurrence of vacuolar degeneration and alteration of the hypnotic profile of the brain. Also, D. suzukii larvae exposed to EOs showed damage to the adipose body, including cytoplasmic vacuolization and irregular morphology of trophocytes with hypnotic nuclei, signaling a possible mechanism of excretion of EOs. This process of vacuolization may indicate that these cells are in the process of dying, as has been demonstrated in larvae of C. quinquefasciatus 60 . Besides, we observed that the fat body nucleus of D. suzukii larvae was divided into smaller, highly condensed fragments when exposed to limonene and advanced disintegration of the brush border and nuclear chromatin condensation of the Malpighian tubules was caused by the EOs of B. milleflora, B. trimera, and B. uncinella. These results corroborate those described for Apis mellifera (Hymenoptera: Apidae )61 and C. macellaria 52 . These physiological disturbances caused by EOs and limonene in D. suzukii larvae are typical of cells submitted to classical apoptosis 62 , consisting of self-destruction of cells into smaller, highly condensed fragments.
The results found in the study of larval and adult D. suzukii clearly demonstrate the toxic activity and sublethal effects of the EOs of Baccharis spp. and limonene, an isolate of these EOs. Furthermore, this study is the first to verify the histological effects of EOs on D. suzukii larvae. This can help to determine the action sites of these compounds on insects. However, considering that these findings has not been fully explained, we are aware that new tests, focused especially on the selectivity of these botanists over natural enemies intentionally released 63,64 and naturally present in the environment 65,66 , may in the future subsidize methods for integrating natural enemies and the development of EO-based biopesticides. Despite this, the use of these substances as such has limitations due to flammability, low dispersion in water, phytotoxicity [67][68][69] . In this sense, the development of formulations based on stable EO reduces these negative aspects and, at the same time, improves the effectiveness against pests and reduces the side effects on the beneficial ones. We are currently conducting work to investigate

Material and methods
Collection of plant material for essential oil extraction. Chemical analysis of essential oils: identification and quantification. We performed GC-MS using a Shimadzu 2030 gas chromatograph coupled to a Shimadzu TQ8040 sequential mass detector (GC-MS/ MS). The GC was equipped with a fused HP-5MS capillary column (film thickness 30 m × 0.25 mm × 0.25 μm) coated with a stationary phase of 5% phenyl-95% dimethylpolysiloxane. Helium was used as a drag gas at a flow rate of 1.0 mL min −1 . The temperature setting was set to increase from 60 to 240 °C at a rate of 3 °C min −1 and held at 240 °C for 10 min. The injector temperature was maintained at 250 °C. The essential oil samples were diluted into a 1% hexane solution, and 1.0 μL of the solution was injected with a partition ratio of 1:30. The mass detector was operated in electron impact mode (70 eV). The transfer line was kept at 260 °C and the ion source at 250 °C. For quantification, essential oils were injected, and the Shimadzu GC 2030, equipped with a flame ionization detector (FID), was operated at 250 °C. Synthetic air was used as a carrier gas at a flow rate of 1.5 mL min −1 , using the same column and conditions described above. The quantification of each constituent was estimated by the FID detector with the corresponding peak area, which was determined using the average of three injections ( Table 1). The identification of the SB components was performed by comparing the mass spectra with those of commercial libraries 70 , as well as by their linear retention rates 71 , after the injection of a homologous series of alkanes (C 8 -C 26 ) under the same experimental conditions, and compared with data in the literature 72 . The structure of limonene was confirmed by injecting a commercial standard solution (Sigma-Aldrich Brazil). www.nature.com/scientificreports/ Breeding and maintenance of Drosophila suzukii. The adults of D. suzukii used in bioassays were in their tenth generation. Breeding was performed using insects collected in the strawberry fields (Fragaria × ananassa Duchesne) in January 2018 in Curitiba, Paraná, Brazil (31° 38′ 20′′ S, 52° 30′ 43′′ W). In the laboratory, the infested strawberries were placed individually in plastic jars (150 mL) with a perforated lid (2 cm in diameter) and covered with cheesecloth containing a thin layer of vermiculite (1 cm). The fruits were kept in an airconditioned room (25 ± 2 °C, 70 ± 10% RH, and 12-h photoperiod) until the emergence of adults. Following emergence, the adults (males and females) were transferred to glass bottles (300 mL) containing an artificial diet (12 mL) 73 . Seven-day-old adults were used in all bioassays, which were deprived of food for 8 h, though they were supplied with water in hydrophilic cotton.
Discriminatory bioassays (initial experiment). In order to evaluate the lethal toxicity of Baccharis spp.
EOs and limonene, initial tests were performed using ingestion bioassays and topical application using discriminatory concentrations on adults of D. suzukii (Table 5). For the ingestion bioassays, 16 adults of D. suzukii (eight couples) were grouped in transparent plastic cages (1 L) inverted in plastic Petri dishes (25 cm diameter). The top side of the cages (i.e., the bottom of the containers) was sealed with a cheesecloth-type fabric to allow gas exchange. Once the solutions (treatment) were prepared, the products were supplied to the flies by capillarity in hydrophilic cotton rolls inside a 10 mL glass bottle. After 24 h of exposure, the treatments were removed and replaced with an artificial diet and distilled water until the end of the evaluation period.
In the topical application bioassay, adults of D. suzukii (ten couples) were separated and placed in transparent glass tubes (1.3 cm in diameter × 10 cm in length), which were closed at the top with hydrophilic cotton. Subsequently, the flies were transferred to a petri dish (9 cm in diameter) lined with filter paper and sedated in ethyl ether for 40 to 60 s to apply the treatments. The solutions (2 mL) were then sprayed using a Potter Tower (Burkard Scientific, Uxbridge, UK) at a working pressure of 0.049 MPa, resulting in an average residue deposition of 1.0 mg cm −2 . After spraying, the insects were placed in transparent plastic cages (1 L) as described above and fed an artificial diet and distilled water throughout the evaluation period.
In both tests, the experimental design was entirely randomized, with 10 treatments containing five repetitions (cages) with 16 adults (eight couples) in the ingestion bioassay and four repetitions (cages) with 20 adults (ten couples) in the topical bioassay. Mortality in each treatment was evaluated at 1 h intervals for the first 24 h after exposure to treatments (HAET) and every 24 h between 24 and 120 HAET. Insects that did not react to the touch of a fine-tip brush were considered dead. The corrected mortality was calculated using Abbott's formula 74 .

Concentration-response curves and average lethal time of the most promising treatments against Drosophila suzukii.
Based on the initial bioassays, the most promising treatments were selected and submitted to a new bioassay to estimate the lethal concentrations that would result in mortality of 50% or 90% mortality among the flies (LC 50 and LC 90 , respectively). Seven concentration ranges were defined for each treatment and exposure type in the bioassay: 25-80 mg L −1 for the EOs of Baccharis spp. (B. anomala, B, calvescens, B. mesoneura, B. milleflora, B. oblongifolia, B. trimera, and B. uncinella) and limonene; 5-75 mg L −1 for spinetoram; and 25-250 mg L −1 for commercial azadirachtin-based bioinsecticide 75 . The exposure and assessment procedures, as well as the mortality criteria, were identical to the initial tests. Four replicates were used in the ingestion bioassays, each containing 20 flies (n = 80) for each insecticide concentration. In the topical bioassays, five replicates were performed with 16 flies per replicate (n = 80) per concentration of each insecticide tested. For the determination of LT 50 values (mean time required to kill 50% of the population) of the treatments on D. suzukii adults, the maximum concentration tested in the bioassays of ingestion and the topical application was used ( Table 5). The experimental design and bioassay procedures were identical to those used in the initial experiments.

Repellent effect against Drosophila suzukii in olfactometer bioassay.
To verify the effectiveness of EOs at repelling females of D. suzukii relative to acetone treatments, we began by placing individual females aged up to 24 h into glass tubes (1.3 cm in diameter × 10 cm in length). In the test, the glass tube containing the female was connected to a double-choice glass olfactometer with a diameter of 8.0 cm and an initial compartment of 20 cm on each side, under fluorescent light (60 W, luminance 290 lx). At the end of one of the olfactometer arms, we placed a filter paper measuring 4 × 10 cm and bent into an accordion shape, which contained 5 µL of an EO of Baccharis spp. (B. anomala, B. calvescens, B. mesoneura, B. milleflora, B. oblongifolia, B. trimera, or B. uncinella) or limonene at the discriminatory concentration (80 mg L −1 of oil). Another filter paper was placed at the end of the other arm (4 × 10 cm), which contained 5 µL of acetone (control). Airflow in the system was supplied at a rate of 0.8 L min −1 from a previously filtered source with active carbon and humidified in distilled water. The olfactometer was washed with neutral soap and hexane after every fourth repetition and then dried in a sterilization oven at 150 °C. After this process, the substances were replaced, and the evaluation www.nature.com/scientificreports/ continued. Each treatment consisted of 40 replicates, each of which consisted of a female of D. suzukii (n = 40). The responses were considered positive (EOs, limonene, or acetone) when D. suzukii females reached the odor source or traveled at least 10 cm inside the olfactory arms and remained there for at least 1 min 53 . Flies that did not move to either of the olfactory arms after one minute of release were discarded. In each treatment, groups of 20 instar III larvae (L3) were fixed in neutral buffered formalin, pH 7.2, at 10% in distilled water for 2 h at 56 °C inside microtubes (2 mL). Acetone (PanReac-UV-IR-HPLC-GPC PAI-ACS, 99.9% purity) was used as the sole negative control. After fixation, the larvae were washed three times in 70% alcohol for 20 min to remove the fixing solution. They were then dehydrated using an increasing alcoholic series (70% to 100%), remaining at each concentration for 30 min. Subsequently, the larvae were diaphanized in xylol for 10 min and transferred to soaking paraffin (overnight), and incorporated in histological paraffin. Five 4 µm thick longitudinal sections were cut with a microtome and placed on microscope slides. Mayer's albumin was applied to the slide under the sections for bonding, after which the slide was dried at room temperature (22 ± 3 °C). Finally, the histological sections were stained with hematoxylin-eosin (H&E). Histological sections were analyzed under the Stemi 508 optical scanning microscope (Carl Zeiss, Germany; 20 or × 40 magnification), and morphological changes in target organs such as brain, fat body, and Malpighian tubules were noted.

Deterrence of oviposition by
Statistical analysis. All bioassays were conducted using a completely randomized design. Generalized linear models (GLM) 76 of the quasi-binomial distributions were used to analyze mortality rate data. In all cases, the fit of the GLM was determined by using the half-normal probability plot with a simulation envelope 77 . When significant differences were found among treatments, multiple comparisons (Tukey test, P < 0.05) via the glht function in the multicomp package with adjusted p values was performed. For comparisons of the average of two treatments in the repellency bioassay, we used the Student's t-test. All of these analyses were carried out using R statistical software, version 2.15.1 78 . A binomial model with a complementary log-log link function (gompit model) was used to estimate the lethal concentrations (LC 50 and LC 90 ), using the Probit Procedure in the software SAS version 9.2 79 . A likelihood ratio test was used to test the hypothesis that the LCp or LTp values (lethal concentration or lethal time at which a percent mortality P is attained) were equal. If the hypothesis was rejected, pairwise comparisons were performed and significance was stated if CIs did not overlap. Finally, the mean lethal time (LT 50 ) was estimated for Probit analysis of correlated data 80 . The percentage repellence (PR) was calculated using the formula 81 : PR (%) = [(Nc − Nt)/(Nc + Nt)] × 100, where Nc was the number of insects present in the negative control (acetone) and Nt was the number of insects present in the treatment (EOs).

Data availability
This article does not report new empirical data or software.