Transgenerational effects from single larval exposure to azadirachtin on life history and behavior traits of Drosophila melanogaster

Azadirachtin is one of the successful botanical pesticides in agricultural use with a broad-spectrum insecticide activity, but its possible transgenerational effects have not been under much scrutiny. The effects of sublethal doses of azadirachtin on life-table traits and oviposition behaviour of a model organism in toxicological studies, D. melanogaster, were evaluated. The fecundity and oviposition preference of flies surviving to single azadirachtin-treated larvae of parental generation was adversely affected and resulted in the reduction of the number of eggs laid and increased aversion to this compound over two successive generations. In parental generation, early exposure to azadirachtin affects adult’s development by reducing the number of organisms, delay larval and pupal development; male biased sex ratio and induced morphological alterations. Moreover, adult’s survival of the two generations was significantly decreased as compared to the control. Therefore, Single preimaginal azadirachtin treatment can affect flies population dynamics via transgenerational reductions in survival and reproduction capacity as well as reinforcement of oviposition avoidance which can contribute as repellent strategies in integrated pest management programs. The transgenerational effects observed suggest a possible reduction both in application frequency and total amount of pesticide used, would help in reducing both control costs and possible ecotoxicological risks.

natural tetranortriterpenoid compound extracted from the neem tree, Azadirachta indica 19 , is considered as one of the most promising plant compounds for pest control in organic agriculture 14,20 . AZA shows variable effects on insects including the model insect Drosophila melanogaster 21,22 . This triterpenoid acts as sterilant, insect growth regulators by disruption of the endocrine system, repellent, oviposition and feeding deterrent by activating bitter sensitive gustatory cells 23,24 . Larval exposure of D. melanogaster to sublethal doses of azadirachtin was found to affects various aspects of their physiology including digestive enzymes 25 and this effect is also further observed in the adults 10 . This pre-imaginal exposure affects not only the physiology and the fitness of flies but also adults oviposition and feeding preference 7,10 .
Most studies on sublethal effects of insecticides are related to continuously or repeated exposure. This exposure provokes a generalized stress and activating a detoxification response such as up-regulated of cytochrome P450 genes which might lead to the detoxification of insecticide and even the development of resistance 26 . Moreover, the up-regulation is thought to provide versatility in environmental adaptation 27 . In botanical insecticide the potential fast desensitization to a feeding deterrent was reported 28,29 . Individual insects initially deterred by feeding inhibitor become increasingly tolerant due to repeated or continuous exposure 29 . Bomford and Isman 15 reported an habitation to pure azadirachtin in the tobacco cutworms which become less sensitive to the antifeedant properties of azadirachtin, but not to a neem containing a same absolute amount of azadirachtin. This might have an important implication to avoid desensitization to commercial neem-based insecticides which contains additional non AZA-compounds 15 . Larval exposure to Neem Azal, a commercial Azadirachtin-rich based formulation, was found to enhance avoidances of this compound in adults of D. melanogaster surviving from previously treated larvae 10,25 . This long-lasting avoidance is related to conditioned aversion and may be related to another mechanism such as sensitization 30,31 which also generally occurs after long term or repeated exposure and may increase avoidance to noxious stimulus 32 . Moreover, increasing evidence has highlighted the critical role of early life experience in adult physiology and behavior in insect 33 . Recent studies have revealed that insect can modulate their behavior on the basis of previous experiences early life and that various insecticide-mediated changes in the directly exposed generation can persist into the subsequent non-exposed generations 34,35 . Previously, we have focused on the impact of larval exposure to azadirachtin on adult's fitness (fecundity, survival) and oviposition site preference of the parental generation of D. melanogaster as a model organism for testing insecticide activity 7 . Current study aimed to evaluate, the possible adverse effects of this prior single exposure to azadirachtin experienced by the preceding generations on life table and oviposition site preference of the filial generations. We monitored the oviposition site preference, fecundity, development, sex ratio, survival and morphological abnormalities of exposed and non-exposed generations. All these parameters were investigated over generations until their restoration to predict the outcome of azadirachtin use on pest management practices.

Results
Fecundity and oviposition site preference. Azadirachtin, topically applied on the 3 rd instar larvae (LD 25 and LD 50 of immature stages) affect fecundity of females by a significant reduction of the number of eggs laid as compared to controls (KW = 24.73; p < 0.001). This reduction was observed over two successive generations (parental and F1), however, the total eggs laid was higher in the unexposed generation (F1) than in parental (P) ones (KW = 50.89; p < 0.001) (Fig. 1). Full restoration of affected fecundity was noted in the second generations (F2).
Results of oviposition preference in the no choice experiments (Fig. 1) revealed a clear preference for oviposition on untreated medium than in azadirachtin-treated ones.
The results concerning the dual choice experiments (Fig. 2) revealed an oviposition preference in control medium than in treated medium for all tested generations (P, F1 and F2). Furthermore, flies previously exposed to azadirachtin (early 3 rd instar larvae) showed a highest aversion to this substance compared to naïve flies and led fewer eggs for the two first generations (P and F1) with a more marked effects for parental generation (P < 0.001).
The oviposition preference index (OPI) of adult females of D. melanogaster exposed, or not, to azadirachtin at larval stage of parental generation were always negative in all generations (Fig. 3).
In the generation P, statistical analysis showed significant differences between OPI of previously treated flies and controls flies with a dose-dependent response (Fig. 3). In addition, for medium 0.1 μg/ml, Mann-Whitney revealed significant effects between LD 25 of the parental generation and the first generation (Mann-Whitney test U = 8; P < 0.001), LD 25 of parental and second generation (Mann-Whitney test U = 20; P = 0.0018) but there was no difference between the first and the second generations (Mann-Whitney test U = 42; P = 0.0887). Similar results were observed for the LD 50 , with significant effects observed between the parental generation and the F1 (Mann-Whitney test U = 19; P = 0.0014), also between P and F2 (Mann-Whitney test U = 34; P = 0.0284) but no difference between F1 and F2 (Mann-Whitney test U = 58; P = 0.4428). For control, there was no difference between all tested generations.
Similar results were obtained for medium 0.25 μg/ml, Mann-Whitney test revealed significant effects between LD 25 of the parental generation and the first generation (Mann-Whitney test U = 25; P = 0.0045), LD 25 of parental and second generation (Mann-Whitney test U = 24; P = 0.0045) but there was no difference between the first and the second generations (Mann-Whitney test U = 66; P = 0.5512). For the LD 50 , significant effects were observed between the parental generation and the F2 (Mann-Whitney test U = 25.50; P = 0.0025), also between F1 and F2 (Mann-Whitney test U = 34; P = 0.0028) but no difference was observed between F1 and P (Mann-Whitney test U = 49; P = 0.1974). There was no difference between controls for all generations.

Analyses of development.
Results from development analysis of D. melanogaster are given in Tables 1 and 2, respectively for parental (exposed) and F1 (non-exposed) generation. Treatment of early third instar larvae at two tested doses (LD 25 and LD 50 ) decreased the number of larvae, pupae and the final number of organisms of parental generation with a dose-dependent relationship as expressed by the FNO which is always negative for the treated series. The development of F 1 D. melanogaster doesn't seem to be affecting by the early treatment of the parental generation. However, the FNO of tested flies (LD 25 , LD 50 and control) in treated medium was significantly lower than in the control medium for both generations. There is no difference between the number of organisms reached the pupae stage and the final number of organism in both generations. In addition, treatment of early third instar larvae increased significantly (p < 0.001) the duration of larval and pupal development as expressed by T 50 , with dose-dependent manner only for the Parental generation (exposed) as compared to controls. There is no difference between the T 50 of the tested flies in both treated and untreated medium.
Larvae, pupae and imagoes of the parental generation showed several types of malformations and anomalies followed by death at each stage of development of D. melanogaster. The most prominent malformations detected are incomplete and malformed imagoes (malformed abdomen and wings), curved and smaller body shape, burned larvae, dead adults inside pupae (Fig. 4).
For the control series no mortality was recorded for both tested generations. For the treated series, the lowest dose (LD 25 ) decline the adult's survival to 49% for males and 36% for females of the P generation versus 94% for males and 84% for females of the F1 generation. The highest dose (LD 50 ) induced more marked effects on adult's survival with 27% for males and 16% for females of the P generation and 81% for males and 64% in females for the F1 generation. Survival of 100% was noted for males and females of the F2 generation.

Discussion
Azadirachtin's impact on reproduction have been reported on different insect species 21,[41][42][43][44][45] . Our study has demonstrated that a single azadirachtin treatment (LD 25 /LD 50 ) of D. melanogaster larvae reduced eggs number affecting negatively the fecundity of surviving females, not only through direct sublethal effects in exposed individuals, but also through transgenerational effects on F1individuals that were never directly exposed to the insecticide.
Oviposition is a complex and critical activity in the life cycle of an insect with a variety of factors that influence both physiology and subsequent behavior, that lead to egg deposition by an insect which tries to ensure www.nature.com/scientificreports www.nature.com/scientificreports/ safety to their progeny. Reduced fecundity and fertility after azadirachtin treatment has been reported in many insects including Spodoptera littoralis, D. melanogaster, Galleria mellonella, Dysdercus cingulatus, Tuta absoluta and Helicoverpa armigera 17,41,[43][44][45][46] and could be correlated to the negative action of azadirachtin on yolk protein synthesis and/or its uptake into oocytes 21 .
Ecdysteroids, JH and insulin/insulin-like growth factor signalling (IIS) regulation are crucial for reproduction of D. melanogaster 47 . Vitellogenesis in females was stimulated under JH action and has led to oocytes development, JH synergic action with 20E and IIS controls the nutrient-sensitive checkpoint necessary for oocytes formation 47 . Consequently, reduced fecundity may be related to the antagonist action of azadirachtin on major hormones controlling the reproductive process (JH/ecdysteroids) 7 .
In Anopheles stephensi, azadirachtin treatment has led to abnormal ovaries structure with a complete arrest of oogenesis, vitellogenesis and vitelline envelope formation impairment, as well as follicle cells degeneration 48 . Ovaries of azadirachtin-treated females of Heteracris littoralis also showed complete shrinkage with suppressed oocyte growth 49 , in addition to mitochondria disintegration and follicular cells destruction 49 . Moreover, Azadirachtin reduced mating success in D. melanogaster flies and negatively affected cyst and oocyte number and size 45 . Its treatment also affected food intake and digestive enzyme activity in the midgut, in these species 10 . www.nature.com/scientificreports www.nature.com/scientificreports/ This may disturb oogenesis and vitellogenesis since ecdysone and JH rates are affected by nutrient availability, which acts as positive regulator on insulin pathway conferring ovaries the necessary signalling for a normal oogenesis 50,51 .
In addition, tested flies of all generations preferred the control medium for oviposition avoiding the azadirachtin ones for the two tested doses and conditions (no-choice and free choice). A low oviposition rate of non-exposed (naïve) flies in azadirachtin-treated areas could be due to the known repellent, deterrent, and locomotor stimulation effects of azadirachtin and other neem-based insecticides, which were reported by Silva et al. 52 for medflies Ceratitis capitata. Valencia-Botín et al. 53 also suggest that the repellent property of neem extracts is the major factor responsible for the reduction of eggs number in Anastrepha ludens (Loew) 53 . The ovipository behavior inhibition may have a valuable impact in pest control.
In addition, flies who have already been treated (third instar larvae of P generation) showed an increased aversion to azadirachtin in comparison to the naïf flies. This continued for two successive generations (P and F1). When oviposition sites were treated with azadirachtin or other neem-based compounds, oviposition repellency, deterrency, or inhibition occurred in several insect species that are able to detect the bioinsecticide on the treated surface 7,14,43,54,55 . The capacity of insects to retain memory from early life exposure affecting the adult response was reported 38,[56][57][58] . In D. melanogaster, females avoided oviposition on sites containing azadirachtin after larval exposure to the bio-insecticide 7 .
Here, we have reported for the first time that the negative effects of a single larval exposure to azadirachtin can also be passed on to the F1 generation (transgenerational effects). Environmental toxicants such as insecticide are able to provoke epigenetic alterations, which can be inherited in the next generations 59 . This may explain the reduced fecundity and oviposition avoidance in the non-exposed generation (F1).
Our study has also demonstrated that azadirachtin applied during the third larval instar of parental generation (LD 25 and LD 50 ) negatively affected various life traits of D. melanogaster, in a dose-dependent manner, as it significantly reduced larval, pupation, and emergence rate of the exposed generation. The biopesticide also significantly prolonged the larval and pupation period of development inducing important delays in immature stages development and affect sex ratios (with fewer females in the offspring) of the same generation. Additionally, the treatment induced morphological alterations of larvae, pupae and adults only in the exposed generation (P generation). The most prominent abnormalities were burned larvae, larva-pupa intermediate, pupa-adult intermediate, deformed wings, smaller body size and deformed abdomen. The recorded malformations finally resulted in insect dead. Similar results were noted in D. melanogaster 37 , Hyalomma anatolicum excavatum 60 and Spodoptera litura 22 . Finally, a decline in adult's survival was noted for the two successive generations with more marked effects among the P generation.
Azadirachtin is known to reduce pupation and eclosion rates of many insects like Aphis glycines 61 , Plodia interpunctella 62 , Aedes aegypti 63 and D. melanogaster 21 . A negative impact of azadirachtin on the immature stages was expected due to its insect's growth disruptor (IGD) action by suppressing haemolymph ecdysteroid and JH peaks 25,64 . Furthermore, azadirachtin is known to cause nucleus degeneration in the different endocrine glands (prothoracic gland, corpus allatum and corpus cardiacum) controlling insects moulting and ecdysis, which could act as generalised disruptor of neuroendocrine system 24 . Azadirachtin alters the growth and molting process of several insects by compromising their survival 7,20,43,65,66 . Lai et al. 67 reported that azadirachtin down regulated the expression of different genes that are linked to hormonal regulation. This could explain the developmental aberrations observed in our results. Azadirachtin is also known to affect Drosophila nutrient intake and metabolism compromising the nutritional signals, which result in a decrease in insect weights and growth rates, and thus www.nature.com/scientificreports www.nature.com/scientificreports/ resulting in smaller body size impacting survival 10,25,37,66 . The male biased sex ratio under azadirachtin treatment was reported in literature 67,68 .
In conclusion, the present study indicated that pre-imaginal exposure to sublethal doses of azadirachtin affects the fecundity, oviposition preference, and the survival of D. melanogaster of parent generation as well as the non exposed F1 generation. The treatment triggered life history traits variation in the P generation.
Results demonstrated that a single azadirachtin application significantly reduced the survival of flies over two successive generations (P: exposed and F1: unexposed) while insects showed clear recovery in the survival rates in the second generation (F2). These findings reflect a long term effects through developmental stage and generations. This effect is consistent in the two first generations could be considered as advantage for pest control by compensating the well-known fast degradation by sunlight and low persistence of azadirachtin in environment (half-life DT 50 : 1.7-25d) 23,69 and suggest a possible reduction both in application frequency and total amount of pesticide used.
Furthermore, the decreased fecundity and survival in P and F1 generations indicated an absence of resurgence induction in offspring, even after full restoration in F2, when parental flies were treated. This translated an absence of hermetic effect, which is considered as a serious problem of exposure to sublethal doses in agriculture.

Larvae
Pupae Imagoes www.nature.com/scientificreports www.nature.com/scientificreports/ In addition, the treatment extended the aversive effect induced by azadirachtin to over two successive generations. This could contribute as push-pull strategies that increase its insecticidal effects in integrated pest management programs.
Fecundity and oviposition site preference. We assessed the egg-laying performances of the females of D. melanogaster using a no-choice test. Three mated females (3 days old) that were pre-exposed to azadirachtin at the larval stage (LD 25 and LD 50 ) were tested for 24 h in a petri dish (Ø = 65 mm) filled with 3 ml medium containing azadirachtin at two concentrations 0.1 and 0.25 μg/ml according to Bezzar-Bendjazia et al. 37 in addition to acetone as control medium. These concentrations were not lethal with the short exposure time (24 h) used. At the end of the test, flies were removed, and the number of eggs laid on each medium was counted. The control medium was used to test the possible effect of azadirachtin on female fecundity. The experiment was repeated 12 www.nature.com/scientificreports www.nature.com/scientificreports/ times for each medium and each generation. Oviposition site preference was measured by means of dual choice experiments. Three fertilized females (3 days old) from controls and treated series (LD 25 and LD 50 ) were allowed to oviposit for 24 h in a free choice egg-laying device. This device consisted of a two petri dishes either filled with control medium (acetone) or treated medium (azadirachtin: 0.1, 0.25 μg/ml). After 24 h, the egg-laying preference was assessed by counting the number of eggs laid in each medium. The test was performed for two successive generations with 12 replicates for each medium and generation.
Oviposition preference index (OPI) defined as (number of eggs on azadirachtin medium -number of eggs on control medium)/total number of eggs was calculated 38 . Development assays. Ten controls or pre-exposed (LD 25 and LD 50 ) mated females (3 days old), named parental generation, were released into an oviposition box containing petri dishes filled with control (acetone) or treated medium (azadirachtin: 0.1 or 0.25 μg/ml) and left to lay eggs for 8 hours. At the end of the test, the flies were removed and a pool of 100 eggs for each experiment was transferred to a new petri dish containing the same medium. For all groups, we monitored the time course of larval development from egg to adult emergence by counting the number of third instar larvae, pupae, imagoes and their sex ratio, expressed as the number of males divided by the total number of emerged insects.
Next, ten parental flies from each condition (controls or treated) were crossed and the experiments were repeated for the non-exposed first generation (F1) as cited above with the same parameters recorded.
Furthermore, the developmental duration of each stage was recorded for the two tested generations expressed by T 50 (time in hours, when 50% of population reached larval, pupal and imaginal developmental stage in vials). All insects were observed under stereo zoom microscope to find any morphological distortions and photographs were taken with Leica Z16 APO.
A factor describing the final number of organisms in comparison to control (FNO) according to Ventrella et al. 39 was determined to compare the results: T C C 100 T = final number of organisms counted in treated medium. C = final number of organisms counted in control medium. Positive values of FNO show that number of organisms was higher in tested groups than within control, negative values mean that the number of individuals was higher in control than in exposed groups.
Survival analysis of adults. Survival analysis was performed according to Linford et al. 40 . For each generation (P: exposed (LD 25 and LD 50 ), F1: non-exposed) newly emerged adults were sexed and housed separately into a plastic vials (15 flies per vial) containing fresh food. Insects were transferred to new vial every 2 days. The flies were kept under observation for 15 days during which mortality was assessed every 24 h. Ten replicates were done for each dose and generation.
Statistical analyses. Data analysis was performed by R studio version 3.5.0 for Mac OS. The results were expressed as the means ± SE for each series of experiments. The homogeneity of variances was checked using Bartlett's test. The Shapiro-Wilk statistic test was used for testing the normality.