Risk Assessment of Insecticides Used in Tomato to Control Whitefly on the Predator Macrolophus basicornis (Hemiptera: Miridae)

Simple Summary The whitefly Bemisia tabaci is a problem in tomato crops worldwide. The use of chemicals is one method to control this pest. Predators from the family Miridae have been used in Europe as biological control agents. We tested the insecticides most often used to control B. tabaci in tomato fields in Brazil for compatibility with the native Brazilian mirid Macrolophus basicornis. The results showed that regarding lethality, buprofezin, cyantraniliprole and spiromesifen were reduced-risk insecticides. Acetamiprid, bifenthrin, etofenprox + acetamiprid and pyriproxyfen + acetamiprid were considered broad-spectrum insecticides. The insecticides were also tested to be classified ecologically and were found to be safe, except for acetamiprid that was moderately toxic. Overall, our findings indicated that it is possible to use M. basicornis as a biological agent to control B. tabaci in tomato crops by means of pest management strategies that are compatible with agrochemicals in current use. Abstract The generalist mirid predator Macrolophus basicornis may contribute to Integrated Pest Management (IPM) of Bemisia tabaci in tomato crops. It is important to know the compatibility of the chemicals used to control this pest with this promising biological control agent. Seven insecticides were tested to investigate their toxicity to the predator. For four of the products, the LC50 for adults were determined. Buprofezin, cyantraniliprole and spiromesifen did not cause lethality and were classified as harmless. Acetamiprid, bifenthrin, etofenprox + acetamiprid and pyriproxyfen + acetamiprid caused acute toxicity and were classified as harmful. LT50 for all harmful insecticides were relatively low, ranging from 1.8 to 3.2 days. Moreover, these four insecticides have low LC50, with acetamiprid (0.26 mg a.i. L−1) as the lowest, followed by bifenthrin (0.38 mg a.i. L−1), etofenprox + acetamiprid (4.80 mg a.i. L−1) and pyriproxyfen + acetamiprid (8.71 mg a.i. L−1). However, the calculated risk quotient (RQ) values demonstrated that these insecticides were mostly ecologically safe for this predator, except for acetamiprid, classified as slightly to moderately toxic. The present study can contribute to the use of M. basicornis as a biological control agent on tomato crops and to compatible use with the insecticides tested, according to IPM strategies.


Introduction
The family Miridae contains a significant number of predator species used in augmentative biological control in tomato crops [1,2]. The genus Macrolophus has been used in Europe to control Tuta absoluta (Meyrick) (Lepidoptera: Gelechiidae) and Bemisia tabaci (Genn.) Biotype B (Hemiptera: Aleyrodidae), the main tomato pests [2]. In Brazil, Macrolophus basicornis (Stal) (Hemiptera: Miridae) has a considerable potential to be reared in Individuals of M. basicornis were obtained from the established rearing colony, more than eight generations old, at the Laboratory of Insect Biology in the Entomology and Acarology Department, "Luiz de Queiroz" College of Agriculture (ESALQ/USP), Piracicaba, Brazil. The insects were originally collected in the state of Minas Gerais, Brazil (21 • 08.596 S and 45 • 03.466 W, 808 m altitude) in tobacco (Nicotiana tabacum L.) fields. The method used was proposed by Bueno et al. [4], in which adults and nymphs were kept on tobacco plants in acrylic cages (60 × 30 × 30 cm) and fed with eggs of Ephestia kuehniella Zeller (Lepidoptera: Pyralidae) offered ad libitum. The cages were kept in a climate-controlled room at 25 ± 2 • C, 70 ± 10% RH and 12:12 h (L:D). Before the experiments, adults were kept in cages with tobacco plants for oviposition. After 48 h, the plants were moved to an insect-free cage. This made it possible to obtain predators with the same age, either third-instar nymphs or adults (<3 days old) from the plants [20].

Insecticides
The commercial insecticides are registered for the control of B. tabaci in Brazilian tomato crops and were tested on M. basicornis at the highest recommended field doses ( Table 1). Table 1. Active ingredient, trade name, chemical group, exposure route, mode of action and field application rate of the principal insecticides used to control Bemisia tabaci in tomato crops in Brazil.

Insecticide Exposure for Testing Acute Toxicity
Five-week-old tomato plants (cv. Santa Clara) grown in greenhouse conditions were sprayed with each of the insecticides listed in Table 1, using a hand-held sprayer (Light Sprayer-Breeze, 500 mL capacity; Guarany; São Paulo, Brazil) until the run-off point (~50 mL per plant). Distilled water was used as the control treatment [31].
After drying for 2 h, the leaves were collected from treated plants and transferred to the laboratory. Each leaf had its petiole inserted into a flask (20 mL) filled with water to maintain turgidity during the bioassay and provided a lid with an opening for the petiole (Supplementary Material Figure S1). Each tube was transferred to a cage (12 cm high × 5 cm diameter) (PET crystal, 500 mL; Copozan, Otávio Dalazen, Brazil), with each unit representing one repetition. In each cage, 15 adults of M. basicornis (<3 days old) or 15 third-instar nymphs were released and the cage was covered with voile fabric to prevent accumulation of toxic gases and retain the insects. M. basicornis individuals were fasted for 24 h before the beginning of the experiments to ensure that they started to feed as soon as they came into contact with the insecticide residues. As an alternative food source for M. basicornis, E. kuehniella eggs (0.4 g) were offered per cage. The design was randomized with 6 replicates per treatment.
M. basicornis were left on the treated leaves for 72 h under controlled room conditions (25 ± 2 • C, 70 ± 10% RH and 12:12 h L: D). After this period, untreated leaves were provided to assess the survival rate and median lethal time (LT 50 ). The insects' survival was assessed every 24 h. The insects were considered dead when they were unable to walk at least the distance of their own body length after being touched with a fine brush.

Determination of LC 50 of Harmful Insecticides
The median lethal concentration (LC 50 ) was estimated for those insecticides that were harmful to adults of M. basicornis in the acute toxicity test (Section 2.3). The procedures for the LC 50 bioassays were similar to the methods in Section 2.3. The design was completely randomized, with 6 replicates per treatment and 15 M. basicornis adults (<3 days old) in each cage. The bioassays were performed with different concentrations per insecticide, below the recommended field concentration ( Table 1), according to procedures described by Finney [32].  50 and the live insects were checked every 24 h. The insects were considered dead when they were unable to move at least the distance of their own body length after being touched with a fine brush.

Statistical Analysis
The data for the total number of live insects per replicate after 24, 48 and 72 h were checked for normality and homoscedasticity using the Shapiro-Wilk and Bartlett tests. If the assumptions of ANOVA were met, one-way ANOVA with Scott-Knott post-hoc (p < 0.05) was used to ascertain differences among treatment means. If the data did not satisfy the normality and variance homogeneity, Kruskal-Wallis non-parametric one-way ANOVA with Dunn with Bonferroni correction post-hoc (p < 0.05) was used through the "ExpDes", "easyanova" and "dunn.test" packages in the R software [33]. The mortality percentage values after 72 h were corrected according to the Schneider-Orelli formula [34]: The data for survival of mirids exposed to insecticides over time were analyzed using Kaplan-Meier estimators (Log-Rank method). The survival curves and the median lethal time (LT 50 ) were compared using the Holm-Sidak test, in SigmaPlot version 12.3 (Systat Software, San José, CA, USA).
The data obtained from the tests to estimate the LC 50 were submitted to a binomial model with the log-logistic regression (drfit) function for dose-response analysis in the statistical program R [33,35].
To assess the ecological risk of the harmful insecticides, the risk quotient (RQ) was calculated from the LC 50 values at 72 h after exposure, based on the formula: RQ = recommended field rate (g a.i. ha −1 )/LC 50 of beneficial insects (mg a.i. L −1 ). To understand the ecological selectivity of the harmful insecticides, the calculated RQ values estimate the possible effect that can occur in the field. According to the results, the insecticides were classified as safe (RQ < 50), slightly to moderately toxic (50 < RQ ≤ 2500) or dangerously toxic (RQ > 2500) [26].
Survival rates for nymphs and adults after 72 h exposure to insecticides showed significant differences among treatments (nymphs: χ 2 = 686.96, df = 7, p < 0.001; adults: χ 2 = 661.1, df = 7, p < 0.001). Buprofezin, cyantraniliprole and spiromesifen were similar to the control (Table 4). In comparison to the control group, the LT 50 values of acetamiprid, etofenprox + acetamiprid, pyriproxyfen + acetamiprid and bifenthrin were reduced by the acute toxicity of these active ingredients, ranging from 1.8 to 2.1 days for nymphs and 2.2 to 3.2 days for adults (Table 4). In the survival curves, nymphs ( Figure 1) were more vulnerable to the harmful insecticides than adults (Figure 2).
The median lethal concentration (LC 50 ) values are shown in Table 5 for acetamiprid, bifenthrin, etofenprox + acetamiprid and pyriproxyfen + acetamiprid, at 72 h after exposure of the adults to insecticide residues. Acetamiprid and bifenthrin had similar LC 50 values and etofenprox + acetamiprid and pyriproxyfen + acetamiprid were less toxic and with similar LC 50 values, with overlapping confidence intervals from 3.28 to 11.25 mg a.i. L −1 ( Table 5).

Discussion
Mirid predators can help to manage B. tabaci in tomato crops [1,2]. The mirid M. basicornis preys on tomato pests in Brazil and may become a biological control agent for use in IPM programs [3][4][5][6]36,37]. Insecticides from different chemical groups and active ingredients for control of B. tabaci are commercially available. These products range from reducedrisk, which rarely harm natural enemies, to broad-spectrum, which are acutely toxic to natural enemies [38][39][40], compromising the implementation of IPM programs. To mitigate incompatibility issues, information is needed on the acute toxicity of the insecticides that are most often used to control B. tabaci and their effects on natural enemies.
According to our findings, buprofezin, cyantraniliprole and spiromesifen were considered reduced-risk insecticides for M. basicornis and classified as harmless according to the IOBC criteria (class 1), with LT 50 values similar to the control treatment. Buprofezin is an insect growth regulator (IGR) that acts on the immature stage of sucking pests by inhibiting chitin synthesis and consequently the insects cannot molt normally [41,42]. Spiromesifen, which inhibits the acetyl CoA carboxylase, derived from tetronic and tetramic acids, interferes with the development, fecundity and lipid biosynthesis of the pest [43]. The diamide cyantraniliprole can act on nymphs and adults of sucking pests, inhibiting muscle contraction when the molecules bind to ryanodine receptors, resulting in starvation, paralysis and death [44,45]. In the present case, these insecticides did not cause acute toxicity to the natural enemy and the survival rate was also similar to the control. The results demonstrated that these insecticides are not harmful to M. basicornis in controlled conditions. Similar results were found when residues of parallel insecticides did not cause high levels of acute toxicity to M. basicornis adults and nymphs [18][19][20][21]. Interestingly, Wanu-men et al. [18] showed that spiromesifen was innocuous to adults of M. basicornis exposed to residues on an inert substrate, but mortality increased in extended laboratory assays (sprayed on tomato leaves). In the present study, spiromesifen on tomato leaves retained the harmless acute effect in controlled conditions. This was elucidated by differences in the concentrations used, contributing to this negative effect. Therefore, at the semi-field level, it does not cause a lethal effect on this natural enemy [18] and probably will not be lethal in field conditions.
Similarly to the results for M. basicornis, buprofezin did not have a lethal effect on adults and nymphs of Pilophorus typicus Distant (Hemiptera: Miridae) under controlled conditions [46] or on Macrolophus caliginosus Wagner (Hemiptera: Miridae) under field conditions [47]. For the predator Deraeocoris brevis (Uhler) (Hemiptera: Miridae), cyantraniliprole was lethal to nymphs and was less toxic to adults [48]. These studies demonstrate the importance of knowing the acute toxicity of insecticides for the integrated use of biological and chemical controls in an IPM program. The physiological effects of reduced-risk insecticides may depend on the sensitivity of a species and its life stages and, therefore, it is important to test each active ingredient on each species of natural enemy [49]. Knowledge of the sensitivity of a species in tropical conditions is important for agriculture, considering that the sensitivity can differ depending on climate, temperature and light incidence [50].
Acetamiprid induces excitation until death, acting on neurons as a competitive modulator of the nicotinic acetylcholine receptor [51,52]. It is a chlorinated neonicotinoid which mainly acts by ingestion, due to its activity inside the plants, allowing systemic translocation in the sap vessels [53]. When sprayed on tomato leaves, it caused 80% mortality in adults of another mirid predator, Nesidiocoris tenuis (Reuter) (Hemiptera: Miridae) after five days of exposure [54]. In addition to acting systemically due to its hydrophobicity, acetamiprid also acts by contact [55,56]. On inert substrates in controlled conditions, acetamiprid caused 100% mortality in M. caliginosus and Orius laevigatus (Fieber) (Hemiptera: Anthocoridae) [57]. Because mirid predators are omnivorous [58], systemic insecticides such as acetamiprid may affect this natural enemy by both the contact and ingestion routes of exposure.
Etofenprox and bifenthrin affect insects mainly through contact exposure [59,60]. Etofenprox and bifenthrin modulate the sodium channel in neuron axons, which keep the insect hyperexcited and also cause death [59]. Some pyrethroids tested on piercingsucking predators (Hemiptera) also caused acute toxicity [61,62], as did bifenthrin in this study. The active ingredient lambda-cyhalothrin is efficient in controlling the B. tabaci pest population in tomatoes, but significantly affected the survival of the mirid predator N. tenuis under laboratory conditions [61]. Deltamethrin demonstrated acute toxicity similar to bifenthrin, which caused 70% mortality in N. tenuis after contact with residues for more than 72 h [62]. Bifenthrin had a similar effect to our results when tested with a full concentration series bioassay in the laboratory, proving highly toxic to adults of Geocoris punctipes (Say) (Hemiptera: Geocoridae) and Orius insidiosus (Say) (Hemiptera: Anthocoridae) [63]. Pyrethroids are harmful to Podisus nigrispinus (Dallas) (Hemiptera: Pentatomidae), with high toxicity to nymphs and adults at the highest recommended rate in soybeans, even when mixed with low-risk insecticides [64].
Most acute toxicology studies do not address the ecological vulnerability of natural enemies to broad-spectrum pesticides and only assess whether they kill the insect, which is insufficient to recommend these pesticides in IPM programs. However, by determining the median lethal concentration (LC 50 ) for the most physiologically harmful insecticides to beneficial insects, it is possible to calculate the risk quotient (RQ) to determine the ecological risks of a given insecticide to a natural enemy. The insecticides evaluated here, especially acetamiprid and bifenthrin, showed quite low LC 50 values. Calculating the RQ values for each formulation and considering the concentration of the active ingredients, most of the RQ values were classified as safe (RQ < 50), except for acetamiprid, categorized as slightly to moderately toxic (50 < RQ ≤ 2500). These results are important to understand both the physiological and ecological risks together, in order to make decisions for IPM recommendations [65,66]. This is the first study with M. basicornis to assess the acute toxicity and RQ values of the insecticides that are most often used to control B. tabaci in tomato crops. Other studies conducted with important natural enemies in different crop systems contributed useful information IPM [26,27,35,[67][68][69]. Nevertheless, the researchers also made clear that certain insecticides tested, although classified as slightly to moderately harmful, should be thoroughly evaluated for inclusion in an IPM program, as they show high acute toxicity to the predator and other species.
Insecticides can act differently in each insect species and it is therefore important to study the pesticide formulations and their effect on the natural enemies that are most frequently found and released in the crops. As an example of the action of the same insecticides on different species, in the case of parasitoids of the genus Trichogramma, neonicotinoids and pyrethroids were tested to determine the LC 50 and to calculate the RQ values [27]. For Trichogramma dendrolimi Matsumura, Trichogramma ostriniae Pang et Chen and Trichogramma chilonis Ishii (Hymenoptera: Trichogrammatidae), the LC 50 values for acetamiprid were 0.32, 1.37 and 0.53 g a.i. ha −1 and the RQ values were 188.8, 44.1 and 114.0, respectively. Therefore, acetamiprid was categorized as slightly to moderately harmful to T. dendrolimi and T. chilonis (class 2) but safe for T. ostriniae (class 1) [27]. In the present study, the LC 50 were also very low for all insecticides tested, similarly to the studies with Trichogramma species, but the RQ values differed for acetamiprid, showing that these specific studies must be considered when assessing the insecticides' risks to a new species of natural enemy.
Taken together, the present results support the hypothesis that some of the insecticides tested were physiologically more harmful than others to the natural enemy. In addition, this study elucidated the ecological risks of those that proved to be physiologically harmful. Physiological and ecological risks must be considered when using IPM tools such as chemical and biological controls. If we consider only the physiological hazard, we eliminate all the other factors that can minimize the effect of these chemicals on non-target organisms in the field. These factors can potentially make pesticides more selective, based on, for example, formulation, placement, dosage and timing [64]. If the ecological risks are considered, there is a chance to match the methods to actual conditions in the tomato fields. It may be possible to use these insecticides with temporal and spatial separation [65]. IPM methods provide better results when most of the tools can be implemented in the field of the crop cycle [70].

Conclusions
The results obtained in controlled conditions for M. basicornis nymphs and adults are important to understand the action of insecticides on this natural enemy. Buprofezin, cyantraniliprole and spiromesifen were considered reduced-risk insecticides, but future studies should assess sublethal and transgenerational effects on this beneficial insect. Acetamiprid, bifenthrin, etofenprox + acetamiprid and pyriproxyfen + acetamiprid were harmful and considered broad-spectrum for M. basicornis. The physiological and ecological classifications for broad-spectrum insecticides were determined for M. basicornis adults and will support future IPM decisions. The RQ data provide insight into the ecological risk assessment for data acquired under more controlled conditions, but this needs to be confirmed with semi-field and field assays. Further studies are necessary to confirm compatibility of the methods with these active ingredients, such as in a greenhouse with regular insecticide spraying, to determine the persistence of the compound residues on tomato plants and the effects on the predator. It is also important to study crop management with these products to gather more accurate information.