Toxicity of nine insecticides on four natural enemies of Spodoptera exigua

Spodoptera exigua, which feeds on various crops worldwide, has natural enemies that are susceptible to the insecticides used against S. exigua. We investigate the toxicity and residue risk of 9 insecticides on the development of H. axyridis, C. sinica, S. manilae and T. remus. S. manilae and T. remus adults were sensitive to all 9 insecticides (LC50 less than 2.75 mg a.i. liter−1), while H. axyridis and C. sinica adults were less sensitive (LC50 between 6 × 10−5 mg a.i. liter−1 and 78.95 mg a.i. liter−1). Emamectin benzoate, spinosad, indoxacarb, alpha-cypermethrin, chlorfenapyr and chlorantraniliprole showed no toxicity on H. axyridis, C. sinica, S. manilae and T. remus pupae with the recommended field concentrations. The risk analysis indicated that chlorantraniliprole is harmless to larvae of four natural enemies and adult of H. axyridis, C. sinica and S. manilae. Emamectin benzoate and spinosad had higher safety to the development of H. axyridis, C. sinica, S. manilae and T. remus with the risk duration less than 4d. Indoxacarb, tebufenozide, chlorfenapyr, methomyl, alpha-cypermethrin and chlorpyrifos showed dangerously toxic and long risk duration on S. manilae and T. remus adults.

Information on the relative toxicity of various insecticides to four natural enemies H. axyridis, C. sinica, S. manilae and T. remus can aid in the development of an integrated pest management strategy for S. exigua.
Toxicity to three T. remus developmental life stages. None of the insecticides affected T. remus larvae survival ( Table 1). The percentage emergence rate of T. remus pupae at the recommended field rate of the insecticides was not significantly different from untreated T. remus pupae (F = 0.16, d.f. = 9, P = 0.99) ( Table 2). All insecticides were toxic to T. remus adults 24 h post-treatment (LC 50 values less than 1 × 10 −5 mg a.i. liter −1 ).
The comparison of LC 95 values of 9 insecticides to H. axyridis, C. sinica and T. remus larvae and H. axyridis, C. sinica, S. manilae and T. remus adult with their field recommended rates is shown in Fig. 1. For H. axyridis larvae, the LC 95 value of emamectin benzoate and spinosad were distinctly higher than its recommended field concentrations, indicating that these insecticide is harmless to the H. axyridis larvae. However, the LC 95 value of tebufenozide was still lower than its residues of the recommended field concentrations occurring 35d after treatment, indicating that tebufenozide had the longest risk duration, followed by indoxacarb (28d), alpha-cypermethrin (7d), chlorpyrifos (7d), methomyl (4d), chlorfenapyr (4d) and chlorantraniliprole (4d). For adult H. axyridis, emamectin benzoate, spinosad, indoxacarb and chlorantraniliprole were harmless with the LC 95 value higher than their recommended field concentrations. Similar to H. axyridis larvae, tebufenozide had the longest risk duration (35d) to H. axyridis adult, followed by chlorfenapyr (21d), chlorpyrifos (7d), alpha-cypermethrin (2d) and methomyl (1d). For C. sinica and T. remus larvae, the LC 95 of 9 insecticides were higher than their recommended field concentrations, indicating that the 9 insecticides are harmless to them. For adult S. manilae and T. remus, the LC 95 value of all 9 insecticides were significantly lower than their recommended field concentrations, indicating that these insecticides would be harmful to them. Meanwhile, tebufenozide had the longest risk duration to S. manilae and T. remus adult (35d), followed by chlorfenapyr (21d), chlorpyrifos (7d), alpha-cypermethrin (7d), methomyl (4d) spinosad (2d) and emamectin benzoate (2d), and the risk duration of chlorantraniliprole and indoxacarb on S. manilae adult (7d and 21d, respectively) shorter than T. remus adult (21d and 28d, respectively).

Discussion
Insecticides may kill natural enemies because of their common physiology 9 . Insecticide evaluations on natural enemies should include not only acute toxicity but also residual toxicity 19 . Under laboratory conditions, we tested the toxicity and residue risk of 9 insecticides to the development of H. axyridis, C. sinica, S. manilae and T. remus. Of the 9 insecticides tested, Indoxacarb, tebufenozide, chlorfenapyr, methomyl, alpha-cypermethrin and chlorpyrifos showed dangerously toxic and long risk duration on S. manilae and T. remus adults, similar situation was also observed on chlorpyrifos to H. axyridis adults and chlorfenapyr, methomyl and alpha-cypermethrin to C. sinica adults. therefore, these six insecticides are not suitable for the control of S. exigua. Chlorantraniliprole, the first commercial anthranilic diamide insecticide, is a potent and selective activator of insect ryanodine receptor (RyRs) that are critical for muscle contraction 20,21 . Activation of the ryanodine receptors in insects affects uncontrolled release of calcium from internal stores in the sarcoplasmic reticulum, causing unregulated release of internal calcium in the cell and leading to feeding cessation, lethargy, muscle paralysis, and ultimately death of the insect 21 . Among the insecticides evaluated, chlorantraniliprole was dangerously toxic and had a long residual (> 21d) activity on T. remus adults, however, the insecticide was safe to larvae and adult of H. axyridis, C. sinica and S. manilae. Brugger et al. 22 reported that chlorantraniliprole had selectivity to the beneficial parasitoid wasps Aphidius rhopalosiphi, Trichogramma dendrolimi, Trichogramma chilonis, Trichogramma pretiosum, Aphelinus mali, Dolichogenidea tasmanica and Diadegma semiclausum 22 . Its use for S. exigua IPM is feasible, and it should be selected according to the target species of the four natural enemies.
Emamectin-benzoate and spinosad were safe to the three life stages of H. axyridis, C. sinica, S. manilae and T. remus and larvae and pupae of S. manilae and T. remus. Though emamectin-benzoate and spinosad were dangerously toxic to adults of S. manilae and T. remus, these two insecticides had short risk duration. This may be caused by emamectin-benzoate and spinosad can penetrate leaf tissues by translaminar movement 23 . It also is important that, Emamectin-benzoate and spinosad are safe to mammals and harmless to other enemies [24][25][26][27] . Therefore, these two insecticides are suitable candidates for suppressing outbreaks of S. exigua.
Our study, C. sinica, S. manilae and T. remus pupae, survived the recommended field rate, probably because the pupae were shielded from insecticide contact by the cocoon. The results indicate that the insecticides can be applied during the pupae of the natural enemies, as discussed previously 28,29 . All the insecticides were non-toxic to T. remus larvae in our study, maybe T. remus larvae were shielded from insecticide contact. Toxicity of the insecticides to S. manilae larvae was likely due to direct host death, because the percentage pupation rate of S. manilae larvae by treated parasitized S. exigua larvae was not significantly different from untreated S. manilae larvae. In the present study, most of the S. exigua were killed directly by insecticides so it was not possible to distinguish whether S. manilae larvae died directly or because their hosts S. exigua larvae were killed.
Release of domesticated natural enemies to control insect has become an important tactic for the management of insect pests in many agricultural crops. Biological control has been increasingly used in crop protection over the past 30 years, with the production of biological control agents also increasing, with more than 130 species of predators and parasitoids on the market in 2000 30 . Laboratory experiment show that S. manilae and T. remus have good control effect against S. exigua 13,31,32 . The outlook for the success of parasitoids S. manilae and T. remus against S. exigua in crops is now more positive 13 . Due to their strong toxic on S. manilae and T. remus adults, 8 out of 9 of the insecticides (emamectin-benzoate, spinosad, indoxacarb, tebufenozide, methomyl, alpha-cypermethrin, chlorpyrifos and chlorfenapyr) as measured in this study should be applied with great caution if releasing adults of S. manilae and all tested insecticides in this study should be used with great caution when releasing adults of T. remus.
The extensive use of insecticides often promotes the development of insect resistance. At present, S. exigua have developed high levels of resistance to emamectin benzoate, cypermethrin and chlorpyrifos 33 . Therefore, the high-efficacy insecticides with minimal impact on natural enemies should be used as alternatives. This study showed important results that will help pest managers to choose the best insecticides to be applied, because products with the lowest impact on biological control agents are the most appropriate for use in IPM programs. However, the impact of insecticides on natural enemies is complex, which requires systematic study to determine sublethal effects on the biology, physiology, and behavior of four natural enemies populations.

Methods
Insects culture. H. axyridis adults and C. sinica adults were obtained from a Brassica oleracea L. var. capitata L field in Tai'an, China. After collection, they were stored separately in plastic insect boxes (23 cm long × 15 cm wide × 9 cm high) with 20-30 adults per box, and maintained under laboratory conditions of 27 ± 1 °C and a photoperiod of 16:8 h (L:D). Both were provided an ad libitum supply of live cotton aphids, Aphis gossypii Glover, (Homoptera, Aphididae) on cotton leaves, and water-soaked cotton ball was supplied as a water supplement. The boxes lined with filter paper disks. The boxes containing adults were checked daily for oviposition. If eggs were found, the adults were transferred to new plastic insect boxes (23 cm long × 15 cm wide × 9 cm high) provided with A. gossypii and water. The boxes containing eggs were checked daily for hatch. After the eggs had hatched and the larvae dispersed from the egg clusters, members of the F 1 generation were placed individually into separate glass scintillation vials (20 mm diameter, 70 mm high), and reared to the desired developmental life stages. Glassvial bioassay toxicity tests were performed using 3d old larvae, 3d pupae and 5d adults of H. axyridis and C. sinica.
All S. manilae and the T. remus populations were provided by the College of Natural Resources and Environment, South China Agricultural University, Guangzhou, China. S. manilae adults were stored in plastic insect boxes (23 cm long × 15 cm wide × 9 cm high) with 50-60 adults per box, and T. remus were stored in 30 mL glass test tubes with 200-300 adults per tube. Both were provided with honey solution and maintained at 27 ± 1 °C with a photoperiod of 12:12 h (L: D). 5 pairs of S. manilae adults were introduced into plastic insect boxes with 2nd instar 100-150 S. exigua larvae. 100-200 T. remus adults were introduced into a glass test tube containing S. exigua 2000-3000 eggs.
Scientific RepoRts | 6:39060 | DOI: 10.1038/srep39060 Parasitized S. exigua larvae or eggs were maintained under laboratory conditions of 27 ± 1 °C, 60-75% relative humidity and photoperiod of 12:12 h (L: D). 3d old larvae and 3d old pupae of S. manilae and T. remus were used in the insect-dip method, and 5d old S. manilae and T. remus adults were used in the glass-vial bioassay.
Insecticides. Nine  Toxicity bioassays. The glass-vial bioassay 34 was used to determine the toxicity of the insecticides to adults of S. manilae. Each insecticide was applied by pipetting 0.5 mL insecticide dissolved in acetone (analytical reagent, purity ≥ 99.7%) into each 22 mL glass scintillation vial (20 mm diameter, 70 mm height). Serial dilutions were used to obtain desired concentrations. Each vial was rolled for several minutes until an even layer of insecticide dried on the inner surface. Control treatment vials only received 0.5 mL of acetone. Vials were used the same day they were coated with the insecticides. All 9 insecticides were used the same method. Ten S. manilae adults were transferred into one vial then the vial was sealed with a layer of gauze. There were three (n = 3) replications were used for each rate of insecticide. After 1 h of exposure, the adults were transferred into insecticide-free vials and supplied with 10% honey solution. After 24 h, the number of dead adult S. manilae were counted, and the dose response (LC 50 ) was calculated for each insecticide. This procedure was used for H. axyridis adults, C. sinica adults, H. axyridis larval, C. sinica larval and T. remus adults, however only 2 H. axyridis or C. sinica adults, 1 H. axyridis or C. sinica larva, and 10 T. remus adults were used per vial.
Insectcide toxicity to S. manilae and T. remus larvae was tested by the insect-dip method. A 100 ml stock solution [diluted with 5% (v/v) acetone in a water solution mixed uniformly with 5% (v/v) Tween-80] was prepared for each insecticide. Serial dilutions were used to obtain desired concentrations. 3d old S. manilae and T. remus larvae with their host were dipped for 3 s in an insecticide solution, placed on filter paper, and then individuals were transferred to separate untreated glass scintillation vials, 1 individual was used per vial. For control test, individuals were dipped in distilled water containing 5% acetone. The parasitized S. exigua larvae and parasitized S. exigua eggs were used to test the insectcide toxicity (direct) to S. manilae and T. remus larvae. The pupation rate and emergence rate of S. manilae and the emergence rate of T. remus were computed after two week to enable them to reach adulthood. ), the recommended field rate was obtained from the e-Pesticide Manual of ICA, MOA, China (http://www.ny100.cn/). Pupae of H. axyridis, pupae of T. remus with S. exigua eggshells, pupae with cocoon of S. manilae and C. sinica were dipped for 3 s in an insecticide solution, placed on filter paper, and then individuals were transferred to separate untreated glass scintillation vials, 10 individuals were used per vial. For control test, individuals were dipped in distilled water containing 5% acetone. The emergence rate of S. manilae and T. remus were computed after one week to enable them to reach adulthood.
All bioassays had 3 replications of 6-9 different insecticide concentrations, and each replication of each concentration included 20 individuals.

Residue determination.
A 100 ml stock solution [diluted with 5% (v/v) acetone in a water solution mixed uniformly with 5% (v/v) Tween-80] was prepared for each insecticide with the recommended field rate. Three pots of cabbage at adult plant stage with leaves blade (ca. 10.0 × 7.0 cm) were grouped and sprayed with insecticide until the plants were completely saturated with the solution. Treated plants were placed outside the greenhouse. All nine insecticides were tested for residue toxicity. Cabbage leaves (ca. 10.0 × 7.0 cm) were collected 0, 1, 2, 4, 7, 14, 21, 28, 35 and 42d post insecticide treatment, rinsing 4 times with Acetone, 10 ml each time, after concentrated in a blowing instrument at 40 °C, added methanol to 1 ml. The concentration of insecticide residue was determined by high performance liquid chromatography (HPLC), using a 5 um Hypersil C18 250*4.6 mm reversed phase column (Diamonsil, America). The mobile phases were methanol:acetonitrile:water (45: 50:5, v/v/v) for emamectin-benzoate and spinosad, methanol:water (80:20, v/v) for indoxacarb, tebufenozide and chlorantraniliprole, methanol:water (90:10, v/v) for chlorpyrifos, alpha-cypermethrin and chlorfenapyr, methanol:water (80:20, v/v) for indoxacarb and methanol:water (50:50, v/v) for methomyl, respectively. The detections were performed at 245 nm for emamectin-benzoate, at 252 nm for spinosad, at 234 nm for indoxacarb, at 289 nm for chlorpyrifos, at 230 nm for alpha-cypermethrin, at 240 nm for tebufenozide, at 261 nm for chlorfenapyr, at 264 nm for chlorantraniliprole and at 234 nm for methomyl. The flow rate was 0.8 ml/min. Ten μ l of test solution was injected into the HPLC system.
The residue dynamics calculated by the residues/retention of water on cabbage leaves. The residues measured by HPLC = D × E × F/G × 1 ml. D is the tested insecticides volume; E is the peak area of the tested insecticides; F is the standard sample concentration; G is the peak area of the standard sample.
Statistical analysis. LC 50 and LC 95 values and slopes were determined by probit analysis using the SPSS program. Survival, mortality, pupation and emergence rates were subjected to arcsine transformation and subsequently analyzed by one-way ANOVA. Means were separated by using Tukey's Student range test (HSD) at P = 0.05 (SPSS13.0 (SPSS Inc, Chicago, USA)).