Myzus persicae Management through Combined Use of Beneficial Insects and Thiacloprid in Pepper Seedlings

Simple Summary Myzus persicae is a worldwide pest causing significant economic loss, especially to vegetables. However, the mainly applied insecticides were not effective, whilst also endangering the safety of pollinators. Harmonia axyridis and Aphidoletes aphidimyza are predators of aphids, but they are costly and affected by temperature and insecticides. We conducted toxicity tests and greenhouse trails to make an effective combination of neonicotinoid insecticides and predators. Both H. axyridis and A. aphidimyza effectively controlled aphids whether combined with thiacloprid or not, at above 20 °C temperature condition. Our results indicated that it is it is necessary to choose H. axyridis or A. aphidimyza to control aphids based on economic and thermal considerations. Practically, thiacloprid could be used either as an emergency option to control aphids’ abundance alone or in combination with natural enemies. Abstract Excessive insecticide application has posed a threat to pollinators and has also increased insecticide resistance of Myzus persicae Sulzer. Therefore, it is urgent to develop an economical and effective strategy, especially for greenhouse vegetables. Firstly, we selected a neonicotinoid insecticide that is specifically fatal to M. persicae but relatively safe to predators and bumblebees by laboratory toxicity tests and risk assessments. Then, we tested the effectiveness of the neonicotinoid insecticide under different temperature conditions. According to the LC50 values and the hazard quotients, thiacloprid met the requirements. Greenhouse trails indicated that thiacloprid was quite efficient, while control dropped to 80% without the application of thiacloprid. As for biological control, Harmonia axyridis effectively controlled 90% of aphids with thiacloprid or not. However, Aphidoletes aphidimyza performed better above 20 °C. Our results indicated that it is cost-effective to control M. persicae with A. aphidimyza in suitable temperature conditions and H. axyridis was more effective at low temperatures. Practically, thiacloprid could be used either as an emergency option to control aphids’ abundance alone or in combination with natural enemies.


Introduction
The green peach aphid, Myzus persicae (Sulzer) (Hemiptera: Aphididae), is a worldwide economically important pest with a host range of over 400 plant species [1]. M. persicae inflicts serious damage on plants, including vegetables grown in the greenhouse, directly

Insects and Insecticides
Initial population of M. persicae colony was established by collecting its population in 2019 with different life ages from pepper plants grown in field near Ji'nan in Shandong Province, China. They were reared on tobacco seedlings in the laboratory under a temperature of 25 ± 2 • C, photoperiod of 16L:8D (h) and 70 ± 5% relative humidity. A. aphidimyza pupae and B. terrestris were obtained from Shandong Lubao Technology Co. Ltd. (a specialized manufacturer of beneficial insects in Jinan, China). Eggs of H. axyridis were bought from Henan Jiyuan Baiyun Industrial Co. Ltd., Jiyuan, China. Both A. aphidimyza and H. axyridis were reared to second instar with pea aphids (Acyrthosiphon pisum Harris) (Hemiptera: Aphididae) while B. terrestris workers reared on the same fresh pollen diet and sugar syrup were used for acute toxicity experiment.

Acute Toxicity Determination
For aphids, the aphid-leaf-dip bioassay was conducted according to Srigiriraju et al. [20] with modifications. Briefly, fresh tobacco leaves with at least 20 aphids (4-5 instar) were dipped for 10 s in the designated concentrations (Table S1), air dried and placed on slightly moistened filter papers in plastic cups. Aphid mortality was assessed at 48 h after exposure and aphids were considered dead when they did not move after lightly touched with a fine paintbrush.
For A. aphidimyza and M. persicae, residual film in glass tubes were conducted according to Lin et al. [21]. The 60 mL circular glass tubes (3 cm diameter, Jinan Huihengtong Co. Ltd., Jinan, China) filled with 1 mL designated concentrations were kept stirred for approximately 2 h to generate residual film. The 20 s instar larvae of A. aphidimyza with M. persicae were introduced to each glass tube. Mortality of the larvae were recorded after 48 h when they remained immobile after being touched with a fine paintbrush. The method to test toxicity of pesticide formulations on H. axyridis was similar to that used for A. aphidimyza, except that only one larva was introduced in each glass tube, and there were 10 tubes for each concentration.
The bumblebee test scheme was modified based on previous research about honeybees [22]. For this, newly emerged workers were collected from the B. terrestris colony, the different concentrations of those selected neonicotinoids with Tween-80 were dropped onto the mesonotum of workers using the microapplicator (Burkard, Rickmansworth, UK). After the liquid become dry, ten workers were placed in an artificial stainless steel nest box (14 cm × 7 cm × 10 cm, Jinan Huihengtong Co. Ltd., Jinan, China), and reared on the sugar syrup for 48 h to obtain the death rate data.
Acute LC 50 bioassays were performed with 5 or 6 doses (Tables S1 and S2) or 8 neonicotinoid insecticides and 3 replication. All the A. aphidimyza and H. axyridis treatments were kept at 25 ± 2 • C, 70 ± 5% relative humidity (RH) and a photoperiod of 16L:8D (h). However, B. terrestris treatments were kept under conditions of 25 ± 2 • C, 60 ± 5% RH and continuous darkness. Obtained data were corrected for control mortality (which was never larger than 10%) using Abbotts's formula before analyses. The slope, LC 50 /LD 50 , 95% confidence interval and LR 50 were estimated, and correlation coefficients was performed.

Risk Assessment Procedures
Assessment procedure for A. aphidimyza and H. axyridis was based on the environmental risk assessment guidelines for non-target arthropods [23,24], which has been described by Lin et al. [22]. The in-field predicted exposure rate (PER in-field ) = the application rate × the multiple application factor (MAF), the hazard quotient (HQ in-field ) = PER in-field / the application rate for 50% mortality (LR 50 ). HQ (in-field) < 2 indicates low risk, high risk which need higher tier tests if not [25,26].
The assessment of B. terrestris followed the HQ mode of EPPO [27]. It was considered to be low risk if HQ ≤ 50, medium risk if 50 < HQ ≤ 2500 and high risk if HQ ≥ 2500. Semi-field tests would be triggered when risk was medium or high.

Greenhouse Efficacy Trial
Greenhouse efficacy trials were conducted on pepper seedlings using cages (90 cm × 90 cm × 90 cm, Zhangjiagang Phoenix red light science and education equipment factory, Zhangjiagang, China) in a greenhouse from 29 October to 25 November 2019. At first, on 29 October, thirty adults of M. persicae were released on the leaves of each pepper seedling (15-20 cm height) by a soft brush. Then, 30 pepper seedlings were set in one cage (90 cm × 90 cm × 90 cm), each treatment had two cages as replicates. Aphid population was recorded at five-day intervals during the study period. The following six treatments were established: (I) Thiacloprid: Pepper seedlings treated with the recommended concentration of acetamiprid (20 mg a.i.·L −1 ); (II) H. axyridis: One newly hatched larva was introduced to each seedling by a soft brush; (III) low-dose thiacloprid and H. axyridis: Pepper seedlings treated with the concentration of LC 50 to M. persicae (0.04 mg a.i.·L −1 ); 1 newly hatched larva of H. axyridis was introduced per 2 seedlings; (IV) A. aphidimyza: Five newly emerged female adults of A. aphidimyza were introduced to each cage; (V) low-dose thiacloprid and A. aphidimyza: Pepper seedlings treated with the concentration of LC 50 to M. persicae (0.04 mg a.i.·L −1 ); three newly emerged female adults of A. aphidimyza were introduced to each seedling; (VI) control group: Water spray only. Thiacloprid or water was sprayed with Amway spray bottles (Amway China, Guangzhou, China) until droplets were evenly distributed on the leaves, ensuring that the inner and outer stems and all leaves were evenly treated. Thiacloprid or water was sprayed at 4:00 p.m. on 31 October and 7 November. Adults of A. aphidimyza were introduced at 4:00 p.m. on 29 October and 5 November, while larvae of H. axyridis were introduced at 4:00 p.m. on 1 November and 5 November.
Due to the low temperature, the aphid stocks in A. aphidimyza and low-dose thiacloprid combined A. aphidimyza treatments were significantly higher; therefore, we conducted these two treatments at a higher temperature from 28 April to 30 May 2020. On 28 April, thirty adults of M. persicae were introduced. Thiacloprid or water was sprayed on 30 April and 7 May and newly hatched larva of H. axyridis or adults of A. aphidimyza were introduced on 1 May and 8 May, while the population of aphids was recorded every fifth day from 30 April. The temperature fluctuation recorded during the tests is shown in Figure S1, the mean temperature was 19.9 • C from 29 October to 25 November 2019, and was 24 • C from 28 April to 25 May 2020.
The M. persicae population reduction rate and percentage reductions compared to a control were computed as following [28]: where e = pest number at day 1, a = pest number after day 1, PT = population reduction rate of treated group, PC = population reduction rate of control group. The total number of larvae, pupae and adults of H. axyridis and A. aphidimyza in each treatment were counted. For the interaction between treatments and temperatures, the 'fitdistr' and 'AIC' functions in R were used to identify error distributions. Then, two-way ANOVA was used to analyze the population numbers of M. persicae. Tukey's honestly significant differences (HSD) was used as post hoc test (p < 0.05) when significant differences were detected.

Acute Toxicity of Neonicotinoids to Insects
The statistical results of the acute toxicity regression equation including the LC 50 together with their 95% confidence intervals with good fit of the data and linear regression models (r 2 > 0.9) are shown in Table 1. According to LC 50 , we found that A. aphidimyza was the most sensitive to the eight neonicotinoids with LC 50 ≤ 0.34 mg a.i.·L −1 . H. axyridis was also significantly more sensitive to most of the tested neonicotinoids than M. persicae except to nitenpyram and thiacloprid, whose LC 50 values were 17.067 and 1.314 mg a.i.·L −1 higher than M. persicae, respectively. B. terrestris was more sensitive to nitenpyram than thiacloprid at 24 or 48 h, the LD 50 values were <0.6 and >17 µg a.i.·bee −1 , respectively ( Table 2).

Risk Assessment of Pesticides to Beneficial Insects in Field
The data of maximum field recommended rates, number of applications and interval of applications in the field were obtained from the China pesticide information network (http://www.icama.org.cn (accessed on 29 July 2019)). As illustrated in Table 3, all of the HQs (in-field) of A. aphidimyza were much bigger than 2, indicating the risks of eight neonicotinoids to A. aphidimyza were high risk. Low risks of nitenpyram and thiacloprid to H. axyridis (HQ < 2) were consistent with acute toxicity results. Two-time points of HQ of nitenpyram to B. terrestris were more than 50; however, the HQ values of thiacloprid were much less than 50 (Table 4). For the purpose of the following experiment, thiacloprid was selected due to its absolutely low risk to H. axyridis, B. terrestris and relatively low risk to A. aphidimyza. Table 3. Risk assessment of eight neonicotinoid pesticides to the natural enemies based on acute toxicity data and field exposure levels.

Pesticides
DT 50 (Days) DT 50 is "half-life of degradation of the pesticide", MAF is "multiple application factor", PER in-field is "in-field predicted exposure rate", LR 50 is "application rate for 50% mortality", HQ is "hazard quotient".

Greenhouse Efficacy Trial
The population size of M. persicae during the trial showed increased aphid population in control groups (Figure 1). The number of M. persicae at the sixth investigation in high temperature (25 May, averaged 24 • C) was 2.3 times of that in low temperature (25 November, averaged 19.9 • C) (p < 0.001). No matter whether under low temperature or higher temperature conditions, the chemical control effect is remarkable at the beginning, the highest control effects were 97.38% and 99.39%, respectively (Table 5). However, as the use of pesticides stopped, the number of insects began to rise, leading to continuous reduction in the control effects. Since the third investigation, H. axyridis controlled the number of M. persicae to 1.13 and 2.68 in low (10 November) and high temperature (10 May) conditions, respectively, and kept the control effects over 96%. There were no significant differences between the two temperature conditions in insecticide and H. axyridis treatments (p = 0.14, p = 1.00). M. persicae population can also be controlled effectively when reduced numbers of H. axyridis were applied together with a low dose of thiacloprid; the control effects increased to over 90% in both low and high temperature conditions (Table 5). When in low temperature, the control effects in groups of A. aphidimyza and A. aphidimyza + thiacloprid were 79.41% and 50.87%, respectively. However, the M. persicae population in groups of A. aphidimyza and A. aphidimyza + thiacloprid continuously declined when the temperature was higher, with control effects of 90.00% and 97.79% on 25 May, respectively. The interactive effects of treatments and temperatures were found to be significant according to two-way ANOVA (Table 6).    Highest numbers of H. axyridis or A. aphidimyza in each treatment were recorded in high temperature conditions (Figures 2 and 3). This situation was more remarkable in A. aphidimyza application cages; there were no more than four A. aphidimyza in low temperature condition, whereas 367 larvae and adults were in high temperature condition. It is worth noting that the numbers of H. axyridis in each condition were much less than originally released (e.g., in H. axyridis treatment, there were only four out of 60 H. axyridis on 25 May).

Discussion
Although, overreliance on neonicotinoids would lead to the development of resistance in the M. persicae population [29], the neonicotinoids are effective to control M. persicae in this study (Tables 1 and 5, Figure 1). Therefore, the dosage and frequency of neonicotinoids should be strictly followed to avoid or slow down resistance of M. persicae. Furthermore, chemical insecticides cannot provide lasting control of M. persicae ( Figure 1). Therefore, it is necessary to integrate neonicotinoids with natural enemies to control M. persicae effectively and sustainably.
Natural enemies of aphids are not necessarily more susceptible to insecticides than their aphid prey [3]. Here we found that nitenpyram and thiacloprid showed more toxicity to M. persicae as compared to H. axyridis (Tables 1 and 3). However, larvae of A. aphidimyza were more sensitive to neonicotinoids (Table 1), and all of the neonicotinoids assessed were higher risk to A. aphidimyza, while the HQ (in-field) of thiacloprid was among the lowest (Table 3). Boulanger et al. [13] summarized that neonicotinoids are generally toxic to A. aphidimyza, while larvae are more sensitive than adults to spray applications of pesticides. It would be relatively safe since the low-dose thiacloprid (0.04 mg a.i.·L −1 ) used in the field trial was lower than the LC50 (0.13 mg a.i.·L −1 ) of A. aphidimyza. Furthermore, dose reduction encourages the appearance of insecticide tolerance genotypes; however, predators would prey on the aphids that have survived the insecticide [3]. In this study, M. persicae were consumed by H. axyridis at both temperatures, and by A. aphidimyza at higher temperature.

Discussion
Although, overreliance on neonicotinoids would lead to the development of resistance in the M. persicae population [29], the neonicotinoids are effective to control M. persicae in this study (Tables 1 and 5, Figure 1). Therefore, the dosage and frequency of neonicotinoids should be strictly followed to avoid or slow down resistance of M. persicae. Furthermore, chemical insecticides cannot provide lasting control of M. persicae ( Figure 1). Therefore, it is necessary to integrate neonicotinoids with natural enemies to control M. persicae effectively and sustainably.
Natural enemies of aphids are not necessarily more susceptible to insecticides than their aphid prey [3]. Here we found that nitenpyram and thiacloprid showed more toxicity to M. persicae as compared to H. axyridis (Tables 1 and 3). However, larvae of A. aphidimyza were more sensitive to neonicotinoids (Table 1), and all of the neonicotinoids assessed were higher risk to A. aphidimyza, while the HQ (in-field) of thiacloprid was among the lowest (Table 3). Boulanger et al. [13] summarized that neonicotinoids are generally toxic to A. aphidimyza, while larvae are more sensitive than adults to spray applications of pesticides. It would be relatively safe since the low-dose thiacloprid (0.04 mg a.i.·L −1 ) used in the field trial was lower than the LC 50 (0.13 mg a.i.·L −1 ) of A. aphidimyza. Furthermore, dose reduction encourages the appearance of insecticide tolerance genotypes; however, predators would prey on the aphids that have survived the insecticide [3]. In this study, M. persicae were consumed by H. axyridis at both temperatures, and by A. aphidimyza at higher temperature.
Many researchers show that wide use of some neonicotinoids will seriously endanger pollinator colonies, such as the honeybee and bumblebee [9,30]. In the past few years, some governments have issued their relevant policies to restrict or inhibit some neonicotinoids use outside of permanent greenhouse structures [31]. At present, the bumblebee pollination technique has increasingly been prevalent all over the world to increase the quality and yield of greenhouse crops. According to the acute toxicity and risk assessment, thiacloprid was low risk to B. terrestris (Tables 2 and 4), thus thiacloprid would be safe to B. terrestris that were globally applied for vegetables pollination in greenhouses.
The control effects of H. axyridis and H. axyridis + thiacloprid were efficient both at lower and higher temperatures (Table 5, Figure 1). H. axyridis is an excellent prospective biological control agent of aphids, especially the fourth-instar larvae and adults, which makes the mass production of commercially available H. axyridis costly [32][33][34]. Furthermore, due to its negative effects on the ladybird diversity in Europe and America, H. axyridis should be released cautiously [12,35]. A. aphidimyza are distributed widely throughout the world except Australia and Polynesia, and one larva preying on less than 10 aphids can complete its development [13]. Under high temperature conditions, the highest number (n = 367) of A. aphidimyza was much higher than H. axyridis (Figure 3), indicating that it can establish a stable population more easily and control M. persicae effectively and sustainably. Thus, A. aphidimyza are a more cost-effective predator and ecological secure when commercially used.
The control effects of A. aphidimyza and A. aphidimyza + thiacloprid at higher temperature were efficient; however, could not control the population of M. persicae when the temperature was lower than 20 • C. Applications of A. aphidimyza resulted in an over 90% reduction rate of M. persicae at average daily temperatures between 20 to 25 • C [17]. But predation, developmental period, survival and fecundity rates of A. aphidimyza decreased at lower temperatures (<20 • C) [36]. Additionally, the egg hatching, larval and adult longevity were negatively influenced when constantly reared at 35 • C [37]. Therefore, the temperature and timing of release are important factors that influence aphid control by A. aphidimyza.
The efficacy of aphid control by A. aphidimyza would be improved when combined with parasitoids Aphidius colemani [38]. Specialist natural parasitoids are often found in fields and greenhouses. When A. aphidimyza were applied alone or integration with lowdose thiacloprid, the natural parasitoids would be protected and contribute to aphid control. A. aphidimyza control effects would be weakened, since they are always the one who will be preyed on by other predators, especially by Orius and ladybirds [39,40]. Therefore, predators must be selected carefully when two or more pests exist in one greenhouse.
To our knowledge, this is the first report on the integrated control of M. persicae, combining the predatory insects H. axyridis and A. aphidimyza with the selected low-risk neonicotinoids. Our results indicate that IPM methods using a low density of predatory insects with low doses of neonicotinoids could control the population density of M. persicae for longer periods, which proved low risk for B. terrestris. Overall, the use of thiacloprid in combination with H. axyridis or A. aphidimyza can provide an effective control of M. persicae. H. axyridis provided a better control effect as compared to A. aphidimyza, especially in low temperature conditions. The cost of the IPM strategy based on the chemical and biological control methods was similar to the cost of chemical control. Whether due to the control effect or input cost, cooperative control of M. persicae in greenhouse vegetables using the IPM methods has practical application prospects for farmers.

Supplementary Materials:
The following are available online at https://www.mdpi.com/article/ 10.3390/insects12090791/s1, Figure S1: temperature recorded every hour during the tests. Where L = low temperature and H = high temperature, Table S1: concentration range used in the acute toxicity determination the aphids and the natural enemies, Table S2: concentration range used in the acute toxicity determination of B. terrestris.  Data Availability Statement: All data generated or analyzed during this study are included in this published article.

Conflicts of Interest:
The authors declare no conflict of interest.