*Article* **Effect of Transgenic Cotton Expressing Bt Cry1Ac or Cry1Ab/Ac Toxins on Lacewing Larvae Mediated by Herbivorous Insect Pests**

**Zheng-Jun Guan 1,2,†, Qiu-Ju Zhou 3,†, Hong Shi 1,†,‡, Zhi-Xi Tang 1, Biao Liu 4,\* and Wei Wei 1,\***


**Abstract:** A simple food chain (plant, insect pests, and predatory arthropods) in an agro-ecosystem was set up here as a model system to elucidate the potential effect of transgenic *Bacillus thuringiensis* (Bt) cotton on non-target organisms. The system included transgenic/non-transgenic cotton, neonate larvae of three herbivorous insects (*Spodoptera exigua*, *Helicoverpa armigera*, and *S. litura*), and predatory lacewing larvae (*Chrysopa* spp.), which represent the first, second, and third trophic levels, respectively. The results showed that transgenic treatments and different densities of prey had significant effects on both body-weight gain of neonate herbivorous larvae and the number of prey captured by lacewing larvae, respectively. It was found that Bt toxin could persist at the third trophic level in lacewing larvae. The diet mixture bioassay showed that body-weight gain of lacewing larvae was significantly affected by various treatments, especially at lower concentrations of plant-expressed Bt toxin in the diet mixture, which caused significant decreases in body-weight gain. In contrast, synthetic Bt toxin at higher concentrations in the diet did not show this effect. Thus, we inferred that Bt toxin indirectly affected the growth of the lacewings and the lacewings may not be susceptible to Bt toxin or are able to metabolize it.

**Keywords:** beet armyworm; cotton bollworm; insect predator; metabolism; non-target effect; trophic level

#### **1. Introduction**

The ecological safety of transgenic crops has been a focus of scientific research and public debate [1–3]. Specifically, the effects of transgenic crops with *Bacillus thuringiensis* (Bt) gene on non-target organisms are of great concern in ecological risk assessment [4]. Bt cotton can significantly reduce damage by insect pests [5]; however, Bt toxin expression and insect resistance in transgenic cotton vary in different organs or tissues during different growth periods [6–8]. Among the target pests of Bt cotton, cotton bollworm (*Helicoverpa armigera*; CBW) is a main focus in scientific reports. CBW is a dominating insect pest for cotton in fields and its damage has caused heavy losses to agricultural production [9]. With the use of Bt cotton, it was reported that the population of CBW had greatly decreased during the growing season [10]; however, Bt cotton did not offer efficient control of CBW in the late growing season [6,11]. Evolved resistance to Bt cotton has been reported in cotton fields [12,13].

Previous studies indicated that Bt cotton can significantly reduce the growth of target pest populations, including body weight, developmental age, pupation, eclosion, behavior, and survival rate [14–16]. It was suggested that Bt cotton can provide better control of CBW populations from the first generation to the third generation, but the effect on the fourth

**Citation:** Guan, Z.-J.; Zhou, Q.-J.; Shi, H.; Tang, Z.-X.; Liu, B.; Wei, W. Effect of Transgenic Cotton Expressing Bt Cry1Ac or Cry1Ab/Ac Toxins on Lacewing Larvae Mediated by Herbivorous Insect Pests. *Plants* **2022**, *11*, 2755. https://doi.org/ 10.3390/plants11202755

Academic Editor: Fermín Morales

Received: 10 July 2022 Accepted: 13 October 2022 Published: 18 October 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

generation was relatively poor [17]. The lethality of Bt cotton to CBW is positively correlated with the expression level of Bt insecticidal toxin [16,18]. In addition, the population boom of secondary insect pests has become a critical problem in cotton fields since the commercialized planting of Bt cotton [18–20]. Beet armyworm (*Spodoptera exigua*; BAW) is a secondary insect pest in cotton fields and its tolerance to Bt toxin could be generated through laboratory selection [21,22]. Although BAW is not a target pest of Bt cotton, it could become an important pest and much more difficult to control, especially given the widespread application of Bt cotton in recent years [23]. This pest is especially not susceptible to Bt toxin expressed in transgenic cotton and the food source of BAW larvae varies [24]. Resistance of BAW larvae to Bt insecticides has been reported previously [21]. Another pest of the same genus, the cluster caterpillar (*S. litura*; CLC), was also found to be less susceptible to Bt toxin [25].

Whether transgenic plants affect natural insect predators has been a hot topic in research and is controversial. Hilbeck and Otto [26] suggested that Bt toxin could negatively affect the growth of predatory lacewing larvae (*Chrysopa cornea*) (Neuroptera: Chrysopidae). Dutton et al. [27] argued that transgenic plants would affect the growth of Lepidopteran insects and, thus, affect the prey quality for lacewings. On the other hand, Bt toxin, expressed in Bt cotton, can be transmitted to higher trophic levels through a non-target pest (e.g., the cotton aphid) and may alter the biology and behavior of a predatory ladybug [28]. In contrast, Tian et al. [29] and Romeis et al. [30] confirmed that Bt toxin did not pose a hazard to the green lacewing, whereas the larvae of both Lepidoptera and cabbage looper (*Trichoplusia ni*) fed on Bt cotton and maize leaves that were exposed to green lacewing larvae (*C. rufilabris*). It was also argued that Bt toxins have no detrimental or adverse effects on natural enemies [31–33]. Bt toxin concentrations decreased considerably from one trophic level to the next in the food web, e.g., Wei et al. [34], except for spider mites [35]. However, those studies were performed without considering the effects of insecticides sprays. Lu et al. [36] demonstrated that bio-control services would benefit from the widespread adoption of Bt cotton, as it was assumed that transgenic Bt cotton would boost the amount of prey consumed by lacewing larvae due to the reduced application of insecticides sprays.

Therefore, there is interest in understanding the relationship between the amounts of non-target insects preyed upon by lacewing and the effects of the Bt plant. In this study, we tried to investigate the effect of Bt cotton on insect pests and the preying ability of lacewing larvae in the presence of transgenic plants, to trace the transmission of Bt toxins through trophic levels and to assay the effect of Bt cotton on the predator mediated by herbivorous insect pests.

#### **2. Results**

#### *2.1. Number and Weight of Prey Consumed by Lacewing Larvae*

In bioassay 1, transgenic or non-transgenic cotton treatments (F1, 50 = 15.71, *p* = 0.0002) and different densities of prey (F1, 50 = 110.82, *p* < 0.0001) had a significant effect on the number of prey captured by lacewing larvae (*C. formosa*); both transgenic cotton treatment of prey and an increase in prey density could boost the number of prey captured by lacewing larvae. In experiment 1, where the density of prey was 10 larvae per cup, the number of prey consumed by lacewing larvae for transgenic Bt and non-transgenic cottontreated prey was 5.4 ± 0.53 and 4.3 ± 0.59 individuals, respectively. In experiment 2, where the density of BAW larvae was 20 individuals, the number of prey captured by lacewing larvae was 14.3 ± 0.60 and 10.0 ± 0.95 for Bt and non-Bt treatments, respectively. Although on different experimental days, the recorded number of prey consumed by the predator varied (F4, 50 = 12.79, *p* < 0.0001), no interaction existed between days and treatments.

In bioassay 2, the number of prey captured by *C. formosa* larvae was different between the two experiments (F1, 40 = 5.00, *p* = 0.031) and the two treatments (F1, 40 = 30.46, *p* < 0.0001) and among different days (F4, 40 = 12.91, *p* < 0.0001). In experiment 1, Bt cotton treatment enhanced the number of prey captured by lacewing larvae from 8.4 ± 0.81 to 11.4 ± 0.81, whereas in experiment 2, the number of prey captured by lacewing larvae for Bt cotton treatment (14.7 ± 0.74) was higher than that for non-Bt-cotton-treated prey (8.7 ± 0.87). In the second experiment, neonate CBW larvae fed actively on cotton disks that were placed at the bottom of the cup after switching of prey from CLC; thus, they were assumed to have been treated by Bt and non-Bt cotton leaves. The change in prey did not alter the number of prey captured by lacewing larvae (F1, 20 = 1.31, *p* = 0.27) in experiment 2.

The results of bioassay 3 are shown in Figure 1. The body weights of neonate BAW larvae were significantly lower for transgenic Bt cotton treatment than non-Bt cotton treatment (F1, 16 = 134.73, *p* < 0.0001, Figure 1a) after the same feeding duration of 24 h on cotton leaves in experiment 1. Remarkably, body weights were not different for BAW larvae feeding on transgenic cotton for 24 h or feeding on non-transgenic cotton for 12 h in experiment 2. The number of BAW prey consumed by lacewing larvae (*C. sinica*) was significantly higher when body weight of prey was low, but the numbers were not different than when body weight of prey was similar (Figure 1b) between Bt and non-Bt cotton treatments. This suggested that Bt cotton treatment could reduce the body weight of prey compared to control and stimulate lacewings to prey on more individuals (F1, 40 = 11.68, *p* = 0.0013).

**Figure 1.** Body weight (**a**) and number (**b**) of neonate beet armyworm (BAW) caterpillars captured by *Chrysopa sinica* larvae. Experiment 1 (Expt 1) included neonate BAW larvae treated for 24 h with transgenic cotton 'Zhong-30' leaves and non-transgenic parent 'Zhong-16' leaves for Bt and non-Bt treatments, respectively. Experiment 2 (Expt 2) included neonate BAW larvae treated for 24 h with transgenic cotton 'Zhong-30' leaves and for 12 h with non-transgenic cotton 'Zhong-16' leaves for Bt and non-Bt treatments, respectively.

#### *2.2. Persistence of Cry1Ac Toxin at Trophic Levels*

Three samples from the 20 transgenic cotton plants (Zhong-30) contained very low concentrations of Bt toxin and, thus, were not used to feed neonate BAW larvae. The average Bt toxin concentration in the remaining 17 cotton plants was 744.6 ± 62.65 ng/g fresh weight (FW), whereas none was detected in Zhong-16. The average Bt toxin content accumulated in the bodies of neonate BAW larvae was 321.7 ± 7.33 ng/g FW (n = 3), which was nearly half of Bt toxin concentration in leaves. Only 4.5 ± 1.93 ng/g FW (n=8) of Bt toxin persisted at the third trophic level in lacewing larvae when measured right after feeding on BAW larvae was stopped. Bt toxin concentration between these three trophic levels was significantly different (χ<sup>2</sup> = 20.32, df = 2, *p* < 0.001). Bt toxin content in lacewing bodies declined (Figure 2) during starvation, up to 24 h but the difference was not significant. This implied a great variation in the metabolism of Bt toxin in lacewing larvae.

**Figure 2.** Changes in Bt toxin content in lacewing larvae (*Chrysopa sinica*) bodies during starvation.

#### *2.3. Effect of Bt Toxin Residual in Larval Body on the Growth of Lacewing Larvae*

No Bt toxin was detected in the leaf samples of Simian-3 (data not shown) and relatively high concentrations of Bt toxins were measured (Table 1) in the third expanded leaves of GK-19 and the results varied among the four times of Bt toxin accumulations (F3, 47 = 5.98, *p* = 0.003). Approximately 0.77~19.48 ng/g Bt toxin was detected in the larval powder of BAW that fed on Bt cotton leaves (Table 1), whereas Bt toxin was not detectable in the larval powder of BAW larvae feeding on Simian-3 (data not shown). Thus, no Bt toxin was presented in the diet mixture for "treatment B" in the following feeding test with lacewing larvae. Bt toxins present in larval powder of BAW decreased 80~90% compared to the leaves. Owing to the possible degradation of Bt toxin stored in the freezer, its concentration was very low at the two earlier accumulations. This caused a very low content in the diet mixture fed to lacewing larvae (Table 1).

Body-weight gain of lacewing larvae was significantly affected by various treatments (F3, 322 = 72.34, *p* < 0.0001). The average weight gain of lacewing larvae in the four synthetic diet treatments ranged from 0.40 to 1.96 mg, whereas the aphid controls ranged from approximately 1.74 to 9.96 mg during the 8 days of the bioassay; the former was obviously lower than the latter (Figure 3). Various treatment durations (i.e., the first, second, third, and fourth 2-day durations) had a significant effect on weight gain of lacewing larvae (F3, 322 = 180.98, *p* < 0.0001), and they also had an interaction with the various treatments (F9, 322 = 8.54, *p* < 0.0001). Overall, both "treatment A" (feeding synthetic diet plus larval powder of BAW fed on transgenic cotton GK-19) and "treatment B" (feeding on synthetic diet plus larval powder of BAW fed on non-transgenic cotton Simian-3) were significantly

different from each other and from other treatments (C and D), whereas no difference was detected between treatment C (treated with synthetic diet plus synthetic Bt pro-toxin) and D (treated with blank synthetic diet only). This result suggested a 5 ng/g concentration of synthetic Bt pro-toxin did not affect the body-weight gain of lacewing larvae, but Bt toxin of lower concentration (Table 1) that was isolated from Bt cotton leaves and was presented in the larval powder could have a negative effect on lacewing larvae. On the fourth and eighth day, the recorded variation between treatments had the same trend with the overall significance, whereas no difference was found among various treatments on the second day, probably due to very low Bt toxin content in the feed. The difference between treatments B, C, and D only bordered on significance.

**Table 1.** Bt toxin contents in third expanded leaves from the plant apex of GK-19 at the six-leaf stage, in larval powder of beet armyworm (BAW) during the four times of Bt toxin accumulations and in the diet mixture feeding to lacewing larvae.


**Figure 3.** Body-weight gain of lacewing larvae (*Chrysopa sinica*) treated with a synthetic diet plus larval powder of beet armyworm (BAW) fed on transgenic cotton (GK-19) (Treatment A), synthetic diet plus larval powder of BAW fed on non-transgenic cotton (Simian-3) (Treatment B), synthetic diet plus synthetic Bt pro-toxin (Treatment C), blank synthetic diet (Treatment D), and aphids (Treatment E), respectively.

In contrast to the significant variation in body-weight gain among the various treatments, different treatments did not result in differences in the growth duration from the first instar to the pupae stage or in the actual weight of the lacewing pupae. Treatment A resulted in the longest growth duration of 16.0 ± 0.3 d, while the aphid control (treatment E) resulted in the shortest growth period of 12.8 ± 0.2 d. The pupal weight of the five

treatments ranged from 10.4 ± 0.5 to 12.7 ± 0.3 mg, where the weight of the aphid-treated pupae was the lowest since they took the least amount of time (less than 13 days) to reach pupation. Treatments A and C had similar and the lowest pupation rates (57% for A, 55% for C), whereas B and D had the highest rates (73% for B, 70% for D); the aphid-treated larvae (E) exhibited an intermediate rate of pupation (65%). The number of lacewing adults that emerged from pupae was recorded, but their individual dates of eclosion were not included. Although less pupae were obtained for the five treatments, the rate of eclosion was not low; in total, 15, 22, 11, 15, and 12 (male and female) adults emerged from the A, B, C, D, and E treatments, respectively (Table 2).

**Table 2.** Pupation and eclosion of lacewing in the five diet mixture treatments: (A) synthetic diet plus larval powder of beet armyworm (BAW) fed on transgenic cotton (GK-19); (B) synthetic diet plus larval powder of BAW fed on non-transgenic cotton (Simian-3); (C) synthetic diet plus synthetic Bt pro-toxin; (D) blank synthetic diet; and (E) aphid feed.


#### **3. Discussion**

While both transgenic Bt and non-transgenic cottons contain secondary metabolites that may adversely affect the growth of insect pest and the integration of foreign Bt gene might alter the metabolite profile in cotton [37,38], our results here showed that the Bt cotton could still provide better control over herbivorous insect pests (e.g., BAW) compared to non-Bt cotton. This study showed that there was a reduced body-weight gain for prey subjected to the Bt cotton treatment and that the number of prey (including CBW, BAW, and CLC) captured by lacewing larvae in transgenic treatments was significantly larger than non-transgenic treatments after 24 h of feeding. However, a feeding duration of 12 h did not result in difference in weight gain; therefore, the number of BAWs captured by lacewing larvae was not affected. The results implied that different feeding durations likely affected the development of prey and eventually changed the number of prey consumed by lacewing larvae. The prey showed growth retardation, weight loss, and small body size after feeding on Bt cotton leaves; thus, the predators had to consume a high number of prey to obtain adequate nutrients [27]. Therefore, the increased number of prey consumed by lacewing larvae might result from the indirect effects of Bt toxin on the nutritional quality of prey, rather than its direct effects.

It was found that the developmental periods of synthetic diet-treated lacewing larvae were longer than the aphid-treated one. Although the four different synthetic diets showed no difference in the developmental period of the lacewing pupae, the growth period of lacewing larvae in treatment A (synthetic diet plus larval powder of BAW fed on Bt cotton leaves) was the longest, i.e., pupation of the lacewing larvae was delayed and its growth cycle was extended. Both treatments A and C, with artificial diets containing either plant-expressed or artificially synthesized Bt toxin, had lower pupation rates than the other treatments. It was inferred that Bt toxin might have a certain influence on the transformation of larvae into adult lacewings. This specific mechanism of effect needs further study. In addition, the plant-expressed Bt toxin resulted in lower body-weight gain, which was not observed in the synthesized Bt toxin treatment. This might be due to the differences in potential insecticidal activity or protein structure between plant-expressed Bt toxin and synthesized Bt pro-toxin. Although it seems that the predatory function of predators could benefit from reduced growth and weight gain (i.e., smaller body size) of insect pest preys in the field, this advantage could become invalid if predators (e.g., lacewing) are directly or indirectly affected while consuming prey feeding on cotton plants expressing Bt toxins.

In our previous study, we found that Bt toxin residue in insect bodies and feces still had insecticidal ability and could affect other organisms exposed [15]. In this present study, Bt toxin residues at the second and third trophic levels were measured and proved. Although the Bt toxin that persisted in lacewing larvae was low, the finding illustrated that the Bt toxin can pass through the trophic levels through the food chain [34,39], which could cause unintended impacts on the natural ecosystem. The result is consistent with Zhang et al. [28], which indicated that Bt toxin expressed in transgenic cotton can be transmitted to higher trophic level into the coccinellid predator *Propylaea japonica* through a non-target pest insect. Furthermore, in our bioassays, as the starvation time prolonged, Bt toxin content in lacewing bodies declined. During its whole growing period, the lacewing could remove waste from the body only before it begins to pupate [40]. Therefore, it was assumed that Bt toxin may be metabolized inside the lacewing body. This phenomenon was also reported as Bt toxin moves through the food chain involving Bt canola (*Brassica napus*), diamondback moth (*Plutella xylostella*), and lacewing larvae (*C. carnea*) [34]. Although BAW is not the main prey of some predators, such as lacewings, the activity of residual Bt toxin that remained in and/or excreted from their bodies and persisted in the environment may influence susceptible organisms [41,42]. Our results here showed that the body-weight gain as well as pupation rate of lacewings could be indirectly affected by Bt toxin expressed in transgenic plants. Nevertheless, this study did not confirm that Bt toxin has no direct toxicity on predatory enemies of insect pests that feed on Bt plants. Direct toxicity is probably not the only factor that has a lethal effect on natural enemies; other growth indicators should also be considered when performing non-target risk assessment of Bt plants [30,43].

Considering the research results presented here, the indirect effects of Bt cotton may be the main factor affecting lacewings. Although Hassanpour et al. [44] showed that lacewing larvae could serve as potential bio-control agents against CBW, Lepidopteran insects are not high-quality prey for lacewings; lacewing larvae had more stunted growth and higher mortality after feeding on Lepidopteran insect larvae compared to other prey species [43,45,46]. In a feeding choice test, lacewing larvae tended to feed on aphids in the field, other than young larvae of Lepidopteran insects [47]. When being attacked by lacewing larvae, young larvae of Lepidopteran insects would fight back; therefore, the lacewings were more likely to become injured and experience higher mortality [26,27]; however, this phenomenon was not observed in the present study due to the limited number of neonate prey larvae used in each assay.

The results indicated that Bt plants might affect the growth and nutritional quality of herbivorous insects. Owing to low-nutrient foods, the lacewing increased prey numbers. At the end, no significant differences in the various growth indicators were observed between transgenic Bt diet treatment and non-Bt treatment [48]. If the differences in nutrition status among the treatments were eliminated, the amount of variation in prey number of lacewings could disappear. Therefore, the quality and quantity of the prey may be only indirect factors affecting growth of the lacewings. Whether Bt toxin has direct toxic effects on the lacewing (e.g., on body-weight gain and pupation rate) still needs to be further examined.

In general, successful suppression of the targeted insect pests in Bt cotton fields was demonstrated. However, the ecological niches of the suppressed targeted insect pests at the second trophic level could be replaced by secondary or non-targeted insect pests, which could be due to the reduction in broad-spectrum pesticides sprays [19], as well as the reduced presence of bio-control services by natural enemies (including arthropod predators), which resulted from direct or indirect impacts by Bt plants and accumulated Bt toxins in their prey. Impact on the arthropod predators and other natural enemies could affect the fundamental bases of IPM.

Bio-control is the cornerstone of integrated pest management (IPM) and natural pest control methods are normally preferred before using the pesticide options. IPM had called on the strategic integration of multiple control tactics [49], while limiting the dependence on one single technology. By many years of experience using genetically engineered crops, including herbicide-tolerant and/or Bt insecticidal traits, resistance inevitably evolved over time [50,51]. Studies demonstrated the benefits of genetically engineered crops expressing Bt toxins; however, managing resistance evolution of insect pests requires the development of novel resistant crops expressing new and multiple Bt toxins or other resistant traits. Therefore, it is hard to define whether the genetically engineered traits are natural host plant resistances or a series of chemical pesticides upgrading. In addition, while the traditional synthetic sprays of Bt pro-toxin insecticides need to be activated by binding to receptors in the midgut epithelium of the herbivorous larvae to become lethal, Bt toxins directly expressed in plant hosts are already in the activated form and might pose adverse impacts to arthropod natural enemies, even at a very low concentration without the presence of receptors in their digestive tract. Although this unintended effect needs to be confirmed in further studies, the potential adverse impacts on natural bio-control agents could add risk of failure to the IPM strategy, especially if there are additional action modes of Bt toxins on natural bio-control agents [26]. Host plant resistance, including traditional breeding and genetic engineering, could remain a critical component of IPM in the near future [49]; however, diverse approaches and control strategies should be employed for sustainable agricultural production.

#### **4. Materials and Methods**

#### *4.1. Plant and Insect Materials*

The plant materials included a Bt cotton (*Gossypium hirsutum*) variety 'Zhong-30' containing a *Cry1Ac* gene isolated from *Bacillus thuringiensis* and its non-transgenic parent 'Zhong-16'; a Bt cotton hybrid 'GK-19' containing a recombinant *cry1Ab/1Ac* gene [52] and its conventional and maternal parent 'Simian-3'. They were planted in a greenhouse at a temperature of 23 ± 2 ◦C and had light conditions set to a photoperiod of 16L:8D. The third expanded leaves from the apex at the six-leaf developmental stage were sampled for further bioassays. In addition, seeds of broad bean (*Vicia faba*) were sown in humid sand and their seedlings were used to feed aphids (*Aphis craccivora* Koch), the prey of lacewing.

The insect materials, including cotton bollworm (CBW), beet armyworm (BAW), and cluster caterpillar (CLC), and their synthetic diets were purchased from Jiyuan Baiyun Industrial Co. Ltd. (Jiyuan, China). The insects were reared in the insectary at the Institute of Botany at the Chinese Academy of Sciences (IBCAS) (Beijing, China) and the second generation was obtained for the study. The adult lacewings (*Chrysopa formosa*) were captured in the Botanic Garden of IBCAS and reared in the insectary to obtain lacewing larvae for bioassays in this study. Chinese green lacewing larvae (*C. sinica*) were kindly provided by Dr. Fan Zhang at the Plant Protection and Environmental Protection Research Institute, Beijing Agriculture and Forestry Academy of Sciences. All insect materials were kept and reared and the bioassays were conducted in the insectary of IBCAS under appropriate rearing conditions (temperature 25 ± 2 ◦C, humidity 50 ± 5% and light 16 h: 8 h). Aphids were reared on the seedlings of broad bean and fed to lacewings regularly during the experiment where applicable.

#### *4.2. Bioassays on Prey Consumption by Lacewing Larvae and Cry1Ac Toxin Transfer through Trophic Levels*

#### 4.2.1. Prey Consumption Bioassays

Three bioassays were designed to investigate the effect of Bt cotton on the number of caterpillars consumed by predatory lacewing larvae (Figure 4 and Appendix A Table A1). Each bioassay contained two experiments. Lacewing larvae of *C. formosa* were used as predator in both bioassay 1 and 2, while *C. sinica* was used in bioassay 3. Both bioassays used transgenic Bt cotton Zhong-30 and non-transgenic cotton Zhong-16 to feed to herbivorous insect larvae (Figure 4).

**Figure 4.** Experimental designs. Bt cotton 'Zhong-30' expressing Cry1Ac toxin and its non-transgenic counterpart 'Zhong-16' were used to study preys consumption bioassays and to measure toxin transfer through trophic levels, while 'GK-19' expressing Cry1Ab/1Ac fusion toxin and its non-transgenic counterpart 'Simian-3' were used to study the impact of toxin on lacewing.

In bioassay 1, neonate BAW larvae were fed by leaves of transgenic Bt cotton Zhong-30 and non-transgenic Bt cotton Zhong-16 for 24 h and then they were fed to the third instar larvae of *C. formosa* in plastic cups with permeable covers on the top and 1% agarose at the bottom. The first experiment contained eight and nine replicates for the Bt cotton-treated and non-Bt cotton-treated prey, respectively. Each replicate had 10 neonate BAW larvae and one lacewing larva. The second experiment contained seven and four replicates for the transgenic-treated and non-transgenic-treated prey, respectively. In each replicate of the second experiment, one lacewing larva was placed with 20 neonate BAW larvae, on either a transgenic or non-transgenic cotton leaf disk.

In bioassay 2, neonate CLC and CBW larvae were allowed to feed for 24 h on the third expanded leaves from the apex of Zhong-30 and Zhong-16, respectively, before they were fed to the third instar larvae of *C. formosa*. In experiment 1, 20 neonate CBW larvae were fed to one lacewing larva in each of the five replicates of both treatments. Experiment 2 contained five and six replicates for non-transgenic and transgenic treatments, respectively. During the first 2 days, 20 individuals of treated CLC larvae were fed to a third instar lacewing larva in each replicate. During the remaining 3 days, the predator was switched to untreated neonate CBW larvae due to a shortage of CLC larvae.

In bioassay 3, neonate BAW larvae were allowed to feed for 12–24 h on the third expanded leaves from the apex of Zhong-30 and Zhong-16, respectively. In experiment 1 of this assay, the larvae were fed for 24 h on both Bt and non-Bt cotton and for 12 h on Zhong-16 in experiment 2. These treated larvae were weighed in a bulk of 50 individuals nine times for each treatment of each experiment and then fed to a third instar larva of *C. sinica* at a density of 20 prey items to one predator. Each treatment of Zhong-16 and Zhong-30 had five replicates in both experiments.

The number of prey captured by lacewings was recorded in each of 5 days of observation for the three bioassays and the density was maintained at the initial levels by adding an appropriate number of caterpillars.

4.2.2. Detecting Cry1Ac Toxin Transfer through Trophic Levels

It was suggested that Bt toxin content in the leaves should be assayed in a rearing bioassay [15]. This bioassay intended to study the movement of Bt toxin through a simple food chain (cotton-BAW-lacewing) (Figure 4). Leaf disks of the third expanded leaves were sampled from each of the 20 transgenic Bt cotton plants (Zhong-30) and their Bt concentrations were measured using ELISA. Leaves of non-transgenic cotton plants (Zhong16) were used as a control. These cotton leaves were used to feed neonate BAW larvae for 24 h; three samples of 50 larvae feeding on 'Zhong-30' were randomly sampled to detect Bt toxin accumulation. A sample of 50 larvae that fed on non-Bt cotton was used as a control. These larvae that fed on Bt cotton leaves were transferred to 32 plastic cups with covers; each cup contained 20 larvae and a third instar lacewing larva. One cup contained BAW larvae feeding on non-Bt cotton leaves and a lacewing larva were used as a negative control. The density of BAW larvae in each cup was maintained by adding an appropriate number of larvae. After 3 days of feeding, lacewing larvae were collected and divided into four groups. The first group of eight lacewing larvae was killed immediately by placing into liquid nitrogen individually; the remaining predators were put in cleaned Petri dishes without any food. The second group of ten lacewing larvae was collected after 4 h of starvation. Seven lacewing larvae in each of the third group and the fourth group were collected individually after 12 h and 24 h of starvation, respectively. All lacewing larvae were stored at −80 ◦C, separately, for further analysis using ELISA Kits (Agdia Inc., Elkhart, IN, USA) to measure the presence of Bt toxin [34].

#### *4.3. Bioassays on the Effect of Cry1Ab/1Ac Toxins Residual on Lacewing Larvae*

Bt cotton 'GK-19' expressing Cry1Ab/1Ac fusion toxin and its non-transgenic counterpart 'Simian-3' were used in this experiment (Figure 4). The third expanded leaves of GK-19 at the six-leaf stage were fed to the second instar BAW larvae for 24 h to accumulate Bt toxin at the second trophic levels. Simian-3-fed BAW larvae were used as controls. Bt toxin concentration was measured in the leaves of both GK-19 and Simian-3 using ELISA test that were fed to BAW larvae. Bt toxin in larval body powders was accumulated for bioassay as described in Shi et al. [15]. Ten of the second instar larvae of BAW were added to a Petri dish for each of the twelve and four replicates for GK-19 and Simian-3, respectively. The larvae were collected after 24 h of feeding and stored separately between GK-19 treatment and Simian-3 treatment at −80 ◦C for >24 h and were ground into dry powder in liquid nitrogen. Part of the larval powder was used to detect Bt toxin concentration and the others was mixed with the synthetic diet [53] of lacewings and fed to the first instar lacewing larvae. The accumulation was repeated four times to generate enough larval powder for further bioassay on the lacewing larvae. Previous research suggests that Bt toxin in synthetic diets persists for 2 days [54]; thus, the mixed diet for the lacewing larvae was re-made and renewed every 2 days and the dry powder that contained Bt toxin obtained at four different accumulations was used in subsequent order to mix with the synthetic diet, respectively.

Five treatments were set up in this feeding bioassay: (A) 30 replicates of synthetic diet feeding tests with larval powder of BAW that fed on transgenic cotton (GK-19); (B) 30 replicates of synthetic diet feeding tests with larval powder of BAW that fed on non-transgenic cotton (Simian-3); (C) 20 replicates of synthetic diet feeding tests with synthetic Bt pro-toxin (1 mg/mL), and the final concentration of Bt pro-toxin in this treatment was 5 ng/g, which was equal to the reported Bt toxin concentration in larval bodies feeding on GK-19 [15]; (D) 20 replicates of blank synthetic diet feeding tests of lacewings; (E) 20 replicates of aphid feeding tests of lacewing larvae. Each replicate contained one lacewing larva and a certain amount of synthetic diet mixture or aphids. The last treatment was used as a control only and not considered for statistical analysis.

A piece of squared sponge (1 cm × 1 cm × 1 cm) was placed inside each plastic container (diameter × height = 11 cm × 8.5 cm) with a permeable cover to feed one first instar lacewing larva for each replicate after absorbing enough diet mixture in liquid in each treatment. Body-weight gain of lacewing larvae was recorded every 2 days. The synthetic diet was replaced by aphids after 8 days of bioassay until lacewing larvae pupated. Survival rate of lacewing larvae was also recorded during the bioassay, as well as the rate of pupation, weight of the pupae, and the developmental days of each larval stage. The eclosion rate of lacewing pupae, the sex ratio and oviposition rate of lacewing adults, and

hatching rate of lacewing eggs were monitored and recorded to study the possible effects of the diet treatments.

#### *4.4. Data Analysis*

Statistical analyses were conducted using SPSS 16.0 software. One-way analysis of variance (ANOVA) was used to analyze the variation in Bt toxin content in leaves of transgenic cotton, lacewing bodies and diet mixture, and the number and weight of prey consumed by lacewing larvae. Chi-square test was conducted to test the difference in Bt toxin content at various trophic levels. The GLM (general linear model) procedure was used to analyze the effects of Bt toxin residual on lacewing. Differences in means between treatments were compared using a least significant difference (LSD) test.

**Author Contributions:** Conceptualization, W.W.; formal analysis, Q.-J.Z., H.S. and Z.-J.G.; investigation, Q.-J.Z., H.S. and Z.-X.T.; writing, Z.-J.G., Q.-J.Z. and W.W.; supervision, W.W. and B.L.; funding acquisition, B.L. and Z.-J.G. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by the China National Special Transgenic Project "Monitoring and controlling technology for ecological risks in nature, grant number: 2016ZX08012005" and the Natural Science Foundation of Shanxi Province, China, grant number: 201701D121089".

**Data Availability Statement:** The data presented in this study are available on request from the corresponding author.

**Acknowledgments:** We thank the three anonymous reviewers whose comments/suggestions helped improve and clarify this manuscript.

**Conflicts of Interest:** The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

#### **Appendix A**

**Table A1.** Bioassays to investigate the effect of Bt cotton on the number of caterpillars consumed by predatory lacewing larvae.


BAW (beet armyworm): *Spodoptera exigua*; CBW (cotton bollworm): *Helicoverpa armigera*; CLC (cluster caterpillar): *Spodoptera litura*. 'Zhong-30' and 'Zhong-16' is the transgenic line and the non-transgenic parent cotton, respectively. 'Tr.' and 'Non-Tr' is the abbreviation for 'transgenic' and 'non-transgenic', respectively.

#### **References**


**Zhentao Ren 1,2, Muzhi Yang 1,2, Haopeng He 2, Yanjie Ma 2, Yijun Zhou 2, Biao Liu 3,\* and Kun Xue 1,2,3,\***


**Abstract:** In order to provide more evidence for the evaluation of the ecological risks of transgenic maize, arthropod population dynamics and biodiversity in fields planted with two kinds of transgenic maize (DBN9868, expressing the *PAT* and *EPSPS* genes, and DBN9936, expressing the *Cry1Ab* and *EPSPS* gene) were investigated by direct observation and trapping for three years. The recorded arthropod species belonged to 19 orders and 87 families, including Aphidoidea, Chrysomelidae, Coccinellidae, Chrysopidae and Araneae. The species richness, Shannon–Wiener diversity index, Pielou evenness index, dominance index and community similarity index of arthropod communities in maize fields were statistically analyzed, and the results showed that (1) the biodiversity difference of arthropod communities between transgenic maize and non-transgenic maize was smaller than that between different conventional cultivars; (2) the differences between ground-dwelling arthropod communities were less obvious than those between plant-inhabiting arthropod communities; and (3) Lepidoptera, the target pests of Bt maize, were not the dominant population in maize fields, and the dominant arthropod population in maize fields varied greatly between years and months. Combining those results, we concluded that the transgenic maize DBN9868 and DBN9936 had no significant effect on the arthropod communities in the field.

**Keywords:** genetically modified organisms; maize; Bt; corn; biodiversity; arthropods; biosafety

#### **1. Introduction**

Genetically modified (GM) crops have been used commercially for more than 20 years. GM maize is one of the most adopted GM crops in the world, with 60.9 million hectares planted in 2019, which is 31% of the global planted area (https://www.isaaa.org/ (accessed on 1 May 2022)). In 2020, three biosafety certificates for GM maize were granted by the Ministry of Agriculture, PR China. Although GM crops have been commercially planted at a large scale in recent decades, their biosafety is still controversial in certain countries [1,2]. One concern about GM crops is the potential negative impact on the diversity and abundance of non-target organisms and, subsequently, on ecosystem functions [3–5]. In maize field ecosystems, non-target arthropods provide important ecological functions such as biological control, recycling of organic matter, and pollination [6,7]. Prior to biosafety certification being granted, local field trials are required to assess the potential adverse effects of GM maize on agroecosystems, including arthropod communities, relative to a non-GM control.

A growing number of studies have revealed the non-target impact of GM maize on arthropod biodiversity in the field. To date, most studies have shown that insect-resistant and herbicide-tolerant traits, either single events or stacked products, do not adversely affect biodiversity, the populations of natural enemies, or other ecologically important non-target arthropods [6,8–16]. In recent years, some long-term studies have also supported this point of view. For example, a farm-scale evaluation of the impact of Bt maize (*Cry1Ab*)

**Citation:** Ren, Z.; Yang, M.; He, H.; Ma, Y.; Zhou, Y.; Liu, B.; Xue, K. Transgenic Maize Has Insignificant Effects on the Diversity of Arthropods: A 3-Year Study. *Plants* **2022**, *11*, 2254. https://doi.org/ 10.3390/plants11172254

Academic Editors: Wei Wei and C. Neal Stewart Jr.

Received: 27 June 2022 Accepted: 26 August 2022 Published: 30 August 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

on non-target arthropods through a three-year study was reported, indicating that Bt maize had no negative effect on hemipteran herbivores (leafhoppers and planthoppers) and hymenopteran parasitic wasps (mymarids) [17]. A three-year study showed that Bt maize (*Cry1Ac*) had no significant negative effects on selected non-target arthropod diversity in the field [18]. Another three-year study also showed that Bt maize (*Cry1Ab* and *Cry2Ab*) was compatible with non-target arthropods [19]. Most studies compared transgenic maize with its parental line, but non-transgenic controls of different varieties were not included in the comparison.

Evaluation of GM crops follows the principles of precaution, familiarity and substantial equivalence, which are concepts arising from the knowledge and experience gained over time [20]. They are based on a series of comparative assessment approaches that consider the GM crop in the context of appropriate non-GM control and the known variation of the crop [21]. The mean values of the key characteristics of the GM crops are assessed against the range of values of the reference varieties or from the literature in the context of known values common for the crop [22]. Those comparative assessments can reveal the potential risks of the cultivation of GM crops [23,24]. This study uses comparative assessment approaches to evaluate the effects of GM crops on arthropod communities in the field. If the mean value of the arthropod diversity of a GM crop is beyond the value range of the cultivated crops, the detected differences are then used to assess whether they could potentially alter arthropod communities.

In this study, the arthropod biodiversity in plots with four kinds of maize was investigated in Northeast China from 2015 to 2017. The maize cultivars were commercial maize XianYu335, non-transgenic maize DBN318, transgenic maize DBN9868 providing tolerance to herbicides, and transgenic maize DBN9936 providing tolerance to herbicides and resistance to lepidopteran pests. In order to evaluate the impact of transgenic maize on arthropod communities, species richness (*S*), Shannon–Wiener index (*H*), dominance index (*C*), Pielou evenness index (*J*) and community similarity index (*C* ) of plant-inhabiting and ground-dwelling arthropods were analyzed. We had the following objectives: (1) to compare the effects of different types of transgenic maize on arthropod communities in the field; (2) to compare the effects of maize on arthropod communities on the plant and on the land; (3) to determine the dynamics of arthropod communities in maize fields in different months and years. The results will provide a statistical baseline and theoretical basis for the ecological risk assessment of transgenic crops.

#### **2. Results**

#### *2.1. Composition of Arthropod Communities*

The arthropod species in maize plots were investigated throughout the growing season in 2015, 2016 and 2017. A total of 89 families from 19 orders were identified and recorded, including insects, spiders and other arthropods (Tables S1 and S2).

A simplified food web, including the main plant-inhabiting and ground-dwelling arthropod species in maize fields, is shown in Figure 1. During all stages of the maize, *Rhopalosiphum maidis* and *Monolepta hieroglyphica* were the main plant-inhabiting herbivores. Lepidoptera herbivores rarely occurred, mainly including the Asian corn Borer (*Ostrinia furnacalis*), the cotton bollworm (*Helicoverpa armigera*) and the oriental armyworm (*Mythimna separata*). Natural enemies included predators and parasitoids. Ladybeetles (*Harmonia axyridis* and *Propylaea japonica*), minute pirate bugs (*Orius sauteri*), lacewings (*Chrysopa pallens*) and spiders (*Misumenops tricupiuatus* and *Synema globosum*) were the most frequent predators on maize. Meanwhile, parasitoids were difficult to find. The detritivores included fruit flies and psocids. Ants are omnivorous insects that protect aphid populations and repel ladybeetles on maize.

**Figure 1.** Important arthropod species in the food webs in maize field in Yitong, northeast China. (**A**) The herbivores included *Rhopalosiphum maidis* (**a1**), *Haplothrips aculeatus* (**a2**), *Monolepta hieroglyphica* (**a5**), *Trigonotylus ruficornis* (**a6**), *Ostrinia furnacalis* (**a7**), *Helicoverpa armigera* (**a8**), *Mythimna separata* (**a9**), eating leaves and buds, and wireworms (Coleoptera: Elateridae) (**a3**), *Anomala corpulenta* (**a4**), *Teleogryllus emma* (**a10**)*,* eating roots and stems; (**B**) the predators and parasites were *Harmonia axyridis* (**b1**), *Propylaea japonica* (**b2**), *Orius sauteri* (**b3**), *Chrysopa pallen* (**b4**)*, Pirata subpiraticus* (**b5**)*,* funnel-web spiders (Araneida: Agelenidae) (**b6**), *Misumenops tricupiuatus* (**b7**), *Tettigonia chinensis* (**b8**), *Trichogramma ostriniae* (**b9**), potter wasps (Hymenoptera: Vespidae) (**b10**) and *Chlaenius bioculatus* (**b11**); (**C**) another related arthropod was *Camponotus japionicus*. The solid arrows indicate predation relationships; the dashed arrows indicate regulatory relationships.

The number of species of ground-dwelling arthropods collected in pitfall traps was less than that of plant-inhabiting arthropods (Figure 1). The cricket (*Teleogryllus emma*) was the most abundant ground-dwelling herbivorous species. The ground-dwelling predators included the wolf spider (*Pirata subpiraticus*) and ground beetles (*Chlaenius bioculatus* and *Dolichus halensis*). The main scavenging insects on the ground were springtails (Collembola).

The cumulative number (summed data from all sampling dates) of arthropods over the growing season was analyzed (Figure 2, Table S3). The most abundant herbivore groups were the aphids (Aphidoidea) and leaf beetles (Chrysomelidae), and important predator groups were ladybeetles (Coccinellidae), lacewings (Chrysopidae) and spiders (Araneae). The abundance of these groups varied highly between different years. For example, in 2015, the number of leaf beetles was the highest, and that of aphids was the second. However, this was reversed in 2016 and 2017, and the leaf beetle population almost disappeared in 2017.

**Figure 2.** From 2015 to 2017, the cumulative number of main arthropod groups in the field of cultivated variety XianYu335 (XY), non-transgenic recipient variety DBN318 (NT), transgenic maize DBN9868 (T1) and DBN9936 (T2) over the growth season. The Y-axis shows the cumulative number of individuals per 100 plants during the growth period, and the X-axis shows the arthropod groups. We only compared XY with NT and NT with T1 or T2 by using a Student's *t*-test. Asterisks denote significant differences, where *p* < 0.05 (\*), *p* < 0.01 (\*\*).

A *t*-test was conducted to compare the total number of main arthropod groups in the plots of XY and NT, NT and T1, NT and T2 (Figure 2). In 2015, Chrysomelidae abundance on NT was significantly higher than that on XY (*p* = 0.047). In 2016, the Aphidoidea abundance on NT was significantly higher than that on XY (*p* = 0.0001). In 2017, Lepidopteran abundance on NT was significantly higher than that on T2 (*p* = 0.005). Other groups showed no significant difference.

#### *2.2. The biodiversity Indices of Arthropods*

Four indices were used for arthropod biodiversity analysis. The species richness (*S*) directly represents the biodiversity in an agroecosystem. The Shannon–Wiener biodiversity index (*H*) reflects species richness and community structure information, which is sensitive to changes in the rare species of a sample [25]. The Pielou evenness index (*J*) describes the relative abundance or ratio of individuals in certain species to the total number of individuals, which is often used in combination with *H* [26]. The dominance index (*C*), based on the Simpson's index, is sensitive to changes in species abundance in the sample and less sensitive to species richness [27,28].

The *S* presented similar dynamics each year, regardless of the type of maize (Figure 3). The dynamics of *S* of the arthropod community in the plots between XY and NT, NT and T1, NT and T2 were compared using a *t*-test. The *S* of plant-inhabiting arthropods on NT was significantly different from that on XY for each pair of comparisons (26 August 2015, *p* = 0.047; 23 June 2016, *p* = 0.016). On 1 September 2016, the *S* of plant-inhabiting arthropods on NT was significantly higher than that on T1 (*p* = 0.013). In 2017, on sample date 7 September (*p* = 0.002), the *S* of plant-inhabiting arthropods on NT was significantly higher than that on T2. In 2015, on the sample dates of 21 July (*p* = 0.016) and 16 September (*p* = 0.047), the *S* of plant-inhabiting arthropods on XY was significantly higher than that on

NT. There were no significant differences between the NT and XY, NT and T1, NT and T2 plots in terms of the *S* of ground-dwelling arthropods at each sampling time.

**Figure 3.** From 2015 to 2017, the species richness (*S*) dynamics of the arthropod community in the plots of cultivated variety XianYu335 (XY), non-transgenic recipient variety DBN318 (NT), transgenic maize DBN9868 (T1) and DBN9936 (T2). (**A**) The *S* of plant-inhabiting arthropods in XY, NT and T1; (**B**) the *S* of ground-dwelling arthropods in XY, NT and T1; (**C**) the *S* of plant-inhabiting arthropods in XY, NT and T2; (**D**) the *S* of ground-dwelling arthropods in XY, NT and T2. Y-axis: Mean ± SE (*n* = 3) of *S* at each sampling point. We only compared XY with NT and NT with T1 or T2 by using a Student's *t*-test. Asterisks denote significant differences, where *p* < 0.05 (\*), *p* < 0.01 (\*\*).

In this study, *H* also had similar dynamics each year, regardless of the maize varieties. There were eight significant differences between XY and NT on different sampling dates and no significant differences between NT and transgenic varieties (Figure 4).

For the Pielou evenness index, there were eight significant differences and two extremely significant differences between XY and NT in Student's *t*-test, while there were two significant differences between NT and T1 and NT and T2, respectively (Figure S1). The *C* presented similar dynamics each year, regardless of the maize varieties. For the dominance index, there were seven significant differences that only existed in comparisons between XY and NT with Student's *t*-test (Figure S2).

**Figure 4.** From 2015 to 2017, dynamics of Shannon–Wiener biodiversity index (*H*) of the arthropod community in the plots of cultivated variety XianYu335 (XY), non-transgenic recipient variety DBN318 (NT), transgenic maize DBN9868 (T1) and DBN9936 (T2). (**A**) The *H* of plant-inhabiting arthropods in XY, NT and T1; (**B**) the *H* of ground-dwelling arthropods in XY, NT and T1; (**C**) the *H* of plantinhabiting arthropods in XY, NT and T2; (**D**) the *H* of ground-dwelling arthropods in XY, NT and T2. Y-axis: Mean ± SE (*n* = 3) of *H* at each sampling point. We only compared XY with NT and NT with T1 or T2 by using a Student's *t*-test. Asterisks denote significant differences, where *p* < 0.05 (\*), *p* < 0.01 (\*\*).

#### *2.3. The Community Similarity Index of Arthropods*

The community similarity index (*C* ) appears to be the best approximation of a linear measure of community relationship [29]. Thus, if the same species were in two communities, the *C* index would be 1.0; if there were no species in common between the two communities, the *C* index would be 0.0. The range from no resemblance to complete identity is appropriately covered by the range from 0.0 to 1.0. The closer the *C* index is to 1.00, the more similar the two communities are. The *C* of arthropod communities between XY and NT, NT and T1, and NT and T2 from 2015 to 2017 was analyzed (Figure 5). On the plants, 67% of the *C* values of "XY vs. NT" were lower than the *C* values of "NT vs. T (including T1 and T2)", and the difference between them was significant according to a chi-square test (*n* = 52, *χ*<sup>2</sup> = 6.231, *p* = 0.013). On the ground, 65% of the *C* values of "XY vs. NT" were lower than the *C* values of "NT vs. T", and the difference between them was not significant according to a chi-square test (*n* = 26, *χ*<sup>2</sup> = 2.462, *p* = 0.117). In total, the community similarity of arthropod between transgenic maize and non-transgenic maize (NT vs. T1; NT vs. T2) was smaller than that between different conventional cultivars (XY vs. NT).

**Figure 5.** Heat maps of the community similarity index (*C* ) of arthropods in the plots of cultivated variety XianYu335 (XY), non-transgenic recipient variety DBN318 (NT), transgenic maize DBN9868 (T1) and DBN9936 (T2) from 2015 to 2017. Blue indicates a low value (0.0) of the community similarity index and red indicates a high value (1.0).

#### **3. Discussion**

There was no evidence to suggest that transgenic crops are toxic to non-target arthropods in the field [10,30–33]. Our study showed that the biodiversity difference of arthropod communities between transgenic maize and non-transgenic maize was smaller than that between different conventional cultivars.

#### *3.1. The Effects of Different Maize Traits on Arthropods*

We hypothesized that the differences between arthropod communities between the different maize lines were not caused by transgenes but by differences in agronomic and phenotypic characteristics. For example, there were obvious differences in the characteristics of the husks of the ear between XY and NT. The husks of XY are relatively tight, while the husks of NT are loose (the looseness means aphids can get into the husks to feed, while their natural enemies, ladybeetles, cannot), which means aphids are outside of the control of ladybeetles and outbreaks occur more frequently. Some studies reported that slightly different agronomic and phenotypic characteristics were detected in the transgenic maize hybrids compared to conventional maize hybrids, but none of these phenotypic differences were expected to contribute to biological or ecological changes that would result in increased pest potential or ecological risks [34]. Therefore, the few significant differences in arthropod biodiversity between GM and non-GM maize plots do not indicate that GM crops have an adverse effect on the ecosystem.

#### *3.2. The Effects of GM Maize on Arthropods*

If the Bt protein is toxic to non-targets, we would expect those effects in Bt maize. For herbicide-tolerant maize, no effects on non-targets are hypothesized. In this study, T1 (*PAT* and *EPSPS* genes) and T2 (*Cry1Ab* and *EPSPS* genes) did not lead to significant differences in non-target arthropod biodiversity. As a result, Bt effects are unlikely to be present. Most of the plant-inhabiting arthropods (lepidopterans, leaf beetles, ladybeetles, aphids, stink bugs, biting insects, mites and thrips) are able to ingest the insecticide Bt Cry toxins expressed in GM crops [35–39]. Some studies have shown the presence of Cry toxins in some ground-dwelling arthropods (ground beetles, spider mites and Diptera larvae) through direct consumption of the maize or its residues [40–42]. However, the available literature does not provide evidence to support the claim that unexpected interactions between different Bt proteins may lead to adverse effects on non-target species [43]. Our results also demonstrated that arthropod communities are not affected by maize expressing *Cry1Ab*, *EPSPS* and *PAT*. This conclusion is consistent with the results of other studies on the effect of Bt maize on arthropods [6,44–50]. In addition, this study showed that the

effect of different maize types on ground-dwelling arthropods was less than that on plantinhabiting arthropods. This may be because many ground-dwelling arthropods, such as springtails, crickets and ground beetles, depend more on the soil or a weedy environment than on maize plants.

Our results are consistent with other field studies that have reported that the abundance of arthropod species varies greatly between years and months [17,18], which might be a consequence of biotic and abiotic elements in complex agricultural ecosystems, including food uniformity, temperature variation, rainfall distribution and the presence of natural enemies [17,51]. In our study, there was no long-term stable food web in maizebased farmland ecosystems during the growing period. Compared with the significant effects of a complex climate, the effects of GM maize on arthropod communities in the field appear negligible.

#### *3.3. The Effects of GM Maize on the Food Web*

Agroecosystems are simple but vulnerable to environmental factors. Multitrophic interactions in communities and food webs are temporary [52]. Food web analyses can provide comprehensive information to understand the effect of GM crops on the structure of ecological communities [12]. Bt maize, expressing *Cry1Ab* insecticidal proteins, kills corn borer and other lepidopteran target pests. In theory, Bt maize can affect the food chain with lepidopterans. In this study, a very low density of lepidopterans was detected in all plots, less than 1% of herbivores. Some studies showed that the abundance of target lepidopteran European corn borer (*Ostrinia nubilalis*) reached a maximum of 0.2% of herbivores [12]. Thus, the effects of GM maize on the food chains and webs with lepidopterans should be assessed.

Aphids do not belong to the leaf-chewing defoliators because of their special feeding behaviors, sucking phloem sap from the leaves and buds. Compared with lepidopterans, Bt maize had little effect on the food chain of aphids and ladybirds. In this study, the differences in aphid populations between maize varieties (XY and NT) were greater than those between GM maize and its counterpart (T1/T2 and NT), which suggested that Bt maize had little effect on the aphid–ladybird food chain. This explains why Bt maize hardly affects arthropod communities in the field. At the same time, there is the question of whether it is necessary to plant Bt maize in areas with low lepidopteran pest densities.

#### **4. Materials and Methods**

#### *4.1. Maize Materials*

Four maize varieties, including non-transgenic cultivated variety XianYu335 (XY), nontransgenic recipient variety DBN318 (NT), transgenic maize DBN9868 (T1) and DBN9936 (T2), were provided by DNB Biotech Co. Ltd. (Beijing, China). XY is a commonly cultivated variety in Northeast China. NT is the non-transgenic recipient for both T1 and T2. T1, expressing the *PAT* (Phosphinothricin *N*-acetyltransferase) and *EPSPS* (5 enolpyruvylshikimate-3-phosphate synthase) genes, is a glyphosate- and glufosinatetolerant variety. T2, expressing the *Cry1Ab* and *EPSPS* genes, is an insect-resistant and glyphosate-tolerant variety. Both DBN9868 (T1) and DBN9936 (T2) were declared safe via safety certificates, and our research played a role in that declaration. The biosafety certificate for DBN9936 (T2) was granted by the Ministry of Agriculture, PR China, on 21 January 2020. The expression levels of Cry1Ab, EPSPS and PAT proteins in maize materials, determined by ELISA, are shown in Table S4, and the methods are described in the Supplementary Materials.

#### *4.2. Planting Plot Design*

A set of four independent experiments was conducted on the test base in Yitong (Jilin Province, China) (125◦26 18.22 E; 43◦14 8.94 N) from June to September of each year, 2015 to 2017. No crops were planted in the field until the experiment started. XY, NT and T1 were planted in 2015 and 2016; XY, NT and T2 were planted in 2015 and 2017. Each

variety was replicated three times on 150 m2 (10 m × 15 m), with corridors of 1.0 m between them (the plots design is shown in Figure S3). The experiment was carried out with a randomized blocks design. Maize was sown with a between-row spacing of 60 cm and a within-row spacing of 25 cm. No herbicides or pesticides were applied during the growth period of maize.

#### *4.3. Investigation of Arthropods*

The diagonal five-point sampling method was used to determine sampling sites. At each sampling site, 10 maize plants were investigated.

The arthropods on the plants were counted by direct observation. The investigations on arthropods were taken once a week from the 50th day after cultivation to harvest. The families, number and developmental stages of all the arthropods on plants were recorded. Some species that occur in large numbers were counted and collected for later identification in the laboratory. In addition, the food consumption behaviors of arthropods were observed in order to map the food web.

The ground-dwelling arthropods were collected with pitfall traps. Samples were taken once every 14 days from the 50th day after planting to harvest. There were five sample sites within each plot and three traps within each sample site. The traps were made of plastic cups (Ø7 cm × 8 cm) holding one-third solution (5% detergent + 10% alcohol + H2O). The traps were embedded in the ground, and the rim of the cup was flush with the ground. Arthropods in the traps were collected the next day and taken to the laboratory for identification and counting.

#### *4.4. Statistical Analyses*

Five indices were used to analyze the diversity of the arthropods: the species richness (*S*), the Shannon–Wiener diversity index (*H*), the Pielou evenness index (*J*), the dominance index (*C*) and the community similarity index (*C* ) [25,26,28,53].

The *S* is the number of species within a defined region.

The *H* of the arthropods was calculated using Equation.

$$H = -\sum\_{i=1}^{S} P\_i \ln P\_i$$

The *J* of the arthropods was calculated as per equation.

$$I = \frac{H}{ImS}$$

The *C* of the arthropods was calculated as per equation.

$$\mathcal{C} = \sum\_{i=1}^{n} \left(\frac{N\_i}{N}\right)^2$$

The *C* of the arthropods was calculated as per equation.

$$\mathcal{C}' = 2w/(a+b)$$

*Pi* is the proportion of individuals belonging to a certain species, genus or family to the total number of individuals. *S* is the total number of species in a community. *Ni* is the number of individuals belonging to a certain species, genus or family and *N* is the total number of individuals in the region. *w* is the number of species shared by two samples, and *a* and *b* are the total numbers of species in each of the respective samples. Since it was not possible to identify all species in the field, the number of families was analyzed.

The mean values and SE of each community were calculated and compared using Student's *t*-test to detect significant differences between XY and NT, NT and T1, and NT and T2. A significance level of *p* < 0.05 was used.

**Supplementary Materials:** The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/plants11172254/s1, Table S1: Sampling dates from 2015 to 2017, Table S2: Orders, family and parts of identified patterns of the arthropod community in maize field, Table S3: From 2015 to 2017, the cumulative number of main arthropod groups in the field of cultivated variety XianYu335 (XY), non-transgenic recipient variety DBN318 (NT), transgenic maize DBN9868 (T1) and DBN9936 (T2), Table S4: The expression levels of Cry1Ab, EPSPS and PAT proteins in maize materials determined by ELISA (*μ*g/g·dwt), Figure S1: From 2015 to 2017, Dynamic of Pielou evenness index (*J*) of arthropod community in the fields of cultivated variety XianYu335 (XY), non-transgenic recipient variety DBN318 (NT), transgenic maize DBN9868 (T1) and DBN9936 (T2), Figure S2: From 2015 to 2017, Dynamic of the dominance index (*C*) of arthropod communities in the fields of cultivated variety XianYu335 (XY), non-transgenic recipient variety DBN318 (NT), transgenic maize DBN9868 (T1) and DBN9936 (T2), Figure S3: Plots design in 2015.

**Author Contributions:** Methodology, B.L. and K.X.; software, Z.R.; investigation, Z.R., M.Y., H.H. and Y.M.; data curation, Z.R.; writing—original draft preparation, Z.R.; writing—review and editing, Y.Z., B.L. and K.X.; visualization, Z.R.; project administration, B.L. and K.X.; funding acquisition, Y.Z., B.L. and K.X. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by Key Laboratory of Ecology and Environment in Minority Areas (Minzu University of China), National Ethnic Affairs Commission (No. KLEEMA202102) and National Project on New Transgenic Crops Breeding (No. 2021ZX08013001-002-007).

**Data Availability Statement:** Data is contained within the Supplementary Material. The data presented in this study are available in [Supplementary Material 2].

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

#### **References**

