A Molecular Marker to Identify Spodoptera frugiperda (JE Smith) DNA in Predators’ Gut Content

Simple Summary This work aimed to build a molecular marker to detect Spodoptera frugiperda DNA in predators’ gut content. The molecular marker developed is highly specific, and it was able to detect S. frugiperda DNA in the gut content of ladybug and earwig predators in field conditions. Our results confirm that generalist predators feed on S. frugiperda in maize fields, and they must be considered in IPM programs for S. frugiperda suppression. Abstract Spodoptera frugiperda is a serious pest of maize and other crops worldwide. The integration of control tactics is recommended for S. frugiperda suppression because reports of insecticide and Btplant-resistance are frequent. Biological control agents would be an alternative to improve S. frugiperda control in agricultural areas. We constructed a species-specific molecular marker to detect S. frugiperda DNA in predators’ gut content and estimated the predation rates of ladybugs and earwigs on S. frugiperda in maize crops. Predators were sampled in Pirassununga, São Paulo state, Brazil, in 2020 and 2021. Using the species-specific molecular marker in laboratory conditions, we estimated the half-life time to detect S. frugiperda DNA in the gut contents of Hippodamia convergens as 6.16 h and Doru luteipes as 25.72 h. The weekly predation rate of S. frugiperda by predators in maize crop varied from 0 to 42.1% by ladybugs and from 0 to 9.2% by D. luteipes. Predation events on S. frugiperda by predators were more frequent during the maize reproductive stage. Our results confirmed that predators might contribute to S. frugiperda suppression in maize fields. However, further studies of prey–predator interactions and agricultural landscapes are essential for a better understanding of predator dynamics in crops.


Introduction
Spodoptera frugiperda (JE Smith) (Lepidoptera: Noctuidae), popularly known as the fall armyworm, is native to the Americas. In 2016, S. frugiperda was detected in Africa and reached the status of a cosmopolitan species after invading and dispersing in regions of Europe, Africa, Asia, and Oceania [1][2][3]. Its polyphagous and voracious feeding behavior, high reproductive rate, and long-distance dispersal capacity make S. frugiperda a serious agricultural pest [4][5][6]. In Brazil, S. frugiperda has been an economically important pest of maize and cotton crops. However, in recent years, this species has increased in abundance in soybean crops, making its management more challenging [7].
Different tactics are used for S. frugiperda control, including insecticides, transgenic plants, and natural enemies. Insecticides are the most common management strategy for S. frugiperda. However, their effectiveness is limited because S. frugiperda larvae feed inside the maize whorl where insecticides may not penetrate [8]. Furthermore, in recent years insecticide-resistant populations of S. frugiperda have often been reported [9][10][11][12]. Transgenic Bacillus thuringiensis (Bt) plants are another worldwide control strategy for S. frugiperda management. However, as with insecticides, S. frugiperda populations resistant to Bt crops have been reported in Puerto Rico, Brazil, Argentina, and the USA [6,[13][14][15].
Although predators are frequent in agricultural areas, the real contribution of these natural enemies to S. frugiperda suppression is still difficult to estimate in the field. One approach used to identify and estimate the predation rate of predators on a group of prey is an analysis of predator gut contents [24][25][26]. In the past, these analyses were carried out by identifying prey body parts found in the predator's gut, although it was difficult to identify the prey to species level [24]. Therefore, molecular tools have been widely used in the analysis of predator gut contents because they allow rapid and precise identification based on the design of species-specific primers, multiplex PCR markers, and, more recently, metabarcoding [24,25,27,28].
Based on the worldwide economic importance and the reports of control failures of S. frugiperda, predators may be important components in the development of IPM plans to suppress this pest. We designed a specific molecular marker to detect the predation rate of generalist predators on S. frugiperda in maize crops. Specifically, earwigs and ladybugs are two groups of predators that are frequently cited as natural enemies of S. frugiperda [16,18], but they are inadequately studied in the field. Our specific aims were: (1) design speciesspecific primers to detect S. frugiperda DNA in the gut contents of generalist predators; (2) estimate the detectability half-life in S. frugiperda DNA in the gut contents of an earwig, D. luteipes, and a ladybug, Hippodamia convergens (Coleoptera: Coccinelidae)-both common generalist predators found in maize crops-and to confirm the utility of this molecular marker; and (3) estimate the predation rate of S. frugiperda for predators collected weekly in maize crops.

Field Detection of S. frugiperda and Predator Collection
Insect monitoring and collections were carried out on the maize farm in Pirassununga. The collections were made in two consecutive fields: in the winter, crop from April to July 2020, a total of 14 weeks of sampling; and in the summer, crop from December 2020 to February 2021, 8 weeks of sampling. In the winter crop, the maize plants developed more slowly due to the low rainfall (46 mm) and temperature (19 • C); consequently, the sampling period was longer than in the summer crop (mean rainfall 217 mm and mean temperature 25 • C). Sampling started in the maize vegetative stage V2 and continued until the harvest. The vegetative stage lasted until week 7 in winter and week 4 in summer. Both crops were cultivated under conventional practices but without insecticide applications for insect control.
A delta trap baited with the sex pheromone Bio Spodoptera (Bio Controle, Indaiatuba, Brazil) was installed in the middle of the area to detect the presence of S. frugiperda during the experiment. The sticky liner was replaced each week and the sex pheromone replaced every 3 weeks or less if the pheromone was depleted. In the laboratory, the moths on the sticky liner were photographed and identified using morphological traits such as body and wing colors.
Earwig and ladybug individuals were manually collected randomly along the plant each week from the maize vegetative stage V2 until the harvest. The maize crop area was divided into three sub-areas of 20 m 2 each, and predators were collected actively to Insects 2022, 13, 635 3 of 11 preserve their gut contents by immediately freezing the specimens. Each predator collected was immediately placed in a 1.5 mL tube and stored in a plastic bag on ice until arriving at the laboratory. At the laboratory, the predators were stored in 99.9% ethanol at -20 • C and taxonomically identified.
To estimate the predation rate by predators on S. frugiperda, the predators' DNA was extracted, and the specific marker to detect the presence of S. frugiperda DNA in the predators' gut contents was applied exactly according to the above method 2.2. Molecular Analysis of Gut Contents 2.2.1. DNA Extraction All insects used in this study were cleaned by submersion in 2% sodium hypochlorite for 2 s, next in 70% ethanol for 2 s, and then in autoclaved distilled water for 5 s and allowed to dry on clean tissue paper. These procedures were carried out to cleanse any external DNA from the insect body.
DNA was extracted using the entire body except for the wings, antenna, and legs, with the CTAB protocol adapted from Corrêa et al. [29]. Each individual was submerged in 500 µL of CTAB buffer, 10 µL of proteinase K 20 mg mL-1 (Invitrogen, Waltham, MA, USA), and 2 µL of β-mercaptoethanol. The samples were incubated at 65 • C for 2 h. Then, 3 µL of PureLinkTM RNase A 20 mg mL-1 (Invitrogen) was added to each sample, which was incubated at 37 • C for 2 h and then at 65 • C for 30 min. The samples were centrifuged at 14,000 rpm for 5 min and the liquid phase was transferred to a new tube, mixed with 500 µL of chloroform and isoamyl alcohol (24:1), and centrifuged at 14,000 rpm for 20 min. The supernatant was transferred to a new tube, and the isoamyl alcohol step was repeated. Then, the supernatant was transferred and mixed with 100% isopropanol and incubated at -20 • C overnight. The samples were centrifuged at 14,000 rpm at 4 • C for 30 min and the supernatant was discarded. The DNA pellet was cleaned with 400 µL of cold 70% ethanol. The cleaning step was repeated with cold 100% ethanol and then the samples were allowed to air-dry. The dried DNA pellet was resuspended in 30 µL of autoclaved distilled water.

Primer Design and Optimization
The S. frugiperda species-specific primers were designed using sequences of the cytochrome c oxidase subunit I (COI) gene available in GenBank (access numbers: KF624877. S. frugiperda DNA in predator gut contents were carried out simultaneously with three control samples: water and DNA predators that did not have contact with S. frugiperda as two negative controls, and a positive control DNA of S. frugiperda. . The amplification (fragment size: 130 bp) was evaluated by electrophoresis on 2% agarose gel. PCR assays to detect the presence of S. frugiperda DNA in predator gut contents were carried out simultaneously with three control samples: water and DNA predators that did not have contact with S. frugiperda as two negative controls, and a positive control DNA of S. frugiperda.

Feeding Studies
Adults of H. convergens were sampled in a maize field in Pirassununga (22°03′59.8″ S 47°25′58.4″ W), São Paulo state, Brazil. The ladybugs were reared in plastic Petri dishes for oviposition and fed with inviable eggs of Ephestia kuehniella Zeller (Lepidoptera: Pyralidae) and a mixture of water, honey, and brewer's yeast in cotton. H. convergens eggs were removed to new containers and held until the larvae emerged. The larvae were kept in glass tubes and fed with inviable eggs of E. kuehniella until they reached the fourth instar. All adults and larvae were reared in Bio-Oxygen Demand (BOD) incubators with a photoperiod regime of 14:10 (L:D) and 60 ± 10% relative humidity (RH) at 26 ± 1 °C.
Doru luteipes adults were collected in the field and maintained in dark plastic containers, with artificial diet that contained 35% cat food, 27% wheat germ, 23% brewer's yeast, 14% milk in powder, 0.5% nipagin, and 0.5% sorbic acid [30]; and an egg box, folded paper, and plastic straws with moistened cotton on one side. The adults oviposited in the moist cotton, and the straws with the eggs and the mother earwigs were maintained in another container until the nymphs reached the second instar. All the containers were kept in a climate-controlled room with 70 ± 10% relative humidity (RH) at 26 ± 1 °C in the dark.
A total of 70 H. convergens fourth-instar larvae and 66 D. luteipes nymphs were starved for 24 h. They were placed in individual Petri dishes and were offered a single S. frugiperda neonate larva for each individual of H. convergens and D. luteipes. The predators were transferred to microtubes containing 99.9% ethanol after 0.5, 2, 4, 12, 24, and 48 h of feeding on S. frugiperda and stored at -20 °C. During the period between the feeding and transfer to microtubes, the individuals were kept with no food in a BOD incubator. Each time group of H. convergens contained 10 individuals. Groups 0-12 h of D. luteipes contained 10 individuals, the 24 h group contained 9, and the 48 h group contained 7. The DNA extraction and PCR to detect the presence or absence of S. frugiperda DNA in the predators' gut contents were carried out according to the methods described above. Additionally, during these bioassays, some positive PCR fragments were sequenced by Sanger sequencing to confirm that the amplified fragment with these primers corresponded only to S. frugiperda.

Feeding Studies
Adults of H. convergens were sampled in a maize field in Pirassununga (22 • 03 59.8 S 47 • 25 58.4 W), São Paulo state, Brazil. The ladybugs were reared in plastic Petri dishes for oviposition and fed with inviable eggs of Ephestia kuehniella Zeller (Lepidoptera: Pyralidae) and a mixture of water, honey, and brewer's yeast in cotton. H. convergens eggs were removed to new containers and held until the larvae emerged. The larvae were kept in glass tubes and fed with inviable eggs of E. kuehniella until they reached the fourth instar. All adults and larvae were reared in Bio-Oxygen Demand (BOD) incubators with a photoperiod regime of 14:10 (L:D) and 60 ± 10% relative humidity (RH) at 26 ± 1 • C.
Doru luteipes adults were collected in the field and maintained in dark plastic containers, with artificial diet that contained 35% cat food, 27% wheat germ, 23% brewer's yeast, 14% milk in powder, 0.5% nipagin, and 0.5% sorbic acid [30]; and an egg box, folded paper, and plastic straws with moistened cotton on one side. The adults oviposited in the moist cotton, and the straws with the eggs and the mother earwigs were maintained in another container until the nymphs reached the second instar. All the containers were kept in a climate-controlled room with 70 ± 10% relative humidity (RH) at 26 ± 1 • C in the dark.
A total of 70 H. convergens fourth-instar larvae and 66 D. luteipes nymphs were starved for 24 h. They were placed in individual Petri dishes and were offered a single S. frugiperda neonate larva for each individual of H. convergens and D. luteipes. The predators were transferred to microtubes containing 99.9% ethanol after 0.5, 2, 4, 12, 24, and 48 h of feeding on S. frugiperda and stored at -20 • C. During the period between the feeding and transfer to microtubes, the individuals were kept with no food in a BOD incubator. Each time group of H. convergens contained 10 individuals. Groups 0-12 h of D. luteipes contained 10 individuals, the 24 h group contained 9, and the 48 h group contained 7. The DNA extraction and PCR to detect the presence or absence of S. frugiperda DNA in the predators' gut contents were carried out according to the methods described above. Additionally, during these bioassays, some positive PCR fragments were sequenced by Sanger sequencing to confirm that the amplified fragment with these primers corresponded only to S. frugiperda.
The feeding results were analyzed by using the package "drc" in R Studio [31,32] to determine the half-life of the DNA in the predators' gut contents.

Primer Specificity and Half-Life for S. frugiperda DNA
The primers Spo_frugi-F and Spo_frugi-R designed for S. frugiperda were highly species-specific in our PCR conditions, and they were able to only amplify the target species when tested with DNA from other species of Spodoptera and other moths, earwigs, and ladybugs ( Figure 2). Only predators that had fed on S. frugiperda in the laboratory (positive

Primer Specificity and Half-Life for S. frugiperda DNA
The primers Spo_frugi-F and Spo_frugi-R designed for S. frugiperda were highly species-specific in our PCR conditions, and they were able to only amplify the target species when tested with DNA from other species of Spodoptera and other moths, earwigs, and ladybugs ( Figure 2). Only predators that had fed on S. frugiperda in the laboratory (positive control) had a positive amplification in the PCR assay.

Insect Field Collection and Predator Gut-Content Analysis
In the winter crop, a total of 388 predators were sampled. Of these, H. convergens was the most abundant predator with 228 samples collected, followed by 124 for C. sanguinea, 47 for H. axyridis, 27 for Eriopis connexa, and 2 for D. luteipes. The predators appeared in week 5 with 3 ladybugs; week 6 with 16 ladybugs; week 7 with 25 ladybugs; week 8 with 78 ladybugs; week 9 with 69; week 10 with 61; week 11 with 40 ladybugs and 1 earwig; week 12 with 54 ladybugs; week 13 with 41 ladybugs; and week 14 with 39 ladybugs and 1 earwig (Figure 3a). The weeks that were most abundant in predators were weeks 8 and 9 when the maize was in the beginning of reproductive stage. The populations of H. axyridis and C. sanguinea fluctuated similarly, reaching their highest in week 8 and decreasing continuously in the following weeks. The population of E. connexa fluctuated constantly during the maize crop cycle (Figure 3a).

Insect Field Collection and Predator Gut-Content Analysis
In the winter crop, a total of 388 predators were sampled. Of these, H. convergens was the most abundant predator with 228 samples collected, followed by 124 for C. sanguinea, 47 for H. axyridis, 27 for Eriopis connexa, and 2 for D. luteipes. The predators appeared in week 5 with 3 ladybugs; week 6 with 16 ladybugs; week 7 with 25 ladybugs; week 8 with 78 ladybugs; week 9 with 69; week 10 with 61; week 11 with 40 ladybugs and 1 earwig; week 12 with 54 ladybugs; week 13 with 41 ladybugs; and week 14 with 39 ladybugs and 1 earwig (Figure 3a). The weeks that were most abundant in predators were weeks 8 and 9 when the maize was in the beginning of reproductive stage. The populations of H. axyridis and C. sanguinea fluctuated similarly, reaching their highest in week 8 and decreasing continuously in the following weeks. The population of E. connexa fluctuated constantly during the maize crop cycle (Figure 3a).
Predation events in the winter crop occurred only in weeks 8, 9, and 10, and all ladybug species fed on S. frugiperda. H. convergens had the most positive results (24 positives) with predation rates varying between 5 and 42%, followed by H. axyridis (6 positives) with predation rates between 11 and 36%. The predation rate for C. sanguinea (four positives) was 18.2% and for E. connexa (two positives) was 25% (Table 1). Insects 2022, 13, x FOR PEER REVIEW 6 of 11 Predation events in the winter crop occurred only in weeks 8, 9, and 10, and all ladybug species fed on S. frugiperda. H. convergens had the most positive results (24 positives) with predation rates varying between 5 and 42%, followed by H. axyridis (6 positives) with predation rates between 11 and 36%. The predation rate for C. sanguinea (four positives) was 18.2% and for E. connexa (two positives) was 25% (Table 1). In the summer crop, a total of 334 predators were sampled, of which 241 were D. luteipes and 93 were ladybugs. The most abundant ladybug species was H. axyridis, with 61 individuals (Figure 3b). The predators appeared in week 4, at the end of vegetative stage, with 31 ladybugs and 22 earwigs; week 5 with 35 ladybugs and 15 earwigs; week 6 with 11 ladybugs and 64 earwigs; week 7 with 15 ladybugs and 109 earwigs; and week 8 with 1 ladybug and 31 earwigs. The most abundant weeks were 6 and 7 for earwigs, and weeks 4 and 5. Predation events were detected in 14 D. luteipes, with the predation rate varying between 1% to 9%. No predation was detected in ladybug individuals. One predation event occurred in the maize vegetative stage, week 4, and three predation events in reproductive stage, weeks 5, 6, and 7, where week 6 had the most predation events (Table 1).

Discussion
We developed a molecular marker for the specific identification of S. frugiperda in predator gut contents. This marker can distinguish S. frugiperda from other Spodoptera species and is useful for identifying and monitoring this cosmopolitan pest by means of a single PCR method. The S. frugiperda-specific primers were based on the cytochrome c oxidase subunit I (COI) gene and were successfully used as a DNA barcoding region in animals [33][34][35][36]. One advantage of the molecular marker developed here is the small fragment size amplified, favoring PCR amplification and identification of insects even when the DNA was degraded, which is an essential trait of a molecular marker for gut-content analysis studies [28,37].
The DNA half-life for detection of S. frugiperda in H. convergens and D. luteipes gut contents differed between predators, with rates of 6.16 h and 25.72 h for H. convergens and D. luteipes, respectively. DNA half-life for detection of other prey in H. convergens gut contents showed a similar detection time for Rhopalosiphum maidis (Hemiptera: Aphydiidae), 8.78 h [28]; and for Diaphorina citri (Hemiptera: Psyllidae), 6.11 h [38]. Here, the DNA half-life detection time in gut contents was estimated for the first time in D. luteipes, and it was longer than in H. convergens.
Differences in DNA half-life detection times in predator gut contents are expected because the DNA half-life is driven by environmental conditions and by biological traits of the predator [39,40]. Insect life stage, metabolism, mouthpart type, and gut size and shape are some variables that affect the process of DNA degradation in the predator gut and, consequently, the DNA half-life detection time [27,38,41]. However, DNA half-life detection assays in laboratory conditions are useful and essential for determining if the molecular marker is efficient in detecting low concentrations of prey DNA in predator gut contents, as we confirmed here for this S. frugiperda marker.
Predation on S. frugiperda was detected in all predator species collected in maize fields using DNA gut-content analysis. However, the higher predation rates were detected in H. convergens and H. axyridis-even higher than in D. luteipes, reported as an efficient predator of S. frugiperda [16]. The higher predation rate detected in these ladybug species compared with D. luteipes is likely even higher, in view of the 4.0x longer DNA half-life in gut contents of D. luteipes compared to H. convergens in laboratory conditions. Therefore, ladybugs must be considered as biological-control agents for S. frugiperda, especially H. convergens and H. axyridis, the two most abundant coccinellid species in our samples.
In the field, generalist predators can feed on a variety of sources such as pollen, decomposing animal and plant tissues, and different prey species. High rates of prey consumption by a predator in the laboratory or greenhouse, as reported for D. luteipes on S. frugiperda [16], may not indicate a high predation rate in the field because it also commonly exhibits scavenging behavior. Alternative prey types or feeding behavior may also affect the predation rate of a predator on a target prey [42,43]. For example, ladybugs may prefer to feed on aphids rather than other prey, as these predators are commonly reported as biological-control agents for aphids [44].
The presence of S. frugiperda in the field was constant throughout the maize cycle. This species damages leaves and cobs, but maize is the most susceptible in the vegetative and initial reproductive stages, when S. frugiperda may feed on the meristem of young plants, stunting or killing them, and damage the kernels, causing enormous losses [45]. However, in both collection seasons, the predators were found in the area from six weeks after the maize sprouted, and their peak was in the maize reproductive stage. Predators are also attracted to the plant by the pollen, nectar, and synomones [46,47]. At the beginning of the reproductive stage, maize produces pollen that attracts insects, including different predator species [48][49][50]. This delay in the arrival of predators, observed in both seasons in comparison to S. frugiperda, may reduce the predators' control efficiency when the maize is most susceptible to the pest. Thus, strategies to advance their arrival and retain the predators in the area should be adopted to improve the biological control of S. frugiperda in the first weeks of plant development [42,51,52].

Conclusions
We successfully constructed a molecular marker to detect S. frugiperda DNA in predator gut contents. Using this molecular marker, we confirmed that earwigs and ladybugs, especially H. convergens and H. axyridis, preyed on S. frugiperda in the maize field. Our study indicated that predators might be considered a reliable component of IPM for S. frugiperda. Therefore, a better understanding of the dynamics of prey and predator in the landscape is crucial for managing predators and implementing conservative and applied biological control programs in maize fields.