Performance of Two Trichogrammatid Species from Zambia on Fall Armyworm, Spodoptera frugiperda (J. E. Smith) (Lepidoptera: Noctuidae)

Simple Summary Two egg parasitoid species (Trichogramma mwanzai and Trichogrammatoidea lutea) from parasitized eggs of fall armyworm (FAW) Spodoptera frugiperda in Zambia were identified by using a combination of both molecular and morphological characters. We compared the parasitism capabilities of the two species with three native Chinese trichogrammatid species (T. ostriniae, T. leucaniae and T. japonicum) using 0- to 2-day-old FAW eggs. Both parasitoid species accepted eggs of all ages tested and completed their development successfully. Trichogrammatoidea lutea females showed the highest parasitism rate of host eggs among the five tested species. T. mwanzai had the shortest developmental time on all test age eggs. Of the five parasitoid species reared on FAW eggs, T. lutea performed the best overall, while T. japonicum was the worst performing of the parasitoids. Abstract The fall armyworm, Spodoptera frugiperda (J.E. Smith), is a noctuid moth native to the tropical and subtropical Americas that has successfully invaded Africa and Asia, where it is has become a serious threat to food security as a pest of cereals and other crops. Biological control is an environmentally friendly means of combating the pest and contributes to an integrated pest management approach. In our study, two egg parasitoid species (Trichogramma mwanzai and Trichogrammatoidea lutea) found in parasitized fall armyworm eggs in Zambia were identified by using a combination of both molecular and morphological characteristics. To evaluate their potential and efficiency on 0- to 2-day-old fall armyworm eggs, we compared their parasitism capabilities with three Trichogramma species native to China (T. ostriniae, T. leucaniae and T. japonicum) under laboratory conditions. The results showed that both parasitoid species would accept 0-, 1- and 2-day-old fall armyworm eggs, and complete their development successfully. Trichogramma mwanzai and T. lutea preferred parasitizing 0- and 1-day-old eggs over 2-day-old eggs. Trichogrammatoidea lutea females supplied with fall armyworm eggs produced the highest parasitism rate of host eggs among the five tested species, while T. mwanzai had the shortest developmental time on all test age eggs. In general, T. lutea was the best performing of the five species when reared on fall armyworm eggs, while T. japonicum was the worst. There were no significant differences, however, in percent emergence in the five test species when reared on fall armyworm eggs.


Introduction
The fall armyworm, Spodoptera frugiperda (J.E. Smith) (=FAW), is a serious crop pest endemic to tropical and subtropical regions of the Americas. It is a species in the family Noctuidae, order Lepidoptera, and was first discovered on the African continent in 2016 [1,2]. It is a destructive pest that has been recorded from 353 host species belonging to 76 plant families, among which Poaceae, Fabaceae, Solanaceae, Asteraceae, Rosaceae, Chenopodiaceous, Brassicaceae and Cyperaceae are the most commonly affected [3,4]. Wide adaptability, strong migratory ability, high fecundity, lack of diapause, rapid development of resistance to insecticides, and voracious appetite have contributed to the difficulties encountered when attempting to manage this serious pest [5][6][7][8][9]. Recently, the species has broken out in Africa [2,10,11] and successfully invaded Europe, including Germany and The Netherlands [11], and has even reached several Asian countries including India, Myanmar, and China [12,13]. The infestation of FAW quickly spanned two continents (Europe and Asia) in just three years [13].
Absence of diapause and favorable climatic conditions in most African countries allow S. frugiperda to complete several generations per year wherever host plants, including offseason and irrigated crops, are available [9,[14][15][16]. Based on estimates published by the Centre for Agriculture and Bioscience International (CABI), in the absence of natural enemies and proper control methods, FAW has the potential to cause maize yield losses of 8.3 to 20.6 t per year, in 12 African maize-producing countries [17]. These losses represent a range of 21-53% of the annual production of maize in the countries involved. In 2017, African countries experienced between USD 2.4 billion to USD 6.1 billion in crop loss due to this species. FAW has also been recorded on other important food crops in Africa including sorghum and millet [11,[17][18][19].
The current management methods used to control FAW rely on intensive use of chemical insecticides, even though long-term insecticide application is known to lead to many serious problems in crops, notably pesticide resistance, negative impacts on non-targeted organisms and environmental degradation [20][21][22][23][24][25][26]. This damage has been responsible for serious problems throughout much of Africa, leading to excessive use of toxic insecticides and small-holder farmers using scientifically unproven methods, for example the application of ash, sand, botanical extracts, and other locally available materials [27]. According to available data, there are currently 121 worldwide species of parasitic wasps distributed in 10 families that can be used against FAW, including Trichogramma pretiosum (Riley), Telenomus remus (Nixon), Campoletis sonorensis (Cameron), Cotesia marginiventris (Cresson), and Chelonus insularis (Cresson) [28][29][30]. Trichogramma egg parasitoids (Hymenoptera: Trichogrammatidae) have been used worldwide as biological control agents because they can control insect pests at an early developmental stage, thereby preventing serious crop damage [31][32][33][34].
Although 12 parasitoid species are currently known to attack fall armyworm in Zambia, no egg parasitoids have been recorded to date [14]. Herein we report the finding of two egg parasitoids of FAW in samples from Zambia. The specific objectives of our study were to: (1) Identify both trichogrammatid species using a combination of external morphological characters including male external genitalia, and molecular analysis; (2) Explore the performance of both trichogrammatid species on FAW eggs of different ages; (3) Compare the bio-control potential of both parasitoid species from Zambia with three local trichogrammatid species (T. ostriniae, T. japonicum, and T. leucaniae) from China. The primary goal of this study was to determine the biological control potential of the above-mentioned parasitoids, and to provide a theoretical basis for their release in IPM programs involving FAW.

Sample Collection and Population Establishment
Masses of fall armyworm (FAW) Spodoptera frugiperda (Smith) were collected from September to October 2019 from maize fields at the China-aid Zambia Agricultural Technol-ogy Demonstration Centre (15 • 21 30 S, 28 • 27 27 E), Lusaka, Zambia. FAW egg masses including parasitized and unparasitized eggs were separated, and each egg mass containing parasitized black eggs was placed into a glass tube (10 cm × 1.5 cm, length × diameter). Emerged parasitoid individuals were used to initiate a small-scale culture under laboratory conditions of 26 ± 1 • C, RH 70 ± 5% and a photoperiod of 14:10 (L:D) h. The parasitoid populations were raised on rice moth (Corcyra cephalonica) eggs and maintained at the above laboratory conditions. Every five generations, the population was rejuvenated by transferring in additional FAW eggs. In this study, three local Chinese Trichogramma species (T. ostriniae, T. leucaniae and T. japonicum) (which are often used as dominant parasitoids to control a variety of Lepidoptera insects) were also tested on the FAW eggs [34]. Trichogramma ostriniae and T. leucaniae were collected from parasitized eggs of Leguminivora glycinivorella (Matsumura) (Lepidoptera: Tortricidae) from Heihe (Heilongliang Prov.) (50.22 • N, 127.53 • E), and T. chilonis was collected from parasitized eggs of Chilo suppressalis (Walker) (Lepidoptera: Crambidae) from Changchun (Jilin Prov.) (43.89 • N, 125.32 • E) in northeastern China in 2011 [35]. These populations were also raised on C. cephalonica eggs, and maintained under laboratory conditions similar to the above.

Morphological Identification
At least 50 parasitoid specimens were initially prepared on microscope slides using the method described by Noyes [36] with some modifications, as follows: prior to dissection, specimens were macerated in 5% KOH in a plastic dish (diameter = 2.0 cm) for 24 h, and then washed two to three times with distilled water. Male genitalia and wings were separated from other parts under a microscope and then dehydrated for 5 min each in 50, 75, 80, 85, 90, 95, and 100% graded alcohols. After allowing for evaporation of alcohol, specimens were fixed in Arabic gum for observation.
To further molecular biological identification, a total of 155 COI public sequences of Trichogramma sp. and Trichogrammatoidea sp. were downloaded, which represented records for the two species on the NCBI and the Bold System databases. We removed some of the sequences that were recorded from the same sample source; only 40 public trichogrammatid COI sequences were used to construct the phylogenetic tree NJ-tree analysis (Supplementary Table S1).

Assessment of Biological Characteristics
Based on the identification results obtained from the above procedures, the performance of the two African trichogrammatid species (Trichogramma sp. and Trichogrammatoidea sp.) and three local Chinese Trichogramma species (T. ostriniae, T. leucaniae and T. japonicum) were assessed when reared on FAW eggs of different ages. Preliminary observations showed that under laboratory conditions FAW eggs hatched within four days. Thus, 0-, 1-and 2-day-old host eggs were selected for the study involving the five trichogrammatid species.
The experiment was conducted under laboratory conditions at 26 ± 1 • C, RH of 70 ± 5%, and a photoperiod of 14:10 (L:D) h. One newly emerged (<8 h old), mated, female parasitoid was introduced into a glass tube with one 0-, 1-or 2-day-old egg mass containing 100 to 120 eggs. The female adults were allowed to oviposit for 24 h. Tests involving individuals that died or were subsequently lost during the experimental period were considered invalid. To avoid host larvae feeding on parasitized eggs, the larvae emerging from host eggs were removed by brush as soon as they hatched. The number of parasitized eggs was recorded under a stereoscopic microscope eight days after the parasitoids were removed. The eggs were checked daily to determine the emergence of parasitoid offspring until no further wasps emerged. For each treatment, the number of parasitized eggs (number of the host eggs parasitized in black), the developmental time (d) (the number of days from exposure of host eggs to the parasitoid to the adult offspring emerging from host egg), percent emergence (the number of parasitized eggs with emergence holes/total number of parasitized eggs × 100), and percent female progeny (number of emerged females/total number of emerged females and males × 100) were recorded. There were 15 replicates for each species tested and for each of the egg ages tested.

Phylogenetic Analysis
Overall, sequences obtained in this study and similar sequences retrieved from Gen-Bank were aligned using ClustalW implemented in MEGA 6.0 software [48]. The COI barcode relationships of all sequences were inferred from the reference sequences using BLAST tool in the GenBank database and Bold Systems database. Initial tree(s) for the heuristic search were obtained automatically by applying neighbor-join and BioNJ algorithms to a matrix of pairwise distances estimated using the Maximum Composite Likelihood (MCL) approach, and then selecting the topology with superior log likelihood value. The tree was drawn to scale, with branch lengths measured in the number of substitutions per site. Codon positions included were 1st + 2nd + 3rd + Noncoding. Branch support was assessed via bootstrapping with 1000 replicates. All positions containing gaps and missing data were eliminated.

Bioassay Analysis
The data sets obtained in each bioassay regarding the number of eggs parasitized, percent emergence, developmental time and percent female progeny were analyzed by a linear model with parasitoid species (5 levels) and host egg age (3 levels) as factors using Tukey's honest significant difference (HSD) test at p < 0.05. Before a GLM was performed, log transformation of number of eggs parasitized was applied, percent emergence and percent female progeny were arcsine square-root-transformed to homogenize variances and were subjected to the Shapiro-Wilk test. All statistical analyses were performed using the SPSS version 20 software package (SPSS Inc., Chicago, IL, USA).

Morphological Identification
The two trichogrammatid species from Zambia were initially identified as Trichogramma mwanzai ( Figure 1a) and Trichogrammatoidea lutea (Figure 1b) based on morphological characters including male genital capsules.

Morphological Identification
The two trichogrammatid species from Zambia were initially identif Trichogramma mwanzai (Figure 1a

Molecular Biological Identification Based on COI
The sequencing of the COI gene of Trichogramma sp. and Trichogrammatoidea sulted in two sequences, a 708 bp fragment of Trichogramma sp. (Accession n MZ7711322) and 716 bp fragment of Trichogrammatoidea sp. (Accession n MZ7711343). The results of the search for similarities by BLAST tool in the NCBI da and Bold Systems database verified that they were the targeted sequences. The res blast showed that the COI sequences of Trichogramma sp. from Zambia showed average identity with published sequences from T. mwanzai in India (Accession n KP142716). The COI sequences of Trichogrammatoidea sp. shared over 98.0% hom identity with the sequences of T. lutea from South Africa (Accession number: KMPO 19). Moreover, NJ-tree ( Figure 2) showed that Trichogramma sp. was close to Trichog mwanzai, and Trichogrammatoidea sp. was nested into the Trichogrammatoidea lutea c The results of molecular biological identification, in combination with morpho evidence, indicated that Trichogramma sp. and Trichogrammatoidea sp. were cons with Trichogramma mawanzai and Trichogrammatoidea lutea, respectively.

Molecular Biological Identification Based on COI
The sequencing of the COI gene of Trichogramma sp. and Trichogrammatoidea sp. resulted in two sequences, a 708 bp fragment of Trichogramma sp. (Accession number: MZ7711322) and 716 bp fragment of Trichogrammatoidea sp. (Accession number: MZ7711343). The results of the search for similarities by BLAST tool in the NCBI database and Bold Systems database verified that they were the targeted sequences. The results of blast showed that the COI sequences of Trichogramma sp. from Zambia showed 96.30% average identity with published sequences from T. mwanzai in India (Accession number: KP142716). The COI sequences of Trichogrammatoidea sp. shared over 98.0% homology identity with the sequences of T. lutea from South Africa (Accession number: . Moreover, NJ-tree ( Figure 2) showed that Trichogramma sp. was close to Trichogramma mwanzai, and Trichogrammatoidea sp. was nested into the Trichogrammatoidea lutea cluster.
The results of molecular biological identification, in combination with morphological evidence, indicated that Trichogramma sp. and Trichogrammatoidea sp. were conspecific with Trichogramma mawanzai and Trichogrammatoidea lutea, respectively.

Performance on Host Eggs
Host age, parasitoid species, and the interaction between these two factors all showed significant effects on the parasitism number, developmental time, and percentage of female progeny of the five egg parasitoid species. There was no effect on the percentage of adult emergence (Table 1).

Performance on Host Eggs
Host age, parasitoid species, and the interaction between these two factors all showed significant effects on the parasitism number, developmental time, and percentage of female progeny of the five egg parasitoid species. There was no effect on the percentage of adult emergence (Table 1).

Parasitism of Trichogramma and Trichogrammatoidea Species on Different Ages of FAW Eggs
The results showed that all five species parasitized FAW eggs ( Figure 3). The egg age of FAW had a significant impact on the number of eggs parasitized by T. mwanzai (F2,42 = 21.797, p < 0.0001), T. lutea (F2,42 = 39.07, p < 0.0001), T. ostriniae (F2,42 = 4.531, p = 0.017), and T. leucaniae (F2,42 = 6.021, p = 0.005). With the exception of T. japonicum, all trichogrammatid species preferred to parasitize 1-day-old eggs over 0-, and 2-day-old eggs. Trichogrammatoidea lutea and T. mwanzai parasitized a significantly higher number of 0-day-old eggs (F4,70 = 12.652, p < 0.0001) and 1-day-old eggs (F4,70 = 64.776, p < 0.0001) than the other three species. T. lutea produced the highest number of parasitized eggs (16.7) on 1-day-old eggs among the five tested species (F4,70 = 64.776, p < 0.0001), followed by T. mwanzai, T. ostriniae, and T. leucaniae, and T. japonicum, which parasitized the fewest hosts in all the egg ages. All of the species studied failed to parasitize more than 20 FAW eggs in a 24 h period, although more than 100 host eggs were provided in every instance.   (Table 2). Conversely, the developmental time of T. mwanzai (F 2,42 = 22.879, p < 0.0001) and T. lutea (F 2,42 = 29.682, p < 0.0001) showed a tendency to decrease as the age of the host egg increased. Trichogrammatoidea lutea had a distinct prolonged developmental time when reared on 0-day-old eggs, while T. mwanzai had the shortest developmental time on 1-, and 2-day-old eggs among all species. Note: Mean ± SE are presented. Means in a column followed by the same lowercase letter and means in a row followed by the same uppercase letter do not differ significantly (p < 0.05) using Tukey's HSD test.

Discussion
Trichogramma and Trichogrammatoidea species reared from parasitized eggs of Spodoptera frugiperda in Zambia were identified based primarily on genitalic and genetic characteristics. The COI sequence of Trichogramma sp. was similar to Trichogramma mwanzai [38] which have been discovered in rice fields in Malawi (Central Africa), and also reported from maize fields in Kenya [38]. Previous reports indicate that T. mwanzai could potentially be used against the eggs of several lepidopterous pests in Africa, such as Chilo partellus Swinhae, Busseola fusca (Fuller), Helicoverpa armigera (Hübner), and Plutella xylostella (L.) [49][50][51]. COI sequencing of Trichogrammatoidea sp. was similar to Trichogrammatoidea lutea described by Nagaraja in his studies on the genus [42]. Most of the lepidopteran hosts were reported to be attacked by T. sp. nr. lutea in eastern Africa [52], and T. lutea in South Africa [53]. Recently, T. lutea has also been found parasitizing FAW eggs in Sadore, Niger [54].
The parasitism and host preference of the egg parasitoids are often determined by the species of parasitoids [55][56][57]. In our study, all species accepted and attacked the FAW egg masses, although T. mwanzai and T. lutea parasitized significantly more FAW eggs (of all ages) than T. japonicum, T. ostriniae, and T. leucaniae. Trichogramma mwanzai and T. lutea parasitized significantly higher numbers of 0-, and 1-day-old eggs than older eggs, suggesting that they preferred these over the older host eggs. Usually, moth scales covering FAW egg masses are quite common and most of the test eggs were covered by scales, constituting a barrier against parasitization [55,58]. Therefore, it is not surprising that all tested trichogrammatid species parasitized fewer than 20% of the FAW eggs provided, with most of the parasitized eggs being those that were not covered by the scales. However, the effectiveness of parasitism by T. mwanzai and T. lutea is quite low compared to Telenomus remus Nixon (Hymenoptera: Platygastridae), a parasitoid native to the Malay Peninsula that has been released in biological control programs against FAW in the Americas [59,60]. In a recent study, T. remus was shown to parasitize more FAW eggs than Trichogrammatoidea sp. due to its ability to overcome the layer of scales on the egg masses [54]. This is consistent with previous findings [61]. Similar results were obtained in another species of Trichogramma, T. pretiosum, on eggs of the closely related noctuid host species, Spodoptera frugiperda and S. litura [57]. Differences in parasitism among species may also be attributed to physical stress of the host, which may affect the oocyte, resulting in increased permeability of the egg [62]. Increased mortality of immature trichogrammatids reported by Calvin et al. [63] and Lund [64] was attributed to decreasing humidity associated with hardening of the eggshells. The relationship between eggshell structure and resistance to desiccation has been studied in several Lepidoptera species [65].
Host age may influence the preference of some parasitoids [66][67][68] and the response to host age is independent of the egg parasitoid species [69]. We observed that, although T. mwanzai and T. lutea accepted all of the host eggs at various ages, in all cases the older host eggs, i.e., 2-day-old, were significantly less parasitized than 0-day-old and 1-day-old eggs. It is common among egg parasitoids that parasitism decreases with increasing host age [68,[70][71][72]. For example, T. chilonis showed a strong preference for 1-2-day-old eggs of P. xylostella [73] or 0-day-old eggs of C. suppressalis [72]. Trichogramma ostriniae preferred to parasitize younger eggs of Mythimna separata (Walker) (Lepidoptera: Noctuidae) at all egg ages, but no significant difference was observed in our case [35]. The rejection of older eggs could be attributed to the change of color, hardening of the chorion, rotation of the host embryo or the sclerotization of the head capsule, making it difficult for the parasitoid to probe the egg [67,74]. In addition, T. japonicum and T. leucaniae tended to have prolonged developmental time on older eggs than on the younger ones (Table 2). Similar observations were reported by Hou et al. [35], who indicated that the development of T. ostriniae was significantly slower with increase in age and that no T. japonicum parasitoids emerged from 3-day-old eggs of M. separata. The possible mechanism underlying these phenomena is the difference in host quality associated with host egg age and the increasing defense capacity with the development of the host egg [75]. In contrast, the developmental time of T. mwanzai and T. lutea in our study showed a tendency to decrease with increasing age. This suggests that host selection and adaptation for some parasitoids could be influenced by the interaction between host age and species [66,67,72]. The Trichogramma species in previous studies exhibit different acceptance for host eggs at different ages [68]. Several previous studies showed that the first 75% of embryological development of the host egg was suitable for trichogrammatid larval development, although the pattern of vulnerability with host age varied among species [67,76].
When the biological control agent is a polyphagous species, it is quite useful to confirm host suitability and parasitism through the test traits (i.e., egg parasitism, adult emergence rate, female progeny, and developmental time) with more than one host species [67,72,77,78]. For example, T. ostriniae from Taiwan, China almost completely ignored Ostrinia nubilalis (Hb.) eggs and only parasitized Sitotroga cerealella (Olivier) eggs [79]. In Kenya, T. mwanzai exhibited a positive preference for Chilo partellus Swinhae eggs over other native host species, but eggs of the silkworm Bombyx mori L. were not attacked [80]. Similarly, in our study, T. mwanzai and T. lutea parasitized and successfully developed in all FAW eggs tested. Recently, we also found that both trichogrammatid species could be effectively produced on the alternative host, Corcyra cephalonica (Stainton) eggs (unpublished data). In general, the ability to control FAW is superior in T. lutea than it is in the other tested species. Nevertheless, our results indicate that the Trichogrammatoidea species show potential for use in biological control programs directed against the FAW in Africa.

Conclusions
We found and identified two species of parasitoids that can be used to potentially control FAW eggs in the field. Both of the parasitoids showed significantly higher parasitism rates on FAW at various egg ages than the three tested species from China. This finding lays the groundwork for the use of T. lutea or T. mwanzai in biocontrol programs in integrated management of the FAW in Africa and China. Likewise, the rearing of T. mwanzai and T. lutea on an alternative host such as Corcyra cephalonica Stainton, should be considered for cost-effective mass production.