Artemisia-derived Actinobacteria Producing Inhibitors of Spodoptera Littoralis via Molecular Modeling Studies


 Insecticide resistance in agricultural pests has prompted the need to discover novel compounds with new modes of action. We investigated the potency of secondary metabolites from seventy endophytic actinobacteria against laboratory and field strains of Spodoptera littoralis (4th instar), comparable to spinetoram (Radiant® 12% SC). Endophytes from Artemisia herba-alba (syn. Seriphidium herba-alba) and Artemisia judaica were highly effective. Chemical profiling of the most potent metabolite was investigated using LC-QTOF-MS-MS technique, and the activity was validated through molecular docking studies. Metabolic extracts from seven actinobacteria (belonging to Streptomyces, Nocardioides, Kitasatospora and Pseudonocardia) have shown lethality to the 4th instar larvae. The metabolite ES2, from Kitasatospora sp., caused significant histopathological and inhibitory effects on 4th instar larvae. Additionally, ES2 caused lesions in the body wall cuticle, indicating a different mode of action than that of spinetoram. Chemical profiling of ES2 have shown presence of cyromazine (molt inhibitor), 4-nitrophenol and diazinon as key constituents. In conclusion, these findings suggest that, secondary metabolites from endophytic actinobacteria inhabiting wild medicinal plants are sustainable source for promising natural biocontrol agents. This is the first illustration of insecticidal activity of the Artemisia spp. microbiome, and the first report on a natural cyromazine synthesized by actinobacteria.


Introduction
The cotton leafworm, Spodoptera littoralis (Boisduval 1833) (Lepidoptera: Noctuidae) is commonly distributed worldwide. This swarming polyphagous foliage feeding insect is one of the most destructive agricultural pests within the subtropical and tropical range (Pasiecznik et al. 2005). It causes considerable damage to plants belonging to 44 different families during the year, including cotton, ornamentals, vegetables, and many other economically important crops. Currently, new technologies based on microbe-and plant-derived insecticides have become essential because S. littoralis has acquired resistance to the traditionally used pesticides (Hazaa et al. 2020).
A promising strategy to nd novel bioactive products is to investigate microbes from non-conventional habitats, such as the endophytes from medicinal plants. The endophyte utilizes unique nutrients that are peculiar of the medicinal plant to produces bioactive metabolites with distinct and novel structures (Tanvir et al. 2019). Endophytic actinobacteria are of great importance as a source of bioactive metabolites, including those for agricultural applications. Streptomyces albus, is an example of endophytes from the drunken horse grass, Achnatherum inebrians which exhibited > 90% mortality in the cotton aphid, Aphis gossypii Glover (Shi et al. 2013). The endophytic Streptomyces sp. from Stemona sessilifolia (a traditional Chinese medicinal plant) can produce 10 new endostemonines compounds with strong lethal activity against Aphis gossypii (Zhao et al. 2020). Worth to mention, the future of endophyte-derived pesticides is not only in exploring novel products for plant protection, but also in developing self-protected plants that are augmented with the potent endophytes in its microbiome. The use of plant microbiome augmentation may allow agriculturists to more sustainably produce crops at a reduced cost (White et al. 2019).
However, only few microbial metabolites have reached the market for pest control (Kvakkestad et al. 2020). Those include, avermectins which is produced by a soil Streptomyces sp.; it acts on gammaaminobutyric acid, GABA, receptors, leading to inhibition of neurotransmission due to an increased ow of chloride ions into the cell. Also, spinosads (a spinosyns mixture, produced by the soil actinobacterium Saccharopolyspora spinosa) which act on nicotinic acetylcholine receptors, causing death due to disruption of the nervous system (Kim et al. 2010). In addition, 30 butenyl-spinosyns of higher potency are produced by Saccharopolyspora pogona (Rang et al. 2020). However, S. littoralis develops resistance due to the widespread and prolonged use of pesticides, pressing a continuous need for alternative fermentation products to control the pest.
This study aimed to investigate the insecticidal potential of the secondary metabolites produced by endophytic actinobacteria isolated from six medicinal plant species in the World Heritage Site of Saint Catherine (WHS No. 954), South Sinai, Egypt. Up to our knowledge, no previous studies exist on insecticides derived from the endophytes of this conservative area. This site can represent a source of unique and diverse endophytes from its endemic plants. The insecticidal activities were assessed on the fourth instar larvae of S. littoralis eld strain, compared to the laboratory strain. Out of seventy endophytic strains, we found high potency for metabolites from seven strains, which were inhabiting two species of Artemisia plant, Family Compositae, (common name: wormwood). The toxicological, histopathological, and biochemical effects of the most potent actinobacterial metabolites were evaluated, as an initial effort to introduce a natural control agent for cotton leafworm. The metabolic extract from the endophyte Kitasatospora sp. ES2 has caused substantial histopathological defects and changes in the metabolism of the larvae, compared to the commercially applied pesticide.

Materials And Methods
Endophytic actinobacteria and culture conditions Seventy actinobacteria strains, previously isolated from six wild medicinal plant species (F. Compositae) in South Sinai, Egypt (El-Shatoury et al. 2013), were studied. The strains were provided by Actinobacteria Lab, Botany Department, Faculty of Science, Suez Canal University. Their selection was based on their published enzymatic activity, plant growth promotion and antimicrobial activity, as shown in Supplementary ( Table S1). The generic identity of the strains and their host plants is illustrated in Table   1. The strains were provided as spore suspensions in 20 % v/v glycerol at -15°C, and were refreshed on starch casein agar, as described by (Kieser et al. 2000). Table 1 Illustration of the endophytic actinobacteria strains and their host plants.
The origin of the seven endophytes that showed the highest effects on the laboratory S. littoralis strain in the initial screen are highlighted in grey a indicates the origin of the four endophytes that showed potent effects on the eld S. littoralis strains.

S. littoralis larvae rearing
The laboratory S. littoralis strain was reared under constant laboratory conditions at the Plant Protection Research Institute, Agricultural Research Center, Zagazig, Egypt. It was reared in an incubator at a temperature of 26 ± 2°C and a relative humidity of 65 ± 10 % RH with a 16 L: 8 D photoperiod, according to (El-Defrawi et al., 1964). The details are shown in Supplementary ( Table S2). The eld strain S. littoralis larvae were collected from the local open eld at Sharqia Governorate. They were transferred to the laboratory and reared for two successive generations as described above. The laboratory strain is reared in the laboratory for several successive generations away from any contamination with pesticides; so, it is guaranteed that it is free from any resistance to the action of pesticides. When experiments are conducted on it, the results of any treatment are assured of con dence; unlike the eld strain that is randomly brought from the eld where exposure to different pesticides is present. It is available and it is uncertain whether the eld strain may enhance the resistance genes for the action of the treated substance or not; therefore, conducting experiments on the two strains gives additional dimensions in evaluating the tested toxicants. Actinobacteria fermentation and metabolites extraction The actinobacterial strains were cultured (2 µL of spore suspension) in 50 mL sterilized starch casein broth and incubated at 28 ± 2°C for 21 days with continuous shaking at 100 rpm. The mycelia were separated by centrifugation at 5000 rpm. The ltrates were extracted three successive times using equal volumes of ethyl acetate. The solvent layers were combined, concentrated, and evaporated to dryness using a rotary evaporator (HS-2005S-N, HAHN SHIN Scienti c Co., Korea) at 40°C. The dried extracts were redissolved in ethyl acetate to prepare a stock concentration of 100 mg/mL and stored at 4°C for bioactivity screening tests.

Toxicity of ES2 metabolite to S. littoralis larvae
The effects of the ES2 metabolite on the growth and development of the S. littoralis 4th instar larvae were evaluated at different concentrations. Toxicity was compared to that of spinetoram (Radiant® 12% SC 100 cm 3 / feddans (4200m 2 ), a 2nd generation of the spinosyns, at LC 50 0.5 mL/ liter (i.e. 0.05% concentration). The commercial Radiant SC 12% (Dow Agros ciences; CAS Number: 187166-40-1) was obtained from the PPRI, Ministry of Agriculture, Egypt. The experimental design included normal group (untreated, receiving distilled water) and negative control group (receiving ethyl acetate). All bioassay assessments were duplicated (each replicate included four larvae) and performed under constant laboratory conditions. A detailed schema for the laboratory and eld S. littoralis treatments is presented in Supplementary ( Figure S1).
Toxicity to 4th instar larvae of the laboratory S. littoralis strain (L-larvae) A series of four concentrations (0.6, 6, 60, 100 mg/mL) of crude metabolites from the 70 actinobacteria strains were prepared. Newly molted 4th instar L-larvae were starved for 3 -4 hours prior to the treatment to clear their alimentary canal and assure quick ingestion of treated leaves. Groups of larvae were transferred to 350 mL sterilized clean glass jars, and all jars were supplied with 7.0 cm lter paper to absorb any surplus moisture. Healthy, untreated leaves of castor, Ricinus communis L., were collected from the experimental eld of the Plant Protection Research Institute. The leaves were washed, cut into equal discs using a cork borer, and impregnated with 50 µL of the corresponding metabolite concentration (i.e. equivalent to 0.03, 0.3, 3 and 5 µg / disc) using a leaf dipping technique. The lethal effects (mortality %) were recorded, daily, and corrected according to Abbott's formula (Abbott 1925). The treated insects were followed up, until the pupation stage.
Toxicity to 4th instar larvae of the Field S. littoralis strain (F-larvae) Metabolic extracts from seven actinobacterial strains, that showed potent activities against the laboratory 4th instar larvae, were further investigated for toxicity to the 4th instar F-larvae, at 100 mg/mL concentration. Lethality was recorded as detailed above.

Phylogenetics of the promising actinobacteria strain
The partial 16S rRNA gene sequence of Kitasatospora ES2 (which produced the metabolite of highest toxicity to the laboratory and eld strains) was analyzed. DNA extraction was performed based on the salting-out method (Kieser et al. 2000), with an additional puri cation step using phenol/chloroform. The 16S rRNA gene of the strain was ampli ed using the universal primer set 27F (5´-AGA GTT TGA TCC TGG CTC AG-3´) and 1492R (5´-GGT TAC CTT GTT ACG ACT T-3´). Ampli cation conditions were according to (Trujillo et al. 2010). Brie y: an initial denaturation step was performed for 9 min at 94°C, followed by 30 cycles of denaturation for 1 min at 95°C, annealing for 1 min at 55°C and extension for 2 min at 72°C. A nal extension step was performed for 10 min at 72°C. Toxic effects Kitasatospora sp. ES2 crude metabolic extract (ES2) The 4 th larval instar of the laboratory S. littoralis strain (L-larvae) were exposed to ES2, the most potent strain, and investigated for morphological defects and histopathological and biochemical changes.

Histopathological examinations of L-larvae
Samples of the treated L-larvae and controls were collected at 72 hours post treatment. They were preserved in 3 mL 10 % formaldehyde (v/v) in sterilized screw-capped tubes, prior to xation, dehydration and embedding in para n wax. Serial sections, at ve microns, were made with a microtome and mounted on clean slides using Mayer's albumin. The sections were stained with Ehrlich's hematoxylineosin (HE) (Ruiz et al. 2004). The histological longitudinal and transverse sections were examined under a light binocular stereomicroscope (NOVEL; NLCD-120, China) using two magni cations (100-X and 400-X).
Biochemical assessments, Larvae-samples preparation Samples (groups of four) of the treated and control L-larvae were placed in clean screw-capped tubes and kept frozen overnight. The frozen samples were homogenized for three min in distilled water (50 mg/1 mL) using a chilled glass Te on tissue homogenizer (ST-2 Mechanic-Preczyina, Poland) surrounded with a crushed ice jacket. Then, they were centrifuged at 8000 rpm for 15 min at 5°C in a refrigerated microcentrifuge (Hettich, Germany). The supernatants, used as enzyme extracts, were stored at -20°C until use in biochemical assays. All biochemical measurements were performed in triplicates. A double beam UV spectrophotometer (Spectronic 1201, Milton Roy Co., USA) was used to measure the absorbance of colored substances. The total protein concentration was determined according to Bradford's method (Bradford 1976).
Acetylcholinesterase (AchE, EC 3.1.1.7) determination. The biochemical activity of acetylcholinesterase, a detoxi cation enzyme, was measured according to (Simpson et al. 1964), using acetylcholine bromide (AchBr) as a substrate. The samples were measured at 515 nm absorbance against a blank (ethanol in phosphate buffer, pH 8.0). The activity was expressed as U/mg protein.
Protease (EC 3.4.21.112) determination. Proteolytic activity was measured as described by (Tatchell et al. 1972), with modi cations, by measuring the increase in free amino acids split from a substrate protein (albumin) during 1 hour of incubation at 30°C. Amino acids were colorimetrically assayed by ninhydrin reagent. The zero adjustment was performed at 570 nm against the reagent blank (100 µL distilled water). The amino acids were expressed as µg D, L-alanine/min/mg protein.

Non targeted metabolomics analysis
Liquid chromatography, combined with quadrupole-time-of-ight high-de nition mass spectrometry (LC-Q-TOF-MS), was used to investigate the chemical constituents of the metabolites from Kitasatospora sp.
ES2 strain. This technique is a powerful tool for the characterization of microbial compounds with similar structures, particularly in the analysis of natural products ).

Molecular docking simulation
Molecular docking aimed to illustrate the virtual mechanism of binding of some reported compounds used as insecticides towards acetylcholinesterase (AchE, PDB=4EY5), lactate dehydrogenase (LDH, PDB=1LDG), and protease (SREBPs, PDB= 5GPD) target proteins, that were freely accessible through the protein data bank. Both proteins and ligands were prepared and optimized according to (Na e et al. 2019), and the molecular docking study was carried out using MOE 2015-10 as the computational software. Each complex was analyzed for 2D interaction images were taken by using MOE visualizing tool. 3D interaction images were taken by Chimera (UCSF).

Statistical analysis
All data were formulated as means ± standard error of the mean (SEM). Data was subjected to normality testing using Kolmogorov-Smirnov at 0.05 level. Accordingly, LDH, protease and AchE were parametric and parametric data analysis applied. One-way ANOVA was applied to assess the difference between treatment groups, ANOVA was followed by Duncan's Multiple Range tests (DMRTs) as a posthoc test at 0.05 level.

Results
Toxicity of actinobacteria metabolites to S. littoralis larvae Toxicity of seventy actinobacteria metabolic extracts to the laboratory S. littoralis strain (L-larvae) was evaluated at four concentrations. We observed signi cant potent and immediate death of the treated larvae, using 100 mg/ml extracts from seven metabolic extracts, at 48 hours post-treatment (Table 2).
However, using lower metabolite concentrations (0.6, 6, and 60 mg/mL), various levels of toxicity were observed. In the latter case, biological effects such as deformed individuals (Fig. 1b1) and abnormalities in pupae (Fig. 1b2, c) were observed and described as latent toxic effects, compared to the controls (Fig. 1a). Detailed toxic effects of the seventy metabolites on the L-larvae, at 100 mg/ml concentration, are shown in Supplementary (Table S4). Our screening has shown highest activities of 7 actinobacteria on the fourth instar larvae of S. littoralis laboratory strain (L-larvae); with ve toxic to the eld S. littoralis strain (F-larvae). Immediate lethal effects were recorded for metabolites produced by actinobacteria of the genera Kitasatospora and Nocardioides. While, latent effects (death after > 72 hrs.) were observed for Streptomyces and Kitasatospora strains. The latent effects varied between deaths of the larvae after ve days of the treatment, to death of the resulted pupae (Table 2).
Identi cation of Kitasatospora sp. ES2 strain Partial 16S rRNA gene sequence of Kitasatospora ES2, which produces the most potent metabolite against the laboratory and eld larvae strains, has shown similarity to Streptomyces spp. at ≤ 96.69 % ( Table 3). The neighbor-joining phylogenetic tree illustrating the a liation of Kitasatospora ES2 and representatives from its phylogenetically closest taxa is shown in Supplementary ( Figure S2). Kitasatospora ES2 has caused 50 % death in the L-larvae after 48 hours post-treatment. Also, some toxic effects have appeared in the treated larvae after 72 hours; those were considered latent effects. It is noteworthy that, some larvae failed to pupate from the larval stage to the pupal stage, leading to the production of larval-pupal or pupal-larval intermediates, as shown in (Fig. 1). The cuticle layers were distorted in the larvae treated with both ES2 (Fig. 2d) and Radiant (Fig. 2g), compared to the untreated control (Fig. 2a, b). Larvae that were treated with the ES2 have shown obviously higher degeneration and ssures in the muscles (Fig. 2c, d, e, f), than those subjected to the Radiant treatment (Fig. 2g). ES2 has, also, caused deformations (Fig. 2c, d, e, f), vacuolization of the hypodermal layer (Fig. 2d, e, f), and separation of the hypodermis from the cuticle layer (Fig. 2c, e). However, decomposition in the hypodermal cells of occurred only in the larvae treated with Radiant SC 12% (Fig. 2g). ES2 has caused obvious gastrointestinal damages in the larval midgut tissue of S. littoralis, as compared to the control (Fig. 3a, b). These damages included: vacuolization (Fig. 3c, d, e), degeneration and necrosis of the epithelial cells and destruction of the cells and their boundaries (Fig. 3c, d, e, f). The microscopically observed changes were more severe than those observed in the Radiant-treated larvae (Fig. 3g).
The application of ES2 and Radiant at 100 and 5 mg/mL, respectively, has caused highly signi cant reductions in lactate dehydrogenase (LDH) and protease activities (p ≤ 0.01), while acetylcholinesterase (AchE) activity was not signi cantly changed, compared to the control. LDH activity in the untreated (control) larvae was 28.40±0.888 U/mg protein, and it has been reduced by 47% and 16%, using ES2 and Radiant treatments, respectively. While, the reductions in protease activity were almost similar after both treatment (37.00±2.081 and 30.33±0.881 µg D, L-alanine/min/mg protein for ES2 and Radiant, respectively), as shown in Table 4. The biochemical measurements, in our study, served to provide a preliminary understanding of the basis of ES2 toxicity.  Table  S5). These compounds were chosen based on their surveyed activity as pesticides, and their structures are structurally highlighted in Fig. (4), showing the aromatic, polar, and nonpolar moieties. They were docked and visualized inside the acetylcholinesterase, lactate dehydrogenase, and protease proteins to highlight their virtual mechanism of binding in terms of binding energy and interactions. The compounds were properly docked inside the active sites of proteins with good binding energies ranging from -11.13 to -14.67 Kcal/mol. As seen in Fig. (5), compound cyromazine formed good binding interactions through its active pharmacophoric groups with the key amino acids of the tested three proteins. Additionally, the interactions for the other compounds inside the tested targets were supported (Supplementary, Figure   S3). The molecular docking studies, agree with the experimental results enhancing the idea of their insecticide activity through acetylcholinesterase, lactate dehydrogenase, and protease inhibition.

Discussion
The lepidopterous pest S. littoralis has developed resistance due to the intensive pesticide use (Rehan and  Notably, the results indicated signi cantly strong effect of ES2 in LDH inhibition; that effect was about 3times higher than the Radiant. LDH is an important glycolytic enzyme involved in carbohydrate metabolism and is a marker for reduced metabolism in the insect (Diamantino et al., 2001). This result compares well with other studies on the LDH inhibition by natural microbial products. For example, Nathan et al. (2006) demonstrated 20 % suppression of the LDH activity in the fourth larval instar Cnaphalocrocis medinalis (the rice leaf folder), following exposure to 3 µg/mL of Bacillus thuringiensis bioinsecticide. A eld experiment showed 55.7% LDH inhibition, when 5th instar larvae of S. littoralis were exposed to Radiant, at 40 liters/feddan (Fahmy and Dahi 2009), causing low nutritional e ciency and less subsequent vital activities in the larvae. The results, also, indicated highly signi cant inhibitory effect of both ES2 and Radiant on protease (64 -70% inhibition). Proteases are involved in the glycolysis process and the neural signals transmission in insects (Cheng et al. 1999). Therefore, we hypothesize that ES2 may act on the insect metabolism and causes death by disrupting the glycolysis process due to inhibition of LDH and protease enzymes. Radiant causes death due to disruption of the nervous system (Krämer and Schirmer 2007), While the main action of ES2 is not expected to target the nervous system, because it has caused nonsigni cant changes in acetylcholinesterase activity. Acetylcholinesterase is the primary cholinesterase in the pest body that plays a major role the transmission of the nerve signals (Casida and Durkin 2013). ES2 may, thus, have a different mode of action than that of spinetoram.
It is reasonable that, the toxicity of ES2, morphological defects and histopathological deformities in the insect is attributed to its unique mixture of several compounds with special regard to cyromazine, as a key constituent. All the previous studies on cyromazine have used a chemically synthesized formula of the compound. Cyromazine has been widely used for pest control in agricultural crops and public health management systems (Subramanian and Shankarganesh 2016 This is the rst illustration of insecticidal activity of the Artemisia spp. microbiome, and the rst report on a natural cyromazine synthesized by actinobacteria. In our next publication, we anticipate reporting the complete chemical properties of the "active ingredients" produced by Kitasatospora sp. ES2 and their expected targets, based on bioinformatic and molecular docking studies.  Figure 1 Lethality and developmental defects in 4th instar laboratory S. littoralis larvae after exposure to Kitasatospora ES2 metabolites a) Control S. littoralis (from left: larva, prepupa, young pupa and mature pupa). b) Effects of the Kitasatospora ES2 crude metabolite; various direct toxicity on larvae (1) and latent effects on pupae (2). c) Morphological effects of the commercial product Radiant SC 12% on larvae.   Highlighted pharmacophoric regions for the compounds with reported insecticide activities.