Volatiles from Cotton Plants Infested by Agrotis segetum (Lep.: Noctuidae) Attract the Larval Parasitoid Microplitis mediator (Hym.: Braconidae)

Herbivore-induced plant volatiles (HIPVs), chemicals produced by plants infested by herbivorous insects, can act as kairomones that recruit natural enemies of the pest herbivore. Agrotis segetum (Denis and Schiffermüller) is a common, important pest of seedling cotton in Xinjiang Province, China, and the braconid Microplitis mediator (Haliday) is an important mortality factor of this pest’s larvae. In olfactometer tests, which included healthy foliage, infested foliage, or infested roots, M. mediator preferred A. segetum-infested cotton plants to healthy cotton plants. In GC-MS analyses of plant-emitted volatiles, we found that compounds emitted increased 14.9- and 13.3- fold after leaf infestation and root infestation, respectively, compared to healthy control plants. The volatiles were mainly p-xylene, nonanal, tetradecane, decanal, benzaldehyde, β-caryophyllene, and humulene, while linalool was only present in the leaf-infestation treatment. In addition, principal component analysis indicated that all 18 compounds were associated with the infested plants, especially β-caryophyllene, p-xylene, and decanal. Based on the above studies and previous functional evaluations of the volatile compounds, it can be demonstrated that these compounds play a crucial role in modulating the interactions between A. segetum and M. mediator and regulating parasitoid behavior. It may be possible to enhance the biological control of A. segetum by M. mediator through the application of HIPVs.


Introduction
Volatiles emitted from leaves, fruits, or flowers not only provide herbivorous insects with cues useful in foraging for nutritional resources, but also play an important role in herbivore oviposition behavior, host orientation, mate location, and mating behaviors [1][2][3]. For example, the preference of Plutella xylostella for cruciferous plants is due to the presence of isothiocyanates (ITCs), and these compounds also stimulate P. xylostella oviposition [4]. Fragrant volatiles emitted by flowers mediate the mirid bug Apolygus lucorum's preference for flowering host plants [5], and temporal shifts in plant volatiles may regulate the host plant foraging behavior of Ap. lucorum adults [6]. Such flower-emitted volatile organic compounds (VOCs), when applied to fields, act as attractants for several mirids, including Adelphocoris suturalis, Ad. lineolatus, and Ad. fasciaticollis [7].

Analysis of Cotton Volatiles
The cotton volatiles emitted from the three treatments (healthy leaves, infested leaves, and infested roots) differed (Table 1). Compounds often associated with air (e.g., toluene and benzene) or laboratory equipment (e.g., siloxanes or phthalates) were not included in our list of putative plant volatiles [18]. Compared to healthy plants, total emissions increased 14.9-and 13.3-fold after leaf and root infestation by A. segetum larvae, respectively. Of the 18 compounds detected, all but (1) β-caryophyllene, (2) γ-chlorobutyrophenone, (3) humulene, and (4) 5,9-undecadien-2-one, 6,10-dimethyl-, (E)-were emitted in significantly higher amounts from infested plants than from un-infested plants. The compound p-xylene had the greatest concentration, followed by nonanal, decanal, benzaldehyde, β-caryophyllene, and humulene. The concentrations of p-xylene, nonanal, and decanal Behavioral responses of female Microplitis mediator adults to leaf and root volatiles induced by Agrotis segetum larval feeding. Female parasitoids had a choice between: (i) leaves infested by A. segetum larvae versus un-infested cotton plants, (ii) roots infested by A. segetum larvae versus uninfested cotton plants, and (iii) un-infested cotton plants versus clean air. "*" means a significant difference at the p < 0.05 level, while "ns" indicates no significant difference.

Principal Component Analysis
PCA identified which volatile compounds dominated the volatile blends from healthy cotton plants or ones whose leaves or roots were infested by A. segetum larvae. Of the 18 components in the volatile blend, the major volatiles were β-caryophyllene, decanal, and p-xylene (Figure 2a), and all 18 compounds emitted from cotton plants contributed to the infestation process (Figure 2b).

Principal Component Analysis
PCA identified which volatile compounds dominated the volatile blends from healthy cotton plants or ones whose leaves or roots were infested by A. segetum larvae. Of the 18 components in the volatile blend, the major volatiles were β-caryophyllene, decanal, and p-xylene (Figure 2a), and all 18 compounds emitted from cotton plants contributed to the infestation process ( Figure 2b).

Discussion
Volatile compounds play indispensable roles in the interactions among trophic levels in native food webs. HIPVs, as key information chemicals, provide herbivore-specific cues to parasitoids and predators [19]. We found that M. mediator was attracted by A. segetuminduced HIPVs. Li [14] showed that A. segetum females preferred to lay eggs on healthy (versus previously infested) cotton plants, and that the HIPVs induced by conspecific larvae on cotton plants had significant repellent effects on oviposition. Therefore, from an integrated pest management (IPM) perspective, leaf HIPVs in this system should both reduce A. segetum attacks and increase M. mediator parasitism of the pest larvae [14].
HIPVs can both repel herbivores and recruit natural enemies of pests [20]. However, comparative GC-MS analyses of headspace volatiles comparing compounds from healthy versus infested plants showed mostly quantitative, not qualitative effects. Linalool was the only compound that was present only in the VOCs of infested leaves. This compound is significantly repellent at high doses to the foraging and oviposition of several pest herbivores [14]. After infestation of a plant, the amount of p-xylene, nonanal, tetradecane, decanal, benzaldehyde, β-caryophyllene, and humulene all increased significantly. However, field and greenhouse experiments are necessary to confirm parasitoid attraction.
Insects can perceive chemical signals related to feeding, mating, and oviposition through diverse chemoreceptor families, including odorant receptors (ORs), and ionotropic receptors (IRs) [21][22][23]. For example, decanal was reported to be involved in the olfactory recognition process of M. mediator by binding strongly to MmedOBP18, which is mainly involved in the short-distance recognition of chemical information from hosts or host habitats [24]. Our results showed that the concentration of decanal emitted was significantly higher from plants with infested leaves (8.14 ± 0.48 µg/mL) compared to plants with infested roots (3.28 ± 0.66 µg/mL) or healthy (un-infested) plants (0.18 ± 0.05 µg/mL), suggesting that decanal plays an important role in the location of hosts and their habitats by M. mediator. In conclusion, plant volatiles induced by A. segetum can bind to both ORs and IRs, affecting both herbivores and their parasitoids ( Table 2). Currently, control of A. segetum in China relies on the application of chemical insecticides [12]. HIPVs play important roles in pest control in agriculture systems [35,36]. In our present study, female M. mediator wasps significantly preferred plants damaged by A. segetum, especially after foliar infestation. This preference may be related to changes in the release of HIPVs by the host plant. The amount of various volatile compounds in cotton increased after herbivore infestation, especially p-xylene, nonanal, tetradecane, decanal, benzaldehyde, β-caryophyllene, and humulene. Our results emphasize the important role of HIPVs in host selection by M. mediator and provide insights that may help improve the biological control of A. segetum through the combined application of HIPVs and the release of parasitic wasps. Future studies may explore parasitoid efficiency under laboratory and field conditions and investigate the effects of co-infestation of H. armigera and A. segetum on M. mediator in cotton.

Plants
Cotton (CCRI49) seeds were obtained from the Institute of Cotton Research of the Chinese Academy of Agricultural Sciences (CAAS) and sown in a greenhouse at Langfang Experimental Station, CAAS, under the following conditions: 26 ± 1 • C, 60 ± 10% RH, 14:10 h (L:D) photoperiod. Plants used for these tests were at the 3-true leaf growth stage.

Insects
Agrotis segetum larvae were reared continuously in a climate chamber under the same conditions described for plant production at the Langfang Experimental Station of CAAS. Second or third instar A. segetum larvae were used for our experiments.
The colony of M. mediator was established from diapausing cocoons provided by the Plant Protection Institute, Hebei Academy of Agriculture and Forestry Sciences. M. mediator larvae were reared in an incubator at 25 ± 1 • C, 60 ± 10% RH, 14:10 h (L:D) photoperiod at the Langfang Experimental Station, CAAS. All female parasitoids used in olfactometer tests were 3-6 d old, mated, and fed with 10% honey solution after emergence. Wasps had no previous oviposition experience or contact with plants before experiments.

Olfactometer Tests
A Y-tube olfactometer was used to evaluate the behavioral responses of 3-6 d old active, mated M. mediator adults (n = 60 females), when offered choices between the odors of (1) cotton leaves infested by four A. segetum larvae for 12 h, (2) cotton roots infested by four A. segetum larvae for 12 h, (3) un-infested plants, and (4) clean air. To ensure that A. segetum fed only on cotton leaves, we made a net bag (20 × 30 cm) with 120 mesh gauze to cover the above-ground parts of the cotton along with four individuals. In order to make A. segetum larvae feed only on cotton roots, we cut circular rings of blow molding paper (diam: 8 cm) to cover the cotton cotyledon stalks and fixed them with plastic rods, so that they could not climb higher. All parts of the equipment were connected with Teflon tubes, and the direction of air flow was from the atmospheric air intake, through activated charcoal, a distilled water humidification device, a gas flow control meter, a glass odor source vessel, and then into the Y-tube test arena, with similar parameters as those in previous studies [37].
One M. mediator adult was introduced to the initial test chamber after the airflow of both arms had been adjusted to 400 mL/min. Wasps that moved 1/3 of the way down a test arm within 5 min and stayed there for more than 10 s were counted as having made a 'choice', while wasps that did not respond as such were discarded and recorded as making 'no choice'. After testing five parasitic wasps, the two arms of the Y-tube were reversed (with respect to their odor source), and after testing 10 wasps a clean Y-tube (washed with 95% ethanol and soaked and rinsed with distilled water and dried naturally at room temperature) was used.

Collection and Analysis of Cotton Volatiles
For volatile collection and identification of blends associated with our treatments, 3-true leaf cotton plants were separated into three groups: (1) leaves infested by four A. segetum larvae for 12 h; (2) roots infested by four A. segetum larvae for 12 h; (3) healthy, undamaged plants. To create infested plant foliage or roots, four larvae (second or third instar) of A. segetum (starved for 4 h) were placed on the whole plant and allowed to feed on foliage or roots for 12 h. Larvae were then removed, and the plants were immediately processed to collect headspace volatiles, which were then analyzed for their components.
Cotton headspace volatiles were collected from 1300 h to 1700 h every day using a dynamic headspace collection method. Cotton plants at the 3-true leaf stage (n = 3 for each treatment) were placed individually in a custom glass chamber (diam: 20 cm; height: 66 cm; Yuansu Glassware Supply Station, Shenzhen, China) and the soil was covered with aluminum foil before cotton volatile collection. An airflow at 500 mL/min passed over the plant, and volatiles were absorbed by 50 mg of Porapak ® Type Q adsorbent (Altech Assoc, Chicago, IL, USA). Then, cotton volatile samples were extracted into 1.5 mL sample bottles using 400 µL HPLC-grade n-Hexane (Aladdin, Shanghai, China) and, finally, were stored at −20 • C until GC-MS analyses.
GC-MS (GC: Agilent 7890A, equipped with a DB-WAX chromatographic column [30 m × 0.25 mm × 0.25 µm]; MS: Agilent 5975C) was used to analyze cotton volatiles, with similar parameters as those in previous studies [37]. The injector temperature for GC analysis was 230 • C, the oven temperature was kept at 50 • C for 1 min, and then raised by 5 • C/min to 180 • C for 2 min, and then increased by 10 • C/min to 230 • C, and held for 2 min. Helium was the carrier gas, at an average flow rate of 1 mL/min. The ion source temperature was 230 • C. The volatile compounds obtained were first identified by NIST 14 and were then compared with standard compounds to carry out qualitative-quantitative analyses.

Statistical Analysis
Chi-square tests were used to analyze the Y-tube olfactometer data to detect differences between the pairs of treatments. χ 2 and P values were calculated, and non-responsive adults were excluded from the analysis. The amounts of each volatile compound emitted under different treatments were compared using one-way ANOVA, followed by Duncan's new multiple range tests. Principal component analysis (PCA) was performed to analyze the patterns of volatiles from different treatments given its ability to reduce the complexity of the data while identifying the features in the dataset that contribute the most to the treatment effects. Chi-square tests and one-way ANOVAs were conducted using SPSS 25.0, while PCA analysis was performed using R 4.0.2 with a 0.05 level of significance.