Fitness consequences of oviposition choice by an herbivorous insect on a host plant colonized by an endophytic entomopathogenic fungus

Several species of entomopathogenic fungi (EPF), often considered as bioinsecticides, are able to colonize and establish a symbiotic relationship with plants as endophytes. Recent studies have demonstrated that insects feeding on endophytically colonized plants could have reduced survival. These newly emerging, but not yet fully understood, ecological roles suggest the possibility that EPF may affect preferences and performance of herbivorous insects. However, such plant-mediated effects and underlying mechanisms are largely unexplored. Here, we examined that the endophytic EPF, Beauveria bassiana, could affect oviposition selection and offspring fitness of Asian corn borer, Ostrinia furnacalis on maize, Zea mays. We observed that O. furnacalis females preferred to lay eggs on B. bassiana-inoculated maize plants. This was attributed to the changes in plant volatile profiles upon endophytic colonization by B. bassiana. Of these plant volatiles, we observed increased amounts of insect-preferred compounds, 2-ethyl-1-hexanol and 3-hexen-1-ol, and decreased amounts of non-preferred compounds β-caryophyllene, naphthalene and α-pinene. This finding suggests that B. bassiana-induced plant volatiles could modulate the interactions between plants and insects. However, fewer O. furnacalis larvae, pupae, and adults survived on the B. bassiana-colonized maize plants and this was correlated with lower plant nitrogen content in these plants. These results indicated that oviposition selection of O. furnacalis did not reflect the maximization of offspring fitness following maize inoculation with B. bassiana. We suggest that EPF-inoculated maize causes a detrimental attraction for O. furnacalis, which should be considered for potential application of “trap plants” when incorporating endophytic EPF within integrated pest management programs.


Introduction
Plants interact with multiple organisms, including herbivorous insects (Shikano et al. 2017) and microbes associated with plants or insects (Behie et al. 2012;Pineda et al. 2013) in natural and managed ecosystems. The study of microbe-plant-insect interactions focuses not only on the mediation of plant-insect interaction by plant-associated pathogens and endophytes, but also includes insect-associated microbes (Gross 2019;Noman et al. 2020). These microorganisms can affect plant abundance, nutritional quality and plant defense responses and are thus important for structuring plant-insect interactions and influencing insect behavior (Schausberger et al. 2012;Biere and Bennett 2013;Görg et al. 2021). This highlights the importance of considering how microbes can influence the interactions between plants and insects when evaluating the ecological Communicated by Paul Becher. and evolutionary consequences of this tripartite interaction (Biere and Tack 2013). The effects also extend to the application of microbes in controlling insect pests and plant pathogens in natural and agricultural ecosystems.
Entomopathogenic fungi (EPF), often solely considered as insect pathogens, have been well studied for over a hundred years as effective biological control agents (Vega et al. 2009). Recent studies demonstrated that some EPF have alternate lifestyles as endophytes (Hu and Bidochka 2021). These fungi are able to colonize plant tissues under natural settings and artificial inoculation (Vega 2018;González-Mas et al. 2019). The newly emerging, but not yet fully understood, discovery provides opportunities for elucidating endophytic EPF-mediated effects on plant-insect interactions and for potentially improving EPF efficacy against insect pests and plant pathogens in integrated pest management systems (Jaber and Ownley 2018).
EPF could mediate the interactions of insects with their host plants in several ways, such as mediating herbivory (Cotes et al. 2020;Russo et al. 2020) or oviposition behavior (Jaber and Araj 2018). For example, there was a reduction in aphids on EPF-inoculated maize plants (Mahmood et al. 2019), and an increase in parasitoid activity on prey was also observed (González-Mas et al. 2019). Oviposition site selection is also critical for both offspring fitness and inclusive fitness (Gripenberg et al. 2010). Endophytic EPF have the ability to promote plant growth and to provide nutrients, and induce secondary metabolites, such as benzoxazinoids (Sui et al. 2020;Rasool et al. 2021). This could consequently shape oviposition behavior of insects by modulating sensory cues, and could affect offspring performance by altering food quality. This stresses the importance of the effects of endophytic EPF on insect oviposition preference and offspring performance.
When selecting the "most suitable" plants for oviposition, female insects undergo complex sensory integration that includes olfactory, visual, and haptic cues during the decision process. Insects firstly navigate toward host plants through olfactory cues, and then identify the host by visual and haptic cues, and finally make an oviposition decision (Renou and Anton 2020;Riffell 2020). For the first step, plant volatiles are important oviposition cues (Gadenne et al. 2016;Webster and Cardé 2017). However, plant volatile profiles are influenced by a variety of biotic and abiotic factors (Islam et al. 2017;Turlings and Erb 2018), and some studies have shown that their production could be modulated after plants are exposed to microbes, such as plant pathogenic fungi, endophytic fungi, and EPF (Sharifi et al. 2018). These plant volatiles, when altered by the presence of plant pathogenic and endophytic fungi, could modify feeding behavior of insects (Rostás et al. 2015) and mites (Schausberger et al. 2012). Whether and how endophytic EPF influence oviposition behavior of insects by altering plant volatile profiles remains poorly understood (Gross 2016;González-Mas et al. 2021).
Generally, offspring fitness benefits from oviposition selection of female insects (Gripenberg et al. 2010;Kohandani et al. 2017). However, the growth and performance of these offspring are often mediated by environmental factors (Quintero and Bowers 2018;Duan et al. 2021). Several studies have demonstrated that insect fitness can be affected after the eggs are laid on the endophytically EPF-colonized host plants (Fernandez-Conradi et al. 2018;Li et al. 2021). This is mainly attributed to a change in food quality or a plant defense response (Cory and Hoover 2006;Saikkonen et al. 2013). EPF, introduced as endophytes into plants, have been shown to alter development and abundance of insects, such as aphids (Mahmood et al. 2019;Qin et al. 2021;Rasool et al. 2021). However, there is a paucity of information on whether and how EPF affect insect offspring performance following oviposition on host plants inoculated by EPF.
Here, using maize, Zea mays, as an important global crop, the Asian corn borer Ostrinia furnacalis, as a pest insect, and Beauveria bassiana as the EPF, we tested the hypothesis that endophytic EPF affects oviposition preference of herbivorous insects by altering plant volatile compound profiles and offspring survival by changing the quality of host plants.

Study organisms
Maize is one of the main crops for human consumption and animal fodder in the world (Fig. S1a), and accounts for more than one-third of China's cereal production (FAO 2016). We chose sweet maize single hybrid, Kennian 1, in this study because it is a major cultivar in China's commercial production, and our previous study has shown that this cultivar could be colonized by B. bassiana (Sui et al. 2020).
B. bassiana [BbHOSD1 (A3)] was isolated from a dead grub (Holotrichia oblita) at the Institute of Plant Protection, Jilin Academy of Agricultural Sciences in 2010 (Fig. S1b). The strain was deposited in the China General Microbiological Culture Collection Center (CGMCCC No. 19373). The fungus was cultured and grown on potato dextrose agar (PDA, Hopebio Spectrum Instruments Co., Ltd., Shanghai, China) for 15-20 days at 26 ± 0.7 °C in the dark, and the conidia were harvested by scraping with a sterile spatula, and then kept the dark storage (about two-three weeks) at 4 °C before use.
Asian corn borer (O. furnacalis, Fig. S1c) is one of the most serious insect pests of maize in China and causes ca. 30% yield losses . The eggs of O. furnacalis were obtained from maize stands in the field, and an O. furnacalis colony was established in the laboratory (air temperature 26.4 ± 1 °C, 70-75% relative humidity, L16:D8) using artificial diet (mixture of soy flour, raw wheat germ, yeast, agar, sorbic acid, ascorbic acid, and multi-vitamins). Adults were used in the oviposition trials, and larvae were utilized in survivability studies.

Experimental design
The B. bassiana inoculation experiment was conducted from mid-May to mid-June, 2017. Maize plants were treated by one of two treatments: (1) maize inoculated with sterilized water containing 0.05% Tween-80 (Dingguo, Beijing, China) solution (control); (2) maize inoculated with a B. bassiana conidial suspension containing 0.05% Tween-80 solution (inoculated). To establish and improve colonization of B. bassiana as an endophyte in maize plants, we used both two inoculation methods for all maize plants in the inoculated treatments; seed immersion and a soil drench inoculation, as conducted in a previous study (Sui et al. 2020). The two methods could enhance the persistence of EPF in soil throughout the experiments (Sánchez-Rodríguez et al. 2018) and lead to systemic colonization of the plants (Wagner and Lewis 2000). Maize seeds were surface-sterilized (dipped in 70% ethanol for 5 min and then immersed in 2% NaOCl for 3 min), and then half of the sterilized seeds were immersed in a sterile 0.05% Tween-80 solution for the control treatments, and the other half of the sterilized maize seeds were immersed in a B. bassiana conidial suspension (1 × 10 8 conidia/mL containing 0.05% Tween-80) for the inoculated treatments. These seeds were immersed for 12 h, and then were sown 6 cm below the surface of 23 g autoclaved peat soil (Humin substrate, Fenghong Co., Jilin, China; 121 °C for 2 h, 0.1 MPa) in a plastic pot (35 cm in diameter and 45 cm in height). After sowing, 200 mL a B. bassiana conidial suspension (1 × 10 8 conidia/mL containing 0.05% Tween-80) was applied four times with 50 mL each time for all maize plants in the inoculated treatments (each 50 mL was used at days 7, 12, 17, and 22, respectively). At the same time, a 50 mL sterile 0.05% Tween-80 solution was applied for plants in the control treatments. There were 20 pots with three maize seeds per pot in each treatment. The plants were grown in the greenhouse (air temperature 27.5 ± 0.8 °C, and relative humidity 63.5 ± 14.2%) with 14L:10D light cycle.

Endophyte assessment
Maize leaf colonization by B. bassiana from 60 maize plants per treatment was assessed on day 25 after sowing. A portion of the fourth entirely fully developed leaflet from each plant was removed, and divided into nine 1 cm 2 sections (Tefera and Vidal 2009). These leaf sections were surface-sterilized with 1% sodium hypochlorite for 3 min, followed by 2 min in 100% ethanol, rinsed in sterile water three times, and then placed on sterile tissue paper in a laminar flow cabinet (Sui et al. 2020). The efficacy of the surface sterilization was verified by plating 50 μL of the last sterile water rinse onto PDA and incubated for ca. 20 days at 26 °C in the dark. Microbial contamination was not detected in the last sterile water rinse. These surface-sterilized leaf sections were placed onto PDA, and incubated for 23-25 days at 26 °C in the dark. Identification of B. bassiana outgrowth from the leaf sections was based on colony and conidial morphology (Fernandes et al. 2006;Fig. S1d), and all 60 plants from each treatment were tested. Colonization rates were calculated as follows: colonization rate (%) = (the number of B. bassiana colonized plants/total number of plants) × 100. In this study, we observed natural B. bassiana endophytism at a rate of 3.3% in control treatments, and colonization rate was 43.3% in B. bassiana-inoculated treatments (Fig. S2). To ensure similar plant growth in each pot, one maize plant per pot in both control and inoculated treatments were utilized for subsequent experiment after two plants per pot were removed. Thus, each treatment had 20 pots with one plant per pot, with 10 pots for trials of oviposition selection, performance of O. furnacalis and maize characteristics, and the other 10 pots for the collection of plant volatile compounds.

Oviposition selection of O. furnacalis for control and inoculated maize
Oviposition preferences of O. furnacalis for host plants were examined by two-choice tests (De Moraes et al. 2001;Rizvi et al. 2016). Mated O. furnacalis females were released into a pyramidal screen cage (120 cm × 60 cm × 120 cm), which contained two plants (one plant from the control and the other from inoculated treatment, that is 20 pots with one plant per pot for the oviposition trial were used) at the 7-8 plant leaf stage when O. furnacalis often lays their eggs in the field. These oviposition cages were placed in the greenhouse with an air temperature of 26.5 ± 0.8 °C and relative humidity of 63.5 ± 14.2%. The position of each pot in each cage was randomly chosen, and the distance between pots was 35 cm. The cages were separated from each other by at least 1.2 m. To ensure mating, one female and one male were put into a jar (12 cm in height and 10 cm in diameter) 24 h before releasing, and mating behavior was recorded using a video camera (SONY HDR-CX405, Japan). A total of 100 mated individuals with ten females per cage were released into the ten cages at 19:00, in consideration of the nocturnal activity of O. furnacalis. After 72 h, the females were removed from these experimental cages. The number of egg masses and eggs laid were counted. Plants and insects were used once, and 10 replicates (cages) were performed simultaneously.

Collection and measurement of maize volatiles
When conducting the oviposition experiments, we simultaneously collected samples of volatile compounds from maize leaves at ambient temperature (26.4 ± 0.6 °C). The plant volatile profiles (10 pots for the control and inoculated treatments, respectively) were sampled using solid phase microextraction (SPME filed sampler 100 μm polydimethylsiloxane; Supelco [Sigma-Aldrich] Bellefonte, PA, USA), and volatile compounds were identified using gas chromatography linked to mass spectroscopy (GC-MC, Agilent 5975; Agilent Technologies, Madrid, Spain). Three entire young leaves (4-6th from the bottom) per maize plant were cut, and were then placed into a Teflon sampling bag that was made of polyperfluoroethylene propylene (70 cm × 50 cm, E-Switch, Du Pont Co, USA), since O. furnacalis preferred to lay eggs on similar leaves (Zhu et al. previous observations). Volatiles from these leaves were first allowed to accumulate within the airspace of the bag for 30 min, and then, the SPME fibers were exposed into the equilibrated bags for 5 h during the night (21:00-2:00) which is the period O. furnacalis lays most eggs. As the control, headspace was sampled from empty bags in parallel to the bags containing leaves. SPME samples were stored in an oven bag in a refrigerator until GC-MS analysis. We analyzed a total of 23 samples (18 plants and 5 ambient controls), while two plant samples were discarded due to damage of the sampling bags. Desorption of volatile compounds of the 23 samples was done directly into the GC injector. A HP-5 column (30 m × 0.25 mm ID, 0.25 μm film thicknesses; J&W Scientific, Folsom, CA, USA) was used with the following parameters: Helium was used as carrier gas at a constant velocity of 1 mL/min, injector was set at 50 °C for 3 min, and the oven temperature increased at 8 °C/min to 260 °C and was held for 8 min. Compounds were tentatively identified by comparing mass spectra and retention times with commercial standards, when available, using the NIST 2008 Mass Spectral Search Program (National Institute of Standards and Technology, Gaithersburg, MD, USA). A commercially available mixture of n-dodecane, n-heptadecane, and eicosane (Sigma-Aldrich) was used to calculate the Kovats retention indices (RI) of the identified volatile compounds. Quantification of the volatile compounds from maize leaves was made relative to known amounts of the internal standards (1-octanol and tridecane). The mean responses of these compounds (mean peak area) were used to determine the emission amount of the identified compounds (Dötterl et al. 2005), and the calculated amount was used for further analysis.

Electroantennogram responses of O. furnacalis to chemical compounds
The electroantennographic (EAG) responses of gravid O. furnacalis females were recorded using the EAG instrument with a data acquisition interface board (Type IDAC-02), a universal single-ended probe (Type PRS-1), and related software (PC-EAG version 2.4) from Syntech (Hilversum, Netherlands). Solutions of the tested chemical stimuli (liquid paraffin as the solvent) at three different concentrations (10 -3 , 10 -2 , and 10 −1 μg/μL), and the antennae of live gravid moths were used for the EAG recordings. Based on significant differences in emission of plant volatile compounds, and the presence or absence of other plant volatile compounds between the control and inoculated treatments, sixteen chemical standards were tested (for a detailed list of chemical compounds see Table S1). However, 1-penten-3one, 3-carene, 1-penten-3-ol, azulene, and 2-ethyl furan were not included since commercial standards were not available. EAG values of O. furnacalis antennae were recorded by using a standard method (Zhu et al. 2016). Solutions were applied (10 μL) to a filter paper strip (5 mm × 60 mm), and the solvent was allowed to evaporate for 30 s before the strip was placed inside a glass Pasteur pipette. Ten microliters of liquid paraffin was used as the control. Stimulations were carried out by applying puffs of pure air (the air was pumped mechanically) for 2 s through a Pasteur pipette containing the filter paper with chemical compounds. The puffs of the stimuli were applied at 30 s intervals in randomized order of each chemical, each chemical stimuli was tested from low to high dosage. Puffs with liquid paraffin were carried out at the beginning and at the end of each tested compound for monitoring the baseline during antennal recordings. Each concentration per chemical compound had 15-20 replicates (i.e., 15-20 antennae of mated females). EAG responses were normalized with respect to the solvent (Sun et al. 2014). Relative EAG values were calculated as the mean response to the sample minus the mean response to the control divided by the mean response to the control.

Oviposition bioassays of O. furnacalis with chemical compounds
To test oviposition preferences of O. furnacalis for a volatile compound, a modified transparent polyethylene cage (15 cm × 8 cm × 8 cm) was used (De Moraes et al. 2001;Huang et al. 2009). The oviposition cage consisted of an oviposition container and a lid (17 cm × 10 cm) with two holes (3 cm in diameter) cut in the lid. Two glass cuvette tubes (3 cm in diameter and 8 cm in height) enclosed the two holes, and the opening of the tubes was directed towards the holes. One hole received filter paper with 10 μL of each tested compound, and the other contained filter paper with 10 μL liquid paraffin as the control. The distance between the two holes was 5 cm. A piece of wax paper was applied to the inner all walls of the container because females would lay eggs on wax paper. Meanwhile, the wax paper that covered the two holes was perforated ten times with a needle to allow volatiles to pass-through into the oviposition cage (Fig. S3). Oviposition bioassays were performed using mated females in the library (air temperature 26.0 ± 0.9 °C, RH 71.3 ± 12.4%, photoperiod 14L:10D). Six chemical compounds, including 3-hexen-1-ol, 2-ethyl-1-hexanol, α-pinene, β-caryophyllene, naphthalene, and caproaldehyde at a concentration of 0.1 ug/ uL were used because of higher EAG values from O. furnacalis in response to them (Table S1). Oviposition bioassays between distilled water and the control (paraffin) were also done to control for the effects of paraffin. One gravid female was put into an oviposition cage, and eggs laid on wax paper were counted after 72 h. Thirty cages were used together for a tested chemical compound and distilled water. New wax paper was used for each test, and these cages were cleaned with 100% ethanol and distilled water before each oviposition bioassay. The oviposition stimulation index (OSI) was calculated using the following formula to determine whether the compounds repelled or attracted females to oviposit (Huang et al. 2009): In this equation, T is the number of eggs on wax paper in the presence of the tested compound, and C is the number on wax paper in the control.

Measurements of O. furnacalis performance
To evaluate offspring fitness of O. furnacalis after eggs were oviposited on the control and inoculated maize plants, we conducted a no-choice rearing experiment that mimicked the situation in which larvae have no possibility of switching to another plants after hatching. Secondinstar O. furnacalis larvae were placed into a container (35 cm × 20 cm × 15 cm in size) for rearing with a fresh leaf and stalk of maize that had been used in the oviposition selection experiment, including plants of the control and inoculated treatments, respectively. The larvae were reared in full-sib groups of forty individuals in one container for one replicate, with ten replicates for each treatment. The larvae were allowed to develop through pupation to adult. Every two days, the maize leaf and stalk material in these containers was replaced with fresh ones, and the numbers of surviving larvae, pupae, and adults were recorded. Surviving ratio of larvae, pupation ratio, and eclosion ratio was calculated for further analysis.

Measurements of maize characteristics
Morphological variables of all 20 maize plants in both the control and inoculated treatments, including plant height, and leaf length and leaf width for the third to fifth leaf in the middle stratum of each plant were measured ). The average leaf length and leaf width of the three leaves per plant was used for further analysis. Total nitrogen and total carbon of maize plants was assessed. Ten plants (including aboveground leaf and stalk) per treatment were collected, and dried in an oven at 80 °C for 48 h. The dried maize plants were then grounded in a Willey mill equipped with a 1 mm mesh screen before chemical analyses. Five samples of 2 mg per plant were assessed for total nitrogen and total carbon using an element analyzer (vario EL cube, ELEMENTAR).

Data analyses
For the oviposition data (number of egg masses, and number of eggs), we used a generalized linear mixed model with a Poisson distribution and a log link function to examine the difference between the control and plants inoculated with B. bassiana. We used a generalized linear mixed model with Gaussian distribution and identity link function to test differences in colonization rate, and the effects of B. bassiana inoculation on the relative emission amounts of each volatile compound, survival ratio of larvae, pupation ratio and eclosion ratio, and the characteristics of maize plants. We used the nlme package for these analyses. Data where there was no relative emission of volatile compounds (the value was zero) was not analyzed due to incomplete values for some compounds. To examine the differences in EAG responses of O. furnacalis among the control (solvent) and three different concentrations for each tested chemical compound, and the relationships between insect performance and chemical properties of maize plants, we used linear mixed models (LMMs) with the lm-function of the vegan package. For post hoc analysis of the above data, we used Tukey's tests with the multcomp package. Principal component analysis (PCA) was used to differentiate plant volatile profiles between the control and inoculated treatments. For the oviposition bioassay data (only the number of eggs), chi-square goodness of fit test was used to determine oviposition preferences of O. furnacalis for distilled water, liquid paraffin, and tested chemical compounds. All data analyses were carried out in R (version 3.6.0 × 64, 2019, The R Foundation for Statistical Computing Platform).

Oviposition selection
Gravid O. furnacalis females preferred to oviposit on maize plants inoculated with B. bassiana (χ 2 = 3.89, df = 1,18, p = 0.049 for egg mass; χ 2 = 103.77, df = 1,18, p < 0.001 for number of eggs). The number of egg masses and the number of eggs laid by O. furnacalis on B. bassiana-inoculated maize were three times as those as that laid on the control maize (Fig. 1).

Oviposition bioassay
The baseline of paraffin (solvent) compared to a blank control (distilled water) in the oviposition cages was tested, and there was no difference in number of eggs laid by gravid O. furnacalis on wax paper of the solvent control (88.45 ± 4.68, mean ± S.E) and blank control (77.8 ± 7.14, mean ± S.E, χ 2 = 0.68224, df = 1,18, p = 0.4088).

Discussions
The interactions among plants, insects, and microbes are extremely complex (Hartley and Gange 2009;Shikano et al. 2017;Noman et al. 2020). Plants that are colonized by microbes may show alterations in volatile compound profiles, which could have positive or negative effects on insect herbivory and insect oviposition behavior (Pineda et al. 2013;Rostás et al. 2015;Contreras-Cornejo et al. 2021). Here, our results indicated that the endophytic EPF, B. bassiana, could affect the interactions of plants with an herbivorous insect by altering oviposition preferences and offspring fitness.
Several studies have shown that plants with associated microbes, such as pathogens and endophytes, can affect insect oviposition behavior (Pineda et al. 2013;Rizvi et al. 2016). Similarly, our study found that B. bassiana alters oviposition selection of gravid O. furnacalis on maize plants after inoculation (Fig. 1). Insects often locate their oviposition site based on sensory integration among visual (color or size), olfactory (smell), and haptic cues (Bruce et al. 2005;Jürgens et al. 2013). Previous studies demonstrated that some herbivorous insects, such as O. furnacalis prefer to oviposit on taller host plants ). However, maize height, leaf length, and leaf width did not change between the control and B. bassiana-inoculated treatments, which indicated that visual cues may play a secondary role when O. furnacalis females search for oviposition sites. Here, we found that colonization by endophytic B. bassiana could induce and/or modify the profiles of volatile compounds from maize leaves, which is consistent with the findings in other plants, such as melon and cotton (González-Mas et al. 2021). These altered volatile profiles may further modulate behavioral responses of some insects. In our study, the results of the EAG responses and oviposition prefer-  Values are means ± S.E. (n = 9). Different small letters indicate significant difference between the control and inoculated treatments using generalized linear mixed models at significance of p < 0.05 # NA indicates that retention index was not identified in this study.
*ND indicates the compound was not detected in this experiment.  (Russo et al. 2020). Our results also indicated that the number of survived O. furnacalis was reduced when they fed on leaves and stems of maize plants inoculated with B. bassiana. Thus, these results show a decline in the performance of herbivorous insects on host plants inoculated with endophytic EPF B. bassiana. The feeding trial suggested that the fungal propagules were not in direct contact with the insects, and that endophytic EPF may alter insect fitness by changing plant quality and plant secondary metabolite profiles (Gange et al. 2019;Rasool et al. 2021). For plant-feeding insects, food quality is a key determinant influencing insect performance, and higher nitrogen content in plant tissues usually enhance insect growth and Table 2 Electroantennographic (EAG) responses of gravid female Ostrinia furnacalis to different chemical compounds at three concentration levels (10 -3 ug/uL, 10 -2 ug/uL, and 10 -1 ug/uL) and the control Values are means ± S.E. Individuals (N indicated the number of insect individuals tested) were examined for each compound at each concentration level using a linear mixed model at significance of p < 0.05 (F value was from the results of linear mixed model). Different small letters indicate significant differences between the control and the three different concentrations of chemical compounds.
*Liquid paraffin was used as control

Compounds
Relative EAG values F value N p value Control* 10 -3 ug/uL 10 -2 ug/uL 10 -1 ug/uL (Z)-3-hexen-  . 3 Oviposition stimulation index (OSI) of Ostrinia furnacalis females in response to different chemical compounds. The OSI ranges from − 100 to 100, and when OSI = 0, this indicates that oviposition preference on the tested compound was equal to that of the control; when OSI < 0, this indicates that the tested chemical compound was a deterrent to oviposition; when OSI > 0, the chemical tested was an attractant for oviposition. Values are means ± SE, and number is sample size. Asterisk indicates significant differences in OSI between the treated (chemical compounds) and the control (Paraffin) using χ 2 tests (p < 0.05). *0.05 < p < 0.01; ***p < 0.001; ns, no significance development (Awmack and Leather 2002;Tack and Dicke 2013). In our study, reduced nitrogen content in B. bassiana-inoculated plants was correlated with a decrease in insect survival (Fig. S5). In addition, Pieterse et al. (2014) suggested that systemic resistance triggered by biological inducers is a main mechanism of how plants defend against pests. A reduction in the number of surviving O. furnacalis larvae, pupae, and adults could also be explained as induced systemic resistance of maize plants inoculated by endophytic B. bassiana (Jaber and Araj 2018). Indeed, the underlying mechanisms that are responsible for the in planta decreased O. furnacalis fitness are unknown, and require further elucidation.
Generally, oviposition selection of female insects maximizes offspring performance based on optimal oviposition theory (OPT) (Gripenberg et al. 2010). However, a "bad" mother selecting plants that are poorer for their offspring fitness is not without precedent (Proffit et al. 2015;Duan et al. 2021). Our results also confirmed this phenomenon of "detrimental attraction" in plants inoculated with B. bassiana, that is, female O. furnacalis prefer laying eggs on maize inoculated with B. bassiana, but the offspring perform worse than insects reared on control plants. The results indicated that EPF-inoculated maize plants could create an ecological trap for O. furnacalis, which is inconsistent with the OPT. A recent study showed a similar example of an ecological trap with the insect Laelia where it preferred to oviposit on invasive plants (Spartina) but their offspring suffered reduced fitness (Sun et al. 2020). For an herbivorous insect with limited larval mobility, oviposition selection of females is crucial because egg-laying sites provide better food sources for their offspring. However, when the host plant, which is preferred by an herbivorous insect, is influenced by changing biotic and abiotic parameters, female insects may make wrong oviposition decisions (Fernandez-Conradi et al. 2018;Garvey et al. 2020). Such ecological traps may not support the OPT theory (Kohandani et al. 2017) since olfactory cues of insects can be modulated by abiotic and biotic factors, including EPF in our study.
Our results indicated that our strain of B. bassiana was able to colonize maize plants when inoculated by the integration of seed immersion and soil drench, and further modified the oviposition selection and offspring fitness of O. furnacalis. The suspension amount of B. bassiana inoculated into the soil in this study differed to the inoculum applied in other studies in which soil was treated with conidial suspensions (Sánchez-Rodríguez et al. 2018). EPF are able to penetrate plant tissue and move throughout the plant from the inoculation site (Wagner and Lewis 2000;Tefera and Vidal 2009). However, the colonization of EPF can be affected by competition with other microbes in the plants or rhizospheric soil (Jaber and Ownley 2018;Canassa et al. 2020). Other factors such as growth conditions (i.e., temperature and relative humidity; Tefera and Vidal 2009), inoculation method, fungal strains, and host specificity can affect EPF colonization of plants (Bamisile et al. 2018;Ambele et al. 2020). Some studies only used one method, i.e., seed immersion or soil drench when performing EPF inoculation (Quesada-Moraga et al. 2014;Donga et al. 2018), and the application method may result in differences in endophytism and resultant physiological effects on host plants (Sánchez-Rodríguez et al. 2018;Ambele et al. 2020). Leaf spraying rarely led to fungal systemic colonization in plants or an effect on plant growth, despite that this method is often used in attempts to inoculate plants (Barta 2018;González-Mas et al. 2021).  Values are means ± S.E. (n = 10). Different small letters above bars indicate significant differences between control plants and inoculated plants (p < 0.05)