To determine the physiological response of winter wheat caused by O. melanopus, mixed-gender adults were collected from an insecticide-free environment. The collection was carried out in early April 2021 in the Zselicszentpál area (Hungary, Somogy county, GPS coordinates: 46°30′84′′N, 17°82′08′′E). Adults were collected using an entomological sweeping net. The temperature range during the rearing of insects in the climate chamber was 20 ± 1°C. Relative humidity was maintained at 60 ± 5%, and photoperiodic setting was 18L:6D, corresponding to the insect vitality optimum (Mazurkiewicz et al., 2019).

In parallel, healthy, damage-free winter wheat seeds were sown in 19-cm diameter plastic pots (80 seeds/pot), which were placed in a Pol-Eco Apartura KK 1450 climate chamber (POL-EKO-APARATURA sp.j. ul. Kokoszycka172C 44–300 Wodzisław Śląski, Poland) at 20°C, 120 μM m^–2 s^–1 light intensity for 16 h as daylight conditions and 16°C; 0 μM m^–2 s^–1 light intensity for 8 h as night conditions. When seedlings reached the 2–3 leaved stage, 5 pots continued to grow without changing their conditions, and 5 pots were treated with 20 images of the model beetles (O. melanopus) per pot under the isolator covered with well-ventilated textiles. Subsequently, after leaf damage, stress analytical evaluation and non-invasive imaging were taken on the sixth day of insect application to determine the extent of oxidative stress and its visual display. First, the non-invasive type of the measurements were conducted starting with ultra-weak photon emission (UPE). After the completion of the UPE measurement, the soil plant analysis development (SPAD) measurement was done. The sequential order of the measurement was chosen to avoid any potential stress on the plants including even the light pressure that the usage of SPAD equipment poses on the leaf blades of the seedlings. Subsequently, the seedlings were sampled for the fresh/dry weight (DW) determination, ferric reducing antioxidant power (FRAP), and malondialdehyde (MDA) measurements when the whole above-ground part of the plants was used. The sampling method was the following: the above-ground part of the seedlings was cut into ∼0.5-cm pieces. After that the pieces were mixed thoroughly to create an average sample for the fresh DW and for antioxidant capacity and lipid peroxidation measurement. The measurements were repeated three times per pot and the chemical analyses were measured in three parallels per each pot.

Similar to other leaf analysis-based experiments (Sorgini et al., 2019; Zhang et al., 2019; Wang et al., 2020), damaged leaves were scanned to objectively determine leaf surface destruction. Five leaves were scanned and the number of pixels per each image were determined and the pixel number of the white background was subtracted, resulting in the pixel number of the whole leaf. Then, pixel points of differently colored areas were designated using the GIMP 2.10.8 software, and thus the pixels of damaged leaves were determined in the percentage of reduction of the photosynthesizing surface (Figure 1).

A total of 1 g of leaves were measured by an Ohaus Discovery DV215CDM (Ohaus Corporation 1.800.672.7722 7 Campus Drive, Suite 310 Parsippany, NJ 07054, United States) analytical scale weighing up to 5 decimal places. Subsequently, leaves were dried in a drying cabinet (Memmert SLE 600, Memmert GmbH + Co. KG, Aeussere Rittersbacher Strasse 38, D-91126 Schwabach, Germany) at 60°C for 24 h. Then the weight of the dried samples was recorded. Fresh/DW ratio was expressed as % and it was calculated by [(wx − w0)/wx] × 100, where the wx is the DW and w0 is the fresh weight.

Malondialdehyde content was determined by the thiobarbituric acid (TBA) reaction with some modifications to the original method of Heath and Packer (1968). Samples of 0.5 g were homogenized with 2 mL of 0.1% trichloroacetic acid (TCA) in cold mortars from which 1.8 mL was transferred to 2 mL microtubes with automatic pipettes. To this solution, 40 μL of 20% butylated hydroxytoluene (BHT) in absolute ethanol was added to stop further lipid oxidation. The solutions were vortexed for 15 s and centrifuged at 18,600 × g for 10 min at 4°C. From the clear supernatant, 0.25 mL was added to 1 mL of 20% TCA containing 0.5% TBA, gently mixed, and briefly centrifuged for 5 s. The solutions were incubated in a water bath (Julabo ED-5M) for 30 min at 96°C. The reactions were stopped by cooling the solutions immediately on ice followed by centrifugation at 10,000 rpm for 5 min. Absorbance at 532 and 600 nm was recorded using a SmartSpec™ Plus spectrophotometer, and MDA concentration was calculated by subtracting the non-specific absorption at 600 nm from the absorption at 532 nm using an absorbance coefficient of extinction, 156 mM^–1 cm^–1. The results were expressed as nmol g^–1 in DW.

Total antioxidant activity is measured by the modified assay of FRAP of Benzie and Strain (1999). The constituents of the FRAP reagent were the following: acetate buffer (300 mM pH 3.6), 2,4,6-tripyridyl-s-triazine (TPTZ) 10 mM in 40 mM HCl, and FeCl₃6H₂O (20 mM). The working FRAP reagent was prepared by mixing acetate buffer, TPTZ, and FeCl₃6H₂O in the ratio of 10:1:1 at the time of use. The standard solution was 10 mM ascorbic acid (AA) prepared freshly at the time of measurement. To 0.1 mL of the supernatant was added 2.9 mL of FRAP reagent in 5-mL screw cap centrifuge tubes, vortexed in a 37°C water bath (Julabo ED-5M, JULABO GmbH 77960 Seelbach/Germany) for 4 min, and the absorbance was measured at 593 nm against a blank with a BIORAD SmartSpec™ Plus spectrophotometer (Bio-Rad Ltd., 1000 Alfred Nobel Drive Hercules, CA 94547, United States). The FRAP values of the samples were determined were of AA equivalent (mmol AA equivalent g^–1 DW) based on the AA calibration curve, as the averages of three independent measurements (Benzie and Strain, 1999; Szõllõsi and Szõllõsi Varga, 2002).

The SPAD index, a non-invasive measurement for relative chlorophyll content, was measured by reading 10 individual points on 10 seedlings of each treatment with SPAD (Soil Plant Analysis Development–SPAD-502; Konica Minolta Sensing Inc., Japan) equipment. In a wheat leaf with ∼10–15 cm long, measurements were taken along the full length of each leaves approximately every 1–1.5 cm.

For measuring UPE, the NightShade LB 985 In Vivo Plant Imaging System (Berthold Technologies GmbH & Co. KG, 75323 Bad Wildbad, Germany) equipped with a sensitive, thermoelectrically cooled slow-scan NighOwlcam CCD device has been employed. The instrument was controlled by the IndiGo™ 2.0.5.0. software. Intensities of light were converted into counts per second (cps) by using the controlling software. The exposure time was kept at 60 s using a pixel binning of 2 × 2. During the duration of taking the images both the “background correction” and the “cosmic suppression” options were enabled to ensure the elimination of high-intensity pixels potentially caused by cosmic radiation. One pot for each treatment of the seedlings to be imaged was placed into the dark imaging box. In the first part of the measurement, DF signal was measured immediately after placing the pots in the dark chamber for 10 min. Thereafter, the samples were continued to be kept in dark to provide sufficient time for dark adaptation, and from the 30th minute, for the pots spent in the dark, bioluminescence data was also acquired for 10 min.

One-way ANOVA was assumed to applied to determine the differences between samples. To do so, two conditions must be fulfilled. Firstly, the samples must be normally distributed, secondly the samples must have homoscedasticity. Shapiro–Wilks test was used to determine the distribution. The null hypothesis was that the sample is normally distributed, which was accepted when the p-value exceeded 0.05. Depending on the distribution, i.e., the sample is normally or not-normally distributed, two methods were applied to calculate the homogeneity of variances: Bartlett test for normal distributions and Flinger–Killeen non-parametric test for non-normal distributions. When the distribution of the two samples was different, the less robust Flinger–Killeen test was used to determine homoscedasticity. Since the results of testing conditions supported the application of one-way ANOVA only in the case of fresh/DW data, Wilcoxon-test was used to prove the differences between samples for the rest of the measured parameters. When both samples (control + infested) have the same distribution, the type of homoscedasticity test is unequivocal. In the case of MDA, the Flinger–Killeen test was chosen with higher reliability for determining the homogeneity of variance.

Table 1 presents the results of the statistical analysis of the measured physiological parameters, such as relative chlorophyll content, antioxidant capacity, and lipid oxidation; and biophoton emission-related parameters: DF and UPE.

The results of the Shaphiro–Wilks test indicated that UPE and MDA values showed normal distribution but the test of the parameters (SPAD, FRAP, and DF) did not. Furthermore since the variance of the samples was so pronounced, the ANOVA was only applicable for the statistical analysis of the fresh/DW ratio values. According to the results for one-way ANOVA and Wilcoxon tests, all parameters showed significant differences (Table 1; more detailed results in Supplementary Material 1).

The damage of O. melanopus showed serious tissue damage during the 6 days of infestation. Mean of leaf surface destruction caused by O. melanopus was 12.280 ± 1.323%, the average of five investigated leaves (Figure 1). The results of the relative chlorophyll content (SPAD index) and fresh/DW ratio are presented in Figure 2. The infestation of O. melanopus manifested as a significant decrease in the SPAD index from 28.2 to 20.6. The Shapiro–Wilks normality test showed that the SPAD data did not show normal distribution (p > 0.05), and thus was analyzed with the Wilcoxon test (Table 1) and revealed a significant (p = 2.2 × 10^–16) 27% decrease in total (Figure 2). The normality test resulted in normal distribution for fresh/dry mass ratio, and consequently the one-way ANOVA was used and proved a significant (p = 5.84 × 10^–16) increase of O. melanopus infested wheat: 24.9%, from 9.064 to 12.071.

The FRAP values reflect that part of the overall antioxidant state of plant tissues that is possible to determine with the method of ferric reducing antioxidant capacity. The Shapiro–Wilks normality test resulted that the antioxidant capacity data did not show normal distribution (p > 0.05), and the Wilcoxon test (Table 1) was used which showed that the increase was significant (p = 3.383 × 10^–6) in the FRAP values of the O. melanopus-infested wheat samples (11.43 μg AA equivalent g^–1 FW) compared with the healthy leaves (6.92 μg AA equivalent g^–1 FW; Figure 3).

Lipid oxidation is a stress indicator of membrane structure- and function-related processes, such as the damage caused by O. melanopus. The results of MDA level measurement showed the decrease in MDA content in the tissues of O. melanopus-infested wheat. The Shapiro–Wilks normality test resulted that the lipid oxidation data did not show normal distribution (p > 0.05), thus the data were analyzed with the Wilcoxon test (Table 1), which proved that the decrease in MDA content was significant (p = 1.478 × 10^–7). The control MDA level was 16.91 nM g^–1 FW and the O. melanopus-infested level dropped to 14.87 nM g^–1 FW, i.e., 13%, which is a statistically significant (p ≤ 0.05) decrease induced by the biotic stressor: O. melanopus.

The results of DF decay of healthy and O. melanopus infested leaves during the first 5 min of the measurement are presented in Figure 4. The images reveal a distinct difference between the healthy and stressed samples. The DF signal of the healthy plants was highly intensive during the first minute of the measurement in both sample types, which was depicted as intensive orange and red pseudocolors in the images. A decreasing tendency of signals was obtained from the second minute of the measurement in heathy plants; however, the photon emission signals arising from the O. melanopus infested leaves were lower as it was indicated by the pseudocolors. This tendency continued during the fourth minute and the photon emission reached low intensity by the end of the fifth minute.

Figure 5 shows the results of UPE both in the original size and due to UPE in an enlarged version of typical areas with considerable UPE signals. The images of UPE signals showed differences between the healthy and O. melanopus-infested samples, and according to the pseudocolor intensity, it indicated an increase of stressed plants in all the first 5 min of the measurement compared to that of the healthy samples.

Besides the acquisition of images, the fluorescence data were also analyzed to derive objective parameters for a precise evaluation of the putative trends and tendencies depicted by the images. DF measurements were followed consecutively by UWLE data acquisition after a completed dark adaptation period of 30 min (Figure 6). This measurement setup enabled a distinction between photosynthetic- and lipid-oxidation-related processes. The Shapiro–Wilks normality test resulted that the both DF and UPE data did not show normal distribution (p > 0.05), and thus was analyzed with Wilcoxon test.

The presented data are the summary of all photon counts during 10 min of data acquisition for both DF and UPE (Figure 6). The results show an opposite trend of the changes in DF and UPE indicated before in the presented images (Figures 4, 5), but both types of measurements resulted in a significant difference in the sum of overall photon count per second values. However, in the case of DF, the infestation of O. melanopus resulted in a 250% significant decrease (p = 0.001152) compared to the control plants from 1.72 × 10⁶ to 6.88 × 10⁵. On the contrary, the direction of the changes of UPE was opposite to that of the DF values and resulted in a significant (p = 2.2 × 10^–17) increase (Table 1). UPE was more than 300% higher than in the case of the control from 7499.688 to 2197.424 (Figure 6).

To identify the temporal dynamics of O. melanopus-induced changes in photon emission, the time-course analysis of the detected photon signals was also conducted besides the evaluation of the sum of the overall cps values. Therefore, both DF and the consequent UPE were measured for 10 min.

The results show a distinctive trend in both the phenomena. DF has a highly sharp decreasing tendency in first few minutes of the measurement, and from the fifth minute the data decays toward zero, but does not reach it. After dark adaptation was completed, the UPE values reached a quasi-steady state that significantly differed according to the sample types for the whole duration of our observations (Figure 7).