Flax seeds (L. usitatissimum, var. Nike) obtained from the Flax and Hemp Collection of the Institute of Natural Fibers and Medicinal Plants in Poland were sterilized in 50% PPM solution (Plant Cell Technology, Washington, DC, United States) for 10 min and rinsed in sterile water three times. Then, they were placed on Murashige and Skoog (MS) medium solidified with 0.9% agar and grown in a phytotron (16 h under the light intensity of 100 μmol m⁻²s⁻¹, 60% relative humidity, and temperature of 21°C, and 8 h in darkness at 16°C and relative humidity of 70%). After 14 days, the seedlings were exposed to a 50 Hz, 500 μT alternating electromagnetic field (EMF) for 30 min under the aforementioned phytotron light conditions. No significant effect of the environment on the experimental results with the applied EMF was observed.

The EMF exposure system used in the experiment consisted of two identical Helmholtz coils positioned vertically at a distance of 216 mm to ensure horizontal orientation of the magnetic flux. The coil inner and outer diameters were 400 mm and 462 mm, respectively, and the width was 40 mm. The Helmholtz coils were made of a 2.12-mm-thick copper wire wound 225 times (Figure 1; Supplementary Figure S2). The output of the autotransformer (ZWE Eltra, Bydgoszcz, Poland) connected to the electric energy supply was an electric source for a sinusoidal 50-Hz alternating current (sinusoidal 50-Hz MF) in the experiment. The current intensity was controlled by the ammeter (Multimeter Fluke 8846A; Fluke, Cleveland, OH, United States). The MF strength in the center of the Helmholtz coils can be calculated using the following formula derived from Biot–Savart’s law: H c a l c u l a t e d = 0.7155 ∙ N ∙ I r = 0.7155 ∙ 225 ∙ I 0.212 = 759 ∙ I , where I is the value of current, r is the radius of coil wires, and N is the number of windings.

Higher harmonic waves in the current were monitored during the experiments (Fluke 43 Power Quality Analyzer, Fluke, Cleveland, OH, United States) and did not exceed 2%. Magnetic induction was verified using the ESM-100 3D H/E Field Meter (Maschek, Germany).

The exposure was carried out on 14-day-old flax seedlings placed in Petri dishes on MS medium in the geometrical center of Helmholtz coils. After the EMF exposure, the plants were grown for 12 h, and then, the seedlings were collected, immediately frozen in liquid nitrogen, and stored at −80°C for further analyses.

Extraction of total RNA from the flax seedling tissue ground in liquid nitrogen was performed using the mirVana™ miRNA Isolation Kit (Thermo Fisher) according to the manufacturer’s protocol. Genomic DNA was removed using DNase I (Thermo Fisher). The Agilent 2100 Bioanalyzer (Agilent RNA 6000 Nano Kit) was used for RNA quality assessment. RNA samples with a RIN value >7.5 were used for RNA-seq library construction. mRNA was isolated from the total RNA by using the oligo (dT) method. After fragmentation, it served for the first cDNA strand synthesis. Synthesis of the second cDNA strand was followed by adapter joining. The cDNA fragments of suitable size were amplified with PCR and sequenced on an Illumina HiSeq 2500 device. In order to filter the low-quality readings, reads with adapters, and reads with unknown bases (N bases >5%), Trimmomatic software was used (Bolger et al., 2014). The clean reads were aligned to the reference genome (Wang et al., 2012) using HISAT2 software (v2.1; https://daehwankimlab.github.io/hisat2/) with the following parameters: “-q—phred64—dpad 0—gbar 99999999—mp 1,1—np 1—n-ceil L, 0,0.15—no-mixed—no-discordant -p 38 -k 10” [https://doi.org/10.1038/nprot.2016.095]. The gene expression level was calculated using featureCounts software in the Subread package (v1.6; http://subread.sourceforge.net/) (Liao et al., 2013). The analysis pipeline is given in Supplementary Figure S1.

Differentially expressed genes (DEGs) between samples were identified using DESeq2 [https://doi.org/10.1186/s13059-014-0550-8]. The parameters used were as follows: “Fold Change > = 2 and Adjusted p-value < = 0.001.” As some transcripts identified against the reference genome had no GO number assigned, we searched the genome database available for black cottonwood (Populus trichocarpa) (Torr. & Gray) using the BLASTx algorithm. The results were filtered with the E-value threshold equal to or less than 1e-40. DEGs were classified based on the GO annotation results and reference genome annotation. GO functional enrichment using the goseq package in R [https://doi.org/10.1371/journal.pone.0246052] was also performed. The false discovery rate (FDR) for each p-value was calculated, and terms with FDR <0.01 were considered significant for our analysis (Love et al., 2014).

Transcript levels were determined using RT-qPCR using the StepOnePlus™ Real-Time OCR System (Applied Biosystems, United States) on the cDNA matrix in three independent repeats. Primers used in the reactions were designed in LightCycler^® Probe Design v2 (Roche, Switzerland) software (Supplementary Table S1). For the reactions, the DyNAmo SYBR Green qPCR Kit (Thermo Fisher Scientific, United States) was used. cDNA was dissolved five times prior to the experiment. The annealing temperature was 57°C. Changes in the transcript levels of the examined genes were presented as relative quantification with regard to the control. Actin gene was used for reference.

Frozen flax seedling tissue (100 mg) was ground in liquid nitrogen and extracted to methanol containing 1% HCl in an ultrasonic bath for 10 min and centrifuged at 12,000 g, 4°C for 10 min (three times to 1 mL). The collected supernatants were combined, dried under nitrogen flow, and re-suspended in 200 μL of methanol. The reaction with 2,2-diphenyl-1-picrylhydrazyl (DPPH) was used to determine the antioxidant capacity of the extracts.

In the DPPH assay, 6 μL of plant extract was added to 200 μL of 0.1 mM DPPH. Following incubation (15 min at room temperature), absorbance at λ = 515 nm was measured. For the control, 200 μL of DPPH solution with 6 μL of methanol was used. Pure methanol was used as the blank sample. The results were expressed as equivalents of trolox antioxidant potential (inhibition of the free-radical reaction expressed as a percentage) calculated according to the following equation: P = 1 − A s A c ∙ 100 % , where P is the antioxidant potential, A_s is the absorbance of the examined sample solution, and A_c is the absorbance of the DPPH radical solution (Brand-Williams et al., 1995).

In this experiment, 2-week-old flax plants grown in sterilized vermiculite with MS medium (35 mL MS per 50 g vermiculite) were exposed to a 50 Hz, 500 μT alternating electromagnetic field (EMF) for 30 min. Immediately after exposure, they were infected with 500 ul of 10-e5 spores of Fusarium oxysporum freshly prepared as follows: F. oxysporum grown on solid PDA medium at 28°C for 2 weeks were rinsed with 2 mL of sterile water, which was then filtered through sterile gauze and moved to an Eppendorf probe. The Fusarium spore count in the suspension was determined under a microscope in the Thoma cell counting chamber. After 10 days of growth, the plants were collected, and the presence of fungal DNA was assayed using the RT-qPCR method with primers specific to fungal murein transglycosylase gene. Pure water was used in the control.

The plants (exposed and not exposed (control) to ELF-EMF and then infected with F. oxysporum) were collected and subjected to clearing with 0.15% TCA in an ethanol: chloroform mixture (4:1; v/v) for 48 h, divided into green parts and roots, and then submitted to staining with safranin and solophenyl flavine 7GFE (0.1% w/v in 0.1 M Tris/HCl, pH 8.5) as follows: the plants were submerged in safranin solution (0.2% w/v safranin in 10% v/v ethanol) for 5 min and then washed three times with water. Subsequently, the plants were stained for 10 min with solophenyl flavine 7GFE (0.1% w/v in 0.1 M Tris/HCl, pH 8.5) and washed again with water (four times) (Knight and Sutherland, 2011). Roots and green parts were separated and mounted on slides, observed under the epi-fluorescence microscope Olympus BX50 using UV excitation light (360–370 nm), and documented using the Olympus DP71 camera and Cell^B software (Olympus Optical Co.). Images, taken at different depths, were processed using Helicon Focus 6.5.2 Pro (Helicon Soft Ltd.) and CorelDRAW 2017 (Corel Corporation). False colors (cyan) were assigned to the F. oxysporum cells.

Significantly enriched promoter motifs were identified separately for statistically significant up- and downregulated genes using HOMER (v4.9, Hypergeometric Optimization of Motif EnRichment, http://homer.ucsd.edu/homer/motif/) with the following parameters: motif length of 8, 10, or 12 bp within 500 bp upstream or downstream from the transcription start site. Motif enrichment analyses were performed relative to the background gene set of all expressed genes.

The raw count matrix generated using featureCounts was normalized with rlog function in the DESeq2 package. Next, these data were used in principal component analysis (PCA). Calculation was performed in R software with prcomp function and visualized using the FactoMineR (Lê et al., 2008) and ggplot2 packages (Wickham, 2009).

All experiments were performed using at least three biological repeats. The results were presented as the mean values ± standard deviation. To assess the significance of changes, one-way ANOVA with Tukey’s post hoc test or the Kruskal–Wallis test was used (Statistica, Dell, Round Rock, TX, United States).

Electromagnetic fields can be considered an abiotic stressor, which induces the alteration in the chemical content of plant cells by oxidative stress (Shabrangy and Majd, 2018). Under a stress stimulus, a plant produces reactive oxygen species (ROS) through several biochemical transformations. ROS include hydroxyl radicals (^•OH), superoxide anions (O₂ ^•−), hydrogen peroxide (H₂O₂), and singlet oxygen (¹O₂) (Mittler et al., 2004). A superoxide anion radical (O₂ ^•−) is usually the first ROS formed by the enzymatic process, autooxidation reaction, and nonenzymatic electron transfer reactions, in which an electron is transferred to molecular oxygen. O₂ ^•− can be generated in the organelles, where the electron transport chains occur (plastids and mitochondria), and membrane-dependent NADPH oxidase (respiratory burst oxidase homolog (RBOH) proteins) systems. The superoxide anion (O₂ ^•−) is the precursor of various ROS because of its instability (a relatively short half-life of 2–4 μs) and strong oxidation/reducibility. The superoxide anion may serve as a substrate for superoxide dismutase (SOD), which catalyzes its dismutation to H₂O₂. Hydrogen peroxide is considered an important redox molecule, given its specific physical and chemical properties, including remarkable stability within cells (half-life of 10⁻³ s) (Mhamdi and Van Breusegem, 2018). H₂O₂ is an important signaling molecule (Nazir et al., 2020) and thus must be strictly controlled under physiological conditions. This can be realized by its conversion to water and oxygen driven by catalase (CAT).

We measured the H₂O₂ levels in several time points in flax exposed to 500 μT, 50 Hz EMF. After an initial increase (by 33% 1 h after exposure (hae)), the level of H₂O₂ decreased and reached 57% of the control at 12 hae and then increased at 24 hae (Figure 2A). The decrease in the H₂O₂ level was also observed in Anthemis gilanica seedlings, although due to a different experimental setup, the measurement was performed at a later time after the exposure (Hassanpour et al., 2021). NADPH oxidase is a membrane-bound enzyme with channel voltage-gating-associated ROS production which is known to be increased by EMF (Friedman et al., 2007). Although it is thought to be indirectly affected by the EMF (Georgiou and Margaritis, 2021), an increase in the transcript level of the corresponding gene (over three times more than the control) was observed at 12 hae (Figure 2B).

ROS in plants can damage cellular components and lead to cell death; thus, in response to oxidative stress, plants have an evolved complex antioxidant defense system, which includes enzymatic (e.g., catalase, peroxidase, and superoxide dismutase) and non-enzymatic antioxidants with polyphenols, terpenoids, and vitamins that work together to scavenge ROS and prevent cellular damage (Xu et al., 2017). We measured the antioxidant capacity of EMF-exposed flax plants’ methanol extracts and observed an increase of over twofold at 1 hae and its gradual decrease in the subsequent time points (Figure 3). Changes in the antioxidant potential correlated well with the changes in the level of oxygen peroxide. When in contact with a stressor, plants develop the early phase (2–6 h) of the stress response (Rapala-Kozik et al., 2012). During this time, ROS are generated in excess. To cope with them, activities of numerous genes are modulated, but before and during this process, the antioxidants already present in the cytoplasm provide the first line of defense. Under normal conditions, small molecular antioxidants like phenolic derivatives (including flavonoids) are present in the cell in the form of glucoside conjugates. Such modification usually decreases their activity but improves solubility in the cytoplasm, thus allowing for their widespread presence in the cell. When stress occurs, recruitment of these dormant ROS scavengers are activated through cleaving the glycosyl group (Brunetti et al., 2013; Behr et al., 2020). Although the time required for genetic expression depends on many factors, such as the type of stimulus and promoter activity, it usually takes hours until a steady-state expression state is achieved (Waters et al., 2017). Considering this, although the 1 hae time point is characterized with greater antioxidant potential, based on the expression of the NADPH oxidase gene, we selected 12 hae as the optimum time point for further analyses.

Transcriptomes of flax (L. usitatissimum L. var. Nike) were sequenced in the seedlings exposed to the 50 Hz, 500 μT ELF-EMF for 30 min and collected at 12 hae compared to the non-treated control. The main purpose of the study was the evaluation of changes (≥2-fold) in the transcript levels. Principal component analysis (PCA) was performed with PCA function in the FactoMineR (Lê et al., 2008) package of R software and visualized using the ggplot2 (Wickham, 2009) package of R software. The PCA results revealed good separation between the ELF-EMF-exposed and control samples (Figure 4) in the PC01 component.

The sequencing data were used for differentially expressed gene (DEG) analysis. The results are given in Supplementary Table S2. Among the 1,882 identified DEGs, almost 1000 were upregulated, while over 800 were downregulated (Figure 5). For quick visual identification of genes with large fold changes, a volcano plot is given in Figure 6. Then, Gene Ontology (GO) analysis was performed on differentially expressed genes.

GO analysis is typically performed after transcriptome sequencing to gain insights into the biological processes, molecular functions, and cellular components associated with the DEGs. By using GO analysis, genes can be annotated based on their functions and categorized into functional groups to better understand the biological mechanisms that underlie the observed differences in gene expression between the conditions being compared, the influence of ELF-EMF in this case. The number of DEGs of GO terms (categories) that were statistically and significantly overrepresented is given in Figure 7; Supplementary Table S3. The most upregulated genes involved in redox processes, response to oxidative stress, and heme binding/iron binding were identified. These results confirm previous findings on the influence of EMF on plants. For instance, among 34 peroxidase or peroxidase-related genes, only four were downregulated. Xiao-ju and Guo (1999) found an increase in the activity of catalase and peroxidase enzymes in tomato seeds pretreated with MF. An increase in catalase activity after a 30-min exposure of wheat plants to a magnetic field with frequencies of 14.3 Hz and 20.8 Hz by 60% and 20% relative to the control level was shown by Mshenskaya et al. (2022). In addition, all the three enzyme gene activities were shown to be increased after EMF treatment in the study conducted by Hassanpour et al. (2021). The authors showed increased levels of phenolic derivatives, which in turn translated into a higher antioxidant potential measured by the DPPH method. On the other hand, Alemán et al. (2014) showed decreased levels of SOD, CAT, and Apx activities in different developmental stages of coffee (except CAT in 4-month-old seedlings) after EMF treatment. The field intensity was 2 × 10e3 μT, but the exposure was only 3 min (4 times and 10 times shorter than those used in our experiments, respectively). Esim et al. (2017) ascribed adverse effects to the ELF-EMF by showing increased H₂O₂ levels and decreased antioxidant enzymes.

In addition to the genes encoding enzymes directly involved in the antioxidant response to ROS, genes involved in low-molecular-weight antioxidants are important for the quenching of free radicals. Surprisingly, as phenylalanine ammonia lyase (PAL) gene expression is very sensitive to various stresses (Gholizadeh and Kohnehrouz, 2010; Jóźwiak-Żurek et al., 2011; Zhan et al., 2022), no PAL gene was found to be considerably (over 2-fold) changed after ELF-EMF exposure, while the activity of PAL, the key enzyme in phenylpropanoid biosynthesis (Kostyn et al., 2012), as well as its encoding gene expression, was shown to be significantly increased in tobacco after EMF treatment (Radmard Titkanlo et al., 2018). We verified the expression of pal gene using RT-qPCR, and its transcript level was generally downregulated in all time points. There was an increase only at 3 hae in the transcript level, reaching 1.48-fold of the control (Figure 8A). However, the expression of the gene encoding the subsequent enzyme cinnamate-4-hydroxylase (C4H), which oxidizes cinnamic acid to 4-coumaric acid, was over 2.5 times higher after ELF-EMF exposure than the control, which was verified using RT-qPCR (Figure 8B). Its expression was the highest at 12 hae and reached 2.24-fold of the control. Moreover, expression of the gene encoding 4-coumarate-CoA ligase (4CL), which converts 4-coumaric acid to its more active form of 4-coumaroyl-CoA, was over 34 times higher. 4-Coumaroyl-CoA is a starting point for the synthesis of a whole gamut of phenolic derivatives, including phenolic acids (caffeic acid, ferulic acid, lignin, lignans, flavonoids, and anthocyanins). It is noteworthy that all the aforementioned three genes are activated by endogenous H₂O₂ (Cheng et al., 2019). No significant changes were observed for the genes involved in lignin biosynthesis (like the two most important cinnamoyl-CoA reductase (CCR) and cinnamyl alcohol dehydrogenase (CAD)). The level of lignin in such young seedlings, especially grown under in vitro conditions, was below the detection limit of the “acetyl bromide” method (Hatfield et al., 1999) (data not shown). On the other hand, the plant’s reaction to oxidative stress, as during infection, can be manifested by enhanced polymerization of cell wall components, including lignin. The resulting tightening of lignin proceeds mainly via “end-wise” polymerization, in which an oxidized monolignol radical cross-couples with a radical formed on the free-phenolic ends of a growing lignin polymer. The ROS necessary for this process is produced by NADPH oxidases and laccases (Cheng et al., 2019; Lee et al., 2019; Tobimatsu and Schuetz, 2019). Our transcriptome analysis showed that both the NADPH oxidase and laccases were upregulated after the exposure to ELF-EMF. Among other polyphenol derivatives with antioxidant potential are flavonoids and anthocyanins. Transcriptome analysis revealed that the expression of chalcone synthase (chs) encoding the key enzyme in flavonoid biosynthesis was over two times higher than that in the non-treated control, which was verified using RT-qPCR (Figure 8C). Flavonol synthase (fls) gene transcription was also increased (4.7 times higher) compared to the control. The increase in the transcript level of flavonoid 3′-hydroxylase gene (f3′h) but not flavonoid 3′,5′-hydroxylase was also noticed, suggesting the redirection of the pathway to quercetin production. The level of dihydroflavonol reductase (dfr) gene transcript was downregulated after exposure to ELF-EMF (0.4-fold of the control). This, together with the overexpression of anthocyanidin synthase (ans) and anthocyanidin glucosyltransferase (ugt) genes, suggests that anthocyanin synthesis from leucoanthocyanidins is preferred. The phenylpropanoid biosynthesis pathway also includes the production route of salicylic acid (SA), a known phytohormone participating in the plant’s response to stress. SA levels were shown to increase after the use of the pulsed magnetic field on raspberry plants (Upadyshev et al., 2021). This phytohormone is known to participate in the activation of many genes involved in the plant’s response to stress, including pathogen attack. Moreover, the methylated derivative of SA (methyl salicylate (MeSA)) is volatile and shown to act as an inducer of plant defense mechanism against pathogens and certain herbivores (Boba et al., 2017; Ninkovic et al., 2021). The methylation of SA is catalyzed by benzoic acid/salicylic acid methyltransferase (BSMT). In our experiment, the BSMT gene transcript level was found to be increased in flax seedlings exposed to ELF-EMF (almost 2.4 times higher than the control). At the same time, the transcript level of salicylic acid glucosyltransferase (SAGT) was decreased. Data on the expression of the PR genes, genes involved in ROS processing, and phenylpropanoid biosynthesis are given in Figure 9 in the form of heatmaps.

All of the aforementioned observations resemble the plant’s response to pathogen attack, and many genes that are activated upon infection are observed to be upregulated also due to the ELF-EMF exposure. Moreover, we found that the number of DEGs of the WRKY transcription factors (TFs) was higher than the control. WRKY transcription factors, mainly induced by SA upon infection, act in a complex defense response network as both positive and negative regulators (Pandey and Somssich, 2009). In addition, due to infection, pathogenesis-related proteins (PRs) are biosynthesized. Among all PR genes identified in the transcriptome of flax exposed to ELF-EMF, the majority was upregulated. For instance, chitinase transcript levels were higher (2- to 57.5-fold) than those of the control. Similarly, β-glucanase transcripts were more abundant in the ELF-EMF-exposed flax seedlings than in the control. Both chitinase and β-glucanase genes are activated in flax under pathogen attack (Wojtasik et al., 2019). Their expression levels were confirmed by RT-qPCR (Figures 8D,E).

Based on the observation that exposure to the ELF-EMF leads to the upregulation of genes usually activated upon pathogen infection, we decided to perform infection tests on flax seedlings exposed to the ELF-EMF compared to the untreated control. The seedlings were infected with F. oxysporum spores directly after exposure. After 10 days, we measured the level of fungal murein transglycosylase gene by RT-qPCR as a marker of the fungal cells already present in the plant. We divided the plant material into the green parts (stems/leaves) and roots. Although no major changes were observed in the roots, the number of F. oxysporum cells in the green parts was significantly lower (Figure 10).

To visually confirm that there is less fungal penetration into the green parts of flax seedlings treated with the ELF-EMF before infection, we performed histochemical staining of the fungus and observed the stem sections under the microscope. Solophenyl flavine staining clearly showed a less number of F. oxysporum cells in the ELF-EMF-exposed stem tissue than in the control (Figure 11).