Transgenerational somatic mutations and epigenetic imprints on selected DNA repair gene promoters in Arabidopsis thaliana due to pesticides

Abstract

• Pesticides are known to induce genome instability. Reports on eukaryotes, including humans, mostly covers chromosomal aberrations and DNA strand breaks. Studies on scoring of the rates of loci‐specific somatic mutations such as point mutations (PM), somatic homologous recombination (SHR), frameshift mutation (FSM), and transpositions due to pesticides are mostly confined to prokaryotes, using Ames test. Analogous eukaryotic model systems though are available for Drosophila, mouse and certain cell lines, reports using these are meagre, possibly due to the lethality of commercially‐used pesticide doses and/or difficulty in assessing large population.

• DNA replication/repair being conserved among higher eukaryotes, the GUS reporter‐based Arabidopsis thaliana system was used to score loci‐specific mutations such as PM, SHR, FSM and transpositions. DNA repair gene expression and DNA methylation were studied using real‐time PCR and methylation sensitive restriction enzyme digestion PCR, respectively.

• Four commercially used pesticides not only increased the rates of PM, SHR and FSM, and caused epimutations on selected DNA repair gene promoters, but also induced heritable transgenerational memory. While SHR rates were significantly high in the untreated progeny, FSM rates were high even in the grand progeny. Expression of 15 DNA repair genes increased under at least one of the pesticidal treatments. Eight of 21 DNA repair gene promoters showed altered methylation. MRE11A promoter showed demethylation in the untreated progeny due to all four pesticides. Malathion‐induced demethylation was retained in the grand progeny also.

• Thus this work gauges the severity of various pesticides on inducing heritable loci‐specific mutations/epimutations in a non‐target system.

Four different pesticides increased the rates of loci‐specific somatic mutations and caused epimutations on selected DNA repair gene promoters, of which frameshift mutation rates and malathion‐induced demethylation of the MRE11A promoter were retained in the grand progeny of Arabidopsis thaliana, a non‐target system.

INTRODUCTION

Pesticides are applied during crop cultivation to reduce yield loss due to various pests. It is known that most pesticides cause physiological and behavioural changes in organisms, and are carcinogenic owing to their mutagenic property (Calviello et al. ²⁰⁰⁶; Franco et al. ²⁰¹⁰; Lushchak et al. ²⁰¹⁸). The carcinogenicity of pesticides is gauged based on their ability to induce genomic instability such as chromosomal aberrations and loci‐specific mutations [point mutations (PM), somatic homologous recombination (SHR), frameshift mutation (FSM) and transpositions]. While the former is tested in a wide range of organisms, calculating the rates of PM, SHR, FSM and transposition is usually done using the prokaryote‐based Ames test. Similar reports using analogous higher eukaryotic mutants (example Drosophila, mouse) are relatively less, probably due to the lethal effect of the pesticides on the organism and/or non‐availability of desirable mutants. As a consequence, studies on transgenerational inheritance of such loci‐specific somatic mutations due to pesticides are lacking for most pesticides. Transgenerational changes are a reflection of altered epigenetic changes in the genome. Reports on pesticide‐induced heritable epigenetic changes are a few.

Previous reports on pesticide‐induced mutations in animal model systems including humans are numerous (Sánchez‐Alarcón et al. ²⁰²¹; Zúñiga‐Venegas et al. ²⁰²²). For example, increased micronuclei and whole genome DNA breaks were seen in a population of soybean farmers exposed to multiple pesticides in Brazil (Benedetti et al. ²⁰¹³). Loci‐specific point mutations correlating with physiological and genomic disorders have also been widely reported by analysing human populations in pesticide‐exposed geographical locations (Teodoro et al. ²⁰¹⁹). For example, increased polymorphism in XRCC1 and OGG1 was seen correlating with increased DNA damage and altered immune system response in a population of tobacco farmers exposed to multiple pesticides in Brazil (Kahl et al. ²⁰¹⁸). In any case, as studies on humans rely only on previously exposed populations due to ethical concerns, rates of pesticide‐induced heritable loci‐specific mutations cannot be measured in humans.

There are a few studies reporting that apart from physiological/genetic abnormalities, certain pesticides induce epigenetic changes as well. For example, the herbicide glyphosate had a toxicological effect on the endocrine system, was carcinogenic, led to impairment in reproduction and development and also induced epigenetic modifications (global DNA methylation, methylation of specific genes, histone modification and differential expression of non‐coding RNAs) (Rossetti et al. ²⁰²¹). Reports also reveal that certain pesticides induce transgenerational physiological and epigenetic changes in organisms. For example, locomotor disability, anxiety behaviour and transcriptomic/epigenomic (DNA methylation) changes in brain cells were observed in the progenies of permethrin‐exposed zebrafish (Blanc et al. ²⁰²¹). Tumour development in the F2 generation of glyphosate‐treated female lineage of rats was high (Kubsad et al. ²⁰¹⁹). Multiple pesticides (dichlorodiphenyltrichloroethane, atrazine, vinclozolin, methoxychlor) induced transgenerational changes in various physiological aspects (testicular/prostate abnormalities and metabolic disorders), behaviour, tumour development and epigenetic aspects (altered DNA methylation patterns), and also altered gene regulation (transcriptome and DNA methyltransferases expression) in mammals (Van Cauwenbergh et al. ²⁰²⁰). The pesticide methoxychlor also promoted the transgenerational inheritance of diseases and sperm epimutations (Manikkam et al. ²⁰¹⁴).

Epigenetic memory can be of intergenerational and transgenerational types (Heard & Martienssen ²⁰¹⁴; Lämke & Bäurle ²⁰¹⁷). Former is the case when progeny of the treated individual though remembers the stress, this memory is lost in the grand progeny. In the latter case, the epigenetic memory persists in the grand progeny and sometimes even in subsequent generations (Luna et al. ²⁰¹²; Slaughter et al. ²⁰¹²). Intergenerational memory could be because of the effect of treatment on parental germ cells, which later gave rise to the offspring (Daxinger & Whitelaw ²⁰¹²; Lim & Brunet ²⁰¹³). Transgenerational memory, on the other hand, is purely exhibited by cells never exposed to the treatment. Both intergenerational and transgenerational epigenetic memory is mostly due to cytosine methylation, mostly in regulatory/promoter regions, which influence gene expression (Lacal & Ventura ²⁰¹⁸; Legoff et al. ²⁰¹⁹; Tuscher & Day ²⁰¹⁹; Casas & Vavouri ²⁰²⁰). Often methylation is faithfully passed on to the daughter DNA during cell division, even after gametogenesis. Such transcription regulation has the potential to lead to physiological and/or behavioural changes which are passed on even to the progeny generations (Toffolatti et al. ²⁰¹⁴; Klosin et al. ²⁰¹⁷; Tsuboi et al. ²⁰²⁰; Brabson et al. ²⁰²¹). Various biotic/abiotic factors such as heat, UV, microbes and soil nutrients induce transgenerational epigenetic inheritance in plants, often imparting altered gene expression and functional plasticity (Doddavarapu et al. ²⁰²⁴; López et al. ²⁰²⁴; Amritha et al. ²⁰²⁵; Virgen et al. ²⁰²⁵).

It is to be noted that considering the enormous variety of pesticides used across the globe, only a few are screened for their transgenerational impacts in eukaryotic systems. This could partially be because of the ethical difficulty in performing such studies in huge populations across multiple generations. Towards this end, the model plant Arabidopsis thaliana has previously offered a high‐throughput screening system for testing the loci‐specific mutagenic rates induced by pesticides such as 2,4‐D and dicamba (Filkowski et al. ²⁰⁰³), rose bengal and aminotriazole (Kovalchuk et al. ²⁰⁰³), paraquat (Kovalchuk et al. ²⁰⁰³; Pecinka et al. ²⁰⁰⁹) and atrazine (Besplug et al. ²⁰⁰⁴). Pesticides with more damaging impacts can be selected for further studies in animal model systems. In any case, results based on such plant systems are also a reflection of the mutagenic impact of pesticides on the genetic diversity of non‐target organisms. Like most animals and humans, plants also are non‐target victims of pesticides, which silently survive by tolerating physiological changes that lead to yield loss in crop plants (Parween et al. ²⁰¹⁶; Sharma et al. ²⁰¹⁹; Mehdizadeh et al. ²⁰²¹). Also, to be recalled is the fact that adverse factors in the environment lead to forced evolution.

We thus scored the rates of four categories of loci‐specific somatic mutations (PM, FSM, SHR and transpositions) due to four commercially used pesticides: malathion, flubendiamide, chlorantraniliprole and carbendazim + mancozeb combination, using the A. thaliana model system. Since the DNA repair genes are required for maintaining genome stability, we also analysed the influence of these pesticides on the expression and epigenetic methylation of promoters of DNA repair genes belonging to various pathways. Based on these results, we subsequently went on to study whether these somatic mutations and epimutations were heritable in plants. Though all the four pesticides induced transgenerational genetic/epigenetic changes, their intensity of damage varied. Our approach seems to be a feasible way of high‐throughput screening of a wide range of pesticides. We believe that there are no such previous reports based on the pesticides that we analysed.

MATERIAL AND METHODS

Arabidopsis growth conditions

A. thaliana seeds were sown on planting material (vermicompost, purchased from Central Plantation Crops Research Institute, Kasaragod, Kerala, India) and kept at 4 °C for 48 h for vernalisation in order to get a synchronous growth. These were then transferred into the plant growth chamber (Percival, USA. Model Ar‐3663) under uniform conditions of light (8000 lx), temperature (24 °C) and humidity (100%) with a 16 h light/8 h dark cycle.

Arabidopsis mutation detector lines

Various detector lines of A. thaliana were used to detect the occurrence of somatic mutations such as PM, SHR, FSM and transpositions. While the PM rates were scored using the Arabidopsis line 747 (Kovalchuk et al. ²⁰⁰⁰), SHR rates were scored using 651 (Swoboda et al. ¹⁹⁹⁴) and R2L1 (Li et al. ²⁰⁰⁴), and FSM and transposition rates were scored using G10 (Azaiez et al. ²⁰⁰⁶) and Tag1 (Frank et al. ¹⁹⁹⁷), respectively. The lines 747 and 651 were kindly provided by Barbara Hohn (Friedrich Miescher Institute, Switzerland). Lines G10 and R2L1 were obtained from Francois Belzile (University of Laval, Canada). The Tag1 line was obtained from Nigel Crawford (University of California, California). These Arabidopsis mutant lines carry an inactivated GUS construct tailored to score the respective mutation upon its reversal to the functional form (Fig. 1).

Fig. 1.

GUS‐based mutation scoring. (a) Mutation detector lines and their respective constructs of inactive GUS, driven by the CaMV 35S promoter (P), tailored to score point mutation (PM), somatic homologous recombination (SHR), frameshift mutation (FSM) and transposition, restoring to the functional GUS after reversal of respective mutations. (b) Parts of Arabidopsis detector lines (after GUS histochemical staining and chlorophyll removal) with tissues harbouring functionally restored GUS, stained as blue spots.

Pesticides and their application

Arabidopsis plants were subjected to pesticides belonging to different categories. Appropriate dosage (Table 1) of these was either sprayed or directly added to the planting material based on the instructions of the manufacturer. For scoring somatic mutation rates, pesticides (full/half dose) were applied on 2‐week‐old plants of the mutation detector lines, and these were taken for GUS histochemical staining after 7 days. For all other analyses, full dose of pesticides was applied to Columbia plants in the beginning of the 4th week and taken for DNA/RNA extraction after 48 h.

Table 1.

Details of various pesticides used in this study.

no.	generic name of the pesticide	trade name of the pesticide	chemical family	class	dosage a
					full	half
1.	Malathion 50% EC	Jaithion 50% EC	Organophosphate	Insecticide	0.1%–0.15%	0.05%–0.075%
2.	Flubendiamide SC480	Fame SC 480	Pthalic acid diamide group or benzene dicarboxamide	Insecticide	50 ml/500 l/ha	25 ml/500 l/ha
3.	Carbendazim 12% + Mancozeb 63% WP	Two in one	Benzimidazole + ethylene bisdithiocarbamate	Fungicide	300 g/200 l/acre	150 g/200 l/acre
4.	Chlorantraniliprole 0.4% W/W GR	Dupont Ferterra 0.4% W/W GR	Antranilic diamide	Insecticide	4 kg/acre	2 kg/acre

^aDosage as per manufacturer's instruction was calculated for a planting tray of size 100 cm².

GUS histochemical staining

Plants were washed in phosphate buffer (pH 7.0) and incubated at 37 °C for 1 h in phosphate buffer containing 1% tritonX‐100. After decanting the phosphate buffer, plants were vacuum infiltrated for 5 min in the staining solution containing 5‐bromo‐4‐chloro‐3‐indolyl glucuronide (X‐gluc), the substrate for beta glucuronidase (GUS). The plants were then incubated at 37 °C in the same staining solution for 24 h. The chlorophyll was removed subsequently by soaking the plants in ethanol (Jefferson et al. ¹⁹⁸⁷; Ilnytskyy et al. ²⁰⁰⁴). Tissues expressing functional GUS due to the mutation‐induced reversion stained blue, and these sectors were observed and counted manually using the stereo binocular microscope (Zeiss, Stemi DV4) (Fig. 1).

Scoring of mutation rates and statistical analysis

For each experiment, around 100 plants were used. The experiments were repeated three times, and the mean values of mutation frequencies for a population of around 300 plants were calculated by the below formula (Shah et al. ²⁰¹⁵).

$$\text{Mutation frequency}=\begin{array}{l}\text{Total number of}\mathrm{GUS}\text{spots}\\ ÷\text{Total number of plants}\end{array}$$

Mutation frequency analysis using GUS spot is count data, which does not follow normal distribution, so non‐parametric Kruskal–Wallis test (Acar & Sun ²⁰¹³) and Mann–Whitney (Wilcoxon) test (Dexter 2013; Harrison et al. ²⁰¹⁹) were used for the determination of the significance of the difference between the means, for sample sizes more than two and two, respectively.

Total RNA/DNA extraction and cDNA synthesis

Total RNA was extracted using the Spectrum Plant Total RNA Kit (Sigma Aldrich, Burlington, Massachusetts, USA), subjected to DNase (Sigma Aldrich, USA) treatment and then quantified using a nano drop microvolume spectrophotometer (Thermo Fisher Scientific, Waltham, Massachusetts, USA). cDNA was synthesised from RNA using the cDNA synthesis kit (Origin Lab, India). Total genomic DNA was extracted using the CTAB method (Rogers & Bendich ¹⁹⁹⁴).

Primers, quantitative real‐time PCR (qRT‐PCR)

The previously published primers of DNA repair genes (Joseph et al. ²⁰²¹), and the genes for normalisation (Joseph et al. ²⁰¹⁹) were used for all PCRs. qRT‐PCR was carried out to find out the expression analysis of the DNA repair genes using syber green master mix (Roche Diagnostics, Indianapolis, Indiana, USA). Fold change of gene expression was calculated using the 2^−ΔΔCt method, where values were obtained by normalising using 18SrRNA as an endogenous control for normalisation. The mean values of gene expression were compared using Tukey's HSD post hoc ANOVA.

Methylation‐sensitive restriction enzyme digestion (MSRD)‐PCR

The genomic DNA was subjected to independent digestion with various methylation‐sensitive restriction enzymes (MSREs) (HpyChIV4, Hpy99I, BsaAI, AciI, ClaI and ApeKI) (New England Biolabs, Ipswich, Massachusetts, USA) as described in Joseph et al. (2021). The primers used were as described in Joseph et al. (2021), where the expected amplicon sizes ranged from 310 to 867 bp and included promoter regions with at least four MSRE sites (same or different). The reaction mixtures were heat‐inactivated at 65 °C for 20 min, and PCRs were performed using this digested DNA as template. The PCR cycle consisted of an initial denaturation at 95 °C for 10 min, followed by 45 PCR cycles of 95 °C for 40 s, 50 °C–60 °C (primer annealing Tm) for 40 s, 72 °C for 40 s and a final extension at 72 °C for 5 min. As the DNA was digested with MSREs, the presence of a band indicated lack of digestion due to the respective MSRE, thereby indicating methylation of the region meant for amplification. Similarly, lack of a band indicated successful digestion and thus absence of methylation. While each promoter had only one primer set, amplification could be obtained using DNA independently digested with one or more MSREs for each treatment. In the case of PCRs involving multiple MSRE‐digested DNA for a single promoter, altered methylation status was declared if the expected banding was obtained in at least one of the digested samples.

RESULTS

Point mutation rates increase after pesticide treatment

PM rates were scored using plants from the Arabidopsis line 747 treated with two different dosages of four different pesticides (Table 1). Though both the doses of all pesticides increased the PM rates, a statistically significant increase was found in the full dosage of malathion, flubendiamide and the combination of carbendazim + mancozeb (Fig. 2a–d). Of these, flubendiamide induced a highest of 5.7‐fold increase, which was followed by malathion (4.4‐fold), carbendazim + mancozeb combination (4.3‐fold) and chlorantraniliprole (3.3‐fold), respectively.

Fig. 2.

Mutation frequencies upon treatment with various pesticides (malathion, flubendiamide, carbendazim + mancozeb combination and chlorantraniliprole) at two dosages, full (A) and half (B), compared with the untreated control (C). (a–d) are for line 747; (e–h) are for line R2L1; (i–l) are for line 651; (m–p) are for line G10; (q–t) are for line Tag1. The vertical line on top of each bar indicates the ± standard error (SE) of the mean. Number on top of the SE bars indicate the respective mean values. x‐axis indicates the samples and treatments. y‐axis indicates the respective mutation frequencies. The symbols, *, ◊, and ■ indicate that these treatments are significantly different at P value (P < 0.05) as determined by Kruskal–Wallis test (Table S1).

Somatic homologous recombination rates increase after pesticide treatment

Significantly increased SHR rates were induced by all the four pesticide treatments in a dose‐dependent manner in line R2L1 (Fig. 2e–h). Similar to PM, significantly increased mutation rates were induced by all treatments, except chlorantraniliprole, in line 651 (Fig. 2i–l). The slight difference in the behaviour of both lines could possibly be because of the position effect of the GUS construct (Swoboda et al. ¹⁹⁹³). Maximum increase of SHR rates (7.8‐fold) was induced by malathion, followed by carbendazim + mancozeb combination (5.99‐fold), flubendiamide (5.17‐fold) and chlorantraniliprole (3.79‐fold) in line R2L1.

Frameshift mutation rates increase after pesticide treatment

FSM rates significantly increased after treatment with all four pesticides in a dose‐dependent manner (Fig. 2m–p). Unlike PM, where the differences were significant at P < 0.05 and in a few cases P < 0.01, in FSM, these were significant at P < 0.001 (Table S1), indicating that FSM were more sensitive to pesticides. Like SHR, malathion induced the highest (19.43‐fold) FSM rates. This was followed by a combination of carbendazim + mancozeb, flubendiamide and chlorantraniliprole with a fold increase of 9.46, 6.84 and 6.35, respectively.

Most transposition rates decrease after pesticide treatment

Unlike the above three mutations, the rates of transpositions significantly decreased in a dose‐dependent manner in most treatments (Fig. 2q–t). Malathion alone induced an increase of transposition rates (1.3‐fold), similar to PM, SHR and FSM. Nevertheless, this increase was much lower in comparison to those in the other three mutations.

The effect of pesticide‐induced mutations is transgenerational

In order to see whether the plants carried the memory of pesticide‐induced stress to their offspring, we calculated the rates of PM, SHR and FSM in the progenies of pesticide‐treated plants. About 300 plants obtained from the seeds collected from 20 parents (experiment repeated thrice, with 100 plants under each treatment) were analysed for each treatment. While the parents were treated with a full dose of pesticides, the progeny generation was not subjected to any pesticides. Though the PM rates were not enhanced in the progeny generation, the rates of SHR and FSM were significantly higher under all treatments when compared to those from untreated control parents (Fig. 3). As the memory in progenies of plants may be intergenerational or transgenerational (Yao & Kovalchuk ²⁰¹¹), we analysed the grand progenies (third generation) in order to see if there was any transgenerational inheritance of pesticide stress on SHR and FSM rates. By contrast to the second generation, the SHR rates significantly reduced under most treatments (Fig. 3) except under malathion, where the rates were comparable to untreated controls. Unlike SHR, the rates of FSMs were still significantly higher under all treatments in the grand progenies as well. However, an overall decrease in all mutation rates in each generation was seen in all pesticide‐treated progeny samples (Fig. 3). Thus, the memory of increased SHR and FSM are intergenerational and transgenerational, respectively.

Fig. 3.

Rates of somatic homologous recombination (SHR) and frameshift mutations (FSM) analysed in the second and third generation progeny of plants from R2L1 and G10 lines, respectively, treated with full dose of various pesticides. (a) Second generation of R2L1 treated with malathion, (b) second generation of R2L1 treated with flubendiamide, (c) second generation of R2L1 treated with carbendazim + mancozeb combination, (d) second generation of R2L1 treated with chlorantraniliprole, (e) second generation of G10 treated with malathion, (f) second generation of G10 treated with flubendiamide, (g) second generation of G10 treated with carbendazim + mancozeb combination, (h) second generation of G10 treated with chlorantraniliprole, (i) third generation of R2L1 treated with malathion, (j) third generation of R2L1 treated with flubendiamide, (k) third generation of R2L1 treated with carbendazim + mancozeb combination, (l) third generation of R2L1 treated with chlorantraniliprole, (m) third generation of G10 treated with malathion, (n) third generation of G10 treated with flubendiamide, (o) third generation of G10 treated with carbendazim + mancozeb combination, (p) line G10 treated with chlorantraniliprole, (q, r) trends of SHR and FSM, respectively, across three generations (I, II and III) under the treatment of various pesticides (as per the key for dashed/solid black lines). C and T are progenies of untreated control and pesticide‐treated plants, respectively. The vertical line on top of each bar indicates the ± standard error (SE) of mean. Number on top of the SE bars indicates the respective mean values. In (a–p), the x‐axis indicates the samples and the y‐axis indicates the respective mutation frequencies. In (q, r), the x‐axis indicates the generations and y‐axis indicates the fold change of mutation rates under various treatments. Asterisks (** and *) indicate that this mean is significantly different from control at P < 0.01 and P < 0.001 respectively, as determined by Mann–Whitney (Wilcoxon) test (Table S2).

DNA repair gene expression shows pesticide‐specific response

As genome stability is maintained by DNA repair proteins, we analysed the effect of various pesticides on the expression of 15 DNA repair genes belonging to five different pathways. The 18S rRNA‐encoding gene was identified as the reference gene for normalisation after comparison with many other genes (Fig. S1). Though expression of most genes seemed to increase overall under the treatment with most pesticides (Table 2), only certain treatments resulted in a significant change (Fig. S2). Flubendiamide significantly increased the expression of maximum of eight genes. On the contrary, malathion significantly increased the expression of a least number of three genes only. Only two genes, MSH4 and PMS2L3 belonging to MMR pathway responded significantly to treatment with all the pesticides (Fig. S2l,m). Overall, the DNA repair genes showed a preferential response to the pesticides (Fig. 4a,b).

Table 2.

Expression and methylation status of various DNA repair genes due to pesticide application.

repair pathway	gene name	methylation status of promoters in untreated control	expression/methylation status due to various pesticides				overall effect of any of the pesticides on methylation/demethylation status of various promoters
			malathion	flubendiamide	carbendazim + mancozeb	chlorantraniliprole
Nucleotide excision repair	RAD 23C	‐	↑ (2.24)	↑ (5.78)	↑** (29.3)	↑ (15.92)	‐
		■	□	□	■	■	Demethylation
	RAD 23 B	‐	↑ (1.85)	↑** (13.85)	↓ (0.21)	↓ (0.42)	‐
		■	■	■	■	■	No change
	XAB2	‐	↑ (1.7)	↑** (6.35)	↑ (2.48)	↑** (9.8)	‐
		■	□	■	■	■	Demethylation
	XPF	‐	↑ (1.82)	↑** (17.58)	↑** (26.51)	↑** (18.68)	‐
		■	■	□	■	■	Demethylation
	RAD23IV	■	□	□	■	■	Demethylation
	RAD4	■	■	■	■	■	No change
Base excision repair	MBD4	‐	↑ (1.45)	↑ (2.79)	↑** (8.48)	↑ (2.5)	‐
		■	■	■	■	■	No change
	TAG 1	‐	↑ (1.6)	↑ (1.4)	↑ (2.06)	↑ (2.33)	‐
		■	■	■	■	■	No change
	P3MAG 1	‐	↑** (10.84)	↑ (1.95)	↑** (11.99)	↑ (4.85)	‐
		□	□	□	□	□	No change
	XRCC 1		↓ (0.21)	↑** (4.06)	↑ (0.85)	↑ (0.72)	‐
		□	□	□	□	□	No change
Homologous recombination	RAD 51	‐	↓ (0.95)	↑ (1.65)	↑ (1.31)	↑ (1.3)	‐
		□	□	□	□	□	No change
	ATM	‐	↑ (4.01)	↑ (2.55)	↑ (11.49)	↑ (4.19)	‐
		□	□	□	□	□	No change
	BRCA2	■	□	□	■	□	Demethylation
	BLM	□	□	□	□	□	No change
	MRE11A	■	□	□	□	□	Demethylation
Mismatch repair	MSH 4	‐	↑** (104.8)	↑** (20.75)	↑** (23.27)	↑** (38.21)	‐
		■	■	■	■	■	No change
	PMS2L3	‐	↑** (33.42)	↑** (84.69)	↑** (42.34)	↑* (32.37)	‐
		■	■	□	■	■	Demethylation
	MSH6	□	□	□	□	□	No change
Non‐homologous end joining	LIG 4	‐	↑ (3.77)	↑** (16.81)	↑ (7.39)	↑ (1.5)	‐
		■	□	□	□	□	Demethylation
	KU70	‐	↑ (1.67)	↑** (16.12)	↑ (3.18)	↑ (1.84)	‐
		□	□	□	□	□	No change
	KU80	■	■	■	■	■	No change

↑ and ↓ indicate increased and decreased expression, respectively. The values below each arrow indicate the fold change of expression. * and ** indicate that the values are significantly different at P < 0.05 and P < 0.01, respectively, based on the Tukey's hsd Post hoc ANOVA. ■ and □ indicate methylated and non‐methylated status of promoters. ‐ indicates not applicable.

Fig. 4.

Venn diagram showing the categorisation of various DNA repair genes and their respective pathways, based on their expression pattern and methylation status under treatment with four different pesticides. (a) Expression (significantly upregulated). (b) Expression (no change). (c) Demethylation.

Epigenetic control of pesticide stress‐induced response of DNA repair genes

Since pesticides altered the expression of most DNA repair genes, we wanted to see if these genes were epigenetically regulated by pesticide stress. We thus subjected the genomic DNA extracted from pesticide‐treated and control plants to MSRD‐PCR of the promoter of 21 different DNA repair genes. In comparison with the untreated control, promoters of eight out of 21 genes showed altered methylation status in any of the treatments (Table 2 and Fig. S3). Promoters of these eight genes were methylated in the untreated control plants, and hence, demethylation was induced by the pesticides (Table 2 and Fig. S3). While MRE11A and LIGIV of HR and NHEJ pathways, respectively, responded similarly to treatment with all the pesticides, others exhibited a preferential response (Fig. 4c). Thus, the promoters of the same gene responded differently to different treatments. Also, the pesticide‐induced expression of DNA repair genes remained increased irrespective of the methylation status of their promoter (Table 2).

MRE11A promoter exhibits transgenerational epigenetic memory under pesticide stress

Since the memory of altered somatic mutations due to pesticides lingered across multiple generations, we wanted to see if this stress could also induce a similar effect on the epigenetic status of the DNA repair genes. There were previous reports that the duration of methylation can vary from a few hours to many generations (Iwasaki & Paszkowski ²⁰¹⁴). Towards this end, we collected the seeds from a set of 20 plants under treatment with each of the four pesticides. Twenty seedlings under each treatment were then germinated into plants, and these were not treated with any pesticide. DNA extracted from these plants was subjected to MSRD‐PCR for the genes, which showed altered methylation status in the parental generation. Promoter of only one gene, MRE11A, of HR pathway, retained altered methylation status in the progeny (second) generation (Table 3 and Fig. S4). This change was due to all four pesticide treatments.

Table 3.

Intergenerational/transgenerational epigenetic status of various DNA repair genes due to pesticide stress.

repair pathway	gene name	generation	methylation status under various conditions					type of epigenetic memory observed
			control	malathion	flubendiamide	carbendazim+ mancozeb (combination)	chlorantraniliprole
Mismatch repair	PMS2L3	I	■	■	□	■	■	No memory
		II	■	■	■	■	■
Homologous recombination repair	BRCA2	I	■	□	□	■	□	No memory
		II	■	■	■	■	■
	MRE11A	I	■	□	□	□	□	Intergenerational and transgenerational
		II	■	□	□	□	□
		III	■	□	■	■	■
Nucleotide excision repair	RAD23C	I	■	□	□	■	■	No memory
		II	■	■	■	■	■
	RAD23IV	I	■	■	□	■	■	No memory
			■	□	■	■	■
		II	■	■	■	■	■
			■	■	■	■	■
	XAB2	I	■	□	■	■	■	No memory
		II	■	■	■	■	■
	XPF	I	■	■	□	■	■	No memory
		II	■	■	■	■	■
Non‐homologous end joining	LIG 4	I	■	□	□	□	□	No memory
		II	■	■	■	■	■

■ and □ indicate methylated and non‐methylated status of promoters, respectively. I, II and III indicate the three consecutive generations, parent, 1st generation progeny and third generation of grand progeny, respectively.

Since the epigenetic memory in the progeny generation of plants could be either intergenerational or transgenerational, the methylation status of the above promoter was analysed in the third generation also. For these seeds were collected from 20 progeny plants under each treatment to get the grand progeny (third generation). Similar to the analysis in the second generation, DNA was extracted from this unexposed third generation and subjected to MSRD‐PCR. Here, also MRE11A exhibited altered demethylation, suggesting the inheritance of transgenerational memory (Table 3 and Fig. S5). This demethylation was retained only due to malathion. Thus, for the other three pesticides, the memory was intergenerational, as the methylated status reverted to normal in the third generation.

DISCUSSION

A plethora of pesticides are used globally, and their hazardous impact on genome stability and physiology is well‐reported (Mahapatra et al. ²⁰¹⁹; Pathak et al. ²⁰²²). Though there are enormous reports on genotoxic effects of pesticides in eukaryotes, these mostly focus on chromosomal aberrations. Reports on loci‐specific mutations (PM, SHR, FSM and transpositions) are a few. Loci‐specific mutations are a boon and a bane; while these are required for evolution (Stern & Orgogozo ²⁰⁰⁸; Barton ²⁰¹⁰; Itoh et al. ²⁰⁰²), these become a burden if their rates are accelerated (Smith et al. ²⁰²³; Vijg et al. ²⁰²³). Hence, before being released into the market, all pesticides undergo a screening for their potential to induce mutations (USEPA). Though screening of pesticides is preferred using mouse model system (Lang 1984; USEPA 1986), assessing its effect on large population across multiple generations is difficult. Hence, other model systems such as Drosophila and plants are recommended (USEPA 1986). For screening of loci‐specific mutations, though Drosophila mutants offer an analogous system to mouse mutants, they are often succumbed to lethality under commercially used doses of pesticide, with a few exceptions such as fungicides (Osgood et al. ¹⁹⁹¹) and herbicides (Kaya et al. ²⁰⁰⁰). Hence, we tested three such insecticides malathion (Mehdi & Qamar ²⁰¹¹), flubendiamide (Sarkar et al. ^{2014, 2018a}), chlorantraniliprole (Zhao et al. ²⁰¹¹; Cuthbertson et al. ²⁰¹⁴) and the fungicide combination carbendazim + mancozeb (Saraiva et al. ²⁰¹⁸; Ziech et al. ²⁰²³) for their potential to induce heritable genetic/epigenetic imprints deploying the Arabidopsis model system. We scored the rates of various loci‐specific molecular mutations and studied DNA repair gene expression and promoter epigenetic status, under the influence of different pesticides. We further assayed the progeny and grand progeny of treated plants and studied the heritability of these mutations/epimutations.

There was an overall dose‐dependent increase in the rates of most loci‐specific somatic mutations (PM, SHR and FSM), except transpositions. While malathion induced highest increase in SHR (7.8‐fold) and FSM (19.43‐fold) rates, the highest PM rates (5.7‐fold) were induced by flubendiamide. Chlorantraniliprole did not significantly increase the mutation rates in many cases, except FSM, and SHR in one of the mutant lines. Previous reports on the mutations induced by the pesticides that we used include chromosomal aberrations, micro nuclei formation, aneuploidy and strand breaks. For example, malathion induced whole genome strand breaks and micronuclei in Vicia faba (Arya et al. ²⁰¹⁷), and chromosome aberrations, stickiness, chromosome bridges in anaphase and binucleate lesions (Sheikh et al. ²⁰²⁰), and DNA strand breaks (Srivastava & Singh ²⁰²⁰) in onion. Similarly, genotoxicity in onion was induced by flubendiamide (Ciğerci et al. ²⁰²³), carbendazim (Verma & Srivastava ²⁰¹⁸) and mancozeb (Fatma et al. ²⁰¹⁸), and in rats by carbendazim + mancozeb combination (Hashem et al. ²⁰¹⁸). Carbendazim induced DNA strand break in earthworms (Huan et al. ²⁰¹⁶) and hepatocytes of mice (Đikić et al. ²⁰¹²), and also induced aneuploidy in hamster oocytes (Jeffay et al. ¹⁹⁹⁶) and mouse bone marrow (Sarrif et al. ^1994b). Like the malathion treatment of our work, 3‐aminotriazole (a herbicide), urethane (a component of pesticides) and other genotoxic substances increased spontaneous retrotranspositions in yeast (Terziyska 2003; Pesheva et al. ²⁰⁰⁸). Malathion when fed along with agar‐sucrose‐medium to certain strains of Drosophila, it induced whole genome transpositions (Salces‐Ortiz et al. ²⁰²⁰). The influence of all treatments, except malathion, on transpositions was strange because their rates were significantly reduced; the reason for which is not understood. The contradictory behaviour of transpositions in our experiments due to treatment with other pesticides was similar to the previous report where the Arabidopsis Tag1 lines were treated with various strains of the bacteria Agrobacterium tumefaciens (Shah et al. ²⁰¹⁵). Transposons exhibit differential responses depending on the nature of stress because transpositions are regulated via stress‐induced regulatory mechanisms involving transcription factors and epigenetic modifications (Capy et al. ²⁰⁰⁰; Slotkin & Martienssen ²⁰⁰⁷; Ito ²⁰²²). It could be possible that such regulatory mechanisms responded differently to the pesticides used in this work. Somatic mutations using the pesticides that we used were previously screened using prokaryotic system. For example, mancozeb (Shukla et al. ²⁰⁰⁴), but not carbendazim (Sarrif et al. ^1994a), induced PM and FSM in the bacteria S. typhimurium. Unlike our results, malathion did not seem to induce somatic mutations such as PM and SHR when analysed using the prokaryotic systems (based on Escherichia coli and Salmonella typhimurium), as well as other eukaryotic systems (based on yeast, mouse cell lines and Drosophila) (Waters et al. ¹⁹⁸²; Imamura & Talcott ¹⁹⁸⁵; Osaba et al. ¹⁹⁹⁹). The lack of detection of somatic mutations due to malathion in the eukaryotic models could be because these tests were done at very low concentrations due to the lethal effect of malathion itself (Osaba et al. ¹⁹⁹⁹). Similarly, mutagenic effects were not detected due to many other pesticides also previously (Waters et al. ¹⁹⁸²; Osaba et al. ¹⁹⁹⁹; Pitchakarn et al. ²⁰²¹). However, there are many reports of other genotoxic effects such as chromosomal aberrations and micronuclei formation due to malathion in mouse, humans and Drosophila (USHHS 2003). Though chlorantraniliprole and flubendiamide were thought to exhibit no genotoxicity based on the tests on bacterial system (USEPA; California Environmental Protection Agency), these did induce DNA damage in rat tissues (Mandil et al. ²⁰²⁰; Omar et al. ²⁰²²). Similar to our work, there are examples of loci‐specific mutation detections due to other pesticides using Arabidopsis and other model systems. For example, increased SHR rates were scored previously in Arabidopsis plants exposed to atrazine (Besplug et al. ²⁰⁰⁴), rose bengal and aminotriazole (Kovalchuk et al. ²⁰⁰³), paraquat (Kovalchuk et al. ²⁰⁰³; Pecinka et al. ²⁰⁰⁹). Similarly, 2,4‐D and dicamba increased both SHR and PM rates in Arabidopsis (Filkowski et al. ²⁰⁰³). Four types of organophosphate insecticides (methyl parathion, azamethiphos, dichlorvos and diazinon) induced a dose‐dependent increase in somatic mutations such as PM and SHR in Drosophila melanogaster (Çakir & Sarikaya ²⁰⁰⁵). Similarly, various herbicides such as clomazone, linuron and simazine (Castañeda‐Sortibrán et al. ²⁰¹⁹), and bentazone, molinate, thiobencarb and trifluralin (Kaya et al. ²⁰⁰⁴) also induced such mutations in Drosophila. The fungicide maneb, an ethylene bisdithiocarbamate, also induced similar mutations in Drosophila (Osaba et al. ²⁰⁰²). FSM rates were increased in E. coli and Salmonella due to Dexon and Nirit (Kada et al. ¹⁹⁷⁴), and in Salmonella due to Napoleon 4EC (Akyil 2021). In any case, we did not come across reports on the quantification of PM, SHR, FSM and transposition rates in higher eukaryotic system due to any of the pesticides that we used.

The pesticides, malathion, flubendiamide, chlorantraniliprole and carbendazim + mancozeb combination are routinely used in India. We used their commercial formulations because these often give results different from their active principles (Mesnage et al. ²⁰¹⁴; Nagy et al. ²⁰²⁰), and they also represent the situational reality because the ecosystem is exposed to such formulations. We used the carbendazim + mancozeb combination because it is a commercial formulation used to tackle broad‐range fungi, and this combination exhibited greater hematotoxicity, hepatotoxicity and genotoxicity in rats than the individual components (Hashem et al. ²⁰¹⁸). A. thaliana mutants were used because they were tolerant to the commercially used doses of pesticides and offered a convenient detector system to screen PM, SHR, FSM and transpositions (Filkowski et al. ²⁰⁰³; Besplug et al. ²⁰⁰⁴). This model plant previously proved to be an effective eukaryotic biosensor model system to screen the rates of multiple loci‐specific somatic mutations (PM, FSM, SHR and transpositions) under various stresses such as harmful chemicals, radiations and pathogens (Kovalchuk et al. ²⁰⁰¹; van Norman & Benfey ²⁰⁰⁹; Sijila & Shah ²⁰¹⁶). As Arabidopsis plants are of small size with a short life cycle (about 6 weeks), these offer an added advantage of analysing a large population across multiple generations. These plants are a highly sensitive and ethically appealing bioassay system to monitor pollution‐induced mutations (Kovalchuk & Kovalchuk ²⁰⁰⁸; Basu & Kovalchuk ²⁰¹⁰).

As there were previous reports of intergenerational and transgenerational effects of pesticides on the physiology of organisms (Sarkar et al. ^2018b; Montaño‐Campaz et al. ²⁰²²; Nicolella & de Assis ²⁰²²), we wanted to analyse the status of somatic mutation rates in progeny. Though PM rates were normal in the progeny, the rates of SHR and FSM were still significantly high under all treatments. It is to be noted that only the parents were treated with pesticides. When it came to the third generation (grand progeny of treated plants), the trends of SHR and FSM rates behaved differently. While SHR rates drastically reduced than the untreated controls under two treatments, the FSM rates were still significantly high. Thus, when comparing the three generations, PM, SHR and FSM exhibited different patterns of inheritance. The reason for the reduced SHR rates is not known, indicating that much study needs to be done in this direction. Similar to our results for the heritability of somatic mutation rates, there are previous examples of other mutagens such as UVC radiation (Boyko et al. ²⁰⁰⁷; Kathiria et al. ²⁰¹⁰), Bleomycin, zebularine (Pecinka et al. ²⁰⁰⁹) and paraquat which resulted in increased SHR frequency in progeny. Molinier et al. (2006) reported that increased SHR frequency triggered by UVC persisted for five generations. There is only one report on pesticide‐induced intergenerational memory, where paraquat led to enhanced SHR rates in the progeny of the treated plants (Pecinka et al. ²⁰⁰⁹). Similar to our reduced SHR rates in the grand progeny, Pecinka et al. (2009) also observed inheritance of reduced SHR rates due to NaCl, mannitol and zebularine. Our results indicate that pesticides are capable of triggering a mutagenic memory that is able to penetrate into the next generation. Thus, the epigenetic memory can be attributed to the maintenance of genome stability also.

As mutations are a consequence of altered genome stability, which is maintained by many DNA repair genes, we analysed the expression of 15 DNA repair genes belonging to five different pathways. Of these, about 12 genes had significantly increased expression in response to at least one of the pesticides. Each gene exhibited a unique pesticide‐specific expression behaviour. While flubendiamide and carbendazim + mancozeb combination brought about a significant increase in the expression of 8 and 7 genes, respectively, only 3 and 4 genes gave a similar response to malathion and chlorantraniliprole, respectively. MSH4 and PMS2L3 belonging to the mismatch repair pathway were the only genes that responded significantly to all four pesticide treatments. Though most genes showed an overall increased expression in all treatment, considering the number of genes and their treatments, only 1/3rd of cases exhibited significantly increased expression. It could be possible that the highest somatic mutation rates were induced by malathion due to compromised DNA repair gene expression. However, we cannot generalise this statement because the same explanation does not fit with the observations by chlorantraniliprole. It increased the expression of least number of four genes after malathion and yet induced least somatic mutation rates. It could be possible that chlorantraniliprole probably affected the expression of an alternate set of DNA repair genes. Hence, this work paves way for more such analysis to be done. Previously, pesticide‐induced expression of a DNA repair gene MSH4 was reported in pig due to azadirachtin (Sheng‐Fei et al. ²⁰¹⁹) and dimethomorph in human Jurkat T cell (Li et al. ²⁰²²). There is a report on the induction of PMS2L3 expression upon various environmental factors that induce oxidative stress in both rodent and human tissues (Sone et al. ²⁰¹⁰). Similar to our report, previous treatment of Arabidopsis with the herbicide paraquat also resulted in a less pronounced enhancement of the expression of DNA repair genes involved in HR (RAD51, BRCA1, ATM and MIM) (Pecinka et al. ²⁰⁰⁹). They also reported that the expression of these genes was much higher when the plants were subjected to other stresses such as UVC, bleomycin and zebularine. In all cases studied by Pecinka et al. (2009), the response of increased expression was retained only to the treated generation and did not transfer to the progeny like that of the somatic mutations. Other previous reports on DNA repair gene expression studies under pesticide treatment include increased development of breast cancer by altering DNA repair genes (Scandolara et al. ²⁰²²). Chronic exposure to pesticides showed a decreased expression of DNA repair genes in humans (Costa et al. ²⁰²¹). These results suggest that genotoxic exposure to pesticides negatively affects DNA repair mechanisms and increases genome instability.

Mutagenic stresses often lead to methylation/demethylation of certain loci or whole genome in plants (Sun et al. ²⁰²²). Differential methylation of promoter regions across the genome was seen in rats subjected to the fungicide vinclozolin (Guerrero‐Bosagna et al. ²⁰¹⁰). In this work, eight out of 21 DNA repair gene promoters showed altered methylation status under treatment with the four pesticides. All cases were of demethylation. Previously, Qiao et al. (2022) reported demethylation of rice OsBR6ox, a member of the brassinosteroid synthetic pathway on exposure to the herbicide atrazine. Though methylation is generally associated with reduced gene expression, there are reports conveying that methylation can enhance the same (Harris et al. ²⁰¹⁸). In our study also, enhanced expression was seen in most cases, irrespective of the methylation status of the promoters. There are many previous reports of pesticide‐induced genome‐wide methylation alterations in humans (Mahna et al. ²⁰²¹; Rohr et al. ²⁰²⁴). Altered DNA methylations associated with genes related to respiratory organs (van der Plaat et al. ²⁰¹⁸; Hoang et al. ²⁰²¹), haematopoietic tissues (Navarrete‐Meneses et al. ²⁰²³), carcinogenesis‐related genes (Zhang et al. ²⁰¹²; Paredes‐Céspedes et al. ²⁰¹⁹), Parkinson's disease (Schaffner et al. ²⁰²⁴) and LINE 1 DNAm transposable element (Alexander et al. ²⁰¹⁷) were reported in humans subjected to occupational exposure to pesticides. However, studies focusing on DNA repair genes alone were not found. Also, the previous reports were not based on the pesticides used in this work.

Similar to the heritable somatic mutations, we went on to see whether pesticide‐induced altered methylation were also inherited. The promoter of one gene, MRE11A (belonging to the HR pathway), retained the altered methylation status (demethylation) in the next generation also. All four pesticide treatments were capable of inducing this intergenerational demethylation response. Of these, only malathion treatment retained demethylation in the grand progeny plants also. There are a few previous reports on transgenerational hypomethylation of specific loci on exposure to various stresses in plants (Lämke & Bäurle ²⁰¹⁷). For example, progeny of TMV infected plants showed locus‐specific hypomethylation in LRR‐containing loci (Boyko et al. ²⁰⁰⁷). Similarly, heavy metal stress induced locus‐specific hypomethylation of transposable elements and certain protein‐coding genes (Ou et al. ²⁰¹²). When it comes to DNA repair gene promoters, intergenerational demethylation in RAD23C, RAD23D, intergenerational methylation in BLM, BRCA2 and transgenerational methylation in CEN2, RAD51 and LIG4 were seen in Arabidopsis plants exposed to certain strains of A. tumefaciens (Joseph et al. ²⁰²¹). Reports on pesticide‐induced transgenerational differential methylation are few. The insecticides, permethrin, N, N‐diethyl‐meta‐toluamide (Thorson et al. ²⁰²⁰) and dichloro‐diphenyl‐trichloroethane (Beck et al. ²⁰²¹), the fungicide vinclozolin and the herbicide atrazine (Nilsson et al. ²⁰¹⁸) induced heritable differential methylations in various disease‐associated genes in rats. Permethrin also induced such changes in genes associated with the glutamatergic signalling pathway in zebra fish (Blanc et al. ²⁰²¹). Linuron, a herbicide, induced transgenerational epigenetic changes in frog pancreatic genes associated with calcium signaling, type 2 diabetes and a biomarker of pancreatic ductal adenocarcinoma (Roza et al. ²⁰²⁴). However, we did not come across any reports on transgenerational differential methylation due to the four pesticides used in this work. The MSRD‐PCR method used here does not cover the entire promoter and other regulatory sequences. Nevertheless, it has revealed the consistency in altered methylation status at certain bases. This indicates that more studies need to be carried out at the whole epigenome level to get a deeper understanding of the impact of various pesticides.

Plants exposed to stress can pass this memory to their offspring, and this phenomenon is of an epigenetic nature (Boyko & Kovalchuk ²⁰¹⁰). Transgenerational responses to stresses are often associated with adaptation (Boyko & Kovalchuk ²⁰¹⁰; Quadrana & Colot ²⁰¹⁶). Dyachenko et al. (2006) reported that Mesembryanthemum crystallinum when exposed to salt stress underwent changes in satellite DNA methylation which helped them to adapt to C4 type carbon dioxide assimilation. Heavy metal stress induced transgenerational adaptation in rice plants (Ou et al. ²⁰¹²). DNA methylation was proposed to be the reason for adaptation to stress by A. thaliana plants and pine trees in the radioactively contaminated areas of Chernobyl (Kovalchuk & Kovalchuk ²⁰⁰³; Kovalchuk et al. ²⁰⁰⁴). Though the purpose of various transgenerational responses observed in our work is not understood, this evidence does warn us about such silent changes occurring in the living systems. Owing to their sessile nature, plants are forced to tolerate undesirable situations and hence are more adapted to survive in adverse conditions. By contrast, non‐target animals/humans are more vulnerable, as evidenced by the physiological changes in pesticide‐prone locations.

CONCLUSIONS

Four pesticides, malathion, flubendiamide, chlorantraniliprole and carbendazim + mancozeb combination, had the potential to increase the rates of loci‐specific somatic mutations (PM, SHR and FSM) in the non‐target organism, A. thaliana. The increased rate of FSM was retained in the unexposed progeny as well as grand progeny. SHR rate was high in progeny, whereas it was suppressed in grand progeny. The expression and promoter methylation status of many DNA repair genes were altered by all four pesticides. Intergenerational demethylation was seen in the MRE11A promoter due to all four pesticides. Malathion alone induced transgenerational demethylation of the MRE11A gene promoter. Thus, our study provides strong evidence for the transgenerational mutagenic and epigenetic impact due to exposure to various pesticides. These findings have significant implications on agriculture and the environment, highlighting the need for greater caution and stricter regulations for the use of pesticides. The research suggests the importance of studying epigenetic changes in addition to genetic mutations, as they play a critical role in shaping the effects of environmental factors on organisms. This work reemphasises that though required for commercial agriculture, this anthropogenic activity be carried out with more caution not only for mankind's safety, but also to preserve the non‐target genetic/epigenetic diversity.

Author Contributions

Conceptualisation: JMS. Performed the experiments: SJ, RR, CD, VB, AEP. Data curation: SJ, AKS, JMS. Formal analysis: SJ, AKS, JMS. Methodology: SJ and JMS. Supervision: JMS. Writing original draft: SJ. Writing review and editing: JMS and AKS. All the authors have read and approved the final manuscript.

Funding information

This research did not receive any specific grant from funding agencies in the public, commercial or not‐for‐profit sectors.

Conflict of interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Supporting information

Table S1. P values of respective treatments in various lines after Kruskal–Wallis test.

Table S2. P values of respective treatments in progenies of various lines after Mann–Whitney test.

Fig. S1. Candidate reference gene stability under four different pesticide treatments, based on (a) Box Plot analysis where, the horizontal lines inside the boxes refer to median Ct values; the upper and lower hinges of the box indicate 75 and 25 percentiles. The whiskers show the largest and smallest Ct values that fall within a distance of 1.5 times the interquartile range from 75 and 25 percentiles. (b) BestKeeper. (c) NormFinder. (d) geNorm. Most stable genes are represented on the left side of the x‐axis, and stability gradually decreases towards right side of the x‐axis in (b–d). Gene stability values are represented on y‐axis in (b–d).

Fig. S2. Expression profile of DNA repair genes belong to various repair pathway. C‐ Control, M‐Malathion, F‐Flubendiamide, Co‐ Combination of Carbendazim+Mancozeb, Ch‐Chlorantraniliprole. Symbols, *, ◊, ■, □, indicate that the respective treatments are significantly different from each other at P < 05 by Tukey's HSD post hoc ANOVA. Vertical line on the top indicates the standard error of the mean of three biological replicate. Number on top of the vertical line shows the respective mean of expression value. Vertical axis shows the mean of expression value. (a–e) represent expression of the nucleotide excision repair genes RAD23C, RAD23B, XAB2, CEN2, XPf; (f–i) represent expression of the base excision repair genes, MBD4, TAG1, P3MAG1, XRCC1; (j, k) represent expression of the homologous recombination genes RAD51, ATM; (l, m) represent expression of the mismatch repair genes MSH4, PMS2L3; and (n, o) represent expression of the non‐homologous end joining genes LIG4, KU70, respectively.

Fig. S3. MSRD‐PCR images of DNA repair gene promoters upon various pesticide treatment. The methylation‐sensitive restriction enzyme (MSRE) used for digestion is shown on left side column of the gel image. The lanes L, N, U, C, M, F, Co, Ch and E indicate 100‐bp ladder, non‐template water control, undigested DNA control, digested DNA of untreated control plant, digested DNA of malathion‐treated plant, digested DNA of flubendiamide‐treated plant, digested DNA of carbendazim+mancozeb combination‐treated plant, digested DNA of chlorantraniliprole‐treated plant and empty lane, respectively. D indicates demethylation; NC indicates no change in comparison with the untreated control. *Denotes presence of nonspecific band, not corresponding to the expected size, as seen in lane U.

Fig. S4. MSRD‐PCR images of DNA repair gene promoters that showed altered methylation patterns in second generation. The methylation‐sensitive restriction enzyme (MSRE) used for digestion is shown on left side column of the gel image. The lanes L, N, U, C, M, F, Co and Ch indicate 100‐bp ladder, non‐template water control, undigested DNA control, digested DNA of untreated control plant, digested DNA of malathion‐treated plant, digested DNA of flubendiamide‐treated plant, digested DNA of carbendazim + mancozeb combination‐treated plant and digested DNA of chlorantraniliprole‐treated plant, respectively. D indicates demethylation; NC indicates no change in comparison to the untreated control.

Fig. S5. MSRD‐PCR image of MRE11A gene promoter showing different methylation patterns in third generation. The methylation‐sensitive restriction enzyme (MSRE) used for digestion is shown on left side column of the gel image. The lanes U, C, M, F, Co and Ch, and L indicate undigested DNA control, digested DNA of untreated control plant and, digested DNAs of plants treated with four pesticides—malathion, flubendiamide, carbendazim + mancozeb combination and chlorantraniliprole, respectively. *indicates presence of nonspecific band, which are seen in multiple lanes (except U) and do not correspond to the expected size (as seen in lane U). D is demethylation. R is reversion to methylation.