Transgenerational changes in plant physiology and in transposon expression in response to UV-C stress in Arabidopsis thaliana

Abstract

Stress has a negative impact on crop yield by altering a gain in biomass and affecting seed set. Recent reports suggest that exposure to stress also influences the response of the progeny. In this paper, we analyzed seed size, leaf size, bolting time and transposon expression in 2 consecutive generations of Arabidopsis thaliana plants exposed to moderate UV-C stress. Since previous reports suggested a potential role of Dicer-like (DCL) proteins in the establishment of transgenerational response, we used dcl2, dcl3 and dcl4 mutants in parallel with wild-type plants. These studies revealed that leaf number decreased in the progeny of UV-C stressed plants, and bolting occurred later. Transposons were also re-activated in the progeny of stressed plants. Changes in the dcl mutants were less prominent than in wild-type plants. DCL2 and DCL3 appeared to be more important in the transgenerational stress memory than DCL4 because transgenerational changes were less profound in the dcl2 and dcl3 mutants.

Abbreviations

(−)

grown under normal conditions; (+), grown under stressed conditions; C₁, the progeny of plants grown under normal conditions in F₀; C₁S₁, the progeny of plants grown under normal conditions in F₀ and exposed to stress in F₁; C₂, the progeny of plants grown under normal conditions for 2 generations; DCL, a Dicer-like protein; dpg, days post germination; ONSEN, a copia-type retrotransposon; S₁, the progeny of plants exposed to stress in F₀; S₁C₁, the progeny of plants exposed to stress in F₀ and grown under normal conditions in F₁; S₂, the progeny of plants exposed to stress for 2 generations; TEs, transposable elements; TSI, transcriptionally silent information; UV-C, ultraviolet radiation C

Introduction

Plant evolution is influenced by frequent changes in the environment. Fluctuations in the environment known as external stresses are continuously forcing plants to adapt and acclimate. The response to stress includes a response at the molecular and cellular level that manifests itself as changes in plant physiology and morphology.¹ These changes often have transient characteristics and disappear after the stress is over. However, sometimes these responses are propagated into the progeny.^2,3 The observed changes are adaptive or maladaptive in nature and are known as the transgenerational response to stress. These changes are often not propagated any further and therefore are not heritable. Some of these modifications, however, can be maintained across several generations, and therefore they are true heritable transgenerational changes. It is impossible to explain such changes in progeny by genetics because mutations are rare and random in nature; therefore they cannot prepare the progeny to face continuous fluctuations in the environment in a timely manner.⁴ In contrast, epigenetic modifications that do not alter DNA sequence but rather change gene expression through DNA methylation, histone modifications and differential expression of small non-coding RNAs are flexible in nature and can help plants respond to the environment and let this memory last.⁵

Since plants are sedentary in nature, they cannot avoid environmental exposures such as fluctuations in temperature or the irradiation with UV light (UV). High temperature stress is often paralleled with UV stress, therefore plants are mostly exposed to UV-A and UV-B. With the depletion of the ozone layer, plants will be exposed to UV even more often. UV light damages plant DNA and alters plant metabolism, including an increase in the production of phenolic compounds and the inhibition of photosynthesis.¹ These changes result in alterations in plant growth, a decrease in plant biomass and seed set.⁶

Our previous experiments showed that exposure to UV-B and UV-C increased the frequency of homologous recombination and point mutations in plants.^7,8 It was also documented that the progeny of exposed plants had similar changes even when these plants were grown under normal conditions.⁹ In previous experiments, we showed that transgenerational changes depended on the function of Dicer-like proteins DCL2 and DCL3.³ We also found that the progeny of salt-stressed plants exhibited changes in DNA methylation and histone modifications.¹⁰

In this work, we attempted to analyze whether the progeny of plants exposed to UV-C would exhibit changes in plant physiology while grown under normal conditions or further exposed to UV-C. Our experiments showed that leaf number decreased and bolting occurred later in the progeny of UV-C stressed plants. We also found that dcl2 and dcl3 mutants were partially impaired in transgenerational changes.

Methods and Materials

Plant growth conditions

Arabidopsis thaliana lines 15d8 (wild-type), dcl2, dcl3, and dcl4 mutants (Columbia ecotype) were germinated and grown in potting soil with vermiculite (at 7:1 ratio) at long day-length (16/8 h, 22°C/18°C, on a day/night shift cycle) under the following light conditions: 32.8 μEm⁻²s⁻¹ of the white light at the wave length with 2 main peaks of 540 and 610 nm and a constant humidity of 65% in 5 × 5 cm pots. At 5 d post germination (dpg), plants were transplanted into pots containing the same soil/fertilizer ratio with a total of 12 plants per pot and 2 pots per sample group, and with a total of 24 plants per treatment group. At approximately one-week post germination, plants were exposed to 4 minutes of UV-C irradiation (G30T8) of 30.5 W and 99 V, and UV output of 13.9 W. Seven days post-stress, leaf size (the length and width) of the third youngest leaf on each plant (24 plants per treatment group) was measured, and rosette leaves were flash-frozen using liquid nitrogen and stored at −80°C. Four samples of approximately 100 mg of tissue were collected from each treatment group.

Bolting was assessed on each plant (excluding plants from which tissue samples were taken) at 4 weeks post germination. Seeds were collected from these plants and photographed under the microscope. Seed length was measured using Image J for approximately 100–200 seeds per treatment.

Real time PCR for the analysis of transposon activity

RNA was isolated from 100 mg of plant tissue using Trizol reagent (Invitrogen) as published before (REF). cDNA was prepared from mRNA using an iScript Select cDNA synthesis kit (Bio-Rad) according to the manufacturer's protocol.

Quantitative real-time PCR was performed using SsoFast EvaGreen Supermix (Bio-Rad) and either the promoter- or gene-specific primers. The amplification of ONSEN and TSI occurred under the following conditions: (1) at 98°C for 2 min for one cycle; at 98°C for 5 s, Tm+1°C for 5 s, for 40 cycles; (2) melt-curve analysis–at 65°C to 95°C for 5 s, with a 0.5°C increment. Tm+1°C was altered according to the primers. Tubulin was used as a control. The primers specific for ONSEN were used: forward–5′ ccacaagaggaaccaacgaa 3′; reverse–5′ ttcgatcatggaagaccgg 3′. TSI- specific primers: forward –5′ accatcaaagccttga-gaagca3′; reverse–5′ ccgtatgagtctttgtctttgtatctt3′. Tubulin-specific primers: forward – 5′ACAGAAGCGGAGAGCAACAT3′; reverse–5′ TCCTCATCCTCGTAGTCACCTT3′. Each reaction was repeated 3 times.

Statistical analysis

Statistical analysis of physiological data including leaf number, length and width, and seed length were performed using R software as previously described.¹¹ A 95% confidence interval was calculated for each measurement under both stressed and normal conditions, and confidence intervals were compared with bootstrap ×10,000 to determine if 2 intervals were significantly different from each other. The response of plants under stress was determined by dividing the measurement of stressed plants by the measurement under normal conditions. If the resulting value overlapped with 1 at a 95% confidence interval (p = 0.05), no significant changes occurred. Comparisons were done between wild-type and dcl mutants in order to determine if the difference between the 2 responses was significant at a confidence interval of 95%. These results were graphed using Prism (Graphpad) software. Bolting time was determined as a percentage calculated by dividing the number of bolted plants by the total number of plants measured, and graphed using Microsoft Excel. Transcription results for transposons were graphed and analyzed using Prism (Graphpad) software. Standard deviations and standard errors of the mean were given from Q-PCR (BioRad) results, and a t-test was performed to determine differences where the significance was given to the result with a P-value ≤0.05.

Results

Experimental set up

Experiments described in this paper started with F₁ seeds collected from the progeny of stressed (S₁) and control (C₁) plants. These seeds consisting of wild-type (15D8) and 3 dcl mutants (dcl2, dcl3 and dcl4) were grown under both stressed (+) and normal (−) conditions, giving rise to 4 groups of plants- C₁−, C₁+, S₁− and S₁+ for each mutant type (Fig. 1). Seeds from these plants (F₂) were collected and germinated, giving rise to the control progeny of control plants (C₂), the stressed progeny of control plants (C₁S₁), the control progeny of stressed plants (S₁C₁) and the stressed progeny of stressed plants (S₂). These F₂ plants were grown under both stressed and normal conditions too, and the corresponding measurements were taken for all resulting 8 groups, as described in Figure 1.

Figure 1.

Experimental set up. Arabidopsis plants were exposed to UV-C (S₀) or sham-treated (C₀). Seeds from these plants were collected giving rise to the progeny of stressed plants (S₁) and the progeny of control plants (C₁). These F₁ seeds were then either stressed with UV-C again or sham-treated, thus producing 4 groups of F₂ plants: S₂, S₁C₁, C₂ and C₁S₁.The same steps were completed for 15D8, dcl2, dcl3 and dcl4 plants.

Changes in the F₁ generation

Seed size

Seeds produced by dcl mutants grown under normal conditions (C₁) were larger than wt seeds, and the difference was significant for dcl2 and dcl3 seeds (Fig. 2A). The response to stress was calculated by dividing S₁ seed length by C₁ seed length. UV-C exposure did not change the size of seeds in wt plants (Fig. 2B). In contrast, UV-C increased seed size in dcl3 and decreased it in dcl2 and dcl4 plants (Fig. 2B).

Figure 2.

The size of seeds in the progeny of UV-C stressed and control plants. (A) The size of F₁ seeds produced by plants grown under either normal condition (C₁) or UV-C stress (S₁). Approximately 100–200 seeds were measured from each group. Confidence intervals of 95% (p = 0.05) were calculated using bootstrap ×10,000. The asterisks (*) are used to indicate mutants that vary significantly in size from wild-type plants (15D8) belonging to the same treatment group. The legend indicates the type of mutant. (B) The ratio of changes in the size of F₁ seeds in response to UV-C. The ratio was calculated by dividing the size of seeds produced by stressed plants (S₁) by the size of seeds produced by plants grown under normal conditions (C₁) using bootstrap ×10,000 and a confidence interval of 95% (p = 0.05). The ratio of 1 indicates no changes in seed size. The asterisks (*) over the bars are used to indicate mutants that vary significantly in size from wild-type plants (15D8) belonging to the stressed group. The asterisks over the lines indicate a significant difference between different groups of the same mutant type as calculated at P ≤ 0.05. The legend indicates the type of mutant.

Changes to leaf number

Leaf number in the progeny of control plants grown under normal conditions (C₁−) was similar among all mutants except dcl2 plants which had more leaves than other plants (Fig. 3A). In C₁ plants exposed to UV-C, leaf number was similar among all plants except dcl3 plants in which it was smaller than in wt plants. A comparison of leaf number between C₁+ and C₁− plants showed that immediate UV-C exposure increased leaf number in all plants except dcl2 in which it was significantly different from that in wild-type plants (Fig. 3B).

Figure 3.

Measurements of leaf number (A), leaf length (C) and leaf width (D) in the F₁ progeny of wild-type plants (15D8) and mutant plants (dcl2, dcl3 and dcl4) exposed to UV-C and control plants. “C₁”–the progeny of plants grown under normal condition in F₀. “S₁”–the progeny of plants exposed to UV-C in F₀. “+” and “−” indicate either exposure to stress or growth under uninduced conditions, respectively. The values indicate a 95% confidence interval calculated using bootstrap ×10,000 as a result of approximately 24 repeats. The asterisks (*) indicate a significant difference either between wild-type and mutant plants in the same treatment group or between the S₁ and C₁ groups of the same mutant type, as calculated at P ≤ 0.05. (B) Changes in response to UV-C stress. The bars represent the ratio of changes in F₁ plants in response to UV-C stress (S+/S− or C+/C−) at a 95% confidence interval, the P-value of 0.05. These ratios are the result of 10,000 bootstrap analysis. The ratios that overlap the value of 1 indicate no significant changes under UV-C stress.

In the non-stressed progeny of stressed plants (S₁−), mutants had more leaves than wt plants (Fig. 3A). A comparison of leaf number between S₁− and C₁− plants showed that the progeny of stressed wt and dcl2 plants had fewer leaves than the progeny of non-stressed plants, although the difference was not significant (Fig. 3A).

Leaf number in the UV-stressed progeny of stressed plants (S₁+) was similar in all plants but dcl4 which had fewer leaves than any other group (Fig. 3A). A comparison of leaf number between S₁+ and S₁− groups showed that all dicer mutant plants had fewer leaves in response to UV-C stress, whereas wt plants did not have any changes in leaf number (Fig. 3B).

Changes to leaf size

Leaf length and leaf width in C₁− were smaller in dcl2 and dcl4 mutants as compared to wt plants (Fig. 3C and D). In contrast, in the C₁+ group, leaf size was similar among all mutants and wild-type plants. A comparison between of C₁+ and C₁− groups showed that leaf size decreased in wt and dcl3 plants and increased in dcl2 and dcl4 mutants in response to UV-C (Fig. 3B). The progeny of UV-stressed wt, dcl2 and dcl4 plants (S₁−) had larger leaves under normal conditions than the progeny of control plants (C₁−), although the difference was significant only for leaf width in wt and dcl2 plants (Fig. 3C and D). In contrast, the progeny of UV-stressed dcl3 plants had smaller leaves. S₁+ plants had leaves of a similar size, except dcl3 plants where leaf length was smaller (Fig. 3C and D). In the progeny of stressed plants exposed to UV-C (S₁+), leaf size decreased in wt plants and did not change in dcl mutants as compared to the non-exposed progeny of stressed plants (S₁−). Changes in leaf width were significantly different in mutants as compared to wt plants in the S₁ group in response to UV-C (Fig. 3B).

Changes to bolting time

Wt plants increased their bolting rate in response to UV-C stress regardless of parental treatment, while dcl2 and dcl3 plants decreased it (Fig. 4). Bolting in dcl4 plants increased in the C₁+ group only. Changes in bolting rates in response to UV-C were more similar between 15D8 and dcl4, and dcl2 and dcl3 plants. Parental treatment affected bolting in plants grown under normal conditions; S₁− had the lower rate of bolting than C₁− plants, suggesting that the transgenerational response to UV-C manifests itself as the decreased rate of bolting in the progeny (Fig. 4).

Figure 4.

The percentage of F₁ plants that bolted at approximately 4 weeks of age. The plants were grown under either normal (−) or stressed (+) conditions. Each treatment group is labeled according to the treatment and mutant type on the horizontal axis. Approximately 24 plants were included in each treatment group. The Y-axis shows the percentage of bolted plants, whereas the X-axis–the type of mutants and treatment.

Changes in transposon expression

The expression of ONSEN was higher in 15D8 than in dcl2 and dcl4 plants but lower than in dcl3 in the non-stressed C₁ group (C₁−) (Fig. 5A). Stress exposure in the C₁ group (C₁+) increased ONSEN expression in 15D8, dcl2 and dcl4 and decreased it in dcl3 plants. The unstressed progeny of stressed plants (S₁−) had a higher expression of ONSEN as compared to the C₁− group in all plants but dcl3 (Fig. 5A). Exposure to stress in the S₁ group (S₁+) did not change ONSEN expression in 15D8 and dcl2 plants but increased it in dcl3 and dcl4 plants.

Figure 5.

The expression of ONSEN and TSI transposons in the F₁ progeny (A, B) in wild-type plants (15D8) and mutant plants (dcl2, dcl3 and dcl4) exposed to UV-C and control plants. The Y-axis shows arbitrary units of gene expression. “+” and “−” indicate exposure to stress or growth under uninduced conditions, respectively. The bars show the standard deviation calculated from 3 technical repeats. The asterisks (*) indicate a significant difference between control (−) and stress (+) plants with the same parental treatment as calculated using a t-test (P ≤ 0.05).

TSI expression was not detectible in wt and dcl2 plants, but it was relatively high in dcl3 and dcl4 plants in the non-exposed C₁ group (C₁−) (Fig. 5B). UV-C exposure increased TSI expression in the C₁ group (C₁+) in wt and dcl2 plants and decreased it in dcl3 and dcl4 plants. The transgenerational response to UV-C (S₁− compared to C₁−) showed that TSI expression increased in all plants but dcl3 where it actually decreased (Fig. 5B). In S₁ plants (S₁+ vs S₁−), TSI expression did not change in wt and dcl3 plants, and it actually decreased in dcl2 and dcl4 plants in response to UV-C.

Changes in the F2 generation

Changes to seed size

All dcl mutants had significantly smaller seeds in the C₂ group (Fig. 6A). All 15D8 seeds produced by plants stressed for either one or 2 generations were larger than those produced by C₂ plants, although the difference was only significant for the C₁S₁ group. Similarly, exposure of C₁ dicer mutant plants to stress resulted in an increase in seed size in the C₁S₁ group in all but dcl4 plants. Both the S₁C₁ and S₂ groups of mutants had significantly larger seeds than the C₂ group.

Figure 6.

The size of seeds produced in F₂ plants. (A) The size of F₂seeds produced by plants grown under either normal conditions (C₁) or UV-C stress (S₁). Approximately 100–200 seeds were measured from each group. Confidence intervals of 95% (p = 0.05) were calculated using bootstrap ×10,000. The asterisks (*) are used to indicate mutants that vary significantly in size from wild-type plants (15D8) belonging to the same treatment group. The legend indicates the type of mutants. (B) The response of F₂ seeds to UV stress was calculated as seed length produced by stressed plants divided by seed length of seeds produced by either non-stressed plants or S₂/S₁C₁ and C₁S₁/C₂. Plants were divided based on F₁ parental treatment under either stressed (S₁) or normal growth conditions (C₁). Approximately 100–200 seeds were measured from each group. The ratios were calculated using bootstrap ×10,000 and a confidence interval of 95% (p = 0.05). The ratio of 1 indicates no changes in size, and bars overlapping with 1 show no significant changes in seed length. The asterisks (*) are used to indicate mutants that vary significantly in size from wild-type plants (15D8) with the same parental treatment. The legend indicates the type of mutants. (C) Measurements of natural variations in seed length in dcl mutants and wild-type plants (15D8) following one (C₁) or 2 (C₂) generations of growth under normal conditions. The bars represent a confidence interval of 95% (p = 0.05) for plant type and a variable calculated by using bootstrap ×10,000. The asterisks (*) indicate mutants that vary significantly in comparison to wild-type plants of the same generation. Approximately 100–200 seeds were sampled from each group. The horizontal axis indicates the generation. The legend indicates the type of mutants.

To estimate the response to stress in theF₂ generation, we divided seed length under stress in the S₂ and C₁S₁ groups by seed length under normal conditions in the S₁C₁, and C₂ groups, respectively. The immediate response to stress resulted in an increase in seed size in all plants but dcl4 (the response in C₁ shows a ratio of C₁S₁ to C₂; Fig. 6B). The response of wt plants in S₂ vs S₁C₁ did not change; in contrast, seed size increased in dcl2 and decreased dcl3 and dcl4 plants (Fig. 6B).

Next we performed a comparison between seed size in the C₁ and C₂ generations and found that wt seeds did not change between 2 generations of plants grown under normal conditions (Fig. 6C). In contrast, seeds of dcl mutants decreased in size when the C₁ generation was compared with the C₂ generation. Seeds of dcl mutants were larger than wt seeds in C₁ but smaller in C₂ (Fig. 6C).

Changes to leaf number

Leaf number was similar among all unexposed plants in the C₂ group (Fig. 7A). Leaf number increased in dcl2 plants of C₂ group after UV-C exposure, but it did not change in other plants (Fig. 7A and B). Leaf number in the C₁S₁ group was similar to that in the C₂ group, although dcl4 plants had more leaves than other plants. Exposure of the C₁S₁ group to UV-C resulted in an increase in leaf number in dcl3, but not in other plants (Fig. 7A and B). The non-exposed S₁C₁ plants had less leaves as compared to the C₂ and C₁S₁ groups of plants; dcl4 plants from the non-exposed S₁C₁ group had less leaves than any other plants. Stressing the plants from the S₁C₁ group did not significantly change leaf number (Fig. 7A and B). Leaf number in the S₂ group was also smaller than that in the C₂ and C₁S₁ groups; mutant plants in the S₂ group had more leaves than wt plants. Leaf number increased in wt plants in the S₂ group in response to UV-C, but it did not change in the mutants (Fig. 7A and B).

Figure 7.

Measurements of leaf number in the F₂ progeny. (A) Measurements of leaf number in the F₂ progeny of wild-type plants (15D8) and mutant plants (dcl2, dcl3 and dcl4) exposed to UV-C and control plants. “C₂”–the progeny of plants grown under normal condition in the F₀ and F₁ generations. “C₁S₁”–the progeny of plants grown under normal condition in the F₀ generation and exposed to UV-C in the F₁ generation. “S₁C₁”–the progeny of plants exposed to UV-C in the F₀ generation and grown under normal condition in the F₁ generation. “S₂”–the progeny of plants exposed to UV-C in the F₀ and F₁ generations. “+” and “−” indicate either exposure to stress or growth under uninduced conditions, respectively. The values indicate a 95% confidence interval calculated using bootstrap ×10,000 as a result of approximately 24 repeats. The legend indicates the type of mutants. The asterisks (*) indicate a significant difference between wild-type and mutant plants in the same treatment group as calculated at P ≤ 0.05. (B) The ratio of changes in leaf number in response to UV-C stress. The ratios were calculated for S₂: S₂+/S₂−, for S₁C₁: S₁C₁+/S₁C₁−, for C₁S₁: C₁S₁+/C₁S₁−, and for C₂: C₂+/C₂ using the 10,000× bootstrap analysis at a 95% confidence interval, the P-value of 0.05. The ratios that overlap the value of 1 indicate no significant changes under UV-C stress. The asterisks (*) indicate a significant difference between different groups of the same mutant type, as calculated at P ≤ 0.05. The legend indicates the type of mutants, and the horizontal axis indicates parental treatment.

Changes to leaf size

Leaf size (length and width) was similar among all F₂ test groups of plants grown under normal conditions, although there was a tendency to smaller leaves in the S₂ group as compared to the C₂ group (Fig. 8A and C). Exposure to UV-C resulted in changes in leaf size in the C₂ and C₁S₁ groups but not in the S₁C₁ or S₂ groups, except wt plants in the S₂ group (Fig. 8B and D). In the C₂ group, it significantly increased in dcl3 plants, whereas in the C₁S₁ group, it increased in wt, dcl2 and dcl3 plants (Fig. 8B and D).

Figure 8.

Measurements of leaf length and width in the F₂ progeny. Measurements of leaf length (A) and leaf width (C) in the F₂ progeny of wild-type plants (15D8) and mutant plants (dcl2, dcl3 and dcl4) exposed to UV-C and control plants. “C₂”–the progeny of plants grown under normal condition in the F₀ and F₁ generations. “C₁S₁”–the progeny of plants grown under normal condition in the F₀ generation and exposed to UV-C in the F₁ generation. “S₁C₁”–the progeny of plants exposed to UV-C in the F₀ generation and grown under normal condition in the F₁ generation. “S₂”–the progeny of plants exposed to UV-C in the F₀ and F₁generations. “+” and “−” indicate either exposure to stress or growth under uninduced conditions, respectively. The values indicate a 95% confidence interval calculated using bootstrap ×10,000 as a result of approximately 24 repeats. The legend indicates the type of mutants. The asterisks (*) indicate a significant difference between wild-type and mutant plants in the same treatment group as calculated at P ≤ 0.05. The ratio of changes in leaf length (B) and width (D) in response to UV-C stress. The ratios were calculated for S₂: S₂+/S₂−, for S₁C₁: S₁C₁+/S₁C₁−, for C₁S₁: C₁S₁+/C₁S₁−, and for C₂: C₂+/C₂ using the 10,000× bootstrap analysis at a 95% confidence interval, the P-value of 0.05. The ratios that overlap the value of 1 indicate no significant changes under UV-C stress. The asterisks (*) indicate a significant difference between different groups of the same mutant type, as calculated at P ≤ 0.05. The legend indicates the type of mutants, and the horizontal axis indicates parental treatment.

Changes to bolting time

A bolting rate continued to decrease in the progeny of plants exposed to UV-C (S₂ plants) when plants were grown under normal conditions (Fig. 9). When S₂ plants were exposed to stress, bolting increased, with the exception of dcl4 plants. Exposure to stress in the non-stressed F₂ plants (C₂) resulted only in minor changes in bolting in wild-type plants and no changes in mutants. Overall, the most dramatic responses to UV were seen in dcl3 plants as they had the largest changes in bolting under stress, with the exception of the S₁C₁ group where the changes were slightly larger in dcl4 (Fig. 9).

Figure 9.

The percentage of F₂ plants that bolted at approximately 4 weeks of age. The plants were grown under either normal (−) or stressed (+) conditions. Each treatment group is labeled on the X-axis according to treatment. The Y-axis shows the percentage of bolted plants. Approximately 24 plants were included in each treatment group.

Changes in transposon expression

In the second generation, wt C₂ plants responded to UV-C exposure with an increase in ONSEN expression, whereas in dcl mutants, the expression decreased (Fig. 10A). In contrast, ONSEN expression did not increase in response to UV-C in the C₁S₁, S₁C₁ and S₂ groups in wt or mutant plants. In wt plants, the expression of ONSEN was higher in the unexposed C₁S₁, S₁C₁ and S₂ groups compared to the C₂ group, but this was not the case in mutant plants (Fig. 10A).

Figure 10.

The expression of ONSEN and TSI transposons in the F₂ progeny (A, B) in wild-type plants (15D8) and mutant plants (dcl2, dcl3 and dcl4) exposed to UV-C and in control plants. The Y-axis shows arbitrary units of gene expression. “+” and “−” indicate exposure to stress or growth under uninduced conditions, respectively. The bars show the standard deviation calculated from 3 technical repeats. The asterisks (*) indicate a significant difference between control (−) and stress (+) plants with the same parental treatment as calculated using a t-test (P ≤ 0.05).

TSI expression did not increase in response to UV-C in any group of wt and mutants. Quite the opposite, TSI expression decreased in the S₂ group of wt plants, the S₁C₁ group of dcl2 mutants and the C₂ and C₁S₁ groups of dcl4 mutant plants (Fig. 10B). TSI expression in the non-exposed wt S₂ group was significantly higher than in the non-exposed C₂ group (Fig. 10B). In contrast, there was no such difference found in mutant plants.

Discussion

Physiological changes in the F₁ generation

The progeny of UV-C-stressed plants tended to have fewer, longer and wider leaves than the progeny of control plants under both UV-C stress and normal growth conditions. In the first progeny (S₁), this trend was less pronounced in dcl2 and not observed in dcl3 and dcl4 mutants. This effect was less pronounced in mutants, and in dcl3 plants, leaf length and width decreased in the progeny of UV-C stressed plants. A similar response was observed in the progeny of plants treated with high temperature, suggesting that some of the signaling mechanisms of transgenerational response are similar in plants exposed to heat and UV-C.¹¹

A previous work on UV-C radiation (200–280 nm) has shown that it impacts growth responses and biomass production and also enhances certain secondary metabolites such as artemisinin and flavonoids.¹² Considering the ability of UV-C stress to reduce the photosynthetic capacity, it may be more important for plants to focus on larger leaves with an increased opportunity for photosynthesis rather than efforts required for the growth of new leaves.¹³ This is especially true considering the fact that UV-C radiation induces changes in plant metabolism.¹²

C₁ plants or the progeny of control plants are expected to have a similar response to stress as the F₀ generation. In this case, C₁ 15D8 plants increased leaf number significantly in response to UV-C and decreased length and width. In contrast, the progeny of most UV-C stressed plants (15D8, dcl2 and dcl4) also had longer leaves under normal conditions than the progeny of control plants. A previous work has shown that leaf area and thickness as well as plant biomass decrease in response to UV-C.¹² Also, the seedlings of Colobanthus quitensis and Deschampsia antarctica grown under ambient and near-ambient UV-B had 25 and 48% smaller total leaf areas, 7 and 16% fewer leaves, and 65 and 82% fewer branches, respectively, than the seedlings grown under reduced UV-B.¹⁴ It seems to be generally accepted that an immediate response to UV light is a decrease in leaf size, but it also often is paralleled by an increase in leaf thickness.¹ Thus, it can be suggested that whereas an immediate response to UV-C may be an increase in leaf number but a decrease in leaf size, the transgenerational response has opposite effects–a decrease in leaf number but an increase in leaf size. Therefore, transmitting a memory that increases leaf area even under normal growth conditions may be beneficial to progeny that could be faced with UV-C stress.

UV-C stress usually resulted in the production of F₁ seeds that were smaller than those produced by control plants. This difference was significant for dcl2 and dcl4 seeds, while dcl3 seeds significantly increased in length in response to UV. Hence, again, dcl3 plants appeared to be different from wt plants in their response to UV.

UV-C stress activated the transition to flowering in Arabidopsis through the production of salicylic acid (SA).¹⁵ If in F₀ plants, flowering was accelerated as a result of UV-C stress, the accelerated development may have prevented F₁ seeds from reaching the same size as those ones produced under control conditions. Further work could determine if smaller seeds have the decreased germination rates. Interestingly, C₁ and S₁ dcl2 and dcl3 seeds were significantly longer than 15D8 ones, despite the fact that in dcl2 and dcl3, seed length was significantly altered in response to UV-C stress, while in 15D8, it was not. Potentially, dcl3 plants were not so negatively affected by stress.

The progeny of UV-stressed plants had a lower bolting rate than the progeny of control plants of the same age, providing further evidence of heritability of bolting time.¹⁶ Interestingly, UV-C stress has previously been shown to result in shorter plants.¹² Though the final height of plants was not measured, it would be interesting to see if the decreased bolting rates corresponded to a lower height in the progeny of UV-stress plants.¹²

UV-C is known to accelerate flowering in Arabidopsis by activating it prematurely to enhance the chances of plant survival in the harmful environment and produce the progeny as a response triggered by the increased SA production. In fact, SA is important even in the non-stressed plants, and those ones that are deficient in SA production flower later.¹⁵ Since in Arabidopsis, bolting corresponded to the elongation of reproductive internodes in the leaf zone and therefore occurred before flowering, UV stress was expected to trigger an earlier bolting in F₁ plants.¹⁷ However, the response to UV was not universal, nor was it based on parental treatment. The most similarity was found among 15D8 and dcl4 plants that bolted earlier in response to UV stress and in dcl2 and dcl3 plants that bolted later. An acceleration of flowering was only observed in response to stress if SA production and accumulation could occur.¹⁵ It is therefore possible that dcl2 and dcl3 plants were not entirely successful in the SA production and accumulation process, resulting in a different response to UV-C. It is curious to note that both mutants were found to be less efficient in repairing DNA damage induced by UV-C.¹⁸

Physiological changes in the F₂ generation

Physiological changes in the F₂ progeny were not as consistent as in the F₁ progeny. The impact of parental stress induced by UV was inconsistent, though the fewest leaves tended to be found following 2 generations of stress (S₂), which corresponds with the results observed in S₁ plants where parental stress decreased leaf number. It is worth noting that the S₁C₁ group of plants also had fewer leaves than the C₂ group but more than the S₂ group. This suggests that a substantial portion of transgenerational memory in the form of changes in leaf number was passed on to the F₂ progeny even when S₁ plants were not stressed by UV-C again. Unlike in F₁, in S_2, parental stress by UV did not significantly impact leaf length or width, although there was a tendency to a decrease in leaf length. A clear correlation exists between an increase in exposure time to UV-C and a reduction in plant growth and biomass.¹⁹

Seed size increased when plants were grown under exposure to UV-C stress for 2 generations, which was in contrast to what was observed in plants exposed to UV-C for one generation. It is possible that the increase was a compensatory mechanism in response to a second exposure to UV-C. All mutant types that had either one or 2 previous generations exposed to UV stress were larger than those grown under control conditions for 2 generations (C₂). It was unexpected that seed size would actually increase in S₂ because S₁ seeds were smaller than C₁ seeds. However, it should be noted that in our previous work, we found that exposure of plants to a low dose of UV-C very early during development also resulted in an increase in seed size in the exposed plants.⁷ Measuring F₀ seeds in order to determine a starting seed size would give the best assessment of changes caused by the increasing number of generations exposed to stress. It would also be interesting to determine if there is a correlation between seed number and seed size, because it is possible that C₁ plants shifted the balance toward fewer, larger seeds, while S₁ plants produced a larger quantity of smaller seeds in response to UV stress.

Positive physiological changes in the progeny of plants exposed to UV-B were observed in Glycine max (L.) Merr plants–an elevated level of N in nodes and a larger seed size were observed in the immediate progeny of stressed plants.²⁰ There is a growing indication that in cases when realistic biological levels of UV-B exposure have been used, especially early during development, the photosynthetic capacity and plant yield have actually increased.²¹ Wargent and Jordan (2013) summarized many cases of beneficial effects of exposure to UV on plant productivity and tolerance.¹ An increased photosynthetic rate in response to UV-B was observed in several species, including Hippophae rhamnoides,²² Hordeum vulgare,²³ Lactuca sativa²¹ and Oryza sativa.²⁴ An increase in yield after exposure to UV-B was found in Fagopyrum tataricum,²⁵ Lactuca sativa.²¹

The percentage of bolting was further decreased in the progeny of plants exposed to UV-C stress for 2 generations (S₂). In F₂ plants, exposure to UV-C increased the bolting rate. This was expected based on F₁ results as well as on the previous work showing that UV-C accelerated flowering in Arabidopsis.¹⁵ Thus, it seems that whereas an immediate response to UV-C is an increase in bolting time, the transgenerational response is opposite–a decrease in bolting time, which may be a compensation for an increase observed in somatic tissues. It is interesting to note that in response to heat stress, in both the stressed plants and in their progeny, bolting occurred earlier, suggesting that responses to heat and UV-C may be different as far as transgenerational responses are concerned.¹¹

An impact of stress on transposon expression in the progeny of stressed and control plants

In general, exposure of parental plants to UV-C stress led to an increased expression of retrotransposons ONSEN and TSI in the progeny of stressed plants grown under normal conditions in comparison to the progeny of control plants. The dcl3 mutant was clearly impaired in such increase in both transposons. When exposed to stress, the progeny of stressed and control plants did not usually vary in TE expression levels. This indicates that although the natural level of transposon expression tended to be higher in the progeny of stressed plants, differences were not significant in S₁ and C₁ under stress.

Transposon expression increased even further in the second generation of stressed (S₂) wt plants grown under control conditions. In contrast, no difference was observed in any mutant in the S₂ group, further suggesting that dcl mutants are impaired in their capacity to change transposon expression in the progeny in response to stress. Our observations of responses to heat stress showed different results–the expression of ONSEN and TSI transposons increased in the progeny of wt and mutant plants exposed to heat for 2 consecutive generations.¹¹

Finally, in most cases, exposure of S₂ plants to UV stress did not significantly change TE expression in comparison to plants grown under normal conditions. There were a few fluctuations in the expression that generally led to a decrease in TE expression under stress. It is possible that since transposon expression was already significantly up-regulated in the progeny, UV-C was not able to increase it any further, which likely resulted in some type of a negative feedback loop counteracting the effects of stress with a decrease in transposon expression. It is possible that the severity of UV-C exposure was not sufficient to have a more profound impact on the plants.

In the past, several studies reported changes in TE expression and TE activation in response to stress. In 1992, Virginia Walbot reported that UV irradiation of maize pollen resulted in a 40-fold increase in reactivation of Mu9 TE in the progeny.²⁶ Exposure to UV-B also resulted in the activation of the expression of mudrA and mudrB transposons 8 h after treatment.²⁷ The expression of copia retrotransposon Reme1 was also induced by UV exposure in melon.²⁸ Ty1-copia retrotransposon in oat was also induced by UV and other abiotic stresses.²⁹

ONSEN expression was also shown to be induced by stress; exposure of Arabidopsis plants to heat resulted in an increase in ONSEN expression in both the stressed tissues³⁰ and their progeny.¹¹ In the work of Ito et al. (2011), no retrotransposition events were found in wt plants, as far as the progeny of stressed plants was concerned. In contrast, the progeny of stressed dcl mutants showed the evidences of new retrotransposition events.³⁰ The authors did not find that ONSEN expression was increased in dcl3 plants grown under normal conditions. This is in contrast to the finding reported in this manuscript. Curiously, in our previous report, we have shown that ONSEN expression was only slightly higher in non-exposed dcl3 plants as compared to wild type plants.¹¹ It is hard to explain the discrepancy in ONSEN expression in dcl3 in these 2 reports. It is possible that some minute changes in growth conditions or plant handling could have resulted in the observed changes in dcl3 mutants.

Heritable changes in MITEs and LTR retransposon activity were observed in rice in response to the spaceflight environment.³¹ The treatment of several rice cultivars with etoposide, a chemical which induces genomic instability including the mobilization of TEs, resulted in the mobilization of mPing transposon in the S₁ and S₂ progeny of treated rice plants.³² TE reactivation was paralleled with DNA hypomethylation at specific genomic regions, suggesting that the epigenetic mechanisms are likely controlling TE expression and activity. The importance of epigenetic regulation and namely siRNA biogenesis for the regulation of transgenerational activation of TEs was demonstrated in the past.³⁰

Conclusion

Here, we have shown that UV-C exposure results in changes in leaf number, leaf size, bolting, seed size and transposon expression. We have demonstrated that some of these changes are passed on to the progeny, and dcl mutants are partially impaired in the transgenerational response to UV-C. Changes in transposon activity are particularly interesting as they have been maintained for 2 generations with and without the presence of stress in the second generation. We noted that exposure to UV-C in the progeny (either S₁ or S₂) did not further increase transposon expression, suggesting that there might be some limiting mechanisms regulating transposon expression. It will remain to be shown whether changes in transposon expression will manifest themselves in new transposon insertions in progeny.

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.

Acknowledgments

We appreciate the help of John Sheriff in statistical analysis. We thank Valentina Titova for proofreading the manuscript.

Funding

Financial support was provided by Alberta Agricultural Research Institute and National Science and Engineering Research Council of Canada Discover grant to Igor Kovalchuk and NSERC MSc scholarship to Zoe Migicovsky.