L. minor (serial number 1007 and ID number 5500) clones were employed in the experiments. Briefly, it was carried out as follows: plants were cultured in sterile 250 mL glass flasks filled with 100 mL of 1/10 strength Hoagland solution (1 mM KNO₃; 0.3 mM Ca(NO₃)2·4H₂O; 0.2 mM MgSO₄·7H₂O; 0.1 mM NH₄H₂PO₄; 1.6 μM FeSO₄.7H₂O; 0.8 μM EDTA diNa-salt 2H₂O; 4.6 μM H₃BO₃; 0.9 μM MnCl₂·4H₂O; 0.03 μM CuSO₄·5H₂O; 0.06 μM H₂MoO₄; 0.08 μM ZnSO₄·7H₂O; pH 5.6) (Van Hoeck et al., 2017). Plant culture prior to the experiment was conducted under sterile conditions, with plants positioned on the surface of the medium exclusively using sterilized inoculation loops under a laminar flow. Then, the experiment was maintained under semi-sterile conditions using transparent sterile round Petri plates placed on top of the pots. This was done to minimize potential contamination and to further reduce evaporation and limit particle deposition while allowing air exchange. The growth conditions were set to 16h day and 8h night photoperiod (150 μmol m^-2 s^-1 at the leaf level), day/night temperatures of 22°C/18°C in a controlled climate chamber (Microclima MC1000E, Snijders Scientific B.V., Tilburg, The Netherlands).

Upfront of each experiment, the plants were initially pre-cultured for two weeks in sterilised transparent pots (Nalgene^®, Thermo Fisher Scientific) in 50 mL of 1/10 strength Hoagland solution, following the experiment phase under similar conditions. It was tested beforehand in a pilot experiment that 50 mL was an appropriate volume for the growth of 7 plants over 7 days, with no growth limitation observed ( Supplementary Figure 1 ). The pre-culture process started with 7 plants, and at the beginning of every following week 7 plants were sub-cultured to new pots with fresh Hoagland medium. For each experimental condition, 15 pots were used as biological replicates.

At the start of the 6-weeks contamination experiments, a ⁹⁰Sr source (SrCl₂, 485895 kBq/L) was diluted in demineralized water and filtered (0.2 µm), added to the respective media, and the pH was adjusted to 5.6 with NaOH. Four different activity concentrations were prepared in 1/10 strength Hoagland medium: β0 = 0 Bq/L (control), β1 = 40 Bq/L, β2 = 400 Bq/L and β3 = 4000 Bq/L. For each time point, 60 sterile 250 mL clear Nalgene^® pots (15 for the control and 15 per each contaminated condition) were prepared with 50 mL of fresh sterile 1/10 Hoagland medium solution per week. The start of each week (Day 0, Week N_1-6) was marked by the transfer of 7 mature plants with 3–4 fronds from the previous week’s culture to new pots, following the respective contamination condition. For the first week, all plants used for every condition came from the pre-culture. Transfer of plants was carried out at the same time during every week. Plants’ growth was followed for the whole duration of the experiment (i.e. 6 weeks). In order to automatize measurements of area growth rates, a 1 cm² sterilized red plastic square reference was placed floating on top of the medium. To prevent excessive evaporation of the medium while yet allowing necessary light transmission and environment exchange, each pot was covered with a clear, sterile Petri dish. To guarantee appropriate homogeneity for all conditions and replicates, the pots were randomly distributed inside the climate chamber. At day 7 of each week, all plants but those transferred to the subsequent week were harvested and immediately analysed (e.g. photosynthetic parameters measurements, pigment content evaluations) or stored at -80°C for future analytical purposes.

Top-view photographs were taken from the same distance, employing the same light conditions and camera settings on days 0, 2, 4 and 7 of each week to track frond growth over the course of the experiment. Images were processed using the open source software EasyLeafArea version 2.0 (Easlon and Bloom, 2014) and ImageJ version 1.8.0 (Schneider et al., 2012), through which the total surface area and number of the fronds was obtained. Subsequently, following the OECD guidelines for toxicity testing of the genus Lemna (Test Method No. 221: Lemna sp), average specific growth rates (ASGR) were calculated for the frond number and area, according to the following formula:

where:

- μ i − j : average specific growth rate from time i to j

- N i : measurement variable in the test or control vessel at time i

- N j : measurement variable in the test or control vessel at time j

- t: time period from i to j

- i: day 0 of each week

- j: day 7 of each week

Normality of the data was evaluated through the Shapiro-Wilk test using R software (R version 4.3.3) (R Core Team, 2023). For data sets that exhibited normal distribution, a two-way ANOVA was performed, followed by post hoc Tukey’s multiple comparison tests to identify significant differences between groups, all carried out using GraphPad software (GraphPad prism version 9.4.1). When normality was not met, the non-parametric Kruskal-Wallis test was applied. In such cases, post hoc comparisons were conducted using Dunn’s test with Bonferroni correction to account for multiple comparisons, also carried out through the GraphPad software. For gene expression data, the Wilcoxon Rank-Sum Test was used, conducted using R software (R version 4.3.3).

L. minor plants were grown for up to 6-weeks under varying activity concentrations of ⁹⁰Sr (40 Bq/L, 400 Bq/L, and 4000 Bq/L) to gain insights into multigenerational dose- and time-dependent effects of chronic β-radiation exposure. The calculated average total dose rates were 0.82 ± 0.07 µGy•h^-1 for β1, 7.69 ± 0.98 µGy•h^-1 for β2 and 72.67 ± 6.90 µGy•h^-1 for β3 and reflect the tenfold differences among β1, β2 and β3 activity concentrations employed in the experiment ( Table 1 ). The total dose rate is the sum of the calculated internal and the external dose rate. Overall, internal absorbed dose rates were one order of magnitude greater than external.

To assess biomass dynamics upon ⁹⁰Sr exposure, both fresh (FW) and dry weight (DW) were measured ( Table 2 ). Statistically significant differences were observed in plants exposed to the intermediate (β2) and highest (β3) activity concentrations during week 3, 4 and 6, with the contaminated plants exhibiting higher values than the control plants. In contrast, concentration ratios of ⁹⁰Sr expressed on both fresh and dry weight and dosimetric parameters (internal, external and total dose rates) remained stable throughout the full experiment.

To assess the impact at the morphological level of 4 different ⁹⁰Sr dose rates, frond area and number of L. minor plants were measured over a 6-week period. Measurements were carried out on days 0 (culture start), 2, 4, and 7 of each week ( Supplementary Figures 2 , 3 ).

The average specific growth rate (ASGR, calculated for both frond area and frond number as described in the OECD guideline 226) provides a quantification for comparisons between different treatments. A statistically significant difference in area ASGR was observed between the control group and the β1 and β2 groups during week 1 ( Figure 1 ), which disappeared during weeks 2 and 3. Statistically significant difference in frond area was again observed in the highest dose rate group (β3) of week 4 and 6, which may reflect a cumulative or delayed effect of ⁹⁰Sr radiation exposure. While no significant difference was observed in week 5, it is possible that the higher heterogeneity of the plants in that week hid a potential difference from control. Following the same pattern observed earlier, the group β3 displayed the greatest difference in frond area from the control group during week 6.

The analysis of the number of fronds revealed a within-week specific increase under contaminated conditions, starting in the second week and persisting until the fifth week ( Figure 2 ). Frond number and frond area showed partially overlapping dynamics, although the timing of statistically significant differences did not coincide exactly. This suggests that while both traits reflect growth responses to ⁹⁰Sr exposure, they may do so with differing sensitivities or temporal patterns. In particular, during the first two weeks, a hormetic-like effect could be detected for the two lowest dose rates (β1 and β2), with slightly reduced growth observed under the highest dose rate (β3). In weeks 4 and 5, this hormetic-like response became more pronounced, including the β3 condition, in a similar way to what happened in the weeks 4–6 to area values. Specifically, increased values were observed for β1 and β2 conditions in week 2, for β2 in week 3, and for all non-control conditions (β1, β2, and β3) in weeks 4 and 5, with week 4 showing comparable changes to the ones seen for area ASGR for the highest condition.

Telomeres, the nucleoprotein structures protecting chromosome ends, play a critical role in maintaining genome stability (Blackbum, 1991; McKnight et al., 2002). Although IR exposure is reported to induce genome instability in plants (Duarte et al., 2023; Kim et al., 2019; Yokota et al., 2010), not much is understood about the influence of IR on telomere length, despite the abundant data documenting this link in mammals (Bains et al., 2019; Biswas et al., 2024; Goytisolo et al., 2000; Ilyenko et al., 2011; Movahedi et al., 2019). To address this gap, relative telomere length measurements across all experimental conditions and time points were carried out. Telomere length remained stable across all experimental conditions and time points ( Supplementary Figure 4 ). Statistical analysis revealed no significant differences in telomere length among the control group and the contaminated groups. This suggests that although β-radiation does not alter the overall length of telomeres, it remains to be investigated whether it is due to damage avoidance mechanisms or if a possible reduction could be compensated by increased damage repair.

The electron transport rate (ETR) represents one of the key photosynthetic endpoints commonly assessed during the analysis of the effects of IR on photosynthetic organisms. For instance, acute γ exposure is reported to cause a decrease in ETR in the algae Chlamydomonas reinhardtii (Gomes et al., 2017) and the aquatic plant Zizania latifolia (Fan et al., 2014), whereas chronic IR has been shown to enhance ETR in Arabidopsis thaliana subjected to low dose rates of α (Biermans et al., 2015) and γ radiation (Vanhoudt et al., 2014). Therefore, to gain insights on L. minor’s response during chronic low dose exposure over multiple generations, the ETR in function of the photosynthetic active radiation (PAR) was assessed over a 6-week exposure period ( Supplementary Figure 5 ). Photosynthetic efficiency in function of the light intensity appears to remain unaltered under long-term ⁹⁰Sr contamination in comparison to the control ( Supplementary Figure 5 ), with the exception of week 2 ( Supplementary Figure 5 ), where there is an observed decline ETR across all the conditions.

IR reportedly affects the concentration of photoactive pigment in plants (Kim et al., 2004; Sengupta et al., 2013). Thus, L. minor’s chlorophyll a, b, and carotenoid levels were measured to gain information about chronic IR responses at the physiological and biochemical level ( Figures 3A–C ). For chlorophyll a, a generally lower content was displayed in plants exposed to ⁹⁰Sr IR during all sampling weeks ( Figure 3A ). The reduction was statistically significant for groups β1 and β2 in week 2 (p < 0.01) and week 4 (p < 0.0002), and for group β3 in week 4 (p < 0.0001) ( Figure 3A ). Despite the lack of statistically significant variations in weeks 1, 3, 5, and 6, the detected trends suggested a potential involvement of chronic IR exposure in chlorophyll a levels.

The results for chlorophyll b ( Figure 3B ) indicated a trend of decreased content induced by chronic low dose exposure over multiple generations, observed across all time points, which was statistically significant for weeks 2, 3, and 4. Chlorophyll b is mostly present in the antennas rather than in photoactive cores, therefore the chlorophyll a/b ratio (Chl a/b) gives an estimate of the proportion of molecules bound by light-harvesting complexes (Walters, 2005). This additional parameter is a common endpoint used to monitor acclimation in plants and was also analyzed in this experiment. Here, acclimation is intended as a short-term, reversible physiological adjustment to environmental changes, distinct from adaptation, which involves long-term, heritable genotypic and phenotypic changes that persist beyond an individual’s lifetime (Collier et al., 2019; Kumarathunge et al., 2019; Mitra et al., 2021; Morgan-Kiss et al., 2006). In the context of this work, such acclimation responses may reflect the plant’s ability to retain function under mild but persistent stress conditions, i.e. a form of eustress (Van Hoeck et al., 2017). To this end, in week 1 and 4, significant increases in Chl a/b were observed only in the β3 groups ( Supplementary Figure 6 ), which might indicate a high-dose threshold for initiating photosynthetic acclimation towards chronic IR in L. minor. This finding could also provide information on the steric restructuring of the photosynthetic apparatus in response to chronic low dose exposure over multiple generations, as a high Chl a/b ratio indicates acclimation to stress conditions, possibly involving reduced antennae size to avoid damage from excessive harvested photons (Walters, 2005).

Regarding the carotenoids ( Figure 3C ), β2 values were significantly lower than the control group (p = 0.005) in week 2. Overall, a consistent trend was observed for all the weeks, with the control group maintaining highest levels and the contaminated groups showing lower levels, as seen for chlorophyll a and b. This trend persisted and became statistically significant for all conditions in week 4, in the same way as reported for chlorophyll a and b ( Figures 3A–C ).

It is reported that shifts of macromolecule levels are linked to stress responses in plants (Dubey and Singh, 1999; Gill et al., 2003; Huang et al., 2013; Pagliuso et al., 2018; Strand et al., 1999; Van Dyck et al., 2024). Therefore, total protein, sugar and starch content were measured to gain understanding about L. minor’s physiological state upon ⁹⁰Sr stress. These measurements have been carried out on control, β2 and β3 plants.

While protein content is associated with stress responses in plants, there is no consensus on whether stress enhances or decreases its levels (Abdul Qados, 2011; Campbell et al., 1981; Doganlar et al., 2010; Gulen and Eris, 2004; Mohammadkhani and Heidari, 2008). In this study, protein content remained relatively stable throughout 6 weeks, with the exception of a slight increase observed in β3 samples during week 5 ( Figure 4A ). Considering the modest magnitude of this increase, it is plausible that this variation partly falls within the range of experimental variability.

Soluble sugar remained largely unchanged, except for increases observed in β2 plants during week 2 and in in β1 in week 3 ( Figure 4B ). This elevation might represent a stress acclimation mechanism which could be linked to photosynthetic performance (Boriboonkaset et al., 2013; Masoudi et al., 2011; Pociecha et al., 2010) or osmotic adjustment (Djanaguiraman et al., 2006; Dubey and Singh, 1999; Watanabe et al., 2000).

Starch accumulation ( Figure 4C ) is reported to increase in L. minor under stress conditions and it has been correlated to lower growth rates (Sree and Appenroth, 2014; Van Dyck et al., 2021). Interestingly, chronic exposure to ⁹⁰Sr did not lead to higher starch levels, which were in fact reduced in β3 plants during weeks 1 and in β1 in weeks 3-6. With the exception of β1 plants, the overall stability of the three evaluated macromolecule levels (total protein, total sugars and starch) over the weeks suggested that L. minor effectively adjusted its metabolism to the presence of this radionuclide, which at the tested concentrations seemed to trigger low stress response.

Glutathione levels, both in their reduced and oxidized forms (GSH and GSSG, respectively), and peroxidases are often employed as antioxidative defence biomarkers. Reduced antioxidants and peroxidase enzymes can mitigate the effects of excess reactive oxygen species generated under stress, including IR exposure (Oztetik, 2015; Riley, 1994). Our investigation into glutathione content and its redox state revealed, however, that there was very little difference comparing control to exposed plants ( Supplementary Figure 7 ). Figure 5 represents the variation of guaiacol peroxidase (GPOD) in weeks 1, 2 and 6 (other weeks were not measured).

The results of GPOD activity showed, with the exception of week 2, a decreasing trend for all the conditions analysed which was statistically significant both in week 1 and 6. Overall, the data underlined time and dose-rate dependence of L. minor’s GPOD activity upon ⁹⁰Sr contamination. Moreover, it is noteworthy that these patterns align with those observed for area ASGR. This correlation may be attributed to peroxidase’s role in cell loosening (Liszkay et al., 2003). Reduced GPOD activity could indicate decreased cell wall stiffening, potentially leading to more elongation growth and increased surface. While direct ROS quantification could offer additional insight, we strategically prioritized endpoints more central to our study objectives, including ⁹⁰Sr uptake and gene expression related to antioxidant responses, given the limited biomass typical of L. minor. Future studies may build on this work by incorporating direct ROS measurements to further deepen understanding.

2DNA methylation is widely known to play a role in the adaptation of plants to chronic IR exposure, modulating the expression of specific genes (Bondarenko et al., 2023; Horemans et al., 2018; Kovalchuk et al., 2003; Laanen et al., 2021; Volkova et al., 2018). Furthermore, glutathione biosynthesis and DNA methylation pathway compete for the common substrate S-adenosylmethionine (SAM) (Ouyang et al., 2020), indicating an interconnection between redox balance and epigenetic regulation. As such, the current work included the evaluation of gene transcript levels that control DNA methylation (repressor of silencing 1 (ROS1), chromomethylase 3 (CMT3), DNA methyltransferase 1 (MET1)), of key enzymes that play crucial roles in plant’s redox metabolism (glutathione reductase (GR), glutamate-cysteine ligase (GSH1), glutathione synthetase 2 variant 1 (GSH2.1), glutathione synthetase 2 variant 2 (GSH2.2)), and chromatin conformation (sirtuin 1 (SRT1), telomere reverse transcriptase (TERT)). The results of the relative gene expression over the weeks are represented in Figure 6 and Supplementary Table 1 .

Results indicate that GR expression is decreased in the β3 group in week 1 and week 4, GSH1 expression is lowered in the β2 group in week 2 and 3, and in the β3 group of week 6, suggesting a potential role for glutathione homeostasis in early onset ⁹⁰Sr-induced stress responses. Additionally, week 3 results show an increased expression of GSH2.1 and GSH2.2, suggesting a possible shift in the stress scavenging system involved. For the β2 and β3 conditions, MET1 and CMT3 levels decreased in weeks 2 and 3, respectively, with TERT also showing decreased levels in week 3. These changes suggested a dose rate and time dependent variation in DNA methylation maintenance and telomere regulation during early to intermediate stages of ⁹⁰Sr exposure.

The available acute (Alkimin et al., 2019; Flores-Rojas et al., 2015; Nunes et al., 2014; Park et al., 2017) and chronic (Radić et al., 2010; Van Hoeck et al., 2017; Xie et al., 2022) IR-related studies in L. minor usually cover up to 1 week of exposure, which for such a fast reproduction plant represents about 2 to 3 clonal generations (i.e. ~2.5 days/generation) (OECD, 2006).

With respect to the field of plant radioecology, Volkova et al. (2022) define the hormesis mechanism as biological responses characterized by faster development, more robust growth, greater stress tolerance, or the build-up of specific chemicals in response to low-dose radiation. In the early phase of exposure (week 1), ⁹⁰Sr exposed L. minor plants showed increased frond area parameters, evident at the lowest dose rates β1 and β2 ( Figure 1 ).

The data further suggested that even the low and environmentally relevant doses employed in the study were sufficient for the plants to experience stress. It is important to discern between “eustress” and “distress”, which are respectively described in L. minor as growth preserving or stimulating and growth inhibiting types of stress (Van Hoeck et al., 2017). Van Hoeck et al., demonstrated that “distress” after a one-week exposure in L. minor begins at dose rates of milligrays per hour (Van Hoeck et al., 2017). Our results suggests that eustress can be induced already at much lower dose rates (e.g. 0.82 ± 0.07 µGy•h^-1, Table 1 ). These results underline the necessity to take into consideration low-dose rates when assessing the ecological impact of radioactive contamination.

The experimental design employed in the present work covered approximately 17 L. minor generations. Here, time dependence refers to the variation in treatment responses across weeks relative to their respective within-week controls. The results suggest a temporal pattern of response to low-level chronic ⁹⁰Sr contamination that may be interpreted as comprising three phases over the 6-week period: the above described early phase (week 1), the mid phase (weeks 2-5), and the long-term phase (week 6). In the mid-phase, the main observable endpoint changes was frond number, which increased consistently throughout all the weeks ( Figure 2 ), and a decline in pigment content ( Figure 3 ). Interestingly, the decrease in pigment content coincided with conditions and time points where frond number increased. On the other hand, the long-term response phase was characterised by a combination of biomarker changes that mirrored week 1 (see Table 3 ). In fact, early and long-term phases shared increased levels of area average specific frond rate ( Figure 1 ), decreased starch content ( Figure 4 ) and decreased GPOD ( Figure 5 ) activity values. However, one difference must be mentioned: while in week 1 only dose rates β1 and β2 were showing increased area ASGR, in week 6 β3 plants showed significant increased values as well. This might be suggesting the presence of a dose rate and time dependent transient hormetic effect, with the plants taking more time to acclimate to higher dose rates.

Through rapid expansion on the water surface, L. minor plants can outcompete other species for access to light and nutrients, effectively monopolizing the available space and resources (Paolacci et al., 2018). Therefore, producing additional fronds quickly under eustress conditions might represent a competitive advantage over other aquatic plants. The observed increase in frond number ASGR during the mid-phase lends support to this hypothesis. However, the relative importance of frond number and area in L. minor’s competitive strategies may vary depending on the specific environmental context (Vasseur et al., 1995).

It was evident that there were re-adjustments in total macromolecular content and in the antioxidant photosynthetic pigment repertoire, as increases in total protein content ( Figure 4A ) were detected in week 5, increases in soluble sugar content ( Figure 4B ) were measured in weeks 2 and 3, decreases in starch content ( Figure 4C ) were seen in weeks 1 and 3-6, and decreases in the photoactive pigments chlorophyll a, b and carotenoids were respectively observed in weeks 2 and 4, in weeks 2-4, and weeks 2 and 4 ( Figure 3 ). Similarly to what happens for frond number ASGR, total macromolecular content and pigment results were chiefly dose rate- and time-dependent. In particular, pigment levels appeared to be decreased in the lowest dose rates (β1 and/or β2) when the first detectable difference emerged in week 2, while all the dose rates showed decreased levels in week 4. Another indication of this time- and dose rate-dependence is that, from week 3 to week 6, only the lowest dose rate (β1) consistently showed decreased levels of starch content. This finding is in line with the current literature reporting that the time necessary for the organism to acclimate is relative to the dose rate into play (Shadley and Wiencke, 1989; Tapio and Jacob, 2007).

Biological systems use a range of feed-forward regulatory processes to preserve epigenetic information for prolonged periods of time (Castel and Martienssen, 2013; Houri-Zeevi and Rechavi, 2017; Martienssen and Moazed, 2015). Among them, we can find the IR-responsive mechanisms (Laanen et al., 2023) of DNA methylation regulation (Agorio et al., 2017). In this study, the expression of several genes involved in DNA methylation regulation was tested, namely DNA Methyltransferase 1 (MET1) and Chromomethylase 3 (CMT3), which are essential to the maintenance of DNA methylation (Chan et al., 2005), and Repressor of Silencing 1 (ROS1), a demethylation agent that prevents excessive methylation and ensures that epigenetic changes are precisely calibrated to the needs of the plant (Halter et al., 2021, p. 1). In addition to DNA methylation, histone modifications epigenetically regulate chromatin structure and function (Miller and Grant, 2013) and are reported to be involved in plants’ IR stress responses (Kim et al., 2019; Raut and Sainis, 2012). Thus, the expression analysis of the histone deacetylase SIRT1 gene, an epigenetic regulator involved in A. thaliana abiotic stress responses (Liu et al., 2017), was included in this study as well. The beginning of the mid phase was marked by an overall change in gene expression of the DNA methylation machinery: while methyltransferases genes showed decreased expression for dose rates β2 and β3, no changes in gene expression were observed for the demethylase ROS1 across all dose rates ( Figure 6 ; Supplementary Table 1 ). Similarly to what Van Antro et al. proposed in their work on the involvement of DNA methylation in L. minor’s temperature stress responses, these results suggested a potential dose- and time-dependent involvement of the epigenetic machinery in IR stress response in L. minor. For further elucidation of the role of epigenetics in this stress response over time, the current analysis should be complemented by western blotting and/or bisulfite sequencing to provide a more comprehensive understanding.

The DNA methylation machinery relies on S-adenosyl methionine (SAM) to exert its functions (Ouyang et al., 2020), and SAM plays a which also constitutes an important precursor for the synthesis of glutathione. Here, the relative expression of Glutamate-cysteine ligase (GSH1), Glutathione synthetase 2 variant 1 (GSH2.1), Glutathione synthetase 2 variant 2 (GSH2.2), Glutathione reductase (GR) was followed over the course of the 6 weeks exposure. Interestingly, despite non-significant decreasing trends of glutathione content ( Supplementary Figure 7 ), the expression patterns of glutathione metabolism genes showed significant variations. Specifically, GR expression decreased in β3 conditions in weeks 1 and 4, GSH1 decreased in β2 conditions in weeks 2 and 3 and in β3 conditions in week 6, whereas GSH2.1 and GSH2.2 increased under β3 conditions in week 3 ( Figure 6 ). Thus, the mid phase showed the lion’s share of responses at the gene expression levels for glutathione metabolism genes for β2- and β3-exposed plants. Just as it happened with epigenetic expression data, they went back to normal levels upon progressing towards the long-term responses phase, highlighting again the long-term transient nature of these responses.

It has been reported that telomere regulation is involved in plant stress responses and aging (McKnight et al., 2002; Sun et al., 2023; Watson and Riha, 2010). Works from Wyatt and Shippen (Ohio.edu, 2022; Russel, 2022) report that plants grown in space, where the background IR levels are higher [e.g. daily dose on the International Space Station is 0.5 mSv, which is about 100 higher than the dose on earth (Cucinotta et al., 2008; Takahashi et al., 2018 De Micco et al., 2023)], have a steady telomere length but significant changes in telomerase enzyme activity.

In our work, we carried out the measurement of telomere length, and analysed TERT (Telomerase Reverse Transcriptase) transcript accumulation, essential for telomere length maintenance (Vaquero-Sedas and Vega-Palas, 2014). The qRT-PCR results of pointed out a time- and dose-dependence in the mid phase, as only β2 and β3 conditions appeared to be downregulated and only in week 3 ( Figure 6 ). In contrast to this result, the telomere length assessment showed stable levels throughout the experiment ( Supplementary Figure 4 ). As the telomerase system is intimately linked to the epigenetic machinery, chromatin 3D structure and DNA methylation in particular (Dogan and Forsyth, 2021; Galati et al., 2013; Jezek and Green, 2019; Moores et al., 2011), further research investigating this relationship could provide valuable insights into the role of the telomerase system in the plant’s response to low level ⁹⁰Sr radiation.

At day 7 of each experimental week, after subculturing, all plants growing in 3 pots were collected for measuring ⁹⁰Sr uptake. To wash off any excess ⁹⁰Sr that could have adhered to the plants, they were transferred to a small sieve and underwent 3 washing steps: 10 minutes in 10 mL 1mM Pb(NO₃)₂ and 2 steps of 10 minutes in 15 mL demineralized water. Plants were then dried and transferred to a weighing vessel for fresh weight (FW) measurement. Plant material was then placed in an oven at 65°C until dry. Plant samples were then weighed (dry weight, DW) and transferred to 20 mL glass vials, where they were ashed for 24h. The internal temperature of the oven was reached in small steps (from 0 to 550°C in 12h) to avoid losing material from the sample. After the samples were completely ashed, the dissolution procedure started by adding 1 mL of 65% nitric acid (HNO₃). Vials were placed on a sand bath until dry, then moved to the oven and left at 550°C overnight. Subsequently, plants were mineralized by adding a solution of 65% HNO₃ and 30% hydrogen peroxide (H₂O₂) in three sequential septs: 0.65 mL, 0.65 mL and 0.70 mL. The samples were then dried, after which 2 mL of 65% HNO₃ was added. This stepwise addition prevented liquid loss due to bubble formation during the process. The drying step was repeated as needed to safely add the full volume required for complete mineralization. The samples were then dried again to ensure thorough preparation for the subsequent analysis. Finally, samples were dissolved in 8 mL of 0.05 M HNO₃, transferred to polyethylene (PE) vials, weighed, and 12 mL of scintillation cocktail (Optiphase Hisafe 3) were added. β measurement was carried out by the Low-level Radioactivity Measurements (LRM) group at SCK CEN (Belgium) employing the following protocol: reference and blank samples were prepared in the same way as samples. The reference solution, a dilution from a certified reference solution from CERCA-LEA, contained a known amount of solution of ⁹⁰Sr/⁹⁰Y (80.68 ± 2.42 Bq ⁹⁰Sr). The samples were kept for one hour in the counter before starting the measurements. The blank and the reference were measured together (1h per vial) with the samples. A Hidex 300SLL liquid scintillation counter was used to perform the measurement.

Dose rate values [µGy•h^-1] were computed by the ERICA tool (Brown et al., 2008) and calculations were based on the “Lemna leaf model”. Fronds were modelled as elliptical cylinders adopting the standard parameters of L. minor geometry described in Van Dyck et al (Van Dyck et al., 2021), which featured dimensions of 4.1 mm in length, 2.7 mm in width, and 0.4 mm in height, with an average frond mass of 1.66 mg. The dosimetry focused solely on the plants’ frond, with the root component being disregarded due to its relatively low dose contribution. For external dose calculations, the plant was considered to be “on water”, indicating its position as floating on the surface at the centre of the pot. Total dose rates were obtained by adding the internal and external dose rate values given by the ERICA tool. For each week and contamination condition, concentration factors (CF) were calculated and implemented according to the following equations:

where A_w is activity in the growing medium [Bq•L^-1], CF is concentration factor [L•kg^-1], DC_int is internal dose conversion coefficient [µGy•h^-1 per Bq•kg^-1], DC_ext is external dose conversion coefficient [µGy•h^-1], DR_int is internal dose rate [µGy•h^-1], DR_ext is external dose rate [µGy•h^-1], DR_total is total dose rate [µGy•h^-1].

DNA was extracted by first grinding 100 mg of frozen sample material, adding 500 µL of CTAB lysis solution (100mM Tris-HCl pH 7.5, 2% CTAB, 1.4M NaCl, 20mM EDTA pH 7.5, 1% NaHSO₃), and then incubating at 65°C for 30 minutes. Next, 500 µL of a 24:1 mixture of chloroform:IAA was added. After a gentle inversion, centrifugation at 12,000 g for 15 minutes followed. After carefully transferring the upper aqueous layer to a fresh tube, isopropanol precipitation step was performed by adding 350 µL of isopropanol, following incubation at -20°C for 15 minutes. The DNA pellet was spun down again (12,000 x g for 30 min) and the supernatant was removed before being washed with 70% ethanol and air-dried. The freshly extracted DNA was then obtained by resuspending the DNA pellet in 50 µL of water, making it available for further molecular analysis. DNA concentration was measured using the NanoDrop ND1000 and DNA integrity was assessed through gel electrophoresis (1% Agar, Tris-Acetate-EDTA (TAE) buffer).

Liquid nitrogen-frozen plant samples (50 mg) were shred using 2.0 mm Zirconia (zirconium (IV)oxide) beads (BioSpec Products) in a Mixer Mill MM 400 (Retsch) for 3 min at 30 Hz. The RNA was extracted using the RNeasy^® Plant Mini Kit (Qiagen, reference 74904), as recommended by the manufacturer. The NanoDrop ND1000 was then used to perform spectrophotometric measurements of RNA quality and quantity. Contaminant DNA was removed with the Thermo Fisher Scientific’s InvitrogenTM TURBO DNA freeTM kit (INVITROGEN, reference AM1907) and TaKaRa Bio’s PrimeScript 1st strand cDNA synthesis kit (TaKaRa Bio, reference 6110A), was used for cDNA synthesis, following the instructions of the manufacturers.

Gene expression analysis was performed by quantitative real-time PCR (qRT-PCR) on a 96-wells QuantStudio3 (Applied Biosystems) using Quantinova^® SYBR^® Green PCR kit (Cat. No. 208056). The qRT-PCR workflow was developed in accordance with the Eleven Golden Rules of quantitative RT-PCR (Udvardi et al., 2008), ensuring data quality and reliability. The analysis was run in 40 cycles using the following program: 5 seconds at 95°C, 12 seconds at 60°C. Comparison of the expression profile of five potential housekeeping genes, namely CYP, MAP2K1, BSL2, SBT3.3 and TUBB5 indicated that the latter was the most stable one across time-points and conditions, and it was used for normalizing target gene levels by the delta-Ct method employing the RefFinder tool (Xie et al., 2012) (see Supplementary Figure 8 ). The selected genes were identified by NCBI BLAST searches (Altschul et al., 1990) of A. thaliana genes against L. minor (serial number 1007 and ID number 5500) genome (Van Hoeck et al., 2015a). Additionally, we performed in silico validation of each primer pair to confirm sequence specificity both NCBI BLAST against L. minor’s transcriptome and the uMelt tool (Dwight et al., 2011). Primer sequences are reported in the Supplementary Table 2 . Relative gene expression levels were calculated using the comparative Ct (ΔΔCt) method (Livak and Schmittgen, 2001). First, the Ct values of target genes were normalized to TUBB5 to obtain ΔCt values. Then, the ΔΔCt was calculated by comparing the ΔCt of each treatment sample to that of the week 1 control. Fold changes in gene expression were subsequently derived using the formula 2^–ΔΔCt.

L. minor plants (5–6 plants per condition, 4 replicates) were harvested, transferred to a 6 well plate with 7.5 mL H₂O demineralized per each well, packed in aluminium foil and let in the dark for at least one hour to ensure dark adaptation. Photosynthetic activity was then analysed using the DUAL-PAM-100 (Walz) machine and its associated software (Dual PAM version 1.19). Measurements were carried out in the dark and included both induction curve (IC) and rapid light curve (RLC) analyses. Immediately prior to the analysis, the measuring cuvette was filled with demineralised water and just enough plants to cover the surface. The full measurement between two different samples was conducted every 15 minutes, resulting in 4 measurements per hour (Schreiber, 2004).

Dimethylformamide (DMF), at a ratio of 0.5 mL per 20 mg of fresh plant material, was used to extract the pigments from each sample, which were then incubated overnight at 4°C in the dark. The absorbance was measured in triplicate at wavelengths of 664, 647, and 480 nm, using 1:1, 1:2, and 1:4 dilutions. Pigment concentrations (carotenoids, chlorophyll a and chlorophyll b) were obtained following the calculations of Wellburn (1994).

The spectrophotometric measurement of oxidized (GSSG) and reduced (GSH) forms of glutathione was conducted as previously described (Queval and Noctor, 2007; Griffith, 1980; Tietze, 1969). Frozen plant samples (50–80 mg FW) were homogenized while kept frozen with 3 zirconia beads in a Retsch Mixer Mill MM400 at 30 Hz for 3.5 minutes. The measurements were carried out following the modifications previously reported (Horemans et al., 2014).

Plant tissue samples (50–80 mg FW) were stored at -80°C prior to analysis. Samples were homogenized with 20 chrome steel beads (Qiagen) and 2 mg of polyvinylpyrrolidone (PVP) in a Retsch Mixer Mill MM400 at 30 Hz for 3 minutes. Subsequently, 750 μL of extraction buffer (0.1 M TRIS; 1 mM Na2-EDTA; 1 mM DTT, pH 7.8) were added and the samples were centrifuged at 13000 rpm for 10 minutes at 4°C. To determine the Guaiacol peroxidase (GPOD) capacity, a solution comprising 110 μL of 0.1 M phosphate buffer, 10 μL of sample extract, and 80 μL of guajacol mastermix, composed of 90 mM guaiacol and 163 mM H₂O₂ mixed in a 1:1 ratio. This mixture was added to each well of a plastic 96-well plate. The kinetic monitoring of the appearance of tetraguajacol was conducted at 436 nm using the Biotek PowerWave XS Microplate Reader.

Telomere length was estimated by qRT-PCR following the protocol from Vaquero-Sedas et al (Vaquero-Sedas and Vega-Palas, 2014). The amplification was performed on a Rotor-Gene Q cycler (QIAGEN), using 5 ng of genomic DNA per vial, 1 µL of each primer (forward and reverse) at a concentration of 10µM, and 10µL of QuantiFast mix with SYBR (QIAGEN, catalog No. 204156) in a final volume of 20 µL. The program was set for holding for 15 minutes at 95°C. Then, 15 seconds at 94°C followed by 15 seconds at 49°C, for two cycles. This was followed by 32 amplification cycles consisting of 94°C for 15 seconds, 62°C for 10 seconds, 74°C for 15 seconds (with signal acquisition for telomeric Ct values), 84°C for 10 seconds, and 85°C for 15 seconds (with signal acquisition for the housekeeping gene). A final extension step was performed at 72°C for 60 seconds. The melt curve analysis followed, with pre-melt conditioning at 72°C for 90 seconds, then a ramp from 72°C to 95°C in 1°C increments with 5 seconds of hold at each step. Each run includes a standard curve of the single-copy housekeeping gene TUBB5 (Zhang et al., 2017) and for the telomeres that were created by serially diluting the total pool of DNA samples from 2 to 0.6 ng. Primers are listed in Supplementary Table 2 . Samples from different stages of exposure were distributed in the same run, and reactions were carried out in triplicate, ensuring that the difference in Ct values between duplicates was kept to less than 0.3. The standard curve’s efficiencies for telomeres and tubulin were both within the permissible standard range. One product per pair of primers was amplified, according to analysis of the melting curve. For each specific condition, the ratio (T/S) of the number of telomere tandem copy repeats to the number of housekeeping genes, was used to calculate telomere shortening. Following the approach described by Vaquero-Sedas et al., relative telomere length was determined using the comparative Ct (ΔΔCt) method (Vaquero-Sedas and Vega-Palas, 2014). For each sample, Ct values were obtained for the telomeric sequence and the single-copy reference gene TUBB5 in separate reactions. The ΔCt was calculated as Ct_TEL – Ct_TUBB5. A calibrator sample (control condition) was included on every plate to normalize across runs. The ΔΔCt was then obtained by subtracting the calibrator’s ΔCt from the sample’s ΔCt. Finally, the relative telomere length was expressed as 2^–ΔΔCt. All reactions were performed in technical duplicates, and primer efficiencies were confirmed to be within the acceptable range.

First, the frozen plant samples in Eppendorf tubes were ground to a fine powder with 15 chrome steel beads (Qiagen) in a Retsch Mixer Mill MM400 at 30 Hz for 3 minutes. 100 mL of extraction buffer were prepared as follows: Tris-HCl (1M pH 8.8) 12 mL, EDTA (0.25 M) 1.6 mL, SDS (10%) 40 mL, glycerol (50%) 20 mL, and distilled H₂O up to 100 mL. 2 mL of buffer were added to the samples, they were mixed well then centrifuged at 12000 rpm for 20 min at 4°C. The supernatant liquids containing total proteins fraction (protein extract) were transferred to clean Eppendorf tubes and stored at -20°C. This total protein content measurement was then carried out following the Bio Rad DC™ protein assay protocol: Reagent A’ was prepared by adding 20 µL of reagent S (C₁₂H₂₅NaO₄S 7%) to each mL of reagent A (NaOH 3%). A standard curve was prepared using a protein standard diluted to concentrations ranging from 0.2 mg/mL to 1.5 mg/mL. 5 µL of standards and samples were pipetted into a clean, dry microplate, followed by the addition of 25 µL of reagent A’ and 200 µL of reagent B (a diluted FOLIN reagent containing less than 1% each of the following ingredients: lithium sulfate, tungstic acid, sodium salt, molybdic acid, hydrochloric acid and phosphoric acid). After a 15-minute incubation at room temperature, absorbance was measured at 750 nm using the Biotek PowerWave XS Microplate Reader. The protein concentration of L. minor samples was determined by plotting absorbance against the concentration of known standards and interpolating the absorbance values of unknown concentration samples.

Plant samples were dried in an oven at 60°C for 3 days. Around 15 mg of dried biomass were weighed into a 2 mL Eppendorf tube and mixed with 3 chrome steel beads. Samples were then shredded using a Mixer Mill MM400 at 30 Hz for 4 minutes. Interfering pigments were extracted with 100% acetone: four extractions of 1 mL each were performed, followed by vortexing and centrifugation for 1.5 minutes at maximum speed. The beads and the supernatant were removed. Sugars were extracted with 80% ethanol: two extractions of 1 mL each were performed, and the entire pellet was transferred to a 15 mL tube for starch extraction, where 3 mL of 80% ethanol were then added. The samples were then centrifuged for 5 minutes at maximum speed (4800 rpm) using a Rotina 420R centrifuge. The supernatant containing soluble sugars was transferred to a new 15 mL tube labelled for sugar extraction. Starch residues were hydrolysed through an addition of 5 mL of 1.1% HCl. The samples were heated in a water bath at 100°C for 30 minutes. 5 mL of demineralized H₂O were added to each sample, resulting in a total volume of 10 mL. Before the spectrophotometric measurements, the following solutions were prepared: anthrone reagent was prepared adding 0.1 g anthrone (C₁₄H₁₀O) in 50 mL of 72% H₂SO₄, starch standards (0 to 10 mg starch/10 mL 1,1% HCl), and glucose standard (0 to 10 mg glucose/10 mL 80% EtOH). The concentration of unknown samples was analysed as follows: 100 mL of test solution (samples or standard solutions) were pipetted into a 2 mL Eppendorf tube and cooled to 0°C on ice. Then, 500 µL of ice-cold anthrone reagent were added, and all samples were vortexed. The samples were then heated for exactly 11 minutes in a water bath at 100°C, cooled rapidly to 0°C on ice, and vortexed again. Absorbance was read at 630 nm using the Biotek PowerWave XS Microplate Reader within an hour. The concentration of starch and glucose of L. minor samples were determined by plotting absorbance against the concentration of known standards and interpolating the absorbance values of unknown concentration samples.