Description

The ability to process environmental information and adjust the phenotype in a way to face daily, seasonal and annual changes (phenotypic plasticity) is critical for survival, growth and reproduction ​of living organisms (Liancourt et al. 2013)​. Seeds respond to environmental cues experienced by mother plants, which can influence germination responses to and synchronization of the new generation of plants with optimal growth conditions and seasons (Donohue et al. 2010; Auge et al. 2017). Seeds from mother plants that experienced cold before flowering or during seed development and maturation show reduced germination—i.e., dormancy is induced (Chen et al. 2014; Burghardt et al. 2016). Temperature effects on progeny germination are mediated by the RNA-directed DNA Methylation pathway (RdDM): cool temperatures experienced by mother plants after bolting induce methylation through RDR6 (RNA-dependent RNA POLYMERASE6) of an ATHPOGON1 transposon located in the promoter region of the ALLANTOINASE (ALN) gene, a negative regulator of dormancy, reducing its expression and inducing seed dormancy (Iwasaki et al. 2019). RdDM is a pathway unique to plants that has been associated with multiple regulatory processes, including transposon silencing, responses to stresses, and development (Erdmann and Picard 2020). As RdDM influences the expression of transgenerational plasticity, we evaluated its role in the integration of the maternal life cycle environmental information and the expression of early developmental traits (germination) in seeds.

The seed environment showed a strong effect on germination: seeds incubated at 18ºC showed a higher germination proportion than those incubated at 24ºC for all genotypes and maternal treatments (Figure 1B). The maternal environment also showed a strong effect on germination as seeds from mother plants grown at warmer temperatures germinated more than those grown at cooler temperatures (Figure 1B). Interestingly, this effect was larger for seeds incubated at 18°C showing the maternal environmental history influences seed responses under certain environments (significant maternal treatment × incubation temperature, Table 2).

The timing of the environmental changes during the mother plant life cycle affected progeny germination, having a significant effect on the expression of genotypic variation (significant maternal treatment × genotype, Table 2). First, as expected, the environment after flowering showed a significant effect on progeny germination for both incubation temperatures and affected the expression of genotypic variability (18-18 vs. 18-24, and 24-18 vs 24-24; Table 3b; Figure 1C). Interestingly, the temperature before flowering affected progeny germination at both incubation temperatures as well: we observed significant interactions between the environment ‘before flowering’ and ‘genotype’ indicating that the maternal environment during vegetative growth, before seeds started developing, influenced the germination response of some genotypes (18-18 vs. 24-18, and 18-24 vs. 24-24; Table 3a; Figure 1C).

Given the strong genotypic effect, we analyzed the influence of the maternal environment for each genotype individually. Temperature before and after flowering experienced by the mother plants and the incubation temperature for each of the genotypes studied interacted in different ways to affect progeny germination for each genotype. Temperature before flowering influenced progeny germination and the expression of this effect depended on incubation temperature for Columbia, rdr6 and dcl234 seeds (significant ‘before flowering’ × ‘incubation temperature’ interaction, Table 4). This means that seeds responded to the maternal environment before flowering only when facing a certain temperature during imbibition; this could be due to conditions used in this study that can be masking other effects (i.e., combination of cooler incubation and warmer maturation environments). On the other hand, for nrpd2a and rdd mutant seeds, ‘after flowering’ temperature influenced the expression of ‘before flowering’ and seed incubation temperature effects on progeny germination (significant ‘before flowering’ × ‘after flowering’, and ‘after flowering’ × ‘incubation temperature’ interactions, Table 4). For these genotypes, the maternal environment during seed maturation affected not only the expression of the early life cycle effects, but also the response of seeds to their own immediate environment. For ago4 seeds, the effects of both before and after flowering maternal environments individually modified how the progeny responded to its immediate environment during imbibition (significant ‘before flowering’ × ‘incubation temperature’, and ‘after flowering’ × ‘incubation temperature’ interactions, Table 4). Finally, the response of rdr2 seeds to the maternal environment showed a response influenced by the environments of both the mother plant and the seeds, showing a complex response to across and within-generation environmental cues (significant ‘before flowering’ × ‘after flowering’ × ‘incubation temperature’ interaction, Table 4).

Lastly, our results suggest that the response of progeny seeds to the thermal maternal environment is not mediated by the RdDM canonical pathway as the rdr2 mutant did not show significant differences to the wild type in any of the conditions tested in this study (Table 5; Figures 1B and 1C). The RdDM non-canonical pathway mutants rdr6 and dcl234, and the demethylases rdd mutant showed reduced germination, while ago4 and nrpd2a germinated more than the wild type (Table 5, Figures 1B and 1C). Most effects were observed for seeds incubated at 24°C—incubation at 18°C could be masking genotype effects due to the inductive nature of that environment. A correct methylation homeostasis seems to be required for seeds to germinate and respond properly to their environment as rdr6, dcl234 and rdd mutants showed an overall effect on germination. This could be caused by direct or indirect effects of RdDM on germination; however, we cannot discern between these possibilities with our experimental design. For other mutants, genotype effects on progeny germination showed specificity to the maternal environment: AGO4 would repress germination when mother plants are grown in warmer temperatures before flowering, while NRPD2A is likely repressing the response to cooler environments perceived after flowering (Figure 1B; Table 5). Whereas the response of nrpd2a mutant seeds is consistent with previous reports for RNA polymerase mutants (Iwasaki et al. 2019), the novel effect of ago4 in impairing the response of the seeds to environmental cues experienced early in the mother plant life cycle needs to be studied further.

Altogether, our results suggest that integration of environmental information across generations through the RdDM pathway can potentially affect the expression of optimal phenotypes for early traits in response to previous generations’ environments and, in consequence, adaptive processes.

TABLES

Table 1. Information of the mutants used in this study.

Table 2. Interaction model between the variables ‘maternal treatment’ (Mat), ‘incubation temperature’ (IT) and ‘genotype’ (GT). Germination proportions were analyzed with logit-linked generalized linear models, and likelihood ratios were tested based on χ². The triple interaction was dropped to increase power as it did not improve the model (based on Akaike Information Criterion, AIC). Table shows Chi-square (χ²), degrees of freedom (df) and p-values. Significance is shown with asterisks: ***, p < 0.001. Source of variation: ‘Mat’, maternal treatment; ‘IT’, incubation temperature; ‘GT’, genotype.

Table 3. Effects of timing of maternal environmental changes (before and after flowering) on germination under each incubation temperature (‘IT’, columns). Maternal effects of temperature changes perceived by mother plants before (‘BF’) or after flowering (‘AF’) with the environment in the other life stage fixed, and their interactions with the variable ‘genotype’ (‘GT’). Environmental effects of a) before (24-24 vs. 18-24, and 24-18 vs. 18-18) or b) after flowering (24-24 vs. 24-18, and 18-24 vs. 18-28). Table shows LR Chi-square (χ²) and significant differences based on p-values: ***, p < 0.001.

Table 4. Maternal treatments and seed environment effects on germination responses for each genotype. Temperature effects perceived by mother plants before (‘BF’) or after flowering (‘AF’), and the interaction with the seed environment (‘IT’) tested for each genotype individually (columns). Table shows LR Chi-square values and significant differences based on p-values: ***, p < 0.001; **, p < 0.01; *, p < 0.05.

Table 5. Genotype effects on the germination responses. Comparison between the Columbia wild type and RdDM pathway mutants within each maternal treatment (Mat: before and after flowering temperatures) and incubation temperature (IT). Table shows the LR Chi-square values and significant differences based on p-values: ***, p < 0.001; **, p < 0.01. See Figure 1c for effect sizes and direction of responses.