Abstract
Desiccation tolerance (DT) has contributed greatly to the adaptation of land plants to severe water-deficient conditions. DT is mostly observed in reproductive parts in flowering plants such as seeds. The seed DT is lost at early post germination stage but is temporally re-inducible in 1 mm radicles during the so-called DT window following a PEG treatment before being permanently silenced in 5 mm radicles of germinating seeds. The molecular mechanisms that activate/reactivate/silence DT in developing and germinating seeds have not yet been elucidated. Here, we analyzed chromatin dynamics related to re-inducibility of DT before and after the DT window at early germination in Medicago truncatula radicles to determine if DT-associated genes were transcriptionally regulated at the chromatin levels. Comparative transcriptome analysis of these radicles identified 948 genes as DT re-induction-related genes, positively correlated with DT re-induction. ATAC-Seq analyses revealed that the chromatin state of genomic regions containing these genes was clearly modulated by PEG treatment and affected by growth stages with opened chromatin in 1 mm radicles with PEG (R1P); intermediate openness in 1 mm radicles without PEG (R1); and condensed chromatin in 5 mm radicles without PEG (R5). In contrast, we also showed that the 103 genes negatively correlated with the re-induction of DT did not show any transcriptional regulation at the chromatin level. Additionally, ChIP-Seq analyses for repressive marks H2AK119ub and H3K27me3 detected a prominent signal of H3K27me3 on the DT re-induction-related gene sequences at R5 but not in R1 and R1P. Moreover, no clear H2AK119ub marks was observed on the DT re-induction-related gene sequences at both developmental radicle stages, suggesting that silencing of DT process after germination will be mainly due to H3K27me3 marks by the action of the PRC2 complex, without involvement of PRC1 complex. The dynamic of chromatin changes associated with H3K27me3 were also confirmed on seed-specific genes encoding potential DT-related proteins such as LEAs, oleosins and transcriptional factors. However, several transcriptional factors did not show a clear link between their decrease of chromatin openness and H3K27me3 levels, suggesting that their accessibility may also be regulated by additional factors, such as other histone modifications. Finally, in order to make these comprehensive genome-wide analyses of transcript and chromatin dynamics useful to the scientific community working on early germination and DT, we generated a dedicated genome browser containing all these data and publicly available at https://iris.angers.inrae.fr/mtseedepiatlas/jbrowse/?data=Mtruncatula.