Genome-wide quantitative identification of DNA differentially methylated sites in Arabidopsis seedlings growing at different water potential

Abstract

Background: In eukaryotes, the combinatorial usage of cis-regulatory elements enables the assembly of composite genetic switches to integrate multifarious, convergent signals within a single promoter. Plants as sessile organisms, incapable of seeking for optimal conditions, rely on the use of this resource to adapt to changing environments. Emerging evidence suggests that the transcriptional responses of plants to stress are associated with epigenetic processes that govern chromatin accessibility. However, the extent at which specific chromatin modifications contribute to gene regulation has not been assessed.

Methodology/principal findings: In the present work, we combined methyl-sensitive-cut counting and RNA-seq to follow the transcriptional and epigenetic response of plants to simulated drought. Comprehensive genome wide evidence supports the notion that the methylome is widely reactive to water potential. The predominant changes in methylomes were loci in the promoters of genes encoding for proteins suited to cope with the environmental challenge.

Conclusion/significance: These selective changes in the methylome with corresponding changes in gene transcription suggest drought sets in motion an instructive mechanism guiding epigenetic machinery toward specific effectors genes.

Figure 1. Growth phenotype of Arabidopsis Col 0 exposed to different water potential.

Upper panel: plants were first grown under constant light on plates containing ¼ MS media with a water potential of −0.25 MPa (ample water control condition). Seven-day-old seedlings were transplanted to new plates that were infused with PEG to obtain the indicated (−0.25, −1.2, and −2 MPa) water potentials as described in Methods. Representative seedlings are shown. Lower panel: After the 3 days treatment, seedlings were transferred to fresh ¼ MS media plates with water potential of −0.25 MPa. After 7 day, representative seedlings are shown.

Figure 2. Comparison of methyl sensitive cut counting profiles in simulated drought and ample water.

Digestion frequencies maps: Reads were mapped to the Arabidopsis reference genome (TAIR 7 release), and visualized in the UCSC genome browser. The local view shows the average digestion frequencies (average number of reads) associated to each one of the surveyed CG sites (track 3) included in the indicated interval of the chromosome 4. Sites overlapping 5′ region of genes were hypersensitive to the four restriction enzymes, indicating a predominance of un-methylated sites. Oppositely, methylated sites located at intergenic regions were consistently mapped with a low number of reads, suggesting high level of methylation. Description of the UCSC genome browser panel: Track 1- average digestion frequency profiles, 4 replicates, −0.25 MPa treatment; Track 2- average digestion frequency profiles, 4 replicates, −2 MPa treatment; Track 3- surveyed CG sites; Track 4- all CG sites; Track 5, TAIR 7 gene description; Track 6- AGRIS cis-regulatory elements.

Figure 3. Genome wide impact of simulated drought in the Arabidopsis methylome.

The coefficient of determination (r²) was used to quantify similitudes in the distribution of methylation among different samples. For the genome-wide estimation, 1,200 CG sites were randomly selected. A) scatter plot comparing the digestion frequencies in two replicates representing the ample water condition. B) scatter plot comparing digestion frequencies of samples treated at different water potential but in the same experimental replication. C) scatter plot comparing the digestion frequencies in two replicates representing the drought condition (−2 MPa). D) A summary of all vs. all pairwise comparisons. C = Control; # = replicate number; D = simulated drought, numbers are the correlation coefficients. Scale below represents heat map for similarity.

Figure 4. Distribution of differentially methylated sites (DMS) in relation to the transcription start sites.

Distribution of DMS in the Arabidopsis genome: Arabidopsis genes were aligned relative to their transcription start sites. The Y-axis represents the relative frequency of DMS computed for 100-bp intervals along the X-axis (red curve) or the relative frequency of all restriction sites used in this study (blue curve). The X-axis represents the relative distances to the TSS.