Active DNA demethylation in plant companion cells reinforces transposon methylation in gametes

Abstract

The Arabidopsis thaliana central cell, the companion cell of the egg, undergoes DNA demethylation before fertilization, but the targeting preferences, mechanism, and biological significance of this process remain unclear. Here, we show that active DNA demethylation mediated by the DEMETER DNA glycosylase accounts for all of the demethylation in the central cell and preferentially targets small, AT-rich, and nucleosome-depleted euchromatic transposable elements. The vegetative cell, the companion cell of sperm, also undergoes DEMETER-dependent demethylation of similar sequences, and lack of DEMETER in vegetative cells causes reduced small RNA-directed DNA methylation of transposons in sperm. Our results demonstrate that demethylation in companion cells reinforces transposon methylation in plant gametes and likely contributes to stable silencing of transposable elements across generations.

Fig. 1

Local DME-dependent demethylation of maternal endosperm chromosomes. (A, C, and E) Transposons were aligned at the 5′ and 3′ ends (dashed lines) and average methylation levels for each 100-bp interval are plotted. (B, D, and F) Kernel density plots trace the frequency distribution of endosperm methylation differences for all 50-bp windows with an informative SNP. A shift of the peak with respect to zero represents a global difference; shoulders represent local differences.

Fig. 2

Local DME-dependent demethylation in the pollen vegetative cell. (A, C, and E) Average methylation in transposons was plotted as in Fig. 1. (B, D, and F) Kernel density plots, as in Fig. 1, of pollen methylation differences in 50-bp windows.

Fig. 3

DME demethylates small, AT-rich euchromatic TEs. (A) Box plots showing absolute fractional demethylation of 50-bp windows within transposons. Each box encloses the middle 50% of the distribution, with the horizontal line marking the median, and vertical lines marking the minimum and maximum values that fall within 1.5 times the height of the box. Differences between size categories are significant (P < 0.001, Kolmogorov-Smirnov test). (B) Significant (P < 0.001, Kolmogorov-Smirnov test) Spearman correlation coefficients between extent of demethylation and chromatin features. (C) Distribution of TEs and significantly differentially methylated cyto-sines (DMCs) (CG context; P < 0.0001, Fisher's exact test) with respect to genes.

Fig. 4

Demethylation activates TEs in companion cells and reinforces methylation in sperm. (A) DME-dependent demethylation of TEs that are maternally expressed in wild-type, but not dme, endosperm (table S4). DMRs (table S2) are underlined. (B) Kernel density plots of the CHH methylation differences between sperm cells from wild-type and dme/+ plants. (C) Kernel density plots of the CG methylation differences between vegetative cells from wild-type and dme/+ plants. (D) Snapshots of CG and CHH methylation in pollen on chromosome 1. Arrows point out loci with decreased CHH methylation in sperm and increased CG methylation in vegetative cells from dme/+ pollen.