Small RNAs guide histone methylation in Arabidopsis embryos

Abstract

Epigenetic reprogramming occurs during gametogenesis as well as during embryogenesis to reset the genome for early development. In flowering plants, many heterochromatic marks are maintained in sperm, but asymmetric DNA methylation is mostly lost. Asymmetric DNA methylation is dependent on small RNA but the re-establishment of silencing in embryo is not well understood. Here we demonstrate that small RNAs direct the histone H3 lysine 9 dimethylation during Arabidopsis thaliana embryonic development, together with asymmetric DNA methylation. This de novo silencing mechanism depends on the catalytic domain of SUVH9, a Su(Var)3-9 homolog thought to be catalytically inactive.

Keywords: DNA methylation; epigenetic reprogramming; histone methylation; plant embryogenesis; small RNA molecules.

Figure 1.

Small RNA-dependent and -independent DNA methylation in embryos. (A) Hypomethylated DMRs were extracted for ago4 ago6 ago9 and suvh4 suvh5 suvh6 mutants and mapped on TE features and broken down in families. Distribution of all transposable elements from the Arabidopsis thaliana genome is presented at the left. (B) Distributions representing methylation ratios of the different contexts are identified above in the different genetic backgrounds indicated below. Ratios were averaged within the genomic regions identified by the DMR analysis. Differences between distributions and the first WT distribution were calculated using a Kolmogorov–Smirnov test, and asterisks represent the significance of the difference. (*) P ≤ 1.0 × 10⁻⁴, (**) P ≤ 1.0 × 10⁻¹¹, (***) P ≤ 1.0 × 10⁻¹⁵. (C) Methylation plot of representative genomic regions with coordinates indicated below and containing TEs. Methylation context is color-coded as indicated above, and genotypes are indicated at the right. Small RNA track was generated using published embryo data (Ye et al. 2016).

Figure 2.

Small RNA-dependent and -independent H3K9me2 in embryos. (A) Distributions of normalized H3K9me2 ratios [log₂(H3K9met/H3)] were calculated over the genomic regions defined by the DMR analysis. Biological replicates are presented for the genomic backgrounds indicated below. Differences between distributions and the first WT distribution were calculated using a Kolmogorov–Smirnov test, and asterisks represent the significance of the difference. (*) P ≤ 1.0 × 10⁻⁴, (**) P ≤ 1.0 × 10⁻¹¹, (***) P ≤ 1.0 × 10⁻¹⁵. (B,C) Immunofluorescence imaging of H3K9me2 (red), H3K4me1 (green), and chromatin (gray) in representative embryos of indicated background. Early embryos from the four-cell to the 16-cell stage are presented. “En” and white arrows point at endosperm cells. (C) DIC images are presented at the left; early stages are presented at the top; and later stages are below.

Figure 3.

Functional complementation of SUVH9. (A) Confocal imaging of mCitrine signal and transmitted light on mature embryo cotyledons of the genotype indicated above. (B) Pictures of 4-wk-old F1 plants resulting from the crosses indicated above. (C) Distributions of normalized H3K9me2 ratios [log₂(H3K9me2/H3)] were calculated over the genomic regions identified by the MACS2 broad peak analysis. (D) Distributions representing methylation ratios of the different contexts are identified above in the different genetic backgrounds indicated below. Ratios were calculated over the genomic regions identified by the MACS2 broad peak analysis. Distances between distributions were calculated using a Kolmogorov–Smirnov test, and asterisks represent the significance of the difference. (*) P ≤ 1.0 × 10⁻⁶, (**) P ≤ 1.0 × 10⁻¹⁰, (***) P ≤ 1.0 × 10⁻¹⁵. The difference is between the distribution identified and the equivalent control unless otherwise represented.