Tandem repeats upstream of the Arabidopsis endogene SDC recruit non-CG DNA methylation and initiate siRNA spreading

Abstract

Plants use siRNAs to target cytosine DNA methylation to both symmetrical CG and nonsymmetrical (CHG and CHH) sequence contexts. DNA methylation and siRNA clusters most frequently overlap with transposons in the Arabidopsis thaliana genome. However, a significant number of protein-coding genes also show promoter DNA methylation, and this can be used to silence their expression. Loss of the majority of non-CG DNA methylation in drm1 drm2 cmt3 triple mutants leads to developmental phenotypes. We identified the gene responsible for these phenotypes as SUPPRESSOR OF drm1 drm2 cmt3 (SDC), which encodes an F-box protein and possesses seven promoter tandem repeats. The SDC repeats show a unique silencing requirement for non-CG DNA methylation directed redundantly by histone methylation and siRNAs, and display spreading of siRNAs and methylation beyond the repeated region. In addition to revealing the complexity of DNA methylation control in A. thaliana, SDC has important implications for how plant genomes utilize gene silencing to repress endogenous genes.

Figure 1.

Epigenomic profile of SDC. Graphical representation of DNA methylation and siRNAs detected in the region of SDC (At2g17690). The SDC ORF is indicated as a light-green bar marked with chevrons. The SDC 5′-UTR mapped by RACE is highlighted by dark green bars. The promoter tandem repeats are marked by the light green block running down the figure. DNA methylation detected by methyl-cytosine immunoprecipitation (mCIP) and hybridization to genomic tiling arrays is highlighted beneath as colored bars in wild type (green), drm1 drm2 cmt3 (orange), and met1 (red) (Zhang et al. 2006). Methylation state of individual cytosines as measured by Solexa sequencing of bisulfite treated DNA is represented as small rectangles. Methylation is shown by shading (dark is methylated) and color-coded according to sequence context (CG, green; CHG, blue; CHH, red) (Cokus et al. 2008). Small RNAs detected by sequencing are represented by blue bars with white chevrons indicating their strandedness (Rajagopalan et al. 2006).

Figure 2.

SDC is necessary and sufficient for drm1 drm2 cmt3 developmental phenotypes. (A) Leaf phenotypes of plants showing either wild-type (Col) flat leaves or mutant leaves that are curled downward, as in drm1 drm2 cmt3. (B) SDC expression detected by Northern blotting and hybridization. Loading was checked by hybridization with a UBIQUITIN10 (UBQ) probe.

Figure 3.

Genetic analysis of DNA methylation and siRNA accumulation at SDC. (A) Graphical representation of SDC with the tandem repeats marked as arrows within the white box. The ORF is marked by the black box and the 5′-UTR by gray boxes. Black lines show the positions of the two bisulfite sequencing products and LNA siRNA probes. (B) DNA methylation analyzed by sodium bisulfite sequencing shown as percentage CG (black), CHG (gray), and CHH (white). The Bisulfite-I region data has been separated into the tandemly repeated (Repeats) region and upstream, unique sequences (Upstream). The upstream sequence contains no CHG sites. The Bisulfite-II region data is labeled as “Downstream.” (C) Detection of SDC siRNAs by Northern blotting and hybridization. SDC LNA probes were used either for the tandem repeat siRNAs (SDC tandem) or for spreading siRNAs (SDC spread). Loading was analyzed by hybridization with a probe to microRNA159 (miR159).

Figure 4.

DRM2 and siRNA-dependent silencing of SDC transgenes. (A) Analysis of DNA methylation using sodium bisulfite sequencing of gSDC transgenes transformed by Agrobacterium into wild-type (Columbia) or drm1 drm2 backgrounds. Percentage methylation is shown for CG (black), CHG (gray), and CHH (white) sites. (B) Analysis of SDC expression by Northern blotting and hybridization in wild-type and mutant backgrounds transformed with gSDC transgenes. Loading was analyzed by hybridization with a UBIQUITIN10 (UBQ) probe. (C) Northern blot analysis of SDC expression in wild-type and mutant plants transformed with gSDC transgenes. (D) Northern blot analysis of SDC expression after drm1 drm2 mutants with unsilenced gSDC were backcrossed to wild type.

Figure 5.

Genetic requirements for establishment of SDC silencing during backcrossing. (A) Leaf phenotypes and (B) SDC expression analyzed by Northern blotting and hybridization in wild-type and mutant plants used in the indicated backcrossing experiments. Loading was checked by hybridization with a UBIQUITIN10 (UBQ) probe. (dd) drm1 drm2; (c) cmt3; (nn) nrpd2a nrpd2b; (k) kyp; (ddc) drm1 drm2 cmt3; (ddcnn) drm1 drm2 cmt3 nrpd2a nrpd2b; (ddk) drm1 drm2 kyp.