SDG7 and SDG8 mediate H3K36me2 and K36me3 deposition

a, Heatmap representation of gene expression levels for class II SDG family genes in specific tissues. b, Diagrams of Arabidopsis class II SDG family proteins. c, 3D structure prediction of the conserved domain of SDG7 and SDG8 by AlphaFold 2. Left, Inter-chain predicted aligned error (PAE) plot for the best-ranked model structures of SDG7 and SDG8 by AlphaFold 2. Right, Structure models of SDG7 and SDG8 predicted by AlphaFold 2. Yellow, green and purple lines indicate the AWS, SET and post-SET domains, respectively. d, Overlay of the crystal structure for ATXR5 SET domain in complex with K36me3 histone H3 peptide (Dark green: 5VAC), SDG8 (pale purple) and SDG7 (purple). The right panels show the interfaces of SDG7 and SDG8 with histone tails. e, Representative photographs of the wild type (WT), sdg7, sdg8 and sdg7 sdg8 at 25 days after germination. Scale bar, 1 cm. f, Plant height of the WT, sdg7, sdg8and sdg7 sdg8(WT: n = 35; sdg7: n = 25; sdg8: n = 25; sdg7 sdg8: n = 18). Significant differences were calculated based on one-way ANOVA tests (p = 1.1 x 10-16). Different letters indicate significant differences based on post-hoc Tukey’s HSD test (p < 0.05: See the Data Source file for all combinations of the exact p-values). g, Top view of stage 14 flowers from soil-grown WT, sdg7, sdg8 and sdg7 sdg8. Scale bar, 0.5 mm. h, Representative photographs showing the roots of WT, sdg7, sdg8 and sdg7 sdg87-day-old seedlings. Scale bar, 1 cm. i, Primary root length of WT, sdg7, sdg8 and sdg7 sdg8 (WT: n = 41; sdg7: n = 33; sdg8: n = 46; sdg7 sdg8: n = 15). Significant differences were calculated based on one-way ANOVA tests (p = 1.1 x 10-16). Different letters indicate significant differences based on post-hoc Tukey’s HSD test (p < 0.05: See the Data Source file for all combinations of the exact p-values). j, Averaged profiles of H3K36me3 (left) and H3K36me2 (right) signal intensity around genes in WT and sdg7 sdg8. k, Heatmaps of H3K36me3 (left) and H3K36me2 (right) accumulation across genes in WT and sdg7 sdg8. l, Scatterplot of H3K36me3 (above) and H3K36me2 (below) levels in WT and sdg7 sdg8. Top, shoot samples. Bottom, root samples. Hypermethylated and hypomethylated genes in sdg7 sdg8 relative to WT are shown in light and dark colors, respectively. m, Selected enriched GO terms determined by their – log10-adjusted p-values. H3K36me3 (top) and H3K36me2 (bottom) hypomethylated genes in sdg7 sdg8 relative to WT. Top, shoot samples. Bottom, root samples. n, H3K36me3 and H3K36me2 ChlP-qPCR at MAF3and At1g01040 in the WT and sdg7 sdg8. ChIP-seq results and the diagram of the genomic loci and PCR amplicons are shown below. Values are means ± SE (n = 4 independent experiments). Significant differences were calculated based on one-way ANOVA tests (K36me3 at MAF3: p = 3.6 x 10-3; K36me3 at DCL1: p = 4.8 x 10-7; K36me2 at MAF3: p = 2.9 x 10-3; K36me2 at DCL1: p = 7.0 x 10-3). Different letters indicate significant differences based on post-hoc Tukey’s HSD test (p < 0.05: See the Data Source file for all combinations of the exact p-values).

SDG7 is present at TSS and TES, whereas SDG8 occupies gene body

a, Averaged profiles of SDG8-GFP and SDG7-VENUS signal intensity around genes in wild type. b, Distribution of peak position for SDG8-GFP and SDG7-VENUS. The percentages of binding peaks at each region type are shown. c, Top 20 GO terms as determined by their –log10-adjusted p-values using SDG8-(left) or SDG7- (right) bound genes. d, Scatterplot showing representative clusters in a two-dimensional space based on semantic similarities between GO terms. Color and size indicate p-value and the frequency of the GO term in the underlying gene ontology annotation database, respectively. Left: SDG8-bound genes. Right: SDG7-bound genes. e, Upset bar plots showing the number of overlapping genes among SDG8- and SDG7-bound genes and H3K36me3- and H3K36me2-hypermethylated genes in the wild type. f, Left: Venn diagram showing the overlap between H3K36me3- and H3K36me2-hypomethylated genes in sdg7 sdg8 and direct SDG8 (top) and SDG7 (bottom) targets. Right, distribution of peak position for SDG8-(left) and SDG7-(right) bound peaks according to their functional genomic distribution, using the same color legend as in b. AnnotatePeak was used for assignment.

SDG7 displaces PRC2 complex from PRE to overcome Polycomb repression

a, De novo motif analysis under each SDG7 peak summit. p-values obtained from HOMER are shown. b, Venn diagram showing the overlap between SDG7-bound cis-elements and all PREs bound by transcription factors. c, Classification of transcription factor families seen binding to SDG7 peak summit and PREs. d, ChIP-seq signals for H3K27me3, FIE, CLF, AZF1, BPC1, and SDG7 at the selected target loci. The gene models are shown as black bars and lines at the bottom of each panel. e, SDG7-HA-GR, CLF-GFP, and H3K27me3 ChIP-qPCR at PREs of PTL, SVP, and AG genes in mock- and dexamethasone (DEX)-treated pro35S:SDG7-HA-GR seedlings. ChIP-seq assay and diagram of the genomic loci and PCR amplicons are shown below. Values are means ± SE (n = 3 independent experiments). Significant differences were calculated based on one-tailed Student’s t-test (SDG7-HA at PTL: p = 0.039; SDG7-HA at SVP: p = 0.046; SDG7=HA at AG: p = 4.7 x 10-3; CLF-GFP at PTL: p = 6.5 x 10-3; CLF-GFP at SVP: p = 0.011; CLF-GFP at AG: p = 0.01; H3K27me3 at PTL: p = 0.040; H3K27me3 at SVP: p = 0.029; H3K27me3 at AG: p = 0.018). f, Representative photographs of sdg7 sdg8, clf and sdg7 sdg8 clf at 25 days after germination. Scale bar, 1 cm. g, Top view of stage 14 flowers from soil-grown sdg7 sdg8, clf and sdg7 sdg8 clf. Scale bar, 0.5 mm. h, Representative photographs of roots from sdg7 sdg8, clf and sdg7 sdg8 clf. Scale bar, 1 cm. i, Plant height in sdg7 sdg8, clf and sdg7 sdg8 clf (sdg7 sdg8 n = 18; clf: n = 10; sdg7 sdg8 clf: n = 22). Significant differences were calculated based on one-way ANOVA tests (p = 1.1 x 10-16). Different letters indicate significant differences based on a post-hoc Tukey’s HSD test (p < 0.05: See the Data Source file for all combinations of the exact p-values). j, Root length in sdg7 sdg8, clf and sdg7 sdg8 clf (sdg7 sdg8: n = 15; clf: n = 29; sdg7 sdg8 clf: n = 46). Significant differences were calculated based on one-way ANOVA tests (p = 1.1 x 10-16). Different letters indicate significant differences based on a posthoc Tukey’s HSD test (p < 0.05: See the Data Source file for all combinations of the exact p-values). k, Volcano plot showing differentially expressed genes (DEGs). Top, WT vs clf. Middle, WT vs sdg7 sdg8. Bottom, sdg7 sdg8 vs sdg7 sdg8 clf. l, Venn diagram showing the overlap between differentially expressed genes identified by RNA-seq. Top, upregulated genes in clf vs WT, downregulated genes in sdg7 sdg8 vs WT, and upregulated genes in sdg7 sdg8 clf vs sdg7 sdg8. Bottom, downregulated genes in clf vs WT, upregulated genes in sdg7 sdg8 vs WT, and downregulated genes in sdg7 sdg8 clf vs sdg7 sdg8. m, Heatmap representation of gene expression levels in wild type (WT), clf, sdg7 sdg8, and sdg7 sdg8 clf at 7 days after germination. The Z-scores are based on FPKM values. n, Interactive graph view of 827 genes (genes shown by while color in Fig. 3m) generated by REVIGO.

SDG8 and PAF1C propagate H3K36 methylation together with RNA Pol II

a, Venn diagram showing overlapping downregulated genes in sdg7 sdg8 vs wild type (WT), ef7 vs WT, vip3 vs WT, vip4 vs WT, vp5vs WT and vip6 vs WT. b, Heatmap representation of gene expression levels for 230 PAF1 C-regulated genes in WT, elf7, vip3, vip4, vip5and vip6. RPKM values are shown. c, Venn diagram showing the overlap between 230 PAF1C complex–regulated genes, upregulated genes in clf vs WT, and upregulated genes in sdg7 sdg8 clfvs sdg7 sdg8. d, ChlP-seq signals for CLF, H3K27me3, SDG8, SDG7, RNA Pol II, RNA Pol II-S2P, RNA Pol II-S5P H3K36me3 and H3K36me2 at the selected target loci and their expression levels in WT, clf, sdg7 sdg8 and sdg7 sdg8 clf. The gene models are shown as black bars and lines at the bottom of each panel. e, Expression levels of COOLAIR, FLC, At4g13572, At4g13575 and UMAMIT14 in WT, clf, sdg7 sdg8and sdg7 sdg8 clf Values are means ± SE (n = 3 independent experiments). f, Molecular mechanism of K27/K36 methylation propagation by PRC2 and SDG8/SDG7. H3K27me3; purple circles. H3K36me3; green circles. H3K36me2; blue circles.

Overlapping expression patterns of SDG7 and SDG8 in Arabidopsis thaliana.

a–h, Expression pattern of SDG7 and SDG8 using proSDG7:SDG7-GUS (a-d) and proSDG8:GUS (e-h) transgenic lines in cotyledons (a,e), shoot apical meristems (b,f), top and side views of primary inflorescences (c,g and d,h, respectively). Scale bars, 500 μm (e, g, h) or 50 μm (f). i–p, Expression pattern of SDG7 and SDG8 using proSDG7:SDG7-GUS (i-l) and proSDG8:SDG8-GFP (m-p) transgenic lines in primary root tips (i,m), stage II lateral roots (j,n), stage IV lateral roots (k,o) and emerged lateral roots (l,p). Scale bars, 50 μm (m-p).

Structural prediction of class II SDG family proteins.

a, Sequence alignment of the conserved domains in SDG7, SDG8 and other SDG members. Blue, green and red lines indicate the AWS, SET and post-SET domains, respectively. Asterisks show conserved amino acids. b, 3D structure prediction of the conserved domains of SDG4, SDG26 and SDG24 by AlphaFold 2. Blue, green and red lines indicate the AWS, SET and post-SET domains, respectively. c,d, Structure prediction of full-length SDG8 and SDG7 by AlphaFold 2. Top, Inter-chain predicted aligned error (PAE) plot for the best-ranked model structures of SDG7 (c) and SDG8 (d) by AlphaFold 2. Bottom left, Multiple Sequence Alignment (MSA) sequence coverage by AlphaFold 2. Bottom right, Predicted local distance difference test (LDDT) per position of the five models by AlphaFold 2.

Isolation of sdg7 sdg8 double mutants.

a, Diagram of the SDG7 and SDG8 loci. The black boxes, gray boxes and short lines inside gene bodies indicate exons, untranslated regions, introns and qPCR-amplified regions, respectively. White triangles indicate the position of T-DNA insertions; black arrowheads denote genotyping primers (sdg7-2 LP, sdg7-2 RP, sdg8-2 LP and sdg8-2 RP); and white arrowheads denote RT-PCR primer pairs. b, c, Genotyping results of WT and sdg7-2 sdg8-2 double mutant. b. SDG7 genotyping. c. SDG8 genotyping. d–g, Cellular organization of the root meristem, as visualized by propidium iodide (PI) staining in wild type (WT; d, f) and sdg7 sdg8 (e, g). h,i, Cellular organization of the columella root cap, as visualized by Lugol staining in WT (h) and sdg7 sdg8 (i). Scale bar, 50 μm. j–l, Representative photographs of the sdg7-2 sdg8-2 (j), sdg7-3 sdg8-2 (k) and sdg7-4 sdg8-2 (l) double mutants at 25 days after germination. Insets: top view of 1st true leaves. Scale bar, 1 cm. m–o, Top view of stage 14 flowers from soil-grown sdg7-2 sdg8-2 (m), sdg7-3 sdg8-2 (n) and sdg7-4 sdg8-2 (o). Scale bar, 0.5 mm. p–r, Representative photograph of the roots from the sdg7-2 sdg8-2 (p), sdg7-3 sdg8-2 (q) and sdg7-4 sdg8-2 (r) double mutants. Scale bar, 1 cm.

Phenotypes of the sdg7 sdg8 double mutant.

a, Top views of the wild type (WT), sdg7, sdg8 and sdg7 sdg8 at 10 days after germination. b, Side views of the WT, sdg7, sdg8 and sdg7 sdg8 at 25 days after germination. c, Number of rosette leaves (left) and cauline leaves (middle), and days to bolting (right) in the WT, sdg7, sdg8 and sdg7 sdg8 (WT: n = 20; sdg7: n = 34; sdg8: n = 28; sdg7 sdg8: n = 26). d, Top view of 5th true leaf in the WT, sdg7, sdg8 and sdg7 sdg8. Scale bars (a, b, d), 1 cm. e, Appearance of palisade cells from 5th true leaves in the WT, sdg7, sdg8 and sdg7 sdg8. Scale bar, 20 μm. f, Leaf size (left), cell size (middle), and cell number (right) in the WT, sdg7, sdg8 and sdg7 sdg8 (WT: n = 24; sdg7: n = 24; sdg8: n = 24; sdg7 sdg8: n = 24). g, Top view of stage 14 flowers from soil-grown WT, sdg7, sdg8 and sdg7 sdg8. Scale bar, 0.5 mm. h, Petal epidermal cells in the WT, sdg7, sdg8 and sdg7 sdg8. Scale bar, 10 μm. i, Petal area (left), petal cell number (middle), and petal cell size (right) in the WT, sdg7, sdg8, and sdg7 sdg8 (WT: n = 50; sdg7: n = 50; sdg8: n = 50; sdg7 sdg8: n = 40). j, Lateral root number (left) and lateral root density (right) in the WT, sdg7, sdg8 and sdg7 sdg8 at 7 days after germination (WT: n = 27; sdg7: n = 33; sdg8: n = 46; sdg7 sdg8: n = 15). Significant differences were calculated based on one-way ANOVA tests (p = 1.1 x 10-16). Different letters indicate significant differences based on post-hoc Tukey’s HSD test (p < 0.05: See the Data Source file for all combinations of the exact p-values).

Molecular defects of the sdg7 sdg8 double mutants.

a–d, GUS staining pattern from the DR5:GUS reporter construct in the WT (a, b) and sdg7 sdg8 (c, d). Seven-day-old (a, c) and 13-day-old (b, d) seedlings are shown. Left, top views of seedlings. Right, Close-up views of cotyledons. Scale bar, 3 mm. e,f, YFP fluorescence pattern from the proCYCB1;2:CYCB1;2-YFP transgene in the shoot apical meristems from WT (e) and sdg7 sdg8 (f). Scale bar, 50 μm. g,h, YFP fluorescence pattern from the proCYCB1;2:CYCB1;2-YFP transgene in the inflorescence meristems of WT (g) and sdg7 sdg8 (h). Scale bar, 100 μm. i-r, GFP fluorescence or GUS staining pattern from the indicated cell markers in the primary roots of WT (i, k, m, o, q) and sdg7 sdg8 (j, l, n, p, r). proWOX5:NLS-GFP (i, j), proPLT2:PLT2-YFP (k, l), proSHR:SHR-GFP (m, n), proCYCB1;2:CYCB1;2-NLS-YFP (o, p) and DR5:GUS (q, r). s-x GUS staining pattern from the indicated cell markers in lateral roots of WT (s, u, w) and sdg7 sdg8 (t, v, x). DR5:GUS (s-v) and proLBD16:GUS (w, x). Stage II (s, t, w, x) and Stage IV (u, v). Scale bars, 50 μm.

Phenotypic and molecular defects in the clf sdg7 sdg8 triple mutants.

a, Top views of the sdg7 sdg8, clf and sdg7 sdg8 clf at 20 days after germination. Scale bar, 1 cm. b, Top views of the 1st leaf in the sdg7 sdg8, clf and sdg7 sdg8 clf at 20 days after germination. Scale bar, 1 cm. c, d, Leaf number at flowering and size of 1st true leaf in the sdg7 sdg8, clf and sdg7 sdg8 clf (sdg7 sdg8: n = 18; clf: n = 20; sdg7 sdg8 clf: n = 17). Significant differences were calculated based on one-way ANOVA tests (p = 1.1 x 10-16). Different letters indicate significant differences based on post-hoc Tukey’s HSD test (p < 0.05: See the Data Source file for all combinations of the exact p-values). e, Hierarchical clustering. The dendrogram of all RNA-seq samples obtained with the R package stat is shown with the bootstrap values on each node. f, Principal component analysis (PCA). The PCA plot illustrates all twelve RNA-seq samples, with replicates represented by the same color.