The ASYMMETRIC LEAVES complex maintains repression of KNOX homeobox genes via direct recruitment of Polycomb-repressive complex2

Abstract

Polycomb-repressive complexes (PRCs) ensure the correct spatiotemporal expression of numerous key developmental regulators. Despite their pivotal role, how PRCs are recruited to specific targets remains largely unsolved, particularly in plants. Here we show that the Arabidopsis ASYMMETRIC LEAVES complex physically interacts with PRC2 and recruits this complex to the homeobox genes BREVIPEDICELLUS and KNAT2 to stably silence these stem cell regulators in differentiating leaves. The recruitment mechanism resembles the Polycomb response element-based recruitment of PRC2 originally defined in flies and provides the first such example in plants. Combined with recent studies in mammals, our findings reveal a conserved paradigm to epigenetically regulate homeobox gene expression during development.

Figure 1.

CLF mediates the trimethylation of H3K27 at KNOX loci. ChIP experiments show that H3K27me3 levels across BP (left panel) and KNAT2 (right panel) are reduced in clf seedlings compared with wild type. Quantitative PCR values (mean ± SE) are shown as percentage of input and calculated from at least three independent biological replicates. Schematic representations of the BP and KNAT2 loci are shown at the top. (Black dashes) Positions of amplicons analyzed in ChIP assays; (red ovals) AS1–AS2 complex-binding sites; (arrow) transcription start site; (light blue boxes) 5′ and 3′ untranslated regions (UTRs); (dark blue boxes) ORF. Values on the X-axis are distances in kilobases relative to the transcription start site.

Figure 2.

The AS1–AS2 complex influences chromatin structure at BP and KNAT2. (A) Levels of H3K27me3 at BP (left panel) and KNAT2 (right panel) are significantly reduced in as1 and as2 leaves compared with wild type. (B) Occupancy levels of the PRC1 component LHP1 at BP (left panel) and KNAT2 (right panel) are likewise reduced in as2. (C) Enrichment levels for the active chromatin mark H3K4me3 across BP (left panel) and KNAT2 (right panel) in wild type, as1, and as2. Values (mean ± SE; n ≥ 3) are normalized to H3 levels (A,C) to correct for possible variation in nucleosome density at BP and KNAT2 or calculated as percentage of input (B). Values significantly different from wild type in at least one of the mutants are indicated. (*) P < 0.05; (**) P < 0.01. Schematic representations of the BP and KNAT2 loci are as in Figure 1. Control experiments establishing the specificity and efficiency of ChIP reactions are shown in Supplemental Figure 1.

Figure 3.

The AS1–AS2 complex physically interacts with PRC2 and directly recruits this complex to BP and KNAT2. (A) ChIP analyses showing CLF occupancy at specific regions of BP and KNAT2. CLF occupancy at these sites is significantly reduced in as1 compared with wild type. (*) P < 0.05; (**) P < 0.01. Values (mean ± SE; n ≥ 3) are relative to the enrichment of CLF at AG. (B) In plants carrying an inducible AS2-YFP fusion (pOlexA:AS2-YFP), MYC-CLF (top panel) and FIE-HA (bottom panel) coimmunoprecipitate with AS2-YFP specifically upon induction. AS2-YFP also coimmunoprecipitates with MYC-CLF (top panel) and FIE-HA (bottom panel) in immunoprecipitation assays with anti-MYC and anti-HA antibody, respectively, but not in control immunoprecipitation assays with IgG. Antibodies used in immunoprecipitation assays are listed at the top, and proteins detected by Western are at the right. (C) Bimolecular fluorescence complementation assays reveal direct physical interactions between AS1 and the PRC2 core components FIE and CLF and between AS2 and EMF2. Cobombardment of functional AS2 and CLF fusion constructs (Supplemental Table 2) yields no fluorescence signal. (Left panels) EYFP signal monitoring protein–protein interactions. (Middle panels) DAPI staining indicating positions of nuclei. (Right panels) Merged images.

Figure 4.

AS1–AS2-binding sites are necessary and sufficient for recruitment of PRC2 activity. ChIP analysis showing that H3K27me3 levels are enriched at the GFP-GUS reporter in lines where the transgene contains wild-type AS1–AS2-binding sites in the promoter. H3K27me3 levels are significantly reduced when the AS1–AS2-binding sites are mutated. (*) P < 0.05; (**) P < 0.01. Two pools comprising three independent transgenic lines were analyzed for each construct. Quantitative PCR values (mean ± SE) are normalized to H3 levels and calculated from three independent replicates. A schematic representation of the reporter lines is shown at the top. A BP promoter fragment spanning nucleotides −2707 to −1088 from the ATG was fused to the 35S minimal promoter and inserted upstream of a GFP-GUS fusion. (Black dash) Position of the amplicon analyzed; (red ovals) AS1–AS2 complex-binding sites; (arrow) transcription start site; (violet box) the 35S minimal promoter; (green box) GFP; (blue box) GUS. Sequence of the wild-type and mutated AS1–AS2-binding sites are shown below. Controls establishing the specificity and efficiency of ChIP reactions are shown in Supplemental Figure 7.