An open EAR for polycomb repressive complexes

Since their discovery over 70 years ago, the polycomb group proteins (PcGs) have been implicated in the epigenetic regulation of genes that drive cell-fate choices and cell identity during development (Schuettengruber et al., 2017). Polycomb repressive complexes (PRCs) mediate transcriptional repression by catalyzing posttranscriptional modifications (PTMs) of histone proteins. In a canonical model, PRC2 first mediates the trimethylation of lysine 27 on histone H3 (H3K27me3). This mark is then bound by PRC1, which catalyzes the ubiquitination of histone H2A on lysine 121 (H2AK121ub). These histone PTMs finally recruit regulatory factors that lead to chromatin reorganization and inhibition of transcription. Nonetheless, PRCs can also function independently, and the exact recruitment mechanisms remain to be fully understood.

Cis-regulatory sequences and trans-acting components such as DNA-binding proteins and transcription factors (TFs) play a role in PRC2 targeting. Recently, VIVIPAROUS1/ABI3-LIKE (VAL)1/2, BMI, and RING1B proteins have been implicated in the recruitment and ubiquitination activity of PRC1 (Yang et al., 2013; Qüesta et al., 2016). Interestingly, H2A ubiquitination at some seed maturation genes precedes PRC2-mediated H3K27me3 deposition, thus challenging the sequential PRC2-PRC1 order of action. To further dissect the relationship between both PRCs, Baile et al. (2021) engineered a synthetic promoter system to tether specific PcG proteins and monitor their effects on histone PTMs and transcriptional repression.

In the designed system, the bacterial LexA operator (LexO) was inserted into the cauliflower mosaic virus (CaMV35S) promoter upstream of the β-glucuronidase (GUS) reporter gene (see Figure). This construct was stably transformed into Arabidopsis thaliana plants which were crossed to various transgenic lines constitutively expressing translational fusions of selected PRC components fused to the LexA DNA-binding domain (BD). Several features were then monitored in progeny lines expressing both transgenes: (1) the recruitment of PcG proteins fused to BD by chromatin immunoprecipitation followed by quantitative PCR (ChIP-qPCR), (2) the transcriptional repression by GUS activity, and (3) the ubiquitination of H2A and the trimethylation and deacetylation of H3 by ChIP-qPCR.

Figure.

Function of VAL1 in recruiting PRC1 and PRC2 to chromatin. A, Recruitment of VAL1 fused to the LexO-BD to the LexO-containing promoter leads to reduced GUS activity and increased levels of H2AK121ub and H3K27me3. B, Proposed model for VAL1-mediated recruitment of PRC1 and PRC2 to chromatin. Adapted from Baile et al. (2021), Figures 1, 2, and 5.

First, the authors tethered PRC1 components to LexO and found that BD-VAL1, but not BD-BMI nor BD-RING1B, reduced GUS expression and increased H2AK121ub and H3K27me3 levels while reducing H3 acetylation (see Figure). In a bmi^−/− mutant background, BD-VAL1-mediated H2AK121ub was lost and H3K27me3 levels were not lost but reduced, suggesting that tethered VAL1 was still able to recruit PRC2 in absence of BMI. Next, Baile and colleagues tested if tethering the TFs KNU, FLC, and ERF10 can recruit PRC2. Indeed, these tethered TFs led to reduced GUS expression and H3 acetylation and increased H3K27me3, while H2AK121ub remained unchanged. Together, these results indicate that TFs can recruit PRC2, but VAL1 is needed for PRC1-mediated ubiquitination.

Many TFs that have been reported to play a role in PRC2 recruitment contain an EAR domain (LxLxL, DLNxP, or DLNxxP). EAR domains mediate interactions with TPL/TPR corepressors or the SAP18 protein, which in turn recruit histone deacetylases (HDAC) to repress transcription (Krogan et al., 2012). Consistently, removing the EAR domain of BD-KNU abolished its action. Furthermore, tethering only the EAR domain was sufficient to induce GUS silencing along with an increase in H3K27me3 and decrease in H3 acetylation. In a nutshell, these data support the idea that the EAR domain is involved in directing both HDAC and PRC2 activities to target genes.

Finally, the authors used their targeting system to prove that the PcG-associated factor EMF1 is involved in PRC2-mediated H3K27trimethylation. In summary, this study nicely illustrates the power of a targeted approach in dissecting complicated protein interdependencies. It will be interesting to see if this method can be developed further, for example, to refine the order of assembly and composition of the PRC complexes by combining tethering, genetics, and proteomics.