PF-477736 Our study shows that deletion
Our study shows that PF-477736 of Dpf2 in mouse ESCs decreased their self-renewal ability and dramatically impaired their differentiation into mesoderm and endoderm while promoting neural ectoderm differentiation. The differentiation defect to meso-endoderm could be rescued by restoring Tbx3 levels in Dpf2 ESCs. We also found that the PRC2 complex subunit Eed oppositely regulates meso-endoderm differentiation compared with Dpf2, also by regulating Tbx3 expression. Mechanistically, Dpf2 and Eed act on two different Tbx3-controlling enhancers. We further demonstrate that Ezh2, another PRC2 subunit, also regulates meso-endoderm differentiation as opposed to Dpf2 but through a distinct mechanism that involves Nanog suppression. Thus, our work uncovers complex mechanisms by which PRC2 subunits and the BAF subunit Dpf2 control differentiation of ESCs.
Discussion In this study, we revealed that the BAF subunit DPF2 is critical for ESC differentiation into mesoderm, endoderm, and neural ectoderm. Moreover, we show that meso-endoderm differentiation defects because of Dpf2 deletion can be rescued by restoring the expression of Tbx3 to normal levels. The differentiation defects of Dpf2 KO ESCs are different from those described for other BAF components (Ho and Crabtree, 2010), in agreement with the notion that different subunits confer different functionalities. Importantly, this study defines a functional downstream target of the BAF complex in ESCs, for which maintenance of expression is important for ESC fate decisions. Our study further revealed an opposing regulation of endoderm and mesoderm differentiation by Dpf2 and Eed. This relationship was supported by the restored expression of endoderm and mesoderm marker genes in Eed/Dpf2 double KO EBs. We postulate that Dpf2 and Eed oppositely regulate endo- and mesoderm differentiation of ESCs via differential control of Tbx3 expression. An antagonistic role of polycomb and BAF complexes has been reported previously through competitive binding of these complexes at the same locus (Wilson et al., 2010, Ho et al., 2011, Kadoch et al., 2017). In contrast, our work shows that Eed and Dpf2 function in meso-endoderm differentiation via their respective interaction at different enhancers, the IE and DE, respectively, at the Tbx3 locus, as summarized in Figure 7. Specifically, the loss of Eed diminished the enrichment of H3K27me3 over the Tbx3 gene, including its IE, which increased the access of OCT4 and SOX2 to the IE, which likely leads to upregulation of Tbx3. The loss of Dpf2 led to an increase of H3K27me3 deposition at the IE of Tbx3 by increasing the access of PRC2, consistent with competitive binding between PRC2 and BAF complexes at the IE. Conversely, the loss of Dpf2 significantly decreased the H3K27ac level and the access of OCT4 and SOX2 at the DE. The decrease in OCT4 binding could precede the drop of H3K27ac because the impaired physical interaction of DPF2 and OCT4 upon loss of Dpf2 may destabilize OCT4 binding. Conversely, because P300 is known to acetylate histone H3K27 (Tee and Reinberg, 2014), another possible scenario for the decrease in H3K27ac is that loss of the direct interaction between DPF2 and P300 leads to the decrease in H3K27ac in Dpf2 ESCs, which, in turn, may affect OCT4 binding. Regardless, the interaction between DPF2, P300, and OCT4 indicates a collaborative regulation of Tbx3 via a chromatin remodeler, chromatin modifications, and critical TFs. Our study also demonstrates that the opposing regulation of targets by Dpf2 and Eed in ESCs is not limited to Eed but extends to the PRC2 subunit Ezh2. However, meso-endoderm markers were repressed in the absence of Ezh2, in contrast to their increase upon Eed deletion. We show that these differences in the differentiation defect are achieved through distinct downstream transcription factors because the opposing effect of Dpf2 and Ezh2 ensued mainly via differential regulation of Nanog.