Figure 7. Working model delineates a non-canonical mechanism by which EED represses gene transcription.
Two mechanisms for EED repression were operative in the postnatal heart. One subset of repressed genes was occupied by EED and H3K27me3 in WT, and EED inactivation reduced H3K27me3 in association with gain of H3K27ac. Upregulation of these genes in EedCKO could be due to a combination of loss of H3K27me3 (canonical mechanism). Loss of EED itself with subsequent gain in H3K27ac might also contribute to regulation of these genes. A second subset of repressed genes was also occupied by EED and H3K27me3 in WT, but H3K27me3 was not reduced by EED inactivation. While H3K27me3 may contribute to the repression of these genes in WT, their upregulation in EedCKO was not attributable to H3K27me3, which was unchanged. Rather, our data suggest that upregulation was directly due to loss of EED itself, with consequent reduction of HDAC activity and gain in H3K27ac. A third subset of genes was occupied by EED but little H3K27me3 in WT. These genes had significant H3K27ac at baseline, which was further increased by Eed inactivation. Thus, these genes may represent a set ‘poised’ for activation; in WT EED occupancy represses these genes by collaborating with HDAC to limit gene activity. EED inactivation reduces HDAC activity, resulting in H3K27ac accumulation and gene upregulation.