T be required to create chromatin states permissive for subsequent expression upon differentiation. This would explain why these genes show no changes in expression in the mutant ES cell lines. Unfortunately, although Eed2/2 is dispensable for the maintenance of pluripotent stem cells, it is required for the differentiation and/or maintenance of multipotent progenitors, CX-5461 site limiting our ability to differentiate Eed2/2 cells in vitro [24,38]. To explore this possibility, we compared the list of genes with altered DNAme caused by loss of H3K27me3, to published datasets describing gene expression differences in neural progenitor cells (NPCs) or MEFs relative to ES cells [39]. Although genes undergoing DNAme changes upon loss of H3K27me3 were not activated in either NPCs or MEFs (data not shown), this does not preclude the possibility that coordinated regulation of PRC2 and DNAme controls gene expression in other differentiated cell types. Recently, it has been demonstrated that within the gene bodies of HCC1954 cells there is mutual antagonism between H3K27me3 and DNAme in an allele specific manner, and that disruption of this mutual antagonism leads to 25331948 misexpression [40]. We are not able to complete a similar analysis because our MeDIP experiments do not provide information about allele-specific methylation or data across the gene body. It is possible that by comparing our ChIP-seq dataset with a more complete analysis of DNAme in Eed2/2 cells derived from polymorphic strains that we may be able to identify important examples of mutual antagonism between DNAme and H3K27me3. Finally, We looked at the expression changes that occur upon loss of either PRC2 or DNA methyltransferase activity and find an interesting overlap in regulation at a number of developmentally important genes and similar effects on gene expression upon loss of either mark. We do not find, however, that the genes that are regulated by crosstalk between these two epigenetic marks correlate with genes that have expression changes in either cell type, so it appears that the crosstalk between the marks is not directly controlling gene expression, at least in undifferentiated ES cells. Future studies examining the effect loss of one epigenetic mark has on the placement of other marks will be important to understand how coordinate regulation of epigenetic modifications effects gene regulation. It is possible that using a similar approach to examine additional marks coordinately regulated with H3K27me3, such as H3K4me3, or with DNAme, such as H3K9me3, will give important insights into gene regulation. Additionally, as high-throughput sequencing technologies continue to increase read number and read length, comprehensive methylome analysis will become more common, giving a greaterability to intersect ChIP-seq datasets with methylome data genome-wide with basepair resolution.Materials and Methods Cell CultureV6.5 [41], Eed2/2 [38] DnmtTKO [25] murine ES cells were cultivated in 5 CO2 at 37uC on irradiated MEFs in DMEM containing 15 FCS, LIF, Penicillin/streptomycin, L-glutamine, NEAA, L-glutamine 2-mercaptoethanol. Cells were passed at least 2 times feeder-free on 0.2 gelatin to exclude feeder contamination. The DNMT deficiency in DnmtTKO cells was validated by qRT-PCR. The inactivating mutation in the Eed2/2 cell line was validated by sequencing, and low EED protein levels confirmed by western blot (Figure S4). For 5-AzaC treatment, v6.5 cells were grown to 60 confluency the.
