the consensus binding motif for OLIG2 in iOLs but not in OPCs (Fig. 2j). 72.2% iOL-specific H3K9me3-enrichment regions, which do not show up in OPCs, contained OLIG2 motif, while the ratio reduced to 51.1% when we considered whole H3K9me3-enrichment regions in iOLs (Fig. 2k). Furthermore, in rat iOLs the enrichment of H3K9me3 was strengthened in loci of genes which are downregulated and have enhanced OLIG2 binding upon differentiation (Fig. 2l, m). The occupancy of SETDB1 along genome was measured by ChIP-seq in primary OPCs and iOLs. SETDB1 occupied at genomic loci encoding genes that are frequently downregulated with enhanced OLIG2 binding upon differentiation (Fig. 2n, o). OLIG2 binding motif was identified by HOMER analysis within HA-SETDB1 peaks among loci of genes downregulated upon OPC differentiation (Supplementary Fig. 2k). To rule out the possibility that SETDB1 was recruited by other transcriptional factors, we performed a series of immunoprecipitation assays in rat iOLs. There was weak or undetectable interaction between SETDB1 and these factors, such as NKX2-2, YY1, SOX10 and ZFP191 (Supplementary Fig. 2l). All the evidence indicates that OLIG2 recruits SETDB1 to govern H3K9me3 distribution for gene repression in early stage of OPC differentiation.
