MGA, L3MBTL2 and E2F6 determine genomic binding of the non-canonical Polycomb repressive complex PRC1.6

Abstract

Diverse Polycomb repressive complexes 1 (PRC1) play essential roles in gene regulation, differentiation and development. Six major groups of PRC1 complexes that differ in their subunit composition have been identified in mammals. How the different PRC1 complexes are recruited to specific genomic sites is poorly understood. The Polycomb Ring finger protein PCGF6, the transcription factors MGA and E2F6, and the histone-binding protein L3MBTL2 are specific components of the non-canonical PRC1.6 complex. In this study, we have investigated their role in genomic targeting of PRC1.6. ChIP-seq analysis revealed colocalization of MGA, L3MBTL2, E2F6 and PCGF6 genome-wide. Ablation of MGA in a human cell line by CRISPR/Cas resulted in complete loss of PRC1.6 binding. Rescue experiments revealed that MGA recruits PRC1.6 to specific loci both by DNA binding-dependent and by DNA binding-independent mechanisms. Depletion of L3MBTL2 and E2F6 but not of PCGF6 resulted in differential, locus-specific loss of PRC1.6 binding illustrating that different subunits mediate PRC1.6 loading to distinct sets of promoters. Mga, L3mbtl2 and Pcgf6 colocalize also in mouse embryonic stem cells, where PRC1.6 has been linked to repression of germ cell-related genes. Our findings unveil strikingly different genomic recruitment mechanisms of the non-canonical PRC1.6 complex, which specify its cell type- and context-specific regulatory functions.

Fig 1. MGA, L3MBTL2, E2F6 and PCGF6 colocalize in 293 cells.

(A) Schematic representation of PRC1.6 core components. (B) Western blot analysis of MGA, L3MBTL2, E2F6 and PCGF6 expression in wild type 293 cells (wt) and in corresponding MGA-, L3MBTL2-, E2F6- and PCGF6-depleted cell clones (MGAko, L3MBTL2ko, E2F6ko and PCGF6ko). Re-probing for tubulin (TUB) controlled loading of extracts. (C) Venn diagrams showing the overlap of MGA, L3MBTL2, E2F6 and PCGF6 binding regions in HEK293 cells. The total number of high-confidence MGA, L3MBTL2, E2F6 and PCGF6 ChIP-seq peaks (≥30 tags, ≥3-fold enrichment over knockout control) and their overlap is shown. (D) A heat map view of the distribution of union MGA, L3MBTL2, E2F6 and PCGF6 peaks in HEK293 cells (n = 8342) at +/- 2 kb regions centred over the MGA peaks. (E) Representative genome browser screenshots of a 0.7 Mb region of chromosome 19 showing co-localization of MGA, L3MBTL2, E2F6 and PCGF6 at the CTC-232P5.1, RFX2, MLLT1 and KHSRP promoters. (F) Distribution of MGA, L3MBTL2, E2F6 and PCGF6 peaks relative to positions -2000 bp upstream to +2000 bp downstream of gene bodies. TSS, transcription start site; TES, transcription end site. (G) ChIP-qPCR analysis of MGA, L3MBTL2, E2F6 and PCGF6 binding to selected promoters. The region -2 kb upstream of the CDC7 promoter served as a negative control. Percent of input values represent the mean of at least three independent experiments +/- SD. (H) Sequence motifs enriched in PRC1.6 binding regions. Logos were obtained by running MEME-ChIP with 300 bp summits of the top 600 union MGA-L3MBTL2-E2F6-PCGF6 ChIP-seq peaks. The numbers next to the logos indicate the occurrence of the motifs, the statistical significance (E-value) and the transcription factors that bind to the motif. Right panel, local motif enrichment analysis (CentriMo) showing central enrichment of the MGA/MAX bHLH and the E2F6/DP1 binding motifs within the 300 bp peak regions. The NRF1 binding motif was not centrally enriched.

Fig 2. MGA is essential for genomic binding of PRC1.6.

(A) Heat map view of the distribution of union MGA, L3MBTL2 and E2F6 peaks in wild type cells (n = 8342) and in MGA-depleted cells at +/- 2 kb regions centred over the MGA peaks. (B) Representative genome browser screenshots showing binding of MGA, L3MBTL2, E2F6 and PCGF6 to the AEBP2, RPA2, RFC1 and SPOP promoters in wild type cells. MGA-depleted cells lack binding of L3MBTL2 and E2F6. (C) Western blot analysis of L3MBTL2, E2F6, PCGF6 and RING2 in wild type HEK293 cells and in two different MGA-depleted clones (cl26 and cl27). The anti-Tubulin blot served as a loading control. (D) L3MBTL2-, E2F6- and PCGF6 transcripts were determined in wild type cells and in MGA-depleted cell clones by RT-qPCR analysis. B2M transcript levels were used to normalize the data across samples, and transcript levels in wild type cells were arbitrarily set to 1. Data represent the average of technical replicates ± SD. (E) ChIP-qPCR data showing lack of L3MBTL2, E2F6, PCGF6, MAX, RING2, RYBP and HP1γ binding to representative PRC1.6 target promoters in MGAko cells, and diminished deposition of H2AK119ub1. The CDC7 -2kb region served as a negative control region. Percent of input values represent the mean of at least three independent experiments +/- SD. (F) PRC1.6 target promoters are not bound by PRC2 and lack H3K27me3. Local levels of EZH2 and H3K27me3 at selected PRC1.6 target promoters in wild type (WT) and in MGAko cells (clones cl26 and cl27) were determined by ChIP-qPCR analysis. Genomic regions known to be bound by canonical PRC1 (FUT9, MYT1 and TSH2B) served as positive control regions. These regions were not bound by MGA (right panel). Percent of input values represent the mean of at least three independent experiments +/- SD.

Fig 3. MGA promotes binding of PRC1.6 by DNA-binding-dependent and DNA-binding-independent mechanisms.

(A) Expression of wild type MGA in MGAko cells rescues binding of PRC1.6. ChIP-qPCR data showing binding of transiently expressed MGA and of endogenous L3MBTL2, E2F6, PCGF6, RING2 and MAX to representative PRC1.6 target promoters. The level of the H2AK119ub1 was not affected. Percent of input values represent the mean of at least three independent experiments +/- SD. (B) Schematic representation of the MGA ΔTbox and bHLH mutants. (C) Western blot analysis of wild type MGA and of the DNA-binding-deficient MGA mutants (ΔT-Box, bHLHmut and ΔTbHLHmut) expressed in MGAko cells. The anti-Tubulin blot served as a loading control. (D) ChIP-qPCR analyses of MGA and L3MBTL2 binding to selected PRC1.6 target promoters in MGAko cells and in MGAko cells re-expressing wild type MGA (MGA WT) or DNA-binding-deficient MGA mutants (MGA ΔT-Box, MGA bHLHmut or MGA ΔTbHLHmut). The error bars denote SD; n = 3.

Fig 4. L3MBTL2 and E2F6 contribute to chromatin binding of PRC1.6.

(A) Western blot analysis of L3MBTL2, MGA, E2F6, PCGF6 and RING2 in L3MBTL2ko, E2F6ko and PCGF6ko cells. Re-probing for Tubulin (TUB) controlled loading of extracts. The Tubulin blots are related to the MGA blots. Uncropped versions of the MGA blots are shown in S4 Fig. (B) Representative genome browser screenshots of a 0.5 Mb region of chromosome 19 showing reduced binding of PRC1.6 components to several promoters in L3MBTL2ko and in E2F6ko cells but not in PCGF6ko cells. (C) Heat map views of the distribution of MGA, L3MBTL2 and E2F6 peaks in wild type cells (n = 8342) and in MGAko, L3MBTL2ko, E2F6ko and PCGF6ko cells at +/- 2 kb regions centred over the MGA peaks. (D) Scatter plots comparing the signal intensity of MGA, L3MBTL2 and E2F6 peaks in wild type cells with the signal intensity of corresponding peaks in L3MBTL2ko (left panels), E2F6ko (middle panels) or PCGF6ko cells (right panels). Normalized ChIP-seq read counts in MGA ChIP-seq peak regions of wild type cells were plotted against the normalized read counts in corresponding peak regions of L3MBTL2ko, E2F6ko or PCGF6ko cells. (E) Left panel, scatter plot showing the correlation between reduced MGA binding and reduced L3MBTL2 binding in E2F6ko cells. Right panel, scatter plot showing the correlation between reduced MGA binding and reduced E2F6 binding in L3MBTLko cells. The top 500 ranked MGA binding sites were used to calculate the fold change of normalized ChIP-seq read counts in L3MBTL2ko and E2F6ko cells relative to wild type cells.

Fig 5. L3MBTL2 and E2F6 recruit PRC1.6 differentially in a promoter-specific manner.

(A) Left panel, scatter plot comparing the extent of reduction (fold change of normalized tag counts) of MGA binding in L3MBTL2ko cells with the extent of reduction in E2F6ko cells. Right panel, scatter plot comparing the extent of reduction of E2F6 binding in L3MBTL2ko cells with the extent of reduction of L3MBTL2 in E2F6ko cells. The E2F6-dependent RNF130 and the L3MBTL2-dependent ZFR promoters are indicated for clarity. (B) Genome browser screenshots of ChIP-seq tracks showing binding of MGA, L3MBTL2, E2F6 and PCGF6 to the RNF130 and ZFR promoters in wild type cells (WT), and in MGAko, L3MBTL2ko, E2F6ko and PCGF6ko cells. (C) ChIP-qPCR analysis of MGA binding to selected promoters in two different L3MBTL2ko (L2ko cl10 and L2ko cl14, upper panel) and in two different E2F6ko (E2F6ko cl1 and E2F6ko cl11, lower panel) cell clones. The CDC7 -2kb region served as a negative control region. Percent of input values represent the mean of at least three independent experiments +/- SD. (D) Expression of L3MBTL2 in L3MBTL2ko cells rescues binding of PRC1.6. Left, Western blot for L3MBTL2. Right, ChIP-qPCR data showing binding of exogenous L3MBTL2 and of endogenous MGA, E2F6 and PCGF6 to representative PRC1.6 target promoters. Percent of input values represent the mean of at least three independent experiments +/- SD. (E) Expression of wild type E2F6 but not of a DNA binding-deficient E2F6 mutant (E2F6mut) in E2F6ko cells rescues binding of PRC1.6. Left, Western blot for E2F6. Right, ChIP-qPCR data showing binding of exogenous E2F6 (wild type or DNA-binding deficient mutant) and of endogenous MGA and L3MBTL2 to representative PRC1.6 target promoters. Percent of input values represent the mean of at least three independent experiments +/- SD. (F) Venn diagram showing the overlap of E2F6 peaks in L3MBTL2ko cells and L3MBTL2 peaks in E2F6ko cells. Logos of the enriched sequence motifs were obtained by running MEME-ChIP with 300 bp summits of the ChIP-seq peaks. (G) GO analyses of biological functions of E2F6-dependent and of L3MBTL2-dependent PRC1.6 target genes. Enriched GO terms were retrieved using Enrichr. p values are plotted in -log2 scale.

Fig 6. The role of PRC1.6 in HEK293 cell function.

(A) Reduced proliferation of MGAko, L3MBTL2ko and E2F6ko cells. Shown are growth curves of wildtype, MGAko, L3MBTL2ko, E2F6ko and PCGF6ko HEK293 cells. Cells were seed at 3x10⁵, and counted and replated at the indicated time points. Cumulative cell numbers were calculated by multiplying the initial cell number with the fold-increase in cell numbers in each interval. (B) Venn diagrams illustrating the overlap of MGA-bound genes and genes down- or up-regulated in MGAko cells. Left circle, genes with ≥2-fold reduced transcript levels in MGAko cells; right circle, genes with ≥2-fold increased transcript levels in MGAko cells. (C) Representative genome browser screenshots of ChIP-seq and RNA-seq tracks illustrating binding of MGA, L3MBTL2, E2F6 and PCGF6 (top tracks) to the CNTD1 and SMC1B promoters, and RNA expression (bottom tracks) of the corresponding genes in three wild type samples (MGA_wt1, MGA_wt2 and MGA_wt3), and in three different MGAko cell clones (MGAko_cl26, MGAko_cl27 and MGAko_cl30). (D) RT-qPCR-based analysis of expression changes of selected genes in MGAko, E2F6ko, L3MBTL2ko and PCGF6ko cells. Transcript levels were normalized to B2M transcript levels, and are depicted relative to transcript levels in wild type cells.

Fig 7. Mga, L3mbtl2 and Pcgf6 colocalize in mouse ESCs and repress genes involved in differentiation.

(A) Venn diagrams representing the overlap of Mga, L3mbtl2 and Pcgf6 peaks in mouse ESCs. The total number of filtered (≥30 tags and ≥3-fold enrichment over IgG control) ChIP-seq peaks and their overlap is shown. (B) A heat map view of the distribution of the top 8000 union Mga-L3mbtl2-Pcgf6 peaks in mouse ES cells at +/- 2 kb regions centred over the MGA peaks. (C) Representative genome browser screenshot of a 100 kb region of chromosome 1 showing co-binding of Mga, L3mbtl2 and Pcgf6 to four promoter regions. (D) Sequence motifs enriched in Mga-L3mbtl2-Pcgf6 binding regions in mouse ESCs. Top, logos were obtained by running MEME-ChIP with 300 bp summits of the top 600 union Mga-L3mbtl2-Pcgf6 ChIP-seq peaks. The numbers next to the logos indicate the occurrence of the motifs, the statistical significance (E-value) and the transcription factors that bind to the motif. Bottom, local motif enrichment analysis (CentriMo) showing central enrichment of the Mga/Max bHLH domain E-box binding motif and the motif that identified MEME Tomtom as a T-box as well as a E2f6 recognition sequence. The Nrf1 motif was not centrally enriched within the 300 bp peak regions. (E) Distribution of Mga, L3mbtl2 and Pcgf6 peaks relative to positions -2000 bp upstream to +2000 bp downstream of gene bodies. TSS, transcription start site; TES, transcription end site. (F) Middle panel, Venn diagram illustrating the overlap of PRC1.6-bound genes and genes up-regulated in Pcgf6ko cells [26] and in L3mbtl2ko cells [13]. Left panel, GO analyses of biological functions of PRC1.6-bound genes that were de-repressed ≥2-fold in Pcgf6ko cells. Right panel, GO analyses of biological functions of PRC1.6-bound genes that were de-repressed ≥2-fold in L3mbtl2ko cells. Enriched GO terms were retrieved using DAVID 6.8. (GOTERM_BP_DIRECT, Functional Annotation Chart). Benjamini values are plotted in log10 scale.

Fig 8. PRC1.6 binding sites partially overlap with cPRC1, PRC2 and ncPRC1.1 binding sites.

(A) ChIP-seq heatmaps of Pcgf6, IgG control, Ring1b (GSM1041372) [34], Rybp (GSM1041375) [34], Cbx6-HA (GSM2610616) [33], Cbx7 (GSM2610619) [33], Pcgf2 (GSM1657387) [56], Suz12 (GSM1041374) [34], Kdm2b (GSM1003594) [6] and H3K27me3 (GSM1341951) [10] peaks in mESCs at +/- 2 kb regions centred over the Mga-L3mbtl2-Pcgf6 peaks. (B) Venn diagrams showing the overlap of high confidence Pcgf6 target genes (location of binding sites between -2.5 kb of TSS and TES) with those of Cbx7 (cPRC1), Suz12 and H3K27me3 (PRC2) and Kdm2b (ncPRC1.1). (C) Genome browser screenshots of ChIP-seq tracks at promoters of representative meiosis-related genes (Dazl, Sycp3, Stk31 and Mei1) (D) Genome browser screenshots of ChIP-seq tracks at cPRC1 target genes (Nkx2-4 and Hoxa7).

Fig 9. Model summarizing PRC1.6 targeting mechanisms.

(1) PRC1.6 is recruited to a subset of target promoters by direct DNA binding of MGA/MAX to E-boxes (CACGTG) and/or T-boxes (TCACACCT). (2) Interaction of L3MBTL2 and HP1γ with methylated histones may promote binding site selection by facilitating and stabilizing binding of MGA/MAX. (3) MGA also acts as a scaffold tethering E2F6 that in turn mediates PRC1.6 binding to E2F6-recognition sites. (4) PCGF6 recruits RING1/2 that deposits the repressive histone mark H2AK119ub1.