Control of developmental regulators by Polycomb in human embryonic stem cells

Abstract

Polycomb group proteins are essential for early development in metazoans, but their contributions to human development are not well understood. We have mapped the Polycomb Repressive Complex 2 (PRC2) subunit SUZ12 across the entire nonrepeat portion of the genome in human embryonic stem (ES) cells. We found that SUZ12 is distributed across large portions of over two hundred genes encoding key developmental regulators. These genes are occupied by nucleosomes trimethylated at histone H3K27, are transcriptionally repressed, and contain some of the most highly conserved noncoding elements in the genome. We found that PRC2 target genes are preferentially activated during ES cell differentiation and that the ES cell regulators OCT4, SOX2, and NANOG cooccupy a significant subset of these genes. These results indicate that PRC2 occupies a special set of developmental genes in ES cells that must be repressed to maintain pluripotency and that are poised for activation during ES cell differentiation.

Figure 1. Genome-Wide ChIP-Chip in Human Embryonic Stem Cells

(A) DNA segments bound by the initiation form of RNA polymerase II or SUZ12 were isolated using chromatin-immunoprecipitation (ChIP) and identified with DNA microarrays containing over 4.6 million unique 60-mer oligonucleotide probes spanning the entire nonrepeat portion of the human genome. ES cell growth and quality control, the antibodies, ChIP protocol, DNA microarray probe design, and data analysis methods are described in detail in Supplemental Data. (B) Examples of RNA polymerase II ChIP signals from genome-wide ChIP-Chip. The plots show unprocessed enrichment ratios (blue) for all probes within a genomic region (ChIP versus whole genomic DNA). Chromosomal positions are from NCBI build 35 of the human genome. Genes are shown to scale below plots (exons are represented by vertical bars). The start and direction of transcription are noted by arrows. (C) Examples of SUZ12 ChIP signals from genome-wide ChIP-Chip. The plots show unprocessed enrichment ratios (green) for all probes within a genomic region (ChIP versus whole genomic DNA). Chromosomal positions, genes, and notations are as described in (B). (D) Chart showing percentage of all annotated genes bound by RNA polymerase II (blue), SUZ12 (green), both (yellow), or neither (gray). (E) Distribution of the distance between bound probes and the closest transcription start sites from RefSeq, Ensembl, MGC, UCSC Known Genes and H-Inv databases for SUZ12 (green line), and RNA polymerase II (blue line). The number of bound probes is given as the percentage of total probes and is calculated for 400 bp intervals from the start site. The null-distribution of the distance between all probes and the closest transcription are shown as a black line.

Figure 2. SUZ12 Is Associated with EED, histone H3K27me3 Modification, and Transcriptional Repression in ES Cells

(A) Venn diagram showing the overlap of genes bound by SUZ12 at high-confidence, genes bound by EED at high-confidence, and genes trimethylated at H3K27 at high-confidence. The data are from promoter micro-arrays that contain probes tiling −8 kb and +2 kb around transcription start. 72% of the genes bound by SUZ12 at high-confidence are also bound by EED at high-confidence; others are bound by EED at lower confidence (Figure S6). (B) SUZ12 (top), EED (middle), and H3K27me3 (bottom) occupancy at NEUROD1. The plots show unprocessed enrichment ratios for all probes within this genomic region (SUZ12 ChIP versus whole genomic DNA, EED ChIP versus whole genomic DNA, and H3K27me3 ChIP versus total H3 ChIP). Chromosomal positions are from NCBI build 35 of the human genome. NEUROD1 is shown to scale below plots (exons are represented by vertical bars). The start and direction of transcription are noted by arrows. (C) Relative expression levels of 604 genes occupied by PRC2 and trimethylated at H3K27 in ES cells. Comparisons were made across four ES cell lines and 79 differentiated cell types. Each row corresponds to a single gene that is bound by SUZ12, associated with EED and H3K27me3, and for which Affymetrix expression data are available. Each column corresponds to a single expression microarray. ES cells are in the following order: H1, H9, HSF6, HSF1. For each gene, expression is shown relative to the average expression level of that gene across all samples, with shades of red indicating higher than average expression and green lower than average expression according to the scale on the right. Cell types are grouped by tissue or organ function, and genes are ranked according the significance of their relative level of gene expression in ES cells.

Figure 3. Cellular Functions of Genes Occupied by SUZ12

(A) Genes bound by SUZ12 or RNA polymerase II were compared to biological process gene ontology categories; highly represented categories are shown. Ontology terms are shown on the y axis; p-values for the significance of enrichment are graphed along the x axis (SUZ12 in green, RNA polymerase II in blue). (B) Selected examples of developmental transcription factor families bound by SUZ12. SUZ12 is represented by the green oval; individual transcription factors are represented by circles and grouped by family as indicated. Examples of transcription factors with defined roles in development are labeled. Transcription factor families include homeobox protein (HOX), basic helix-loop-helix domain containing, class B (BHLHB), HOX cofactors (MEIS/EVX), distal-less homeobox (DLX), Forkhead box (FOX), NEUROD, GATA binding protein (GATA), runt related transcription factor (RUNX), paired box and paired-like (PAX), LIM homeobox (LHX), sine oculis homeobox homolog (SIX), NK transcription factor related (NKX), SRY box (SOX), POU domain containing, classes 3 and 4 (POU), early B-cell factor (EBF), atonal homolog (ATOH), hairy and enhancer of split protein (HES), myogenic basic domain (MYO), T-box (TBX), caudal type homeobox (CDX), and iroquois homeobox protein (IRX).

Figure 4. SUZ12 Occupies Large Portions of Genes Encoding Transcription Factors with Roles in Development

(A) The fraction of SUZ12 target genes associated with different sizes of binding domains. Genes are grouped into four categories according to their function: Signaling, Adhesion/migration, Transcription, and Other. (B) Examples of SUZ12 (green) and RNA poly-merase II (blue) binding at the genes encoding developmental regulators TBX5 and PAX6. The plots show unprocessed enrichment ratios for all probes within a genomic region (ChIP versus whole genomic DNA). Genes are shown to scale below plots (exons are represented by vertical bars). The start and direction of transcription are noted by arrows. (C) Binding profiles of SUZ12 (green) and RNA polymerase II (blue) across −500 kb regions encompassing HOX clusters A–D. Unprocessed enrichment ratios for all probes within a genomic region are shown (ChIP versus whole genomic DNA). Approximate HOX cluster region sizes are indicated within black bars.

Figure 5. SUZ12 Binding Is Associated with Highly Conserved Regions

(A) SUZ12 occupancy (green) and conserved elements are shown at NKX2-2 and adjacent genomic regions. The plots show unprocessed enrichment ratios for all probes within this genomic region (SUZ12 ChIP versus whole genomic DNA). Conserved elements (red) with LoD scores > 160 derived from the PhastCons program (Siepel et al., 2005) are shown to scale above the plot. Genes are shown to scale below plots (exons are represented by vertical bars). A higher resolution view is also shown below. (B) Enrichment of conserved noncoding elements within SUZ12 (green) and RNA polymerase II (blue) bound regions. The maximum nonexonic PhastCons conservation score was determined for each bound region. For comparison, the same parameter was determined using a randomized set of genomic regions with the same size distribution. The graph displays the ratio of the number of bound regions with that score versus the number of randomized genomic regions with that score.

Figure 6. Preferential Activation of PRC2 Target Genes during ES Cell Differentiation

(A) Fold enrichment in the number of genes induced or repressed during ES cell differentiation. The change in gene expression is given as the log(2) transformed ratio of the signals in differentiated H1 cells versus pluripotent H1 cells and is binned into six groups. The upper limit of each bin is indicated on the x axis. The two lines show genes transcriptionally inactive in ES cells (absence of RNA polymerase II) and bound by SUZ12 (green) and genes transcriptionally inactive in ES cells and repressed by other means (blue). In both cases, fold enrichment is calculated against the total population of genes and normalized for the number of genes present in each group. (B) Expression changes of genes encoding developmental regulators during ES cell differentiation. Expression ratio (differentiated/pluripotent) is represented by color, with shades of red indicating upregulation and shades of green downregulation according to the scale shown above. Genes are ordered according to change in gene expression, with genes exhibiting higher expression in pluripotent ES cells to the left and genes exhibiting higher expression in differentiated cells to the right. Genes bound by SUZ12 in undifferentiated ES cells are indicated by blue lines in the lower panel. (C) Fold enrichment in the number of genes induced or repressed in SUZ12-deficient mouse cells. The change in gene expression is given as the log(2) transformed ratio of the signals in Suz12-deficient cells versus wild-type ES cells. The two lines show genes transcriptionally inactive in human ES cells (absence of RNA polymerase II) and bound by SUZ12 (green) and genes transcriptionally inactive in human ES cells and repressed by other means (blue). In both cases, fold enrichment is calculated against the total population of genes. (D) Gene expression ratios (log base 2) of Suz12 target genes in differentiated human H1 ES cells relative to pluripotent H1 ES cells (x axis) and in Suz12-deficient mouse cells relative to wild-type mouse ES cells (y axis). Upper right quadrant: genes upregulated during human ES cell differentiation and in Suz12-deficient mouse cells; lower right: genes upregulated during ES cell differentiation and downregulated in Suz12-deficient cells; lower left: genes downregulated during ES cell differentiation and in Suz12-deficient cells; upper left: genes downregulated during ES cell differentiation and upregulated in Suz12-deficient cells. (E) SUZ12 binding profiles across the gene encoding muscle regulator MYOD1 in H9 human ES cells (green) and primary human skeletal myotubes (gray). The plots show unprocessed enrichment ratios for all probes within a genomic region (ChIP versus whole genomic DNA). Genes are shown to scale below plots (exons are represented by vertical bars). The start and direction of transcription are noted by arrows. (F) Suz12 binding profiles across the gene encoding LHX9 in H9 human ES cells (green) and primary human skeletal myotubes (gray). The plots show unprocessed enrichment ratios for all probes within a genomic region (ChIP versus whole genomic DNA). Genes are shown to scale below plots (exons are represented by vertical bars). The start and direction of transcription are noted by arrows.

Figure 7. SUZ12 Is Localized to Genes also Bound by ES Cell Transcriptional Regulators

(A) Transcriptional regulatory network model of developmental regulators governed by OCT4, SOX2, NANOG, RNA polymerase II, and SUZ12 in human ES cells. The ES cell transcription factors each bound to approximately one-third of the PRC2-occupied, developmental transcription factor genes. Developmental regulators were selected based on gene ontology. Regulators are represented by dark blue circles; RNA polymerase II is represented by a light blue circle; SUZ12 is represented by a green circle; gene promoters for developmental regulators are represented by small red circles. (B) SUZ12 occupies a set of repressed developmental regulators also bound by OCT4, SOX2, and NANOG in human ES cells. Genes annotated as bound by OCT4, SOX2, and NANOG previously and identified as active or repressed based on expression data (Boyer et al., 2005) were tested to see if they were bound by SUZ12 or RNA polymerase II. Ten of eleven previously identified active genes were found to be bound by RNA polymerase II at known promoters, while eleven of twelve previously identified repressed genes were bound by SUZ12. Regulators are represented by dark blue circles, RNA polymerase II by a light blue circle, and SUZ12 by a green circle. Gene promoters are represented by red rectangles.