Polycomb Group Protein Pcgf6 Acts as a Master Regulator to Maintain Embryonic Stem Cell Identity

Abstract

The polycomb repressive complex 1 (PRC1) is a multi-subunit complex that plays critical roles in the epigenetic modulation of gene expression. Here, we show that the PRC1 component polycomb group ring finger 6 (Pcgf6) is required to maintain embryonic stem cell (ESC) identity. In contrast to canonical PRC1, Pcgf6 acts as a positive regulator of transcription and binds predominantly to promoters bearing active chromatin marks. Pcgf6 is expressed at high levels in ESCs, and knockdown reduces the expression of the core ESC regulators Oct4, Sox2, and Nanog. Conversely, Pcgf6 overexpression prevents downregulation of these factors and impairs differentiation. In addition, Pcgf6 enhanced reprogramming in both mouse and human somatic cells. The genomic binding profile of Pcgf6 is highly similar to that of trithorax group proteins, but not of PRC1 or PRC2 complexes, suggesting that Pcgf6 functions atypically in ESCs. Our data reveal novel roles for Pcgf6 in directly regulating Oct4, Nanog, Sox2, and Lin28 expression to maintain ESC identity.

Figure 1. Pcgf6 is Essential for the Maintenance of ESC Identity.

(A) Distribution plot showing z score of Oct4-EGFP signal changes with various siRNA treatments. Oct4-EGFP ESCs were transfected with indicated siRNAs (siPcgf6, siOct4, siNanog) and EGFP signal was detected and measured by flow cytometry. Non-targeting control siRNA serves as negative control (NTC). Z score was calculated by normalizing to NTC. (B~E) Scatter plot showing mRNA expression changes in Pcgf6- or Oct4-depleted CCE ESCs one or two days after siRNA transfection. Cells were transfected with Pcgf6 or control siRNAs for 3–5 h, and mRNA expression was analyzed by microarray 1 or 2 days later. Green dots indicate genes showing reduced expression (<−0.6 on log₂ scale) in Pcgf6- or Oct4-depleted compared with control cells. Red dots indicate increased genes (>0.6 log₂) and black dots indicate non-differentially-expressed genes with Pcgf6 or Oct4 knockdown. The number of differentially expressed genes is shown in the box in each panel. (F) Table showing gene ontology analysis (GO) of decreased genes in Pcgf6-depleted ESCs. 2983 decreased genes at day 1 post Pcgf6-depletion in ESCs were analyzed using DAVID. Probability is represented as Score (−log₁₀ p-value). (G) Heat map showing mRNA expression profiling in CCE ESCs depleted of Pcgf6 or Oct4. Cells were transfected with Oct4, Pcgf6, or control siRNAs for 3–5 h, and mRNA expression was analyzed by microarray 1 day later. Results in Oct4 and Pcgf6 knockdown (KD) cells were normalized to cells treated with control siRNA and then subjected to K-mean clustering analysis. Color bar is on a log₂ scale. Red arrow indicates the small proportion of genes showing increased expression in Oct4 KD cells, but which are decreased in Pcgf6 KD cells. (H) Venn diagram showing overlap between genes suppressed by Pcgf6 or Oct4 KD. mRNA expression was analyzed by microarray 2 days after transfection of CCE ESCs with siRNA, and results were normalized to cells treated with control siRNA. The data show that 97.3% of genes that were reduced by Oct4 KD were also reduced by Pcgf6 KD.

Figure 2. Pcgf6 Regulates the Core Circuitry and Maintains Pluripotency in ESCs.

(A) Micrographs showing abnormal EB morphology of differentiating Pcgf6-mESCs. CCE ESCs were stably transfected with Pcgf6 or EGFP (control), and in vitro differentiation was assessed by monitoring EB formation over 15 days. Images were taken using phase contrast microscopy from day 12 cultures of control cells and 4 individual clones overexpressing Pcgf6. Scale bar = 100 μm. (B,C) Expression of ESC-specific (B) and differentiation marker (C) genes during in vitro differentiation of EGFP- or Pcgf6-overexpressing CCE ESCs. mRNA levels were measured on days 0, 3, 6, 9, 12, and 15 and are presented as the ratio of mRNA levels relative to the highest expression level during differentiation. Four biological replicates are shown in four different color lines and control (EGFP-overexpressing CCE) is in black dotted line.

Figure 3. Pcgf6 Enhances Induced Reprogramming Efficiency.

(A) Reprogramming efficiency of Oct4-EGFP MEFs transduced with OSKM and either Pcgf6 or DsRed (control) retroviruses. EGFP+ colonies were counted 2 weeks post-transduction, and the results are expressed as colony number relative to control DsRed-miPSCs. Results are the mean ± SEM of n ≥ 5. *p < 0.05 by Student’s t-test. (B) Reprogramming efficiency of human BJ fibroblasts transduced with OSKM and either Pcgf6 or DsRed (control) retroviruses. AP+ colonies were counted 3–4 weeks post-transduction, and the results are expressed as for (A). Results are the mean ± SEM of n ≥ 5. *****p < 0.000005 by Student’s t-test. (C) Reprogramming efficiency of Oct4-EGFP MEFs depleted of Pcgf6. Cells were treated with non-targeting (siNT) or Pcgf6-specific siRNA for 3–5 h and then transduced with OSKM retroviruses. EGFP+ colonies were counted 2 weeks later, and the results are expressed as for (A). Results are the mean ± SEM of . ****p < 0.00005 by Student’s t-test. (D–F) Images of Pcgf6-hiPSCs derived as in (B). Cells were cultured on feeders for at least 4 passages before imaging of: cell morphology (D), AP staining (E), and Tra-1-81, SSEA4, Tra-1-60, and Nanog immunofluorescence staining (F). Scale bars = 200 μm (D,E) or 50 μm (F). (G) Histological staining of mesodermal, endodermal, and ectodermal tissues in teratomas formed from Pcgf6-hiPSCs. Cells were injected under the kidney capsule of SCID mice, and teratomas were removed and analyzed ~10 weeks later. Scale bar = 50 μm. Images were taken using either phase contrast microscopy or fluorescent microscopy. (H) Normal karyotype of Pcgf6-hiPSCs. Cells were cultured for ~2 months on feeder cells before karyotype analysis. Chromosomes in the metaphase of 20 cells were counted, and 4 cells were examined for G-band staining with band resolution at 450–525. All cells contained normal karyotypes with 46 chromosomes including XY, and no clonal abnormalities were detected.

Figure 4. Pcgf6 Directly Regulates Key Pluripotency Factors in ESCs, Including Oct4, Sox2, Lin28b, and Nanog.

(A) Heat map representations showing the chromatin-binding distribution of various epigenetic regulators in CCE ESCs. For Pcgf6, ChIP-seq experiments were performed with two different antibodies, as indicated (Ab#1 and 2; See Methods). ChIP-seq data for H3K4me3, H3K27me3, Suz12, Ezh2, and Rnf2 were obtained from the Bernstein lab, and data for Rbbp5 and Wdr5 were obtained from the Ihor lab. Regions decorated with active (H3K4me3), repressive (H3K27me3), and bivalent poised/repressive (both H3K4me3 and H3K27me3) chromatin marks are indicated by the vertical bar at left. (B) Pcgf6 binds promoter regions of key pluripotency genes in ESCs. ChIP-seq data for Pcgf6 (this study) were uploaded to the Integrative Genomics Viewer browser and compared with ChIP-seq data for H3K27me3, Suz12, Ezh2, and Rnf2 (as above) and Oct4, Rbbp5, Wdr5, and Wdr5-FL (Ihor lab32). The scale of each genomic region is indicated at the top of each panel. The TSSs of Sox2, Oct4, Nanog, Lin28a, Myc, and Lin28b target genes are shown as black squares with the arrow showing transcription direction. Green indicates Pcgf6 binding regions; orange indicates regions with histone modification markers; blue indicates regions bound by Oct4, PRC1, PRC2, and TrxG components. (C) Heat map showing that changes in gene expression caused by Pcgf6 knockdown is recapitulated by knockdown of Pcgf6 targets. CCE ESCs were transfected with siRNAs targeting various Pcgf6-bound genes (named at the top), and RT-qPCR was performed ~24 h later to detect expression of a panel of Pcgf6-bound genes (listed at right). Expression levels were normalized to cells treated with the non-targeting siRNA (Control) and displayed as a heat map. The hierarchical tree was created by Cluster and visualized by Java TreeView. The blue hierarchical lines and yellow rectangle indicate the major cluster. (D) Proposed model for the central role of Pcgf6 in modulating the ESC core circuitry. Pcgf6 activates the key regulators Oct4, Sox2, Nanog, and Lin28 in ESCs in a noncanonical fashion and directly regulates the expression of other factors, including Kpna2, Rfwd3, Nup88, Rpa2, Tmpo, and Snurf, which positively regulate the ESC core circuitry.