Bivalent chromatin and PRC1-mediated repression of neural genes is retained in hemangioblast-committed cells. (A) Analysis of modified histones at the promoters of ES- (Sox2, Oct4, and Rex1; blue bars), neural- (Math1, Nkx2-2, Nkx2-9, and Sox1; red bars), mesodermal- (Ikaros, Bra/T, Flk1, and Myf5; green bars), and control genes (Gapdh, black bar) by immunoprecipitation of chromatin from FACS-sorted GFP⁺/Flk1⁺ (hemangioblast) cells using antibodies specific for H3K4me3 (top panel) or H3K27me3 (bottom panel). Enrichment was measured by quantitative real-time PCR and is presented relative to total H3. As a control, IgG (open bars) was used for ChIP in parallel (values for most genes were low and barely visible). Data are mean ± SD of 4 experiments. Threshold levels, based on the enrichment at negative control loci (Myf5 for H3K4me3, Flk1 for H3K27me3), are indicated in gray. (B) Sequential ChIP analysis of GFP⁺/Flk1⁺ (hemangioblast) cells using first anti-H3K4me3 followed by a second round of immunoprecipitation using anti-H3K27me3 antibody (filled bars) or nonspecific IgG control (open bars). Genes are color-coded as in panel A. Data are mean ± SD of 3 experiments. Threshold levels, based on the enrichment at a negative control locus (Flk1), is indicated in gray. (C) Domain-wide profiling of histone methylation along the Nkx2-2 locus; the positions of a conserved upstream control element (gray), the transcription start site (arrow), coding regions (black), UTR (light gray), an adjacent gene (Rik), and PCR amplicons are indicated in the top panel, where scale = 1 kb. Enrichment of H3K4me3 (top) and H3K27me3 (bottom) across the Nkx2-2 locus was assessed by ChIP using IgG as controls (white bars, barely visible) in ES and GFP⁺/Flk1⁺ (hemangioblast) cells. Data are mean ± SD from 3 experiments. (D) ChIP analysis of GFP⁺/Flk1⁺ (hemangioblast) samples using anti-Ring1b antibody (filled bars) or IgG (open bars). Enrichment levels were measured by real time-PCR and are expressed relative to 10% input. Genes are color-coded as in panel A. Data are mean ± SD from 3 experiments. Threshold levels, based on the enrichment at the expressed (Ring1b⁻) Flk1 locus, are indicated in gray. (E) Strategy for withdrawal of Ring1a/Ring1b activity from CreER_T2 ES cells, or from differentiated hemangioblasts, based on tamoxifen treatment (3 days). (F) Quantitative reverse-transcription PCR analysis of candidate gene expression (color-coded as in panel A), in Ring1a^−/−/Ring1b^fl/fl/Cre-ER_T2 undifferentiated ES cells (top panel) or embryoid body-derived Flk1⁺ hemangioblast cells (bottom panel) 3 days after addition of 800nM tamoxifen (Tam) to delete Ring1b (colored bars) versus untreated controls (white bars). Values were normalized to a housekeeping gene (Hmbs) and expressed as fold change relative to untreated. Data are mean ± SD from 3 experiments. *Significantly up-regulated expression in tamoxifen-treated samples (P < .05, Student t test). (G) Hematopoietic colony assays were performed in triplicate for untreated (−) and tamoxifen-treated (+) samples; FACS-sorted Flk1⁺ Ring1a^−/−/Ring1b^fl/fl/CreER_T2 cells were replated in differentiation medium (Exp1 and Exp2) or blast medium (Exp2*) with or without tamoxifen for 3 days before dissociation and scored 6 days after replating in semisolid hematopoietic medium. *Significantly different numbers of colonies in treated compared with untreated samples (P < .05, Student t test).