Exploring the Mechanism of ETO2-Dependent Transcriptional Regulation in Erythroid Cells

Developmental control mechanisms often utilize multimeric complexes containing transcription factors, coregulators, and additional non-DNA binding components. It is challenging to ascertain how such components contribute to complex function at endogenous loci. We recently analyzed the function of components of a complex containing master regulators of hematopoiesis (GATA-1 and Scl/TAL1) and the non-DNA binding components ETO2, the LIM domain protein LMO2, and the chromatin looping factor LDB1. We revealed that ETO2 and LMO2 regulate distinct target gene ensembles in erythroid cells. Furthermore, it was found that ETO2 commonly represses GATA-1 function via suppressing histone H3 acetylation, and also regulates methylation of histone H3 at lysine 27 (H3-trimeK27) at select loci, which suggested that ETO2 might be an important determinant of the erythroblast epigenome (Fujiwara et al. PNAS. 2010). Here, we investigated the role of ETO2 in the epigenetic regulation of erythroid genes.

CBFA2T3 mRNA (which encodes ETO2 protein) was cloned into pcDNA3.1 (Clontech) and Flexi HaloTag vector (Promega), and ETO2 was transiently overexpressed in K562 cells using Amaxa nucleofection technology™ (Amaxa Inc.). For ETO2 knockdown, pGIPZ lentiviral shRNAmir (Open Biosystems) was used. Quantitative ChIP analysis was performed using antibodies for acetylated H3K9 (abcam), trimethyl H3K27 (Millipore), ETO2, c-Myc (Santa Cruz), and HaloCHIP™ system (Promega). To obtain human primary erythroblasts, CD34-positive cells isolated from cord blood were induced in liquid suspension culture. For transcription profiling, human whole expression array was used (Agilent), and the data was analyzed with GeneSpring GX software.

First, we conducted microarray analysis to characterize ETO2 target gene ensemble using erythroid cell line (K562 cells). The analysis demonstrated that 598 genes were downregulated in the ETO2-overexpressed cells (> 2 fold). To test what percentages of ETO2-repressed genes could be direct target genes of GATA-1 or GATA-2 in K562 cells, we merged the microarray results with ChIP-seq profile (n= 5,749 and n=21,167 for GATA-1 and GATA-2 ChIP-seq, respectively) (Fujiwara et al. Mol Cell. 2009), and demonstrated that 23.1% and 40.5% of ETO2-repressed genes contained significant GATA-1 and GATA-2 peaks in their loci, respectively. Gene Ontology analysis among ETO2-repressed genes revealed significant enrichment of genes related to “oxygen transporter” and “hemoglobin complex” (p=0.00128), corresponding to HBG, HBB, HBE, HBA, HBQ, HBM and HBZ. We also confirmed that shRNA-mediated knockdown of ETO2 de-repressed globin genes in K562 cells. Quantitative ChIP analysis confirmed endogeneous and exogeneous ETO2 protein occupancy at beta-globin locus control region (LCR) and alpha-globin HS-40 in K562 cells. Furthermore, the overexpression significantly increased H3-trimeK27 and reduced acetylated H3K9 at these loci. Co-immunoprecipitation analysis revealed the interaction of ETO2 with EZH2/SUZ12, known as components of histone H3K27 methyltransferase complex, polycomb repressor complex 2 (PRC2), implying that the complex might be involved in ETO2-mediated transcriptional repression. To test if ETO2-mediated repression of globin genes is also observed in primary erythroblasts, we conducted shRNA-mediated knockdown of ETO2 in cord blood cell-derived primary erythroblasts, and demonstrated that ETO2 knockdown significantly de-repressed HBB and HBA expression. We are currently analyzing the mechanism of ETO2-dependent transcriptional repression and how ETO2-dependent histone marks are established in erythroid cells.

In conjunction with the evidence that ETO2 binds histone deacetylases and associates with GATA-Scl/TAL1 complex that binds epigenetic modifiers, our results suggest that ETO2 appears to have important roles in establishing the erythroblast epigenome.

No relevant conflicts of interest to declare.