Figure 7.
ATF4 is a direct target of RUNX1 in hematopoietic cells. (A) Analysis of publicly available RUNX1 ChIP-sequencing data shows that RUNX1 binds to the promoter region of ATF4 (black box) in human and mouse hematopoietic cells. The top panel shows ChIP-sequencing tracks for RUNX1 (blue), H3K27ac (black), and H3K4me3 (gray) around the ATF4 locus in human CD34+ HSPCs. The lower panel shows ChIP-sequencing tracks for RUNX1 (blue) in duplicate around the Atf4 locus in mouse EML hematopoietic precursor cells. (B) Control Runx1+/+ (blue) and Runx1KTAMK/KTAMK (red) LSK stem/progenitor cells were harvested, and ChIP analysis was performed by using a specific antibody against H3K27Ac or control immunoglobulin G (IgG). The amount of genomic DNA in the ChIP and input samples was measured by quantitative polymerase chain reaction using specific primers targeting the mouse Atf4 promoter or an intergenic region (n = 3 mice; 11-16 weeks old, from 3 independent experiments). (C) Luciferase (Luc) activity analysis showing the activation of the human ATF4 promoter (–463 bp to 82 bp of the transcription initiation site) by RUNX1 in human HL-60 promyelocytic leukemia cells. The Luc activities by wild-type (wt) RUNX1 and its binding partner CBFβ (light green), mutant (mut) RUNX1 R233K and CBFβ (light green), and mut RUNX1 R237K and CBFβ (red) were normalized to the activity by an empty vector (blue) (n = 2 independent experiments). (D) Luc activity analysis using a series of deletion/mut constructs in human HL-60 promyelocytic leukemia cells. The Luc activity by RUNX1 and CBFβ (blue) was normalized to the activity by an empty vector (red). Representative data from 1 of 2 independent experiments are shown. Note that 2 upstream open reading frames (uORFs) of ATF4 are known to negatively regulate mRNA translation. The region between −313 bp and 82 bp (Luc #3 wt) is sufficient for the full activation of the human ATF4 promoter, and mutagenesis in the second putative RUNX1 binding site (Luc #3 mut-2) impairs the reporter activity. The schematic representation of reporter plasmids is shown in the left. Two putative RUNX1 binding sites are depicted as black boxes, and the mutated binding sites are shown as white boxes. NS, not significant.