GATA-1-Specific Coactivator MED1/TRAP220, Transcriptional Mediator Subunit, Mediates Erythroid Differentiation in K562 Erythroleukemia Cells.

The multi-protein complex TRAP/Mediator is a subcomplex of RNA polymerase II holoenzyme and acts as the end-point integrator of a variety of activators and intracellular signalings. The Mediator activates transcription in concert with a group of recruited general transcription initiation factors. Among the circa 25 subunits, the MED1/TRAP220 subunit was originally demonstrated by us as a key component that acts as a ligand-dependent nuclear receptor-specific coactivator, and proved to play an important role in retinoic acid-dependent granulocytic and 1,25-dihydroxyvitamin D₃-dependent monocytic differentiation of both normal hematopoietic precursor and acute promyelocytic leukemia cells, via retinoic acid receptor (RAR) and vitamin D receptor (VDR), respectively. MED1 was recently re-identified as a specific coactivator for GATA family activators and found to function in the process of GATA-1-mediated normal erythroblast differentiation. In this study, we analyzed the role of MED1 in erythroid differentiation of malignant erythroblastic (leukemic) cells. We used K562 human erythroleukemia cells as a model since GATA-1-mediated erythroid differentiation of K562 is efficiently induced with hemin. When K562 cells were exposed to 50nM hemin for 5 days, 44% of cells became positive for benzidine staining and the expression of the erythroid lineage-specific (and GATA-1-targeted) genes such as β-globin, γ-globin, porphobilinogen deaminase (PBGD) and 5-aminolevulinate synthase (ALAS-E) were strongly induced. The Mediator subunits including MED1 were also induced strongly together with these erythroid-specific genes in this process. Since the precedent study has demonstrated that Mediator subunits are induced during nuclear receptor-mediated myelomonopoiesis, and since genes whose products have physiological significance are often induced, Mediator might be employed in GATA-1-mediated erythroid differentiation of erythroleukemia cells as well. Then we focused on the GATA-1-specific coactivator MED1 among the Mediator subunits. In order to analyze the role of MED1, we first reduced the expression of endogenous MED1 in K562 cells with the MED1 knockdown small hairpin RNA (shRNA) expression vector. When the transiently knocked down cells (approximately 25% of the control) were treated with hemin, the number of benzidine-positive cells were half of the control in 3 days, indicating that hemin-induced erythroid differentiation was strongly inhibited. Further, the expression of direct target genes for GATA-1 such as β-globin, γ-globin, PBGD and ALAS-E were significantly reduced by approximately 70% in knockdown cells. Moreover, when the knockdown cells were cultured for 3 days without hemin, the basal expression levels of the erythroid lineage-specific genes such as β-globin, γ-globin and PBGD were significantly reduced by approximately 30%. We next asked if overexpressed MED1 might enhance erythroid differentiation. Since the MED1 domain near the N-terminus (a.a. 622 to 701) interacts with GATA-1 and that the N-terminal domain that contains two LxxLL nuclear receptor recognition motifs (NR box; aa. 592 to 703) is sufficient for in vitro and most in vivo transcription events, we tested the N-terminal truncation of MED1, N-MED1 (aa. 1 to 703), for overexpression. When N-MED1 was stably expressed in K562 cells after selection with G418, MED1 expression was boosted at least 2.2 folds of the control transfectant cells. When these cells were treated with hemin, benzidine-positive differentiated cells were 3-fold reduced and the expression of the erythroid differentiation marker genes (above) were approximately 2-fold increased in N-MED1 cells after one day. However, after 3 days, the number of benzidine-positive cells and the expression of erythroid marker genes in N-MED1 cells were saturated and caught up by the control cells. Thus, it appears that overexpressed MED1 enhances the erythroid differentiation in K562 cells. Together with the finding that GATA-1 is physically associated with MED1 and that MED1 is co-occupied with GATA-1 in the enhancer region, these results indicate that MED1 functions as a GATA-1-specific coactivator in erythroid differentiation of not only normal erythroblasts but erythroleukemic cells. MED1 also appears to be a key coactivator for two distinct lineages of normal and malignant hematopoiesis, namely myelomonopoiesis and erythropoiesis.