CoCoA/CCAR1 Pair-Mediated Recruitment of Mediator Complex Indicates Novel Pathway for the Function of GATA1 in Erythroid Differentiation

Abstract 1302

The mammalian multi-protein complex Mediator, originally identified by ourselves as a nuclear receptor-specific coactivator complex, is a phylogenetically-conserved subcomplex of the RNA polymerase II holoenzyme and serves as an end-point integrator of diverse intracellular signals and transcriptional activators. The 220-kDa Mediator subunit MED1 is a specific coactivator not only for nuclear receptors but for GATA family activators, and serves as a GATA1-specific coactivator that is essential for optimal GATA1-mediated erythropoiesis. In this study, we show a novel nuclear signaling pathway for MED1 action in GATA1-induced transcriptional activation during erythroid differentiation. First, we identified the amino acid residues 681–715 of human MED1 (MED1(aa.681-715)) to be responsible for the direct interaction with GATA1. When MED1 in K562 human erythroleukemic cells was knocked down during hemin-induced erythroid differentiation, the erythroid differentiation was significantly attenuated as assessed by an erythroid differentiation score defined by the number of cells positive for benzidine staining, and the expressions of the GATA1-targeted and erythroid differentiation marker genes, β-globin, γ-globin, PBGD and ALAS-E, were prominently attenuated. However, overexpressions of the N-terminal MED1 truncations without and with nuclear receptor recognition motifs, MED1(aa.1–602) and MED1(aa.1–703), respectively, but neither of which could bind to GATA1 (above), prominently enhanced erythroid differentiation of K562 cells. Luciferase reporter assays by using the human γ-globin promoter and Med1^−/− mouse embryonic fibroblasts (MEFs) showed that these N-terminal MED1 truncations rescued GATA1-mediated transactivation, indicating that MED1(a.a.1–602) served as the functional interaction surface for GATA1. Hence, a putative bypass for GATA1-MED1 pathway appears to exist, and is expected to interact with the N-terminus of MED1. As a candidate bypass system, we tested both the recently reported bypass molecule for a nuclear post-activator signaling, CCAR1, and its partner coactivator CoCoA. CCAR1 was reported by others to bypass the estrogen receptor-mediated transactivation by a simultaneous binding of CCAR1 with the estrogen receptor and the N-terminus of MED1. Functionally, serial luciferase reporter assays by using the γ-globin promoter and MEFs demonstrated cooperative transactivation by combinations of GATA1, CCAR1, CoCoA and/or the N-terminus of MED1, but the transactivation mediated by the N-terminus of MED1 was not as prominent as the one mediated by the full-length MED1. An overexperssion of CCAR1 or CoCoA in K562 cells prominently enhanced both the GATA1-mediated erythroid differentiation and the expressions of the GATA1-targeted genes. Next, the mechanisms underlying the CCAR1- and CoCoA-mediated GATA1 functions were analyzed by serial GST-pulldown and mammalian two-hybrid assays, and the following results were obtained. (i) The N-terminus of CCAR1 interacted with the C-terminus of CoCoA. (ii) The N-terminus of MED1 interacted with both the N- and C-termini of CCAR1. (iii) While the N-terminal zinc-finger domain of human GATA1 (GATA1(a.a.204–228)) is known to bind to the well-known GATA1 partner FOG1, intriguingly, the C-terminal zinc-finger domain of GATA1 (GATA1(a.a.258–272)) interacted with all three of the following cofactors; MED1 (MED1(aa.681–715)), CCAR1 (at the C-terminus) and CoCoA (at both the N- and C-termini). The affinity of CoCoA to bind to GATA1 appeared to be a little higher than the other. Thus, the GATA1(a.a.258-272) zinc finger appears to serve as a docking surface for multiple coactivating proteins, where both MED1 and CoCoA/CCAR1 pair can interact, probably in a competitive manner, or perhaps simultaneously. Here, both CoCoA/CCAR1 as a pair and CCAR1 by itself can serve as a bypass. Finally, ChIP assays of hemin-treated K562 cells showed that GATA1, CCAR1/CoCoA and MED1 were all recruited onto the γ-globin promoter during transactivation. Taken together, besides a direct interaction between GATA1 and MED1, the CoCoA/CCAR1 pair appears to relay the GATA1 signal to MED1. The multiple modes of mechanisms for transcription mediated by the GATA1-MED1 axis might contribute to a fine tuning of the GATA1 function, not only during erythropoiesis but also in other GATA1-mediated homeostasis events, within a living animal.