Silencing of γ-Globin Gene Expression during Adult Definitive Erythropoiesis Is Mediated by a GATA-1 Repressor Complex.

Abstract

Autonomous silencing of gene expression is one mechanism operative in the control of human β-like globin gene switching. Experiments using variously truncated Aγ-globin genes linked to LCR sequences suggested that a region of the Aγ-globin gene between -730 to -378 relative to the mRNA CAP site may function as an adult stage-specific silencer element. A marked copy of the Aγ-globin gene (Aγm-globin) was inserted between LCR 5′ HS1 and the ε-globin gene in a human β-globin locus yeast artificial chromosome (Aγm 5′ ε β-YAC). The Aγm-globin gene was autonomously silenced in Aγm 5′ ε β-YAC transgenic mice, even in the absence of an adult β-globin gene. A -730 to -378 deletion of the Aγm-globin gene was introduced into the Aγm 5′ ε β-YAC to produce a Δ1s Aγm 5′ ε β-YAC. Transgenic lines containing intact β-globin loci expressed the Δ1s Aγm-globin gene in embryonic yolk sac, fetal liver, and adult blood. To further delineate the function of the Δ1s fragment, transient transfection assays and protein-DNA interaction assays were performed. The Δ1s fragment was found to act as a repressor of a constitutively active SV40 promoter in K562 cells. DNaseI footprinting analysis and electromobility shift assays demonstrated GATA-1-binding at a site -570 bp upstream of the Aγ-globin CAP site. Recently generated β-YAC transgenic mice containing a T>G point mutation at the -570 GATA site of the normally-located Aγ-globin gene displayed a HPFH phenotype. Together, these data suggested that the -730 to -378 Aγ-globin gene region contains a silencer element at the -570 GATA site that binds a GATA-1 repressor complex during the adult stage of definitive erythropoiesis to silence expression of the Aγ-globin gene. Previous studies suggested that when GATA-1 functions as a repressor, it interacts with components of the MeCp1/NuRD complex. This complex may remodel chromatin into a repressed state, leading to silenced Aγ-globin gene expression during adult definitive erythropoiesis. The presence of components of the MeCP1/NuRD complex was assessed in uninduced (γ-globin repressor present) and induced (γ-globin repressor absent) erythroid cells (K562 and KU812) and non-erythroid cells (HFF) by Western blot analysis using an antibody to Mi2, which is a component of the NuRD complex. Mi2 protein was observed in erythroid cells when the levels of γ-globin were low (uninduced K562 or KU812 cells), whereas only a weak signal was detected when γ-globin expression was induced in these cells. The Mi2 signal in the HFF cells was even weaker. Chromatin immunoprecipitation (ChIP) using fetal liver samples from day E12 and E18 conceptuses of wild-type β-YAC transgenics showed that GATA-1, FOG-1 and Mi2 proteins co-localize to the -570 GATA site of the Aγ-globin gene in samples where γ-globin is silenced (E18 fetal liver), but not in samples where γ-globin is expressed (E12 fetal liver). Our data strongly suggest that the MeCP1/NuRD complex interacts with GATA-1 protein to form a repressor that may be involved in silencing Aγ-globin gene expression. In addition, we show that GATA-1, FOG-1 and Mi2 are recruited to the analogous -567 GATA site of Gγ-globin, in a pattern that parallels that of Aγ-globin. However the binding of these proteins to Gγ-globin is weaker than that observed for Aγ-globin. These data suggest that GATA-1-mediated repression is common to both γ-globin genes, but that other mechanisms function in the differential regulation of the two γ-globin genes.