The Hbo1-Brd1/Brpf2 HAT Complex Is Required for Erythropoiesis In Fetal Liver

Abstract 3872

Bromodomain-containing protein 1 (Brd1, initially designated as BR140-LIKE; BRL) contains a bromodomain, two plant homology domain (PHD) zinc fingers, and a proline-tryptophan-tryptophan-proline (PWWP) domain, three types of modules characteristic of chromatin regulators. Recently, BRD1 appeared to belong to the BRPF family which includes BRPF1, BRD1/BRPF2, and BRPF3. Among them, BRPF1 is known to be a subunit of the MOZ H3 histone acetyltransferase (HAT) complex. BRD1 has been proposed to be additional subunit of the MOZ H3 HAT complex on the analogy of BRPF1. However, its molecular function remains elusive. To elucidate the biological functions of BRD1, we generated Brd1-null mice and found that they die in utero. Brd1^-/- embryos were alive and recovered at nearly the expected Mendelian ratio at 12.5 days postcoitum (dpc) but died by 15.5 dpc. Brd1^-/- embryos at 12.5 dpc were pale and the cell number of fetal livers, in which fetal hematopoiesis occurs, was decreased to about 20% of the control. Cytological analysis revealed that Brd1^-/- fetal livers had profoundly fewer erythroblasts at maturation stages beyond proerythroblasts compared to wild-type fetal livers. Flow cytometric analysis of Brd1^-/- fetal livers revealed a significant accumulation of CD71⁺Ter119^- proerythroblasts and a reduction in CD71⁺Ter119⁺ and CD71^-Ter119⁺ maturating erythroblasts. A drastic increase in AnnexinV⁺ apoptotic cells was detected in the CD71⁺Ter119⁺ and CD71^-Ter119^- cell fractions in Brd1^-/- fetal livers. These findings suggested that severe anemia caused by compromised differentiation and/or survival of erythroblasts accounts for embryonic lethality of Brd1^-/- embryos. To understand the mechanism underlying defective erythropoiesis in Brd1-null embryos, we performed biochemical analyses and found that Brd1 bridges the HAT, HBO1 but not MOZ, and its activator protein, ING4, to form an enzymatically active HAT complex. Forced expression of Brd1 promoted erythroid differentiation of K562 cells, while Brpf1, which preferentially binds to MOZ, had no significant effect. Correspondingly, depletion of Hbo1 by Hbo1 knockdown perturbed erythroid differentiation of mouse fetal liver progenitors. Of note, the level of global acetylation of histone H3 at lysine 14 (H3K14) was specifically decreased in Brd1-deficient erythroblasts. These results collectively implied that acetylation of H3K14 catalyzed by the Hbo1-Brd1 complex has a crucial role in fetal liver erythropoiesis. To identify the downstream targets for the HBO1-BRD1 complex, we performed the ChIP-on-chip analysis in K562 cells and found that BRD1 and HBO1 largely co-localize on the genome, especially on the promoters of erythroid transcription factor genes. ChIP analysis revealed that acetylation of H3K14 at the promoters of erythroid transcription factor genes, including Gata1, Gata2, Tal1, Stat5a, and ETO2, were profoundly diminished in the Brd1-deficient erythroblasts. Among these target genes, we focused on Gata1, which plays a central role in erythropoiesis, and carried out complementation experiments with Gata1 using a Gata1 retrovirus. Exogenous Gata1, but not Bcl-xL, efficiently improved proliferative capacity and survival of Brd1-deficient erythroid progenitors and also restored, at least partially, their impaired differentiation. These results clearly showed that the Hbo1-Brd1 complex is required for the acetylation of H3K14 at the promoters of erythroid transcription factor genes, thereby is crucial for erythropoiesis in fetal liver.