To address the mechanism by which the human globin genes are activated during erythropoiesis, we have used a tiled microarray to analyze the pattern of transcription factor binding and associated histone modifications across the telomeric region of human chromosome 16 in primary erythroid and nonerythroid cells. This 220-kb region includes the ␣ globin genes and 9 widely expressed genes flanking the ␣ globin locus. This unbiased, comprehensive analysis of transcription factor binding and histone modifications (acetylation and methylation) described here not only identified all known cis-acting regulatory elements in the human ␣ globin cluster but also demonstrated that there are no additional erythroid-specific regulatory elements in the 220-kb region tested. In addition, the pattern of histone modification distinguished promoter elements from potential enhancer elements across this region. Finally, comparison of the human and mouse orthologous regions in a unique mouse model, with both regions coexpressed in the same animal, showed significant differences that may explain how these 2 clusters are regulated differently in vivo.
IntroductionGlobin gene expression plays a central role in normal erythropoiesis, and abnormalities of globin synthesis underlie the clinically important thalassaemia syndromes. A long-term goal in the treatment of thalassaemia has been to identify ways in which to manipulate gene expression to redress the imbalance in ␣ and  globin chain synthesis. Progress in this area would be enhanced by understanding in detail how these genes are normally regulated. To address this, it is important to identify all the key cis-acting sequences controlling ␣ and  globin expression, the transcription factors that bind them, and the accompanying modifications of chromatin at different stages of differentiation.To identify and characterize the cis-acting elements controlling ␣ globin expression we have previously used a variety of approaches, including comparative genomics, 1,2 a wide range of chromatin analyses, 3,4 transgenic experiments, 5,6 and the analysis of expression in interspecific hybrids. [7][8][9] Upstream of the 4 ␣-like globin promoters (, ␣D, ␣2, and ␣1) lie 4 multispecies conserved sequences (MCSs) associated with erythroid-specific DNase1 hypersensitive sites (HS), referred to, in human, as HS-48, HS-40, HS-33, and HS-10 ( Figure 1A). The globin genes, the upstream MCSs, and several widely expressed flanking genes all lie within a region of conserved synteny spanning 130 kb of the telomeric region of human chromosome 16 (16p13.3) ( Figure 1A). Functional analyses suggest that HS-40 is the most important of the MCSs. 9 Recently, to develop a relevant experimental system for studying human ␣ globin expression and to evaluate the roles of individual elements, we established a humanized mouse model in which the 117-kb region containing the human ␣-like globin genes and all of their known regulatory elements (the ␣ globin regulatory domain) completely replaces the corresponding, orthologou...