Cellular and transgenic models have provided important insights into the molecular mechanisms required for mammalian globin gene expression. The homology of the human and murine α- and ζ-promoter sequences coupled with in vivo transcriptional analysis suggested that the genes were controlled in a similar manner in murine and human erythroid cells. The identification of 4 upstream erythroid-specific DNaseI hypersensitive (HS) sites in both species confirmed the overall structural similarity of the α-loci. Support for a conservation of function was provided by the demonstration that some HS sites enhance transcription significantly when linked with a globin promoter in cis. These studies resulted in a model in which the sequences required for developmental-specific expression reside in gene-proximal promoters, the upstream HS being required for high-level tissue-specific expression.

A key test of this model is provided by Anguita and colleagues (page 3450), in which they determine the function of a key upstream HS site at the endogenous murine α-globin locus. Previous work by this group using a chromosome 16 somatic hybrid cell line demonstrated that disruption of hHS −40, a single HS site located 40kb upstream of the human α-globin genes, resulted in complete loss of α-globin gene expression. In the present study, homologous disruption of the apparently functionally equivalent murine HS site, mHS −26, was performed with the expectation of a resultant severe α-thalassemic phenotype in animals homozygous for the deletion. In contrast to the human studies, homozygotes have relatively normal erythropoiesis except when subjected to anemia-inducing phenylhydrazine stress. This result is reminiscent of similar studies deleting individual HS sites in the murine β-globin locus control region (LCR) (Bender et al, Mol Cell. 2000;5:387-393). Several potential explanations for the species-specific difference in outcome can be entertained. These include the unique telomeric location of the human locus when compared to the interstitial position of its murine ortholog, a modest divergence in factor binding sites between the mHS −26 and hHS −40 regions, and/or the confirmed differences in enhancer action of hHS −40 and mHS −26, as measured in murine erythroleukemia cells as reported here. An equally attractive hypothesis, linking many of these observations, is that 2 or more HS sites may be required for appropriate murine α-globin gene expression. In contrast, hHS −40 may be sufficient at the human locus. Further experimentation at the endogenous loci will be required to distinguish these possibilities. Indeed, it is likely that further Darwinian surprises are in store that will remind us that murine and human globin biology may not be one and the same.