Abstract
The active elements of the β-globin LCR are necessary for high-level expression of the linked globin genes. These active elements are contained within regions of unique, erythroid-specific chromatin structure characterized by the formation of positioned nucleosomal arrays, areas of approximately 150–200 bp which are highly accessible for protein-DNA interactions (DNase I HSs) and hyperacetylation of histones H3 and H4. At a DNA level, each of the LCR HS cores contain NF-E2, EKLF/Sp1 and tandem, inverted GATA binding elements. Our lab has shown that all of these transcription factor-binding sites are necessary for HS formation. We have previously proposed a model of HS formation in which Sp1 and/or EKLF first bind to the core regions and alter local chromatin structure allowing GATA-1 and NF-E2 to bind and stabilize the HS core structure. Additional factors could then bind to the core and flanking regions. An important aspect of HS core formation that our experiments have not previously addressed is the mechanism of histone acetylation. Based on our previous findings and the fact that GATA-1 and NF-E2 can both interact with histone acetyltransferases (HATs), we hypothesized that histone acetylation at the HS requires the binding of EKLF/Sp1 followed by GATA-1 and/or NF-E2 and therefor should be a late event in HS formation. To test this hypothesis, we have generated a series of 11 artificial HS core elements based on the structure of human HS4. These plasmid-based constructs contain different combinations of the six binding sites in their normal arrangement and spacing: (EKLF/Sp1)-(NF-E2)-(EKLF/SP1)-(GATA-1 x 2)-(EKLF/Sp1). These constructs were stably transfected into MEL cells and assayed for their ability to mediate histone H3 and H4 acetylation as determined by the ChIP assay. The endogenous murine β-globin LCR HS3 served as a positive internal control for histone acetylation. Previous results from our lab showed that all six factor-binding sites within the HS core are required for HS formation. However, when assayed for histone acetylation, the NF-E2 and the tandem inverted GATA sites are able to independently direct H3 and H4 acetylation to between 10 and 50% of that seen at the native HS site. Combination of the NF-E2 and GATA-1 sites results in 89% of normal H4 acetylation. Full HH4 acetylation (98% of the wild-type HS) is only seen with all six factor binding sites. These results indicate that NF-E2 and GATA-1 direct most of the acetylation of LCR HS core histones but that EKLF binding is also likely to contribute. They also indicate that extensive histone acetylation can occur without HS formation. Finally, in contrast to our model of HS formation, GATA-1 and NF-E2 appear able to direct the acetylation of HS core histones without the presence of binding sites for Sp1 or EKLF.
Author notes
Corresponding author