NLI/Ldb1, is a widely expressed nuclear factor that we have shown is required for long range interaction between the beta-globin LCR and beta-globin promoter. The C-terminal LIM-interaction domain (LID) of Ldb1 interacts with LMO2 which provides association of Ldb1 with chromatin through DNA-binding partners GATA-1 and TAL1. It is known that the N-terminal dimerization domain (DD) of Ldb1 participates in homodimerization of the protein in vitro. We proposed that the DD domain of Ldb1 plays a key role in long range interaction between the LCR and beta-major gene promoter in vivo. To confirm this hypothesis, Ldb1 knock down in MEL cells was rescued by expression of HA tagged Ldb1 from a transgenic construct. Transgenic Ldb1 interacts with the LCR and beta-globin promoter and rescues GATA1/TAL1/LMO2 protein complex binding, along with beta-globin gene expression. Then several deletions of Ldb1 missing short conserved sequences in DD were expressed in the background of endogenous Ldb1 knock down MEL cells. Most of these mutated proteins can interact with the LCR and beta-major gene promoter but are unable to activate beta-globin gene expression. To show a direct role of the DD domain in beta-major globin gene activation, a fusion protein that contains LMO2 and the DD domain of Ldb1 (LMO-DD) was expressed in the Ldb1 KD background. LMO-DD fusion protein can interact with the LCR and beta-globin promoter and rescue GATA1/TAL1 protein binding along with activation of beta-globin gene expression. Coimmunoprecipitation of a truncated Ldb1 protein consisting of only the DD domain with endogenous LDB1 confirms that homodimerization of LDB1 in cells occurs by interaction between DD domains. These experiments demonstrated that the DD domain of LDB1 is necessary and sufficient for beta-globin gene activation. In other experiments, we explored the mechanisms underlying enhancer blocking by insulators. We found that human beta-globin HS5, the orthologue of the CTCF dependent chicken HS4 insulator, has intrinsic, portable enhancer blocking activity that is manifest though chromatin loop formation. To investigate whether the looping activity of CTCF sites is a general property of these sites in the genome, we carried out chromatin conformation capture (3C) on CTCF sites over 2 Mb on chromosome 11 encompassing the beta-globin locus and flanking olfactory receptor genes. We next investigated the 3-dimensional folding properties of a polymer model of chromatin using these established looping interactions of the insulator protein CTCF comprehensively determined in erythroid cells, where the locus is actively transcribed, and in non-erythroid cells, where it is silent. Our results indicate that cell type specific chromosomal interaction frequencies mediated by CTCF in erythroid cells are spatially organized to favor contacts between the beta-globin locus control region (LCR) and genes, which are known to occur in vivo. By contrast, fewer dominant interactions in non-erythroid cells are shown to specifically drive the beta-globin genes away from the LCR, suggesting a mechanism for transcriptional control. Thus, experimental data and appropriate polymer models provide a physiological meaningful description of chromatin dynamics below the megabase scale. In particular, we posit that the biological function of the beta-globin locus relies, at least in part, on a polymer-based biophysical mechanism where CTCF long-range interactions functionally affect spatial distances between control elements and gene promoters.
Showing the most recent 10 out of 21 publications
