Abstract

The activity of the β-globin locus is regulated by the locus control region (LCR) which in humans and mice is comprised of six DNaseI hypersensitive sites (HSs) located upstream of the β-like globin genes. Hispanic thalassemia, a naturally occurring deletion of the LCR plus 25kb upstream results in the failure to activate the β-globin locus at the levels of chromatin structure, transcription and replication. In order to examine how the HSs interact to regulate the endogenous β-globin locus, we have utilized homologous recombination for the mutational analysis of the endogenous murine β-globin LCR in embryonic stem cells, followed by the generation of mice. Previously we reported that deletion of the endogenous LCR by homologous recombination (ΔLCR) does not completely silence expression of the β-like genes, and has no measurable effect on nuclease sensitivity, promoter hypersensitive site formation or core histone hyperacetylation. Thus the LCR provides a necessary enhancer-like activity. In addition, while loss of the LCR leads to only a slight decrease in pre-initiation complex formation and Pol II binding to the promoter, there is a significant decrease in downstream polymerase and this correlates with a basal level of ser-5 phosphorylation of Pol II. To determine if the decrease in downstream Pol II observed along Δthe LCR allele is due to decreased release of polymerase from the promoter, or downstream polymerase pausing, KMnO4 in vivo foot-printing was done.

Comparison of the 5′ end of WT and ΔLCR β-maj globin genes reveal similar patterns consistent with pol II pausing on both alleles, and suggesting that the LCR stimulates elongation by releasing promoter proximal paused polymerases. To further characterize LCR mediated activation and determine what characteristics vary with the level of transcription we have continued our analysis of mice with ΔLCR allele expressing at 1–4% of wild-type (WT), and mice with a deletion of HS 2 and 3 (Δ23) expressing at 30% of WT. The Δ23 allele was chosen as it demonstrates an intermediate transcriptional phenotype and is one of several double HS deletions that demonstrate that the LCR HSs contribute additively to globin gene transcription. To determine if histone modifications other than acetylation vary with LCR mutations and is associated with the level of transcription, the state of K4-tri and K79-di histone methylation was assayed along exon 3 of the β-maj gene of WT, Δ23 and ΔLCR alleles. While in some systems enrichment of methylation of K4 and K79 are associated with a permissive or activated state, we find increases in histone methylation with deletion of the LCR, and an intermediate degree of methylation with deletion of HS 2 and 3. Thus, surprisingly, increasing levels of transcription correlate with decreases in histone methylation. Finally, as enhancers increase the probability rather than the rate of expression we have analyzed WT, Δ23 and ΔLCR alleles with single allele transcription assays. RT-PCR reveals all Δ 23 and ΔLCR alleles express mRNA at 30 and 4% of WT respectively, consistent with bulk RNA analysis and suggesting that individual alleles are not permanently silenced. In contrast, primary transcript RNA FISH analysis demonstrates that the mutant alleles are less likely to be expressed, thus LCR HSs may affect the probability as well rate of transcription.

Author notes

Corresponding author