Abstract
Despite intense investigation, the mechanisms by which human γ- to β-globin developmental gene switching occurs have yet to be fully elucidated. Based on studies in many systems, including human clinical trials with 5-Azacytidine and deoxyazacytidine, methylation has been thought to play an important, and more significantly, reversible role in γ-globin gene silencing. One mechanism by which DNA methylation is likely to effect γ-globin gene expression is through site-specific modification of CpG residues in the promoter regions of the γ-globin genes. For example, CpG methylation status has been proposed to mediate the developmentally-specific binding of Sp1 and the stage selector protein (SSP) complex to the proximal γ-globin promoters. We began this study to determine whether there were other CpG residues in the regions of the γ-globin promoters whose methylation status correlated with γ-gene silencing and thus might also serve as “molecular switches” regulating transcription factor binding, local histone acetylation and globin gene expression. To determine the methylation patterns of the γ-globin promoters, we first purified erythroid cells from human fetal liver (FL) and adult bone marrow (BM) using anti-glycophorin magnetic beads. DNA from the purified cells was subjected to bisulfite modification. The regions of the γ-globin promoters were PCR amplified and subcloned into plasmids. Individual plasmids were then sequenced to determine the methylation status of promoter regions. PCR primers were used which allowed determination of methylation status for CpGs at positions −249, −158, −52, −49. +6, +18, and +49 of the G and Aγ globin genes. An additional CpG at +210 was detectable with the Gγ primers. So far we have analyzed Gγ promoters from three FL and three BM samples. Aγ has been analyzed from two FL and three BM samples. An average of 10 clones have been sequenced for each sample. When results for samples within each condition (i.e., Gγ in FL) were combined for analysis, we see the expected increase in methylation of CpG residues in the Gγ promoter from 38% of all sites in the FL to 73% in the adult BM. This difference increases from 30% in FL to 88% adult BM when the CpGs at −158 and +210 are excluded. Combined methylation at these sites only increases from 7 to 21% between FL and BM and thus does not correlate well with changes in gene expression. Looking at the data another way shows a shift from most of the FL clones (76%) having 0 or 1 sites methylated in the Gγ promoter to 78% of the clones having 6,7 or 8 methylated sites in adult cells. While these results fit with the paradigm that methylation is associated with gene silencing, we saw a very different picture for Aγ. Because the promoter regions have nearly identical sequences, are located very close to each other and are similarly regulated, we expected their methylation patterns to be similar. However, for Aγ 13% of promoter CpGs are methylated in the FL cells but this increases to only 22% in adult erythroid cells. Maximal Aγ promoter methylation occurs at the +6 and +18 CpGs which reach only 33 and 36% methylation in adult erythroid cells. 86% of FL Aγ clones are methylated at only 0–2 promoter CpG sites. This does not change at 83% in adult cells. These results indicate differential methylation of the two human γ-globin genes and suggests that simple promoter methylation is not the primary mechanism of γ-globin gene silencing.
Author notes
Corresponding author
This feature is available to Subscribers Only
Sign In or Create an Account Close Modal