DNA Methylation

DNA methylation is found primarily at cytosine bases adjacent to guanine bases, referred to as a cytosine-phosphate-guanine dinucleotides ( CpGs ). Non- CpG DNA methylation has also been observed in some tissues, but will not be discussed here in detail [2]. The process of DNA methylation is catalyzed by a family of highly conserved enzymes called DNA methyltransferases (DNMTs). In mammals, there are three principal DNMTs: the de novo methyltransferases, DNMT3A and DNMT3B, add methyl groups to unmethylated DNA, while DNMT1 is responsible for the maintenance of DNA methylation at hemi-methylated sites during cell division [3]. Until recently, the removal of DNA meth- ylation marks was highly puzzling, as there was evidence for active demethylation, but the enzymes and process responsible remained elusive. In the past few years, new discoveries have shown that DNA demethylation is primarily controlled by 10-11 translocation methylcytosine dioxygenase (TET) enzymes, which catalyze the first step in active demethylation through the base excision repair system [1,3,4].

DNA methylation is highly genomic region-specific, partly because CpGs themselves are non-randomly distributed across the genome. CpGs are globally underrepresented in genomic sequences, but are enriched in regions called CpG islands: sequences that possess higher than average CpG density and are generally lowly methylated. Interestingly, CpGs are also very dense at many repetitive elements, therefore such high density regions tend to be methylated and silenced to avoid transposition [5, 6]. Nearly 70 % of gene promoters are near a CpG island, and this proximity results in an important relationship between promoter DNA methylation and gene expression [7]. Early models of this relationship posited that high levels of DNA methylation at promoter-associated CpG islands were associated with low levels of expression, and vice versa [8-10] (Fig. 3.1). However, this model is relatively simplistic and we are now beginning to understand that the relationship is much more complex [11-13]. For instance, reverse causation is possible and gene expression levels can influence levels of DNA methylation [12, 14]. In addition, DNA methyla- tion marks in genomic locations other than promoters have equally importance and, in some cases, very different associations with gene expression [12]. For example, both regions of low CpG density directly adjacent to CpG islands, known as “island shores”, and enhancer regions do not follow the high DNA methylation/low gene expression model. In island shores, higher levels of DNA methylation are often associated with high levels of gene expression, while at enhancers these relationships have been inconsistent and thus far are poorly understood [1, 5-7,14,15].

The underlying genetic sequence can also influence both gene expression and DNA methylation. Reminiscent of earlier work examining the linkage of allelic variation and gene expression, multiple studies have identified SNP/CpG pairs at which genotype is associated with methylation level [8-10,12,16-18]. These CpG sites are referred to as methylation quantitative trait loci (mQTLs), and a number of different mechanisms have been hypothesized for their function. For example, the SNP partner may disrupt genetic elements responsible for creating boundaries between methylated and unmethylated genomic regions [11-13,19]. Alternatively, sequence changes may alter gene expression levels, which in turn influence DNA methylation levels [14]. Regardless of the mechanism by which this relationship functions, it is very important to consider genetic regulation in the analysis of DNA methylation effects.

DNA methylation patterns at promoter- associated CpG islands and gene bodies in active

Fig.3.1 DNA methylation patterns at promoter- associated CpG islands and gene bodies in active (above) and inactive (below) genes. Each gene indicates exons (black), introns (white), CpG islands (dark grey) and shores (light grey). Below each gene the drawing indicates where DNA methylation is high or low across the gene

Finally, the relationships between DNA methylation, genotype, and gene expression are dynamic and may change over the life span. For example, it appears that genotype-DNA methylation correlations are susceptible to age-related differences. From early to later life, more than half of the associations between genotype and DNA methylation change [20]. Conversely, correlations between DNA meth- ylation and gene expression appear to be more consistent over the ageing process [21]. Continuing research is required to understand the entire complexity of the role of DNA methylation in genetic regulation across the lifespan.

< Prev   CONTENTS   Source   Next >