In 1975, Riggs17 reviewed the role of DNA methylation on X inactivation, and Holliday and Pugh18 reviewed the role of DNA modification and gene activity on development. Both reviews proposed that DNA methylation acted as a regulator of gene expression in eukaryotic cells. Riggs pointed out that DNA methylation in eukaryotes had not been examined in the light of accumulating data on changes in the regulation in Escherichia coli involving bacterial DNA methylases and their potential for bringing about permanent changes in regulation. Arthur Riggs proposed that DNA methylation of cytosine might have a regulatory role in eukaryotes from the known effect of methylation of adenine on DNA-binding proteins in E. coli. Holliday and Pugh noted that the methylation of adenine in DNA in bacterium might present a very different phenotype from one without methylation. They suggested that the same ordered control of gene transcription could be achieved without changes in the DNA sequence by methylation of cytosine followed by its deamination to thymine as discussed by Scarano.13 Riggs and Holliday and Pugh presented convincing arguments that the methylation of cytosine in CpG doublets merited a more thorough biochemical and genetic study in a higher (eukaryotic) organism.
By 1980, a general consensus had been reached: that methylation of cytosines of CpG doublets was characteristic of genomic DNA, that a deficiency of CpG doublets in genomic DNA was probably due to instability of 5mC through its mutation to thymine, and that the distribution of CpG doublets in genomic DNA was not random. Additional studies then began to reveal connections between DNA methylation and gene expression. In one early study, treatment of a variety of cell lines with the methylation inhibitor 5-azacytidine revealed that a large number of genes were reactivated.19 However, the chemical mechanism by which cytidine analogs altered at the 5 position perturbed established methylation patterns was not clear. Subsequently, cytidine analogs that had been altered at the 5 position, such as 5-azacytidine and 5-aza-20-deoxycytidine, became important tools for studying the role of demethylation in gene expression, but it was not realized until investigators in Jaenisch's laboratory showed that incorporation of 5-aza-20-deoxycytidine into DNA led to covalent trapping of the DNA methyl-transferase enzyme. As a result, the cells were depleted of methyltransferase activity and underwent DNA demethylation that led to reactivation of the associated gene.20
In an attempt to explain the nonrandom distribution of CpGs, genomic DNAs across various nonvertebrates and vertebrates were compared. They showed that DNA methylation in nonvertebrate genomes was confined to a small fraction of nuclear DNA. This compartmentalization, termed ''low-density methylation,'' was also found in eukaryotes. In fact, most DNA, some 98%, in both non-vertebrate and vertebrate genomic DNA was methylated at low density. But in the remaining 2% of genomic DNA, regions of high-density DNA methylation existed, and their existence was particularly evident in vertebrate DNA. The regions rich in CpG nucleotides were designated as ''CpG islands.''21 More recent studies have shown the CpG islands are commonly found at the promoters of genes, in exons, and the 3'-regions of genes.22 But why the patterns of methylated DNA cytosines in vertebrates and nonvertebrates differed so strikingly remained an enigma.
Questions regarding the significance of DNA methylation engaged the attention of many investigators throughout much of the 1980s. Suppression of transposed elements and other "parasitic" elements had been demonstrated in various model systems, and investigators believed this was the ancestral function of DNA methylation. Studies outside the animal kingdom, as in slime molds and filamentous fungi, provided the strongest evidence in favor of this hypothesis. For example, Rothnie and colleagues23 had shown that DNA methylation of a single transposable element prevented damaging transposition events, and they inferred that DNA methylation of the foreign element repressed its function and prevented damaging transposition events. Other investigators demonstrated in different model systems that fully infectious proviruses were rendered harmless by methylation,24 or that demethylation of DNA by 5-azacytidine reactivated quiescent proviral genomes.25 Based on this information, Bird proposed that DNA methylation in invertebrates was more concerned with suppression of "selfish" elements that could disrupt gene structure and function, whereas vertebrates had retained the so-called ancestral function of DNA methylation, but had also adapted this process as a repressor of endogenous promoters of genes.
Was this article helpful?