Mechanistic study on the DNA methyltransferase DNMT3A

Abstract

The phenotypical and functional diversity of mammalian cell types can be attributed to a large extent to epigenetic signals that determine and stabilize gene expression profiles. One of the most important types of epigenetic signals is DNA methylation. This modification is set early in development by the de novo DNA methyltransferases DNMT3A and DNMT3B, and is found predominantly at the C5 position of cytosine bases in a CpG dinucleotide context. The accurate setting of DNA methylation patterns is critical for normal development and is determined by the precise recruitment and control of DNMT activity on chromatin. In this work, four main directions of research were undertaken, with the ultimate goal of shedding novel mechanistic insights into the mechanism of DNMT3A, its regulation by chromatin signals and interaction partners, as well as the dysregulation of this enzyme in cancer. Furthermore, the potential of DNMT3A to generate 3-methylcytosine as a side reaction was explored. The DNA methyltransferase DNMT3A has been shown to multimerize on DNA and to form large multimeric protein/DNA fibers. However, it has also been postulated that this enzyme can methylate DNA in a processive manner, a property incompatible with fiber formation. By using a dedicated set of biochemical experiments, I was able to show that the DNA methylation rate of DNMT3A increases more than linearly with increasing enzyme concentration on a long DNA substrate, but not on a short 30-mer oligonucleotide, which cannot accommodate DNMT3A polymers. Methylation experiments over a range of enzyme concentrations and with substrates containing one or two CpG sites did not provide evidence for a processive mechanism. The addition of a catalytically inactive DNMT3A mutant was found to increase the DNA methylation rate by DNMT3A on the long substrate but not on the short one. Together, these data clearly indicate that DNMT3A binds to DNA in a cooperative reaction and the formation of protein/DNA fibers increases the DNA methylation rate. These results contribute mechanistic insights into the mode by which DNA methylation patterns are established during development. The second project dealt with characterizing the effects of the R882H exchange on DNMT3A. The R882H mutation is found in the DNA binding interface of DNMT3A and is frequently observed in acute myeloid leukemia (AML). By establishing a double-tag affinity purification system, I was able to show that the mutation only leads to a minor reduction in overall DNA methylation activity in mixed R882H/wildtype DNMT3A complexes. However, a pronounced change in flanking sequence preference of the DNMT3A-R882H mutant was found. Accordingly, a substrate designed to contain the target CpG site flanked by sequences preferred by R882H was better methylated by the variant than by the wildtype enzyme. Together, these data strongly argue against a dominant-negative effect of the R882H mutation and rather propose a site-specific gain-of-activity effect. These findings are in agreement with a recently determined structure of DNMT3A in complex with DNA and they might explain the high prevalence of this specific point mutation in AML. The third project was built on previous data from the lab, documenting a strong and direct interaction between the ADD domain of DNMT3A and the TRD domain of the 5mC reading protein MECP2. These experiments revealed that through its binding, MECP2 allosterically stabilizes the autoinhibitory conformation of DNMT3A, resulting in a strong inhibition of enzymatic activity in vitro. The interaction between these two proteins and its associated inhibition could be disrupted by unmodified histone H3. In my work, I further validated the interaction between the ADD and the TRD domains by size exclusion chromatography. Also, by generating cell lines with stable over-expression of MECP2, I could show that MECP2 inhibits DNMT3A activity in cells. Together, the data from this study offer unprecedented insights into the regulation of DNMT3A by the combined action of chromatin modifications and interaction partners. Accordingly, depending on the modification status of the H3 tail at the target site, MECP2 can act as either a repressor or activator of DNA methylation. The last project dealt with the coevolution between DNA methylation and DNA repair systems, a very exciting topic that was addressed in close collaboration with the laboratory of Dr. Peter Sarkies (MRC London). By performing in vitro methylation experiments with the catalytic domain of DNMT3A, I could show that, in addition to 5mC, DNMT3A can also introduce 3mC, a modification which represents an alkylation damage of DNA. This study provides a new evolutionary perspective on the loss of DNA methylation that is observed in many species.