DNA methylation in chromatin - complexes and mechanisms

The DNA in the eukaryotic genome is packaged into chromatin whose basic repeating unit is the nucleosome consisting of 147bp of DNA wrapped around an octameric histone core. The histone octamer is composed of two copies of each of the four histone proteins H2A, H2B, H3 and H4. DNA methylation is an epigenetic mechanism that is involved in various important processes in the cell such as differentiation and proliferation, transcriptional regulation, genomic imprinting, X-chromosome inactivation, silencing of repetitive elements, maintenance of genomic stability and DNA repair. In the mammalian genome, DNA methylation occurs almost exclusively in the context of CpG di-nucleotides and is brought about by three DNA cytosine-5-methyltransferases that use S�adenosyl-methionine (SAM) as methyl-group donor. Due to Dnmt1�s preferential activity towards hemi-methylated DNA and the fact that it restores methylation patterns on the newly synthesized daughter strand during replication, it is referred to as the maintenance DNA methyltransferase, whereas the de novo DNA methyltransferases Dnmt3a and Dnmt3b introduce new methylation marks in the genome. Dnmt3L is structurally related but catalytically inactive and serves as a cofactor for Dnmt3a and Dnmt3b. Importantly, DNA methylation is indispensible for mammalian embryogenesis and aberrant DNA methylation is often found concomitant with cancer. Dnmt1 was shown to be implicated in the formation of tumors, however the underlying mechanisms especially the question of Dnmt1 targeting remain largely unknown. The goal of this thesis was to identify new Dnmt1 complexes from native tissues that would allow comparison of complex composition in tumors. ICBP90 (UHRF1) and USP7 were identified as interacting proteins from co-immunoprecipitation experiments. During the course of this study, UHRF1 was described not only to interact with Dnmt1 but to recruit Dnmt1 to replication foci during late S-phase. In vivo and in vitro immunoprecipitations revealed different possible complexes, namely Dnmt1/ICBP90, Dnmt1/USP7 and ICBP90/USP7. Furthermore, a possible trimeric complex of USP7 binding with two different domains to both Dnmt1 and ICBP90 was established. Notably, chromatin immunoprecipitation demonstrated the existence of different Dnmt1/ICBP90/USP7 complexes at four different loci in vivo, however the function in chromatin related processes awaits further investigation. Interestingly, ICBP90 and USP7 are endowed with antagonistic enzymatic activities. ICBP90 exhibits autoubiquitinylation and ubiquitinylation activity towards histone H3, and USP7 in vitro. On the contrary, USP7 was able to target ICBP90 and histones H2A, H2B and H3 for deubiquitination in vitro whereas global levels of ubiquitinylated H2A and H2B were not changed upon knockdown or over-expression of USP7. Binding of ICBP90 and Dnmt1 to USP7 did not influence the in vitro activity of USP7. Moreover, Dnmt1 was ubiquitinylated by ICBP90 in vitro, but Dnmt1 protein or ubiquitinylation levels were not affected by USP7 over-expression or knockdown in vivo. Future research will focus on the role of histone ubiquitinylation/deubiquitination in transcriptional repression and silencing of repetitive elements in heterochromatin. In another project, the properties of Dnmt binding to DNA and mono-nucleosomes and the principle mechanisms of DNA methylation in chromatin by the de novo DNA methyltransferases were in focus. It could be shown that Dnmt3a and Dnmt3b2 stably associated with DNA >35bp in length, albeit longer DNA fragments were preferred indicative of a cooperative binding process. Furthermore, binding to DNA or mono-nucleosomes was highly dynamic and the interaction of mono-nucleosomes with the de novo DNA methyltransferases did not disrupt mono-nucleosomes. Dnmt3a generally bound comparably well to DNA and mono-nucleosomes with different DNA linker length whereas Dnmt3b2 preferentially bound to free DNA and mono-nucleosomes with long linker DNA. In vitro DNA methylation assays, either performed with radioactive SAM or non-radioactive one, but following single-molecule analysis with bisulfite treatment clearly demonstrated that Dnmt3a and Dnmt3b2 could not methylate DNA within the nucleosome but only linker DNA. This indicated that the DNA strand facing opposite the histone octamer did not represent a target for methylation and that nucleosomes constitute a major restriction for DNA methylation. Further experiments will address the role of chromatin remodeling enzymes in this process. Dnmt3L, a stimulatory factor for Dnmt3a and Dnmt3b, was shown to bind to non-methylated H3K4. Therefore, the effect of Dnmt3L on binding to DNA and nucleosomes by Dnmt3a and Dnmt3b was analyzed. Dnmt3L itself neither bound to DNA nor to mono-nucleosomes in EMSA experiments. Addition of Dnmt3L to Dnmt3a and Dnmt3b enhanced DNA binding and modified the binding behavior towards nucleosomes. Interestingly, recombinant Dnmt3L was tightly associated with nucleosomes when purified from Sf21 insect cells. Sucrose density gradient analysis confirmed this observation as Dnmt3L was distributed over the whole gradient with nucleosomal species of different weight. However, when endogenous nucleosomes were substituted for nucleosomal templates of various sizes or �naked� DNA Dnmt3L entered the gradient by 1/3rd, although the peak fractions migrated at higher densities. To unravel the reasons for the stable association of Dnmt3L with endogenous nucleosomes, future work will concentrate on the identification of possible loading factors, specific posttranslational histone modifications and nucleosomal architecture