Project Summary/Abstract The mammalian de novo DNA methyltransferase, DNMT3A is essential for postnatal development of the brain, control of body size and for the regulation of hematopoiesis. Mutations in the gene are commonly found in age- associated clonal hematopoiesis of indeterminant potential and are drivers for certain leukemias making it important that we understand how the enzyme functions in living cells. All human cancers contain DNA methylation anomalies, however mutations in DNMT3A are relatively rare in solid tumors, suggesting that altered enzyme regulation might be responsible for these changes. Although we know in exquisite detail how DNMT3A methylates naked DNA, we know virtually nothing about how this occurs in the context of nucleosomes, the fundamental building blocks of chromatin. The focus of this application is to address this issue using new insights recently developed in the lab. We will concentrate on the role of a truncated isoform, DNMT3A2, normally expressed during embryonic development but overexpressed in most solid cancers. Our novel cryo-EM structure shows how this enzyme partners with an accessory protein DNMT3B3 to bind to nucleosomes through the acidic patch – a totally unexpected discovery. We will follow up on this work using biochemical, cryo-EM, cellular and mouse studies to gain a more precise understanding of how DNA methylation works in the context of chromatin. A mouse model in which Dnmt3a2, but not the more widely studied longer isoform, Dnmt3a1, has been knocked out will be used to test for causality in the process of immortalization and carcinogenesis in genetic and chemical mouse models of cancer. This comprehensive approach will help in our understanding of how this fundamental process works in normal and transformed cells. Studies on how DNA methylation marks are interpreted in cells has long been confined to comparison between fully and completely unmethylated CpG dyad states. Because of the flexibility in the R35 funding mechanism, we serendipitously discovered that CpG hemimethylation can either stimulate or inhibit binding by CTCF depending on which duplex strand is methylated. We will investigate the prevalence of hemimethylation in cancer cells and investigate its effects on binding of other transcription factors and chromatin structure. Differential binding may play a role in asymmetric cell divisions in cancer cells. We hope to answer three critical questions: 1) Does DNMT3A2 overexpression contribute to carcinogenesis? 2) How does de novo methylation occur in a nucleosomal context? 3) What are the potential biological roles for hemimethylation?