DNMT3B maintains mCA landscape and regulates mCG status of bivalent promoters in human embryonic stem cells. DNMT3B maintains mCA landscape and regulates mCG status of bivalent promoters in human embryonic stem cells

In mammalian cells, DNMT3B is known as a de novo DNA methyltransferase. However, its preferential target sites for DNA methylation are largely unknown. By studying CA methylation (mCA) and various histone mark distributions in human embryonic stem cells (hESC), we set up a connection between mCA, H3K36me3, and DNMT3B. We found that mCA, H3K36me3 and DNMT3B signals in hESC are distributed in a 3-level pattern: low level at promoter region, intermediate level before first splicing junction and high level afterward. Knocking out DNMT3B (KO) in hESC demolished mCA, and this further confirmed that DNMT3B is the main enzyme which maintains mCA. Furthermore, the mCA could not be maintained after we deleted the H3K36me3 binding domain, PWWP domain, in DNMT3B. This observation suggests that DNMT3B maintains mCA through its histone interaction. This is the first study that describes the detailed landscape of mCA and its connection with DNMT3B. In contrast to mCA, we only observed a minor reduction of global mCG level from 83.7% to 79.2% after DNMT3B KO. 5256 de novo hypomethylated regions (dnHMRs) were found in the KO cell, in which 57.8% overlapped annotated promoter sites. Intriguingly, most of these promoter dnHMRs were bivalent, possessing both H3K4me3 and H3K27me3. We call them spurious bivalent promoters. Gene ontology (GO) analysis associated spurious bivalent promoters with development and cell differentiation. Overall, we found the importance of DNMT3B for shaping mCA landscape and maintaining the fidelity of bivalent promoter landscape in hESC. Overall design: DNMT3B is either knock-out or mutated in H1 hESC by CRISPR-Cas9 gene editing strategy. DNA methylation were profiled through WGBS. DNMT3B binding and histone mark status were profiled by ChIP-seq.

在哺乳动物细胞中，DNMT3B被公认为从头DNA甲基转移酶（de novo DNA methyltransferase），但其介导DNA甲基化的偏好性靶位点在很大程度上仍未明确。本研究通过对人类胚胎干细胞（human embryonic stem cells, hESC）中的CA甲基化（mCA）以及多种组蛋白修饰分布进行分析，建立了mCA、H3K36me3与DNMT3B之间的关联。研究发现，hESC中的mCA、H3K36me3与DNMT3B信号呈现三级分布模式：在启动子区域处于低水平，在第一个剪接接点之前处于中等水平，而在剪接接点之后则处于高水平。在hESC中敲除DNMT3B（KO）后，mCA信号完全消失，这进一步证实DNMT3B是维持mCA水平的核心酶类。此外，若删除DNMT3B中的H3K36me3结合结构域——PWWP结构域，则无法维持mCA水平。上述结果表明，DNMT3B通过与组蛋白相互作用来维持mCA水平。本研究是首个阐明mCA详细分布图谱及其与DNMT3B关联的研究。与mCA不同，DNMT3B KO后，全基因组mCG水平仅从83.7%小幅下降至79.2%。在KO细胞中共鉴定到5256个从头低甲基化区域（dnHMRs, de novo hypomethylated regions），其中57.8%与注释的启动子区域重叠。值得注意的是，这些启动子区域的dnHMRs大多为二价结构域，同时带有H3K4me3与H3K27me3两种组蛋白修饰，我们将其称为“伪二价启动子”。基因本体（Gene Ontology, GO）富集分析显示，伪二价启动子与发育及细胞分化过程密切相关。综上，本研究证实了DNMT3B在塑造hESC中mCA分布图谱以及维持二价启动子分布图谱保真度方面的重要作用。实验设计：通过CRISPR-Cas9基因编辑技术，在H1 hESC中构建DNMT3B敲除或突变细胞系。采用全基因组亚硫酸氢盐测序（WGBS, whole genome bisulfite sequencing）分析DNA甲基化水平，通过染色质免疫共沉淀测序（ChIP-seq, chromatin immunoprecipitation sequencing）检测DNMT3B结合情况与组蛋白修饰状态。