Establishment Of A New Pipeline For RRBS Data Ananlysis And Application Of Identification And Characterization Of Allele Specific Methylations In Cancer Cells

DNA methylation is one of the important epigenetic modifications. It plays an important role in embryonic development, cell differentiation, X chromosome inactivation, Genomic imprinting and tumorigenesis. The alteration of DNA methylation pattern was belived to be one of the most important characteristics of genome of cancer cells. Global DNA hypomethylation and site-specific hypermethylation was found in many cancer cells. However, the molecular mechanism of the regulation of DNA methylation is unclear. Reduce representation bisulfite sequencing(RRBS) technology was invented to analyze the whole genome level of DNA methylation in recent years. Because of the complexity of the data analysis of RRBS, the existing software can not fully meet the needs of the researches. The funcation of the de_novo DNA methyltransferase DNMT3A in the abnormal DNA methylome in tumor cells is not clear yet. Firstly we applied RRBS to detect the changes of DNA methylome after over-expressing DNMT3A in lung cancer cell A549. By improved data analysis process, we found that the whole genome DNA methylation was increased by 9.2% when we over-expressing DNMT3A. And the increased level of methylation in different region had no significant difference. Furthermore, these elevation of DNA methylation were mostly taken place at the partialy methylated CpG sites,but not de_novo sites. In addition, the methylation level of repetitive sequence was also elevated, in a similar level of the increase of whole genome DNA methylation. We also found that overexpression of DNMT3A can lead to the denovo hypermethylation of some gene promoters, including RFPL3, TFAP2E etc, which may participate in inhibiting the proliferation of tumor cells. The above results showed that DNMT3A was involved in the maintenance process of DNA methylation in tumor cells, and could regulate the methylation level of some tumor related gene promoter region. Next, we developed a new process to include analyses of continuous CpG methylation linkage in a single molecule, the linkage analysis of DNA methylation and single nucleotide polymorphism, and the correlation between DNA methylation and gene expression. By using the process, we found that there were widespread allele specific DNA methylation(ASM) in tumor cell lines. In order to reveal the characteristics of the ASM, we analyzed the RRBS data of 3 tumor cell lines as well as three human placenta RRBS data downloaded from public database. The results showed that besides known imprinted differentially methylated region(DMR) and X chromosome, more ASMs were located on other genomic region. Unlike the ASMs in the placenta, the ASMs in tumor cells had few overlap, which indicated a minor functional requirement of these ASMs during tumorigenesis. In order to further analyze the molecular mechanism of the genesis and maintenance of ASMs, we constructed some cell lines with modified DNA methylation regulating proteins, including over-expressing UHRF1, knocking down UHRF1, over-expressing DNMT3A, as well as over-expressing the mutant of DNMT3A. By analysing RRBS data, we found that UHRF1 was required for the maintenance of the ASMs, but the generation of new ASMs is mediated by DNMT3A, which also regulated by the combination of DNMT3A and unmodified histon H3K4. The number of ASMs was significantly increased when over-expressing DNMT3A mutants that deliminated the combination of H3K4 and DNMT3A. We speculated that the genesis of ASMs depended on de_novo DNA methylationit may related to the de_novo DNA methylation. Finally, we analyzed the RNAseq of tumor cells and found it had no obvious correlation between the methylation of ASMs and gene expression level. All above results, revealed that the ASMs were regulated by the de_novo and maintenance DNA methylation system in tumor cells. These ASMs had no obvious regulatory function of gene expression. And the ASM locations with genome differed a lot across different cancer cell types, which suggested that these ASMs might be a feature of abnormalty of DNA methylation regulation in cancer cells. Further analyses in clinical tumor samples are required to reveal the function of these ASMs. In conclusion, this study has developed a new RRBS data analysis process. By using this process, we found that DNMT3A was involved in the regulation of abnormal DNA methylation. Furthormore, a lot of ASMs in tumor cells was identified and preliminarily characterized. The improved RRBS analysis process may be applied to more DNA methylation regulatory mechanism researches. Our results in this study could help for further investigations of molecular mechanism of regulation of DNA methylome in tumor cells.