Detection of DNA Cytosine Methylation with Biological Nanopore Sensor

JOANNA SOYRING, CHRISTOPHER J. MORAN, MELISSA D'ALIA, ELIJAH A. JORDAN, TRANG A. VU

Cancer is the second leading cause of death in the US, estimating for 1660 deaths a day in 2019, but the cancer mortality rates can be reduced by early stage detection. Using a biological nanopore, we develop a nanoscale biosensor to investigate DNA methylation, which is a prominent biomarker for cancer screening at an early stage, risk assessment, and personalized medicine. DNA methylation, the addition of a methyl group to a cytosine, is a wellknown covalent epigenetic modification that takes place in human cells. Aberrant hypermethylation can lead to inhibition of promoter activity and prevent the binding of transcription factors, repressing gene expression. Aberrant DNA methylation has been dubbed “a hallmark” of cancer due to decreased expression of genes involved in cell cycle inhibition. Nanopore technology offers a wide range of applications in single-molecule assay and analysis with high throughput and high sensitivity. More importantly, the biological nanopore-based sensor can work with a small sample volume, allowing the extracted gene promoter methylations to be directly and accurately detected without amplification. A biased external voltage across the nanopore drives ion flow, resulting in ionic current through the nanopore. Transports of biomolecules (such as DNA and RNA) through the pore produce an ionic current blockage signature that can be identified and quantified and, thus, used to characterize the molecules. In the current project, we utilized a biological nanopore sensor to investigate among unmethylated-, methylated- and methylated-DNA bound with MBD2 protein. The MBD2 protein bound to methyl sites on the DNA can enhance the detection of the nanopore, changing the characteristics of the current blockage and allowing for better distinction of the DNA methylation. An 85% current blockage is characterized as a “level 4 blockage,” and signifies the unzipped dsDNA region translocating through the pore. Significant differences in the current blockages were seen at this level between the methylated and the unmethylated DNA as well as the MBD2 bound and unmethylated DNA. This is evidence in support of the ability of biological nanopores to detect differences in the characteristics of the DNA. Together with our collaborators, we are exploring the possibility of using Machine Learning algorithms, especially Supervised and Deep Neural Network, in obtaining an accurate classification of different DNA methylation statuses.