Development of a Protein-Based Sensor for the Direct Detection of DNA Methylation

Abstract

Epigenetic phenomena are essential for the regulation of gene expression and consequently, cell fate. At the forefront is the methylation of DNA at cytosine in CpG nucleotides, which causes histone remodeling and eventually leads to gene repression. This is particularly important in cancer, as DNA methylation is used to repress tumor suppressor genes. Our purpose is to develop a rapid and direct approach for detecting DNA methylation to enhance DNA diagnostic systems. We began by developing a protein-based sensor that utilizes the capabilities of split-protein reassembly, tethering target-detecting domains to different halves of split-firefly luciferase. By translating these fusion proteins in vitro, we were able to rapidly profile a panel of proteins known to recognize methyl-CpG sites and found the protein MBD1 to have over 90-fold preference for methylated DNA over unmethylated DNA. The next-best protein, MBD2, demonstrated below 30-fold preference. We chose to further study MBD1, using our system to execute a complete alanine scan of the 69-residue domain. We identified five loss-of-function alanine mutations that were consistent with previous literature and also found an additional seven residues that are necessary for recognition of methylated cytosine. We also carried out a small-scale alanine scan of MBD2 and found similar results, further supporting the validity of in vitro alanine scanning. Finally, using the results from the alanine scanning experiments, we are proceeding towards the development of a library of 3×10⁷ MBD constructs to which we intend to apply a selection process via phage display to identify improvements in binding affinity, allowing us to engineer new MBD variants for cancer diagnostics.