Sequence-specific DNA binders for nucleotide resolution analysis genomic 5-methylcytosine by cell imaging

Abstract

5-methylcytosine (5mC) is a fundamental epigenetic modification in mammalian genomes involved in development, cell differentiation and genomic imprinting. In addition, aberrant DNA methylation patterns are responsible for the pathogenesis of many diseases including neurodegenerative disorders, cardiovascular affections and cancer. In this thesis, we describe the development of a novel method for image-based analysis of 5mC using pairs of fluorescent Transcription Activator Like-Effectors (TALEs). These DNA binding proteins can recognize specific sequences of canonical or epigenetically modified DNA via modular repeats that interact with nucleobases in a one-to-one correspondence. We employed fluorescent TALE pairs that differ only in the repeat responsible for recognizing cytosine (C) at CpG dinucleotides and the fluorophore fused to them (either eGFP or mCherry). By using the 5mC selective repeat HD in one of the TALEs, we can detect differences in methylation level, while the universal binder repeat G* in the other TALE is not responsive to 5mC and allows to detect local changes in chromatin compaction. This way it is possible to analyze 5mC independently of potential differences in target accessibility. We applied our method using recombinantly expressed and purified TALE pairs in cellular stains to image SatIII DNA. This pericentromeric DNA is the origin of nuclear stress bodies (nSBs), exhibits aberrant methylation in several cancers and remains challenging to study by conventional methods due to its highly repetitive nature. We proved the applicability of our method to study 5mC differences in user-defined repetitive sequences with single nucleotide and strand resolution. Furthermore, we correlated the methylation status of SatIII with the presence of heat shock factor 1 (HSF1) at its recognition sequence after stress, revealing a role for 5mC in HSF1 recruitment as initial step of nSB formation in a subpopulation of cells. Finally, we constructed and screened a library of size-reduced TALE repeats to identify potential 5mC binders. We found that RVD NH* binds selectively to 5mC, but not C and its application in combination with HD TALEs allows for improved imaging with higher dynamic range. These studies offer a promising imaging tool for studying 5mC function in repetitive sequences and its interplay with other imageable chromatin-interacting proteins with nucleotide, strand, locus and cell resolution.