Isolation of Human Genomic DNA Sequences with Expanded Nucleobase Selectivity

We report the direct isolation of user-defined DNA sequences from the human genome with programmable selectivity for both canonical and epigenetic nucleobases. This is enabled by the use of engineered transcription activator -like effectors (TALEs) as DNA major groove-binding probes in affinity enrichment. The approach provides the direct quantification of 5-methylcytosine (SmC) levels at single genomic nucleotide positions in a strand-specific manner. We demonstrate the simple, multiplexed typing of a variety of epigenetic cancer biomarker SmC with custom TALE mixes. Compared to antibodies as the most widely used affinity probes for SmC analysis, i.e., employed in the methylated DNA immunoprecipitation (MeDIP) protocol, TALEs provide superior sensitivity, resolution and technical ease. We engineer a range of size-reduced TALE repeats and establish full selectivity profiles for their binding to all five human cytosine nucleobases. These provide insights into their nucleobase recognition mechanisms and reveal the ability of TALEs to isolate genomic target sequences with selectivity for single 5hydroxymethylcytosine and, in combination with sodium borohydride reduction, single 5-formylc-ytosine nucleobases.