Non-canonical DNA structures at eukaryotic centromeres can resolve the CENP-B paradox

Epigenetic mechanisms are thought to specify the identity of centromeres, yet the degree to which DNA sequence contributes to centromere determination remains unresolved. For example, the conserved CENP-B protein binds the 17-bp “CENP-B box” motif found in α-satellite arrays at most human centromeres. Although CENP-B is required for artificial centromere function, it is non-essential and some species lack CENP-B boxes entirely. To address this “CENP-B paradox,” we determined the distribution of CENP-B boxes in primates and directly demonstrated the absence of CENP-B at Old World Monkey centromeres. We found extensive interspecific variation in α-satellite arrays including abundance of <10-bp dyad symmetries and of non-B-form DNA structures. We confirmed the presence of non-canonical DNA at human centromeres and neocentromeres and detected similar structures at mouse, chicken and yeast centromeres. To resolve the CENP-B paradox, we propose that CENP-B enhances non-canonical DNA formation, thus providing a general basis for centromere specification. We used Cleavage under targets and Release using nuclease (Cut-and-Run), a chromatin profiling strategy in which antibody-targeted controlled cleavage by micrococcal nuclease releases specific protein-DNA complexes into the supernatant for paired-end DNA sequencing.