Hacking the centromere chromatin code: dissecting the epigenetic regulation of centromere identity

Abstract

The centromere is a specialized chromatin domain that serves as the assembly site for the mitotic kinetochore structure, thereby playing a fundamental role in facilitating the maintenance of the genetic information. A histone H3 variant termed CENP-A is specifically found at all active centromeres. Beyond this, however, little is known about how and to which extent the chromatin environment of centromeres modulates and contributes towards centromere identity and function. Here, I have employed a novel Human Artificial Chromosome (HAC), the centromere of which can be targeted by fusions to the tet repressor, to characterize the chromatin environment underlying active kinetochores, as well as to specifically probe the role of this environment in the maintenance of kinetochore structure and function. My data demonstrate that centromeric chromatin resembles the downstream regions of actively transcribed genes. This includes the previously unrecognized presence of histone H3 nucleosomes methylated at lysine 36 within the chromatin underlying functional kinetochores. Targeted manipulation of this chromatin through tethering of a heterochromatin-seeding transcriptional repressor results in the inactivation of HAC kinetochore function concomitant with a hierarchical disassembly of the structure. Through an even more selective engineering of the HAC centromere chromatin, I have provided evidence supporting a critical role for nucleosomes dimethylated at lysine 4 on histone H3 in facilitating local transcription of the underlying DNA. Tethering of different chromatin-modifying activities into the HAC kinetochore collectively reveals a critical role for both, histone H3 dimethylated on lysine 4 and low-level, non-coding transcription in the maintenance of the CENP-A chromatin domain. On one hand, repression of centromeric transcription negatively correlates with the maintenance of CENP-A and ultimately results in the loss of kinetochore function. On the other hand, increasing kinetochore-associated RNA polymerase activity to within physiological levels for euchromatin is associated with rapid loss of CENP-A from the HAC centromere. Together, my data point towards the requirement for a delicate balance of transcriptional activity that is required to shape and maintain the chromatin environment of active centromeres.