Centromeres in nuclear architecture

The centromere is a unique region of the chromosome that defines the site of kinetochore assembly and is essential for the faithful segregation of the sister chromatids into the 2 daughter cells. While the kinetochore only forms as a transient structure during mitosis, the centromere persists as a stable domain throughout the cell cycle, bound by proteins of the constitutively centromere associated network (CCAN).¹ Interphase centromeres have been described to cluster together and associate with the periphery of the nucleolus depending on cell cycle stage, cell type, and organism.^2,3

Analyzing the localization of centromeres during the cell cycle in Drosophila cells by visualizing Cenp-A^CID, the histone H3 variant that defines centromere identity,⁴ we discovered that centromeres function as an important anchor for higher-order chromatin structure in interphase. Following a candidate approach, we were further able to describe a network of proteins that mediate the localization of centromeres in the nucleus during interphase.⁵

About seven years ago, the nucleolus protein Nucleophosmin 1 (NPM1) emerged as a promising candidate for centromere anchoring, when it was identified as a direct interactor of Cenp-A with unknown function.⁶ In our recent work we identified a member of the Nucleophosmin/Nucleoplasmin protein family in Drosophila melanogaster, the Nucleoplasmin like protein (NLP), as an important regulator of subnuclear centromere organization during interphase. It does so by interacting with Cenp-A^CID, the insulator protein CTCF and the nucleolin homolog Modulo.

In Drosophila Schneider S2 cells and larval hemocytes centromeres are clustered at the periphery of the nucleolus. We observed that during interphase these clusters co-localize with NLP, CTCF, and the periphery of Modulo. Co-IP experiments confirmed the interaction between the 3 proteins and Cenp-A^CID. Knockdown experiments and fly mutants showed that loss of any of the three proteins leads to a disruption of centromere clustering and detachment of the centromeres from the periphery of the nucleolus. To dissect the individual functions and the interdependencies, we asked whether tethering of the individual proteins via the LacI/LacO system is sufficient to facilitate either clustering and/or nucleolus association of a plasmid carrying a LacO array. While NLP and CTCF together appear to be sufficient to induce clustering, only Modulo was able to mediate tethering of the plasmids to the nucleolus. Cenp-A^CID itself was also able to recruit NLP and CTCF and thereby induce clustering of the LacO containing plasmid. Clustering is therefore an intrinsic property of centromeric chromatin and not dependent on a specific DNA sequence. Notably, the clusters induced by tethering of Cenp-A^CID dispersed during mitosis, while tethering of NLP resulted in the stable formation of plasmid clusters throughout the cell cycle (unpublished observation). This indicates that the release of NLP from chromatin before mitotic entry might be one mechanism by which the cell ensures that all individual centromeres are accessible for the formation of kinetochores and the attachment to the mitotic spindle.

While intact heterochromatin is dispensable for centromere clustering, the declustering of centromeres resulted in a disruption of the spatial organization of heterochromatin. Concomitantly, this also led to a partial loss of silencing, suggesting that the cell tethers centromeres to the nucleolus to help heterochromatin-mediated transcriptional repression. This could be achieved either through spatial proximity to the nucleolus and/or the accumulation of repressive heterochromatin factors at the repetitive DNA elements underlying heterchromatin. Interestingly, declustering of centromeres results in the formation of several individual heterochromatin domains of smaller size, but not in their complete dispersion. Clustering of centromeres therefore seems to be only one of the pathways contributing to the higher-order organization of constitutive heterochromatin. Alternative pathways could be linked to the methylation status of histone H3 at lysine 9 (H3K9me1/2/3), which might be directly recognized by components of other nuclear sub-compartments, such as the nuclear periphery.^7,8

In summary, our work identifies NLP as a new nuclear organizer, which mediates chromatin interactions in trans. We would like to propose that clustering and tethering of specialized chromatin domains (e.g., centromeres, telomeres) serves as a mechanism to spatially organize associated heterochromatin domains, allowing proper silencing of transposons and protection against recombination of repetitive DNA elements. (Fig. 1)

Figure 1. The clustering of centromeres (green circles) facilitates the stable organization of pericentric heterochromatin (red) around the nucleolus (light green). NLP and CTCF link centromeres by interacting with the centromeric histone H3-variant CENP-A, while Modulo mediates the tethering to the nucleolus.

Padeken J, et al. Mol Cell. 2013;50:236–49. doi: 10.1016/j.molcel.2013.03.002.

10.4161/cc.26697