Figure 1. General organization of human centromeric sequences. The centromere is identified as the primary constriction on the chromosome flanked by pericentric heterochromatin. A closer look at the centromere shows that there are two types of -satellite repeats found at and near human centromeres: monomeric -satellite repeats (left) and higher-order -satellite arrays consisting of several monomers repeated as a multimeric unit (right). Monomeric repeats are found in ‘domains’ in the pericentric regions and display less homogeneity in sequence identity than higher-order -satellite arrays at the centromere. Frequent insertions of mobile elements disrupt monomeric -satellite domains while higher-order -satellite arrays can span megabases of DNA largely uninterrupted.

Figure 2. Higher-order -satellite arrays host both heterochromatin and centromeric domains. The centromere is packaged into a discontinuous array of nucleosomes containing CENP-A or canonical histone H3 flanked on either side by heterochromatin. The localization of ‘foundation’ proteins is shown enlarged in a close-up view. CENP-B associates throughout the array bound to its CENP-B box, while CENP-A and CENP-C are found only at a subset of the region bound by CENP-B. Centromeric regions packaged into H3-containing nucleosomes show marks associated with ‘active’ chromatin typical of euchromatin (K4-Me). In contrast, neighbouring heterochromatin is packaged into H3-containing nucleosomes with a different mark (K9-Me) and localize the prototypical heterochromatin protein, HP1.

Figure 3. ‘Centromere-drive’ model. The sex chromosomes provide an exaggerated example of centromere-drive since the X and Y chromosomes do not recombine and the satellites populating each respective centromere are different (shown as different colours). Step 1, the centromere variant has a selective advantage by virtue of its ability to attract more microtubules and gain a favourable position in the asymmetric female meiosis. However, this centromere variant results in unequal tension across the centromeres of the paired sex chromosomes in male meiosis. Unequal tension can potentially increase the rate of nondysjunction in males, resulting in sterility. Step 2, sterility effects in the male will provide strong selection on any allele that restores meiotic parity and male fertility. Repeated bouts of drive and suppression will be observed as positive selection among centromere or heterochromatin proteins. Shown here is an example of selection favouring alleles that no longer recognize the ‘old’ centromeric satellite (on the X) resulting in constriction of the centromere to only the new satellite variant and restoring relatively equal microtubule binding (while the Y centromeric satellite still retains its binding ability).