This result suggests a reversal of the conventional loading hierarchy

e established epigenetically by a pair of inner-centromerespecific nucleosome modifications, phosphorylation of histone H3 threonine 3, and histone H2A threonine PubMed ID:http://www.ncbi.nlm.nih.gov/pubmed/19836572 120, both of which are required for Aurora BCPC recruitment to inner centromeres. First, the BIR domain of Survivin interacts with the ART motif at the very N BIRB796 site terminus of H3 specifically when T3 is phosphorylated. Second, Shugoshin binds to H2A T120-Phos and then directly binds to Borealin. The histone marks themselves therefore coexist at, and in part establish the epigenetic identity of, the inner centromere, and it is the intersection of these nucleosomes that form the Aurora BCPC binding sites where Aurora B functions in tension sensing. H2A T120-Phos is generated by Bub1 kinase, which has a well-understood mechanism of innercentromere association. In contrast, H3 T3-Phos is generated by Haspin kinase, but the mechanism of Haspin recruitment to inner centromeres is not known, and therefore, our understanding of the molecular requirements for tension checkpoint activation is incomplete. There is evidence that Topo II may play some role in Aurora B activation, suggesting a possible function of Topo II at the inner centromere of kinetochores, but there is no mechanistic insight into this possible link. Here, we determine a novel inner-centromere function of the Topo II CTD. In budding yeast mitosis, we find that both the CTD and the lysine residues that become SUMOylated are essential for the recruitment of Aurora B/Ipl1 to inner centromeres. This mechanism also requires Haspin kinases and histone H3 threonine 3 phosphorylation. This Topo IIdependent pathway and the Sgo1-dependent pathway are each important for Ipl1 recruitment to inner centromeres in yeast mitosis. Therefore, both H2A and H3 tails work in concert to form the inner-centromerespecific nucleosomal binding site for Ipl1. Moreover, the CTD of Topo II has a noncatalytic inner-centromere function in mitosis, which can explain the importance of the CTD for faithful chromosome segregation, distinct from the catalytic activity of Topo II. 652 JCB Volume 213 NumBer 6 2016 Results Yeast kinetochores assemble properly in top2 mutants In mitosis, Topo II is most abundant at kinetochores, but no function within this complex has been revealed. A previous study of yeast Topo II mutants analyzed the attachment of a single chromosome to the mitotic spindle in prometaphase and found no gross defect in biorientation. To extend this analysis, we asked if kinetochores assemble correctly in loss-of-function top2-4 mutants. We reasoned that proper localization of an outer-kinetochore protein, Nuf2-GFP, serves to indicate assembly of the underlying inner-centromere chromatin and central regions of the kinetochore. To identify mitotic cells, we simultaneously observed Spc110-mCherry, a component of spindle pole bodies. In both wild type and top2-4, grown at the nonpermissive temperature for top2-4, Nuf2-GFP localized to the clustered kinetochores in almost all metaphase cells, each cluster of sister kinetochores being in line with the spindle axis. Thus, there is no gross defect in kinetochore structure, and thereby the inner centromere, in the absence of Topo II catalytic activity. Second, we assessed a functional aspect of kinetochore integrity by asking if the SAC protein Mad2 can be recruited to kinetochores in top2-4 cells. In untreated top2-4 metaphase cells, we observed the expected localization of Mad2-GFP at nuclear pore compl