Cells don't load new CENP-A (green) until they finish mitosis.
Centromeres serve as platforms for the kinetochores that are essential for divvying up chromosomes during mitosis. CENP-A's presence at centromeres, where it replaces the standard-issue H3 histone, poses several puzzles. On page Jansen et al. answer the “when” question. The researchers applied a recently developed technique called SNAP tagging, which affixes a long-lasting label to a protein. The method allowed the team to pin down CENP-A's location and measure its turnover.
The logical time for cells to begin inserting fresh CENP-A into chromosomes is after S phase, to counter the dilution of the protein during DNA duplication. But what the researchers found was one of those surprises “that makes you jump to the roof,” says lead author Lars Jansen. The addition doesn't begin until G1, after cells have gone through mitosis. Further experiments indicated that cells have to pass through mitosis before they can start incorporating new CENP-A. Why mitosis is essential for this process isn't certain. But the results reveal that chromosomes complete division with a reduced complement of CENP-A.
Maddox et al. tackle the “how” question on page . A previous screen of nematode genes had identified ∼250 that are necessary for chromosome separation. By eliminating the corresponding proteins individually, the researchers pinpointed one that was crucial for kinetochore construction: KNL-2. Without it, CENP-A does not join the chromosome. And KNL-2 stays away if the histone is absent, suggesting that their interaction directs CENP-A to the correct location.
The researchers also found that the human version of KNL-2 homes in on the centromeres late in mitosis and early in G1, which jibes with the findings from Jansen et al. KNL-2 carries a DNA-binding Myb domain that might provide clues to another mystery about CENP-A: how the protein determines where on the chromosome to settle. 
