Credit: STOCKBYTE

Accurate chromosome segregation during mitosis depends on the correct attachment of kinetochores to spindle microtubules. Kinetochores often first bind to the side of microtubules and then establish end-bound attachments, but how exactly this transition is regulated is unclear. Cheerambathur et al. now find that crosstalk between two complexes involved in kinetochore–microtubule attachments, NDC-80 (nuclear division cycle 80) and RZZ (ROD–ZW10–Zwilch), mediates this process. The authors visualized chromosome dynamics during the first division of Caenorhabditis elegans embryos to assess the formation of end-bound microtubule attachments. RZZ normally recruits dynein to kinetochores via Spindly, and together these proteins mediate initial lateral microtubule capture. The authors found that co-depletion of RZZ and Spindly allowed the formation of end-coupled attachments with only a slight delay, similar to that seen in the absence of dynein (which accelerates microtubule capture). However, if only Spindly was removed (and consequently dynein), end-bound attachments and chromosome segregation were inhibited, which suggests that RZZ inhibits the formation of end-bound attachments mediated by NDC-80.

an Aurora B-independent mechanism to control kinetochore–microtubule attachments

By performing two-hybrid analysis with truncated versions of NDC-80 and ROD-1, and partially reconstituted RZZ complexes, Cheerambathur et al. found that RZZ directly interacts with NDC-80 via the ROD-1 amino-terminal domain and the NDC-80 N-terminal basic tail. The NDC-80 basic tail is normally phosphorylated by Aurora B, and this inhibits the interaction of NDC-80 with microtubules (which involves an adjacent CH domain). Here, the authors found that ROD-1 binding to the NDC-80 tail suppressed NDC-80 binding to microtubules in vitro independently of Aurora B.

To investigate the mechanisms underlying RZZ-dependent NDC-80 inhibition in vivo, the authors used mutant versions of NDC-80, either lacking the basic tail or bearing a mutated tail that mimics Aurora B-dependent phosphorylation (that is, a version that fails to bind to microtubules in vitro). NDC-80 that lacks the tail was resistant to persistent RZZ-mediated inhibition when Spindly was removed. Importantly, phosphorylation-mimicking NDC-80 responded to Spindly depletion (and thus RZZ inhibition) in vivo in the same way as wild-type NDC-80: it could form end-coupled microtubule attachments (although with a delay) in the absence of RZZ, but this ability was entirely suppressed when Spindly was removed. Thus, RZZ regulates NDC-80 by interacting with its N-terminal tail, but this effect is independent of the direct contribution that the tail makes to NDC-80 microtubule binding in vitro.

This works uncovers an Aurora B-independent mechanism to control kinetochore–microtubule attachments that depends on crosstalk between NDC-80 and RZZ. This may be a direct mechanism to prevent NDC-80-mediated end-coupled attachment during the initial phase of lateral capture and thus might minimise the risk of a kinetochore being erroneously linked to both spindle poles, leading to chromosome missegregation.