The linking of plate tectonics and evolutionary divergence

Whereas cohesin cleavage alone did not produce any detectable effects on engaged centrioles, Cdk inhibition, in contrast, was sufficient to induce centriole disengagement even in the absence of proper chromosome disjunction. Upon p27 injection, centriole disengagement was observed with a similar kinetics to the disengagement observed in the TEV+p27 experiments (Figure 1). Our previous experiments revealed that Cdk inactivation in metaphase-arrested embryos was not accompanied by prompt separase activation, as sister chromatids did not move apart during induced mitotic exit [6]. Our results therefore also raise the possibility that separase is not universally involved in centriole disengagement. In agreement, previous studies in Drosophila failed to detect any centrosome duplication defects in separase mutant embryos [8]. While Cdk inhibition was sufficient to trigger centriole disengagement, no further separation of sister centrioles could be observed. This finding suggests that even in Drosophila embryos, where centriole disengagement is immediately followed by centrosome separation, these are mechanistically different processes: centriole disengagement appears to depend on a drop in Cyclin-B–Cdk activity whereas centrosome separation is likely to depend on a subsequent rise of cyclin B levels and/or DNA replication. In summary, in contrast to the recent observation in mammalian cells, our experiments support the idea that centriole engagement does not depend on the integrity of the cohesin complex, at least in Drosophila embryos. In agreement, recent studies propose that cleavage of a novel centrosomal substrate for separase — pericentrin/ kendrin — is required for centriole disengagement [9]. Importantly, our experiments further demonstrate that centriole disengagement during mitotic exit, as many other aspects of this key transition, can be negatively regulated by Cdk activity. This supports a role for Cdk1 in preventing premature centriole disengagement in Drosophila early embryos. Further experiments will be required to investigate whether this results from a direct Cdk-dependent phosphorylation of centrosome components or rather an indirect consequence of changing pericentriolar organization or microtubule forces, as recently suggested [10].