In wild-type syncitial embryos, the migration of centrosomes along the nuclear surface before their activation to nucleate microtubules co-incides with localised fenestration of the nuclear envelope beneath them.

This pore remains open until completion of karyokinesis, whereupon the envelope re-seals and the centrosomes re-attach. Thus, the microtubules emanating from each centrosome are able to interdigitate and drive anaphase elongation, culminating in karyokinesis. As indicated in the appropriate panels, the association of the centrosomes fails during prophase in polo mutants and between anaphase and karyokinesis in PP2A-related mutants. The attachment is represented in the cartoons as a sphere around the centrosome to reflect our ignorance of the nature of the attachment regulated by these enzymes at each stage. The defect could be an inability to physically generate/recruit anchors on the centrosomes or an inability to activate anchors at the nuclear surface. For example, it is known that dynein is required for centrosome anchorage. Alternatively, the anchors may be activated at the centrosome and transported to the envelope along microtubules. The question mark above the monopolar spindle cartoon in the Polo reduction box reflects the uncertainty of the detail of the phenotype: fenestration or attachment? It is currently unclear whether the defect at this stage is truly an attachment defect. It could arise from a problem in generating the holes in the envelope to enable the microtubules emanating from the centrosome to meet one another and form the spindle. If the microtubules can not penetrate the envelope to establish a bipolar spindle, the force arising from polymerising microtubules repeatedly pushing against the intact envelope would push the centrosome away. This contrasts with the normal situation where the microtubule-driven forces arising from bipolar spindle formation would pull the two poles towards one another and keep the centrosomes attached. If fenestration is defective, then the “attachment” defect is rather a polo-directed fenestration deficiency. If so, then it is just the PP2A-related regulated events that represent a true attachment defect.