The force-producing mechanism for centrosome separation during spindle formation in vertebrates is intrinsic to each aster.

A popular hypothesis for centrosome separation during spindle formation and anaphase is that pushing forces are generated between interacting microtubules (MTs) of opposite polarity, derived from opposing centrosomes. However, this mechanism is not consistent with the observation that centrosomes in vertebrate cells continue to separate during prometaphase when their MT arrays no longer overlap (i.e., during anaphase-like prometaphase). To evaluate whether centrosome separation during prophase/prometaphase, anaphase-like prometaphase and anaphase is mediated by a common mechanism we compared their behavior in vivo at a high spatial and temporal resolution. We found that the two centrosomes possess a considerable degree of independence throughout all stages of separation, i.e., the direction and migration rate of one centrosome does not impart a predictable behavior to the other, and both exhibit frequent and rapid (4-6 microns/min) displacements toward random points within the cell including the other centrosome. The kinetic behavior of individual centrosomes as they separate to form the spindle is the same whether or not their MT arrays overlap. The characteristics examined include, e.g., total displacement per minute, the vectorial rate of motion toward and away from the other centrosome, the frequency of toward and away motion as well as motion not contributing to separation, and the rate contributed by each centrosome to the separation process. By contrast, when compared with prometaphase, anaphase centrosomes separated at significantly faster rates even though the average vectorial rate of motion away from the other centrosome was the same as in prophase/prometaphase. The difference in separation rates arises because anaphase centrosomes spend less time moving toward one another than in prophase/prometaphase, and at a significantly slower rate. From our data we conclude that the force for centrosome separation during vertebrate spindle formation is not produced by MT-MT interactions between opposing asters, i.e., that the mechanism is intrinsic to each aster. Our results also strongly support the contention that forces generated independently by each aster also contribute substantially to centrosome separation during anaphase, but that the process is modified by interactions between opposing astral MTs in the interzone.