(Fig 2) GFP-Centrosomin localizes to centrosomes. GFP-Centrosomin fusion protein (GFP-Cnn) was expressed in the early embryo. The movie shows two cleavage cycles.

(Fig 3) Cnn has a dynamic relationship with the centrosome. Emerging as 'flares', GFP-Cnn particles move back and forth from the centrosome. The time code in seconds is shown at the lower left corner of the frame. The arrows indicate the direction of flare particle movement.

(Fig 6) Flares are dependent on microtubules. GFP-Cnn embryos were injected with the microtubule-destabilizing drug colchicine to inhibit the formation of microtubules. Flare particles surround the centrosome, but no longer move back and forth as they do in the untreated embryos. Incipient flares appear to emerge from the centrosome but do not bud off as flare particles.

Movie 5A-D. (Fig 7) Flare movement is dependent on microtubule dynamics rather than on microtubule-based motors. Live embryos were treated with paclitaxel (Taxol) to stabilize microtubules. GFP-tubulin embryos treated with Taxol (B) exhibit intense spindle microtubules, bright asters, and have an arrested cycle as compared to mock control GFP-tubulin embryos (A). Flare activity was minimal in Taxol-treated GFP-Cnn embryos (D) as compared to the mock control animals (C).

(Fig 8) Flare particles are restricted by the actin cage. Rhodamine-actin was injected into GFP-Cnn embryos to view the dynamics of the centrosomes and the actin cytoskeleton live.

(Fig 9) Flare activity is not wholly dependent on actin assembly. GFP-Cnn embryos were injected with the actin-destabilizing drug cytochalasin-D. In the absence of the actin cytoskeleton, the nuclear and centrosome divisions were not inhibited. The drug apparently does not affect flare activity. The centrosomes flared under these conditions and appeared to transfer flare particles more readily from one centrosome to another.