G·A·B·B·A

Cytoskeleton microtubules undergo a reversible metamorphosis as cells enter and exit mitosis to build a transient mitotic spindle required for chromosome segregation. Centrosomes play a dominant but dispensable role in microtubule organization throughout the animal cell cycle, supporting the existence of concurrent mechanisms that remain unclear.
Here we investigated microtubule organization at the entry and exit from mitosis, after perturbation of centriole function in Drosophila S2 cells. We found that several microtubules originate from acentriolar microtubule-organizing centers (aMTOCs) that contain γ-tubulin and require Centrosomin (Cnn) for normal architecture and function. During spindle assembly, aMTOCs associated with peripheral microtubules are recruited to acentriolar spindle poles by an Ncd/dynein-dependent clustering mechanism to form rudimentary aster-like structures. At anaphase onset, down-regulation of CDK1 triggers massive formation of cytoplasmic microtubules de novo, many of which nucleated directly from aMTOCs. CDK1 down-regulation at anaphase coordinates the activity of Msps/XMAP215 and the kinesin-13 KLP10A to favor net MT growth and stability from aMTOCs. We further show that microtubule nucleation from aMTOCs also occurs in cells containing centrosomes. Our data reveal a new form of cell cycle–regulated MTOCs that contribute for MT cytoskeleton remodeling during mitotic spindle assembly/disassembly in animal somatic cells, independently of centrioles.
In this study we also investigated the dynamic behavior of spindle-associated γ-tubulin upon microtubule depolymerization conditions such as cold and colchicine treatment. We found that under these conditions, γ-tubulin is recycled into aggregates that are resistant to depolymerization. Curiously, upon physiological depolymerization conditions such as in anaphase, γ-tubulin has a similar behavior and forms identical aggregates.
Finally, we conducted a genome-wide RNAi screen in Drosophila S2 cells to identify a complete list of molecular players that are involved in mitotic spindle assembly, when functional centrosomes are not present. We found that most genes required for normal spindle assembly are also required for acentrosomal spindle assembly, reinforcing the idea of a unique molecular pathway for spindle assembly. Importantly, we have identified 16 novel proteins that play a role in spindle assembly since their depletion leads to a variety of spindle defects. Together, the results presented in this thesis provide a thorough functional characterization of the process of acentrosomal spindle assembly and open the way to explore the specific role of new proteins involved in this mechanism.