The cytoskeleton plays a variety of roles during the cell cycle, none more dramatic than the formation of a bipolar mitotic spindle and the subsequent cleavage of one cell into two. Proper centrosome separation is a prerequisite for positioning the bipolar spindle. Although studies demonstrate that microtubules and their associated motors drive centrosome separation, the role of actin in centrosome separation remains less clear. Studies in tissue culture cells indicate that actin- and myosin-based cortical flow is primarily responsible for driving late centrosome separation, whereas other studies suggest that actin plays a more passive role by serving as an attachment site for astral microtubules to pull centrosomes apart. Here we demonstrate that prior to nuclear envelope breakdown (NEB) in Drosophila embryos; proper centrosome separation does not require myosin II but requires dynamic actin rearrangements at the growing edge of the interphase cap. Both Arp2/3- and Formin-mediated actin remodeling are required for separating the centrosome pairs before NEB. The Apc2-Armadillo complex appears to link cap expansion to centrosome separation. In contrast, the mechanisms driving centrosome separation after NEB are dependent of the actin cytoskeleton and compensate for earlier separation defects. Our studies show that the dynamics of actin polymerization drive centrosome separation and this has important implications for centrosome positioning during processes such as cell migration, cell polarity maintenance, and asymmetric cell division.

Another vital role for spindle formation is in positioning the site of cleavage following anaphase separation of DNA. Rappaport's experiments with sand dollar embryos showed that cleavage furrow positioning is determined by the relationship between the spindle and the actin cortex. In his embryos, astral microtubules, which extend out to the cortex were primarily responsible for initiating a furrow, however, smaller somatic cells seem to position the furrow through the overlapping antiparallel central spindle. This balance between astral and central spindle influences is not well understood however. In the early Drosophila embryo, nuclei divide within a syncytium yet invaginate cortical actin and membrane, encompassing them, in order to complete mitosis in close proximity to neighboring nuclei. These furrows are considered natural Rappaport furrows since they form at astral microtubule overlap. Upon cellularization, the furrow positioning seems to shift from astral microtubule-based to central spindle-based. Our findings show that during the syncytial divisions, key conserved central spindle components Centralspindlin complex, Polo, and Fascetto (Prc1) all localize to regions of overlap astral microtubules during furrow formation. Given that the central spindle does not induce formation of conventional cytokinesis, finding that all of these components, plus the chromosomal passenger complex (Aurora B and INCENP), also localize to the central spindle was unexpected. The lack of furrow formation at the central spindle then is explained by the fact that the syncytial divisions rely on a maternally supplied form of RhoGEF, RhoGEF2, lacking the specific domains that localize zygotically expressed RhoGEF (Pebble) to the central spindle. RhoGEF2 instead localizes to the overlap astral microtubules of the syncytial divisions. Thus, in spite of proper localization of many key furrowing components to the central spindle in syncytial embryos, the failure of RhoGEF to localize to the central spindle may preclude formation of conventional cleavage furrows bisecting the spindle. In support of this idea, we bypass the need for RhoGEF by injecting constitutively active Rho into the syncytial embryos. This generates ectopic furrows strikingly similar to conventional cleavage furrows that form perpendicular to the central spindle during the syncytial divisions. While metaphase furrow formation is myosin independent these Rho-induced ectopic furrows, like conventional furrows, require myosin in addition to microtubules. These studies demonstrate that the early Drosophila embryo is primed to form furrows at either the overlapping astral microtubules or central spindle with the shift to the latter being driven in large part by a corresponding shift from maternal-to-zygotic forms of RhoGEF.

My studies predict that the delivery of RhoGEF2 to the metaphase furrows must be different than the mechanism that localizes Pebble to the central spindle (the Centralspindlin complex). Recently, it has been shown that RhoGEF2 localization to the metaphase furrows requires vesicle trafficking from the recycling endosome (RE). This vesicle trafficking is regulated by the Rab11-GTPase in the RE and its associated effector, Nuclear Fallout (Nuf). Previous observations of Nuf in the early embryo show that it accumulates at the RE from interphase to prophase during the time when furrows are being made. At prophase, Nuf is phosphorylated, which coincides with its diffusion away from the RE. I will present evidence that Nuf localization is regulated by its phosphorylation state and that the mitotic kinase, Polo directly phosphorylates Nuf, inhibiting its localization at the RE, decreasing vesicle trafficking to the furrow. I propose that this mechanism serves as a major component in timing the formation of a furrow and may provide valuable insight into timing cytokinesis in general.

Finally, the regulation of actin dynamics in cytokinesis has been well studied in terms of actin-interacting proteins such as Cofilin and Profilin. However, the direct modifications of actin and microtubules are similarly important for stable furrow ingression and abscission. Here I will present a newly characterized gene push pop that potentially indicates methylation of actin or tubulin as a previously unappreciated mechanism of regulating the cytoskeleton as well as other potential mitotic proteins in the events of cytokinesis.

The thesis work presented here promotes a broader understanding of cytokinetic furrow timing and positioning. On one hand, both centrosome separation and central spindle signaling are vital for proper furrow positioning. On the other hand, vesicle trafficking and folate metabolism are required for the proper timing and maintenance of a furrow during the cell cycle. A deeper understanding of both these processes of cytokinesis will provide valuable insight into the mechanisms of cell division and potentially how they are perturbed in tumorigenesis.