The syncytial divisions of the Drosophila melanogaster embryo lack some of the well established cell-cycle checkpoints. It has been suggested that without these checkpoints the divisions would display a reduced fidelity. To test this idea, we examined division error frequencies in individuals bearing an abnormally long and rearranged second chromosome, designated C(2)EN. Relative to a normal chromosome, this chromosome imposes additional structural demands on the mitotic apparatus in both the early syncytial embryonic divisions and the later somatic divisions. We demonstrate that the C(2)EN chromosome does not increase the error frequency of the late larva neuroblast divisions. However, in the syncytial embryonic nuclear divisions, the C(2)EN chromosome produces a 10-fold increase in division errors relative to embryos with a normal karyotype. During late anaphase of the neuroblast divisions, the sister C(2)EN chromosomes cleanly separate from one another. In contrast, during late anaphase of the syncytial divisions in C(2)EN-bearing nuclei, large amounts of chromatin often lag on the metaphase plate. Live analysis of C(2)EN-bearing embryos demonstrates that individual nuclei in the syncytial population of dividing nuclei often delay in their initiation of anaphase. These delays frequently lead to division errors. Eventually the products of the nuclei delayed in anaphase sink inward and are removed from the dividing population of syncytial nuclei. These results suggest that the Drosophila embryo may be equipped with mechanisms that monitor the fidelity of the syncytial nuclear divisions. Unlike checkpoints that rely on cell cycle delays to identify and correct division errors, these embryonic mechanisms rely on cell cycle delays to identify and discard the products of division errors.

In a number of embryonic systems, centrosomes that have lost their association with the nuclear envelope and spindle maintain their ability to duplicate and induce astral microtubules. To identify additional activities of free centrosomes, we monitored astral microtubule dynamics by injecting living syncytial Drosophila embryos with fluorescently labeled tubulin. Our recordings follow multiple rounds of free centrosome duplication and separation during the cortical division. The rate and distance of free sister centrosome separation corresponds well with the initial phase of associated centrosome separation. However, the later phase of separation observed for centrosomes associated with a spindle (anaphase B) does not occur. Free centrosome separation regularly occurs on a plane parallel to the plasma membrane. While previous work demonstrated that centrosomes influence cytoskeletal dynamics, this observation suggests that the cortical cytoskeleton regulates the orientation of centrosome separation. Although free centrosomes do not form spindles, they display relatively normal cell cycle-dependent modulations of their astral microtubules. In addition, free centrosome duplication, separation, and modulation of microtubule dynamics often occur in synchrony with neighboring associated centrosomes. These observations suggest that free centrosomes respond normally to local nuclear division signals. Disruption of the cortical nuclear divisions with aphidicolin supports this conclusion; large numbers of abnormal nuclei recede into the interior while their centrosomes remain on the cortex. Following individual free centrosomes through multiple focal planes for 45 min after the injection of aphidicolin reveals that they do not undergo normal modulation of their astral dynamics nor do they undergo multiple rounds of duplication and separation. We conclude that in the absence of normally dividing cortical nuclei many centrosome activities are disrupted and centrosome duplication is extensively delayed. This indicates the presence of a feedback mechanism that creates a dependency relationship between the cortical nuclear cycles and the centrosome cycles.

The Drosophila compound entire second chromosome, C(2)EN, displays paternal transmission well below Mendelian expectations (Novitski  et al. 1981). Because C(2)EN stocks also show higher-than-expected rates of zygotic lethality, it was proposed that this reduced paternal inheritance might be wholly or partially due to postfertilization events. Efforts to investigate this phenomenon have been hampered because the progeny of crosses between C(2)EN-bearing individuals and those with normal karyotypes die during embryogenesis. We have circumvented this obstacle by employing fluorescence in situ hybridization to directly karyotype early embryos from crosses involving C(2)EN-bearing individuals. This analysis reveals that the distortion in paternal transmission is established before fertilization. Moreover, measurement of the sperm ratios within both the male and female reproductive organs demonstrates that C(2)EN-bearing sperm are selectively lost after sperm transfer to the female and before storage of sperm in the seminal receptacles and spermathecae. Our results are consistent with a model of meiotic drive in which aberrations occuring early in meiosis lead ultimately to sperm dysfunction.