Local density determines nuclear movements during syncytial blastoderm formation in a cricket

Animal embryos pass through an early stage called the blastoderm, in which cells are arranged in a continuous layer at the periphery of the embryo. Despite the broad evolutionary conservation of this embryonic stage, the cellular behaviours that lead to blastoderm formation vary across animals, and the mechanisms that regulate these behaviours are poorly understood. In most insects, pre-blastoderm development begins as a syncytium: that is, many nuclei divide and move throughout the single shared cytoplasm of the embryo. Then these syncytial nuclei must move from their scattered positions within the cytoplasm to form a single layer at the cortex. Recent work showed that in the fruit fly Drosophila melanogaster, some of these early nuclear movements are caused by pulses of cytoplasmic flows that are coupled to synchronous divisions. Here, we show that the cricket Gryllus bimaculatus has an altogether different solution to the problem. We quantified nuclear dynamics during the period of syncytial cleavages and movements that lead to blastoderm formation in G. bimaculatus embryos with transgenically labeled nuclei. We found that: (1) cytoplasmic flows were unimportant for nuclear movement, and (2) division cycles, nuclear speeds, and the directions of nuclear movement were not synchronized across the embryo as in D. melanogaster, but instead were heterogeneous in space and time. Moreover, several aspects of nuclear divisions and movements were correlated with local nuclear density. We show that previously proposed models for the movement of D. melanogaster syncytial nuclei cannot explain the behaviours of G. bimaculatus syncytial nuclei. We introduce a novel geometric model based on asymmetric local pulling forces on nuclei, which recapitulates the density-dependent nuclear speeds and orientations of unperturbed G. bimaculatus embryos, without invoking the common paradigms of localized polarity cues or cell lineage as determinants of nuclear activity. Our model also accurately predicts nuclear behavior in embryos physically manipulated to contain regions of atypical nuclear densities. We show that this model can be used to generate falsifiable predictions about the dynamics of blastoderm formation in other insect species.