The Polo-like kinase, Plk, has multiple roles in regulating mitosis. In particular, Plk1 has been postulated to function as a trigger kinase that phosphorylates and activates Cdc25C prior to the activation of cyclin B-Cdc2 and thereby initiates its activation. However, the upstream regulation of Plk1 activation remains unclear. Here we have studied the interplay between Plk1 and Cdc2 through meiotic and early embryonic cycles in starfish. Distinct kinases, cyclin B-Cdc2, MAPK along with cyclin B- and/or cyclin A-Cdc2 and cyclin A-Cdc2, were unique upstream regulators for Plk1 activation at meiosis I, meiosis II and embryonic M-phase, respectively, indicating that Plk1 is not the trigger kinase at meiotic reinitiation. When Plk1 was required for cyclin B-Cdc2 activation, the action of Plk1 was mediated primarily through suppression of Myt1 rather than through activation of Cdc25. We propose that Plk1 can be activated by either cyclin A- or cyclin B-Cdc2, and its primary target is Myt1.

In starfish, fertilization occurs naturally at late meiosis I. In the absence of fertilization, however, oocytes complete meiosis I and II, resulting in mature eggs arrested at the pronucleus stage, which are still fertilizable. In this study, we isolated cDNAs of starfish cyclin A and Cdc2, and monitored extensively the cell cycle dynamics of cyclin A and cyclin B levels and their associated Cdc2 kinase activity, Tyr phosphorylation of Cdc2, and Cdc25 phosphorylation states throughout meiotic and early embryonic cleavage cycles in vivo. In meiosis I, cyclin A was undetectable and cyclin B/Cdc2 alone exhibited histone H1 kinase activity, while thereafter both cyclin A/Cdc2 and cyclin B/Cdc2 kinase activity oscillated along with the cell cycle. Cyclin B-, but not cyclin A-, associated Cdc2 was subjected to regulation via Tyr phosphorylation, and phosphorylation states of Cdc25 correlated with cyclin B/Cdc2 kinase activity with some exceptions. Between meiosis I and II and at the pronucleus stage, cyclin A and B levels remained low, Cdc2 Tyr phosphorylation was undetectable, and Cdc25 remained phosphorylated depending on MAP kinase activity, showing a good correlation between these two stages. Upon fertilization of mature eggs, Cdc2 Tyr phosphorylation reappeared and Cdc25 was dephosphorylated. In the first cleavage cycle, under conditions which prevented Cdc25 activity, cyclin A/Cdc2 was activated with a normal time course and then cyclin B/Cdc2 was activated with a significant delay, resulting in the delayed completion of M-phase. Thus, in contrast to meiosis I, both cyclin A and cyclin B appear to be involved in the embryonic cleavage cycles. We propose that regulation of cyclin A/Cdc2 and cyclin B/Cdc2 is characteristic of meiotic and early cleavage cycles.

Graphical Abstract Highlights d E2F expression during cell division, differentiation, and quiescence is measured in vivo d E2F3A, E2F8, and E2F4 accumulate sequentially in the nucleus of cycling cells d E2F3A-4 nuclear accumulation controls gene expression during cell-cycle exit d Deep learning tools are applied to nuclear segmentation of complex mammalian tissues In Brief The study of E2Fs in vivo has been challenging. Cuitiñ o et al. reconstruct the spatiotemporal expression of E2F activators (E2F3A) and canonical (E2F4) and atypical (E2F8) repressors during the mammalian cell cycle and propose that orchestrated accumulation of different E2F combinations control gene expression in proliferating (E2F3A-8-4) and differentiating (E2F3A-4) cells. SUMMARYOrchestrating cell-cycle-dependent mRNA oscillations is critical to cell proliferation in multicellular organisms. Even though our understanding of cellcycle-regulated transcription has improved significantly over the last three decades, the mechanisms remain untested in vivo. Unbiased transcriptomic profiling of G 0 , G 1 -S, and S-G 2 -M sorted cells from FUCCI mouse embryos suggested a central role for E2Fs in the control of cell-cycle-dependent gene expression. The analysis of gene expression and E2F-tagged knockin mice with tissue imaging and deep-learning tools suggested that post-transcriptional mechanisms universally coordinate the nuclear accumulation of E2F activators (E2F3A) and canonical (E2F4) and atypical (E2F8) repressors during the cell cycle in vivo. In summary, we mapped the spatiotemporal expression of sentinel E2F activators and canonical and atypical repressors at the singlecell level in vivo and propose that two distinct E2F modules relay the control of gene expression in cells actively cycling (E2F3A-8-4) and exiting the cycle (E2F3A-4) during mammalian development.

Neural stem cells are generally considered to be committed to becoming precursor cells before terminally differentiating into either neurons or glial cells during neural development. Neuronal and oligodendrocyte precursor cells have been identified in several areas in the murine central nervous system. The presence of astrocyte precursor cells (APCs) is not so well understood. The present study provides several lines of evidence that CD44-positive cells are APCs in the early postnatal mouse cerebellum. In developing mouse cerebellum, CD44-positive cells, mostly located in the white matter, were positive for the markers of the astrocyte lineage, but negative for the markers of mature astrocytes. CD44-positive cells were purified from postnatal cerebellum by fluorescence-activated cell sorting and characterized in vitro. In the absence of any signaling molecule, many cells died by apoptosis. The surviving cells gradually expressed glial fibrillary acidic protein, a marker for mature astrocytes, indicating that differentiation into mature astrocytes is the default program for these cells. The cells produced no neurospheres nor neurons nor oligodendrocytes under any condition examined, indicating these cells are not neural stem cells. Leukemia inhibitory factor greatly promoted astrocytic differentiation of CD44-positive cells, whereas bone morphogenetic protein 4 (BMP4) did not. Fibroblast growth factor-2 was a potent mitogen for these cells, but was insufficient for survival. BMP4 inhibited activation of caspase-3 and greatly promoted survival, suggesting a novel role for BMP4 in the control of development of astrocytes in cerebellum. We isolated and characterized only CD44 strongly positive large cells and discarded small and/or CD44 weakly positive cells in this study. Further studies are necessary to characterize these cells to help determine whether CD44 is a selective and specific marker for APCs in the developing mouse cerebellum. In conclusion, we succeeded in preparing APC candidates from developing mouse cerebellum, characterized them in vitro, and found that BMPs are survival factors for these cells.