key: cord-103055-071a5b0x authors: Rathbun, LI; Aljiboury, AA; Bai, X; Manikas, J; Amack, JD; Bembenek, JN; Hehnly, H title: PLK1- and PLK4-mediated asymmetric mitotic centrosome size and positioning in the early zebrafish embryo date: 2020-04-14 journal: bioRxiv DOI: 10.1101/2020.04.13.039362 sha: doc_id: 103055 cord_uid: 071a5b0x Factors that regulate mitotic spindle positioning have been elucidated in vitro, however it remains unclear how a spindle is placed within the confines of extremely large cells. Our studies identified a uniquely large centrosome structure in the early zebrafish embryo (246.44±11.93μm2 mitotic centrosome in a 126.86±0.35μm diameter cell), whereas C. elegans centrosomes are notably smaller (6.75±0.28μm2 mitotic centrosome in a 55.83±1.04μm diameter cell). During early embryonic cell divisions, cell size changes rapidly in C. elegans and zebrafish embryos. Notably, mitotic centrosome area scales closely with changing cell size compared to changes in spindle length for both organisms. One interesting difference between the two is that mitotic centrosomes are asymmetric in size across embryonic zebrafish spindles, with the larger mitotic centrosome being 2.14±0.13-fold larger in size than the smaller. The largest mitotic centrosome is placed towards the embryo center in a Polo-Like Kinase (PLK) 1 and PLK4 dependent manner 87.14±4.16% of the time. We propose a model in which uniquely large centrosomes direct spindle placement within the disproportionately large zebrafish embryo cells to orchestrate cell divisions during early embryogenesis. During early embryogenesis, rapid cell divisions increase the number of cells in an embryo to ensure that proper tissue and organ formation can proceed during later development. However, it remains unclear how the mitotic spindle is able to position itself within the confines of a cell that is disproportionately large. Previous studies have proposed that spindle size scales to cell size in embryos [1, 2] . However, this poses the question whether the mitotic spindle adapts to the rapidly changing cell size during early embryonic cell divisions. This study aims to understand the previously unknown mechanism by which cell division is regulated during early development in extremely large cells. One proposed model is that large embryonic cells use acentrosomal microtubule nucleation sites so that astral microtubules can reach the cortex in large cells [3] . The mitotic centrosome/spindle pole assembles the microtubule-based spindle, and one spindle pole consists of two centrioles surrounded by pericentriolar material (PCM) that contains microtubule nucleation sites [4] . Typically, astral microtubules emanate from the centrosome and project towards the cell cortex, where they anchor and facilitate pulling forces to position the spindle and undergo cell division [4] . In the proposed acentrosomal model, microtubule nucleation sites exist outside of the centrosome through branched microtubules, positioning astral microtubules closer to the cell cortex in large cells [3] . However, we find that large dividing zebrafish embryo cells have notably large mitotic centrosomes that scale with cell size. We hypothesize that large centrosomes are used to assist astral microtubules in reaching the cortex in large cells. We present a model where mitotic centrosome size scales with cell size, and that this scaling requires Polo-Like Kinase (PLK) 1 and PLK4. transitioning to an asynchronous wave [6] . During the first five cell divisions, blastomeres create a cellular monolayer on top of the yolk. Each division during this stage occurs perpendicular to the plane of the previous division, leading to the construction of a monolayer grid (model, Figure 1B ) [6] . This is clearly visualized through the use of a fluorescent microtubule transgenic zebrafish line (EMTB-3xGFP) [7] , where the 16-cell stage embryo contains mitotic spindles oriented perpendicular to the previous division at the 8 In early development, rapid rounds of division result in a stark decrease in cell size during the cleavage stage [8] . We measured cell area during the first five cell cycles in each respective embryo (1-to 5-cell stage embryo in C. elegans, 8-to 128-cell stage in zebrafish). Since imaging and quantitative analysis is difficult to obtain from the 1-cell to 4-cell stage zebrafish embryo, we focused on the 8-to 128-cell stage. Both organisms had a significant decrease in cell area during these divisions. In C. elegans, the change in cell area became less drastic over time. In contrast, the decrease in cell area remained constant in zebrafish embryos out to 128-cells (Supplementary Figure 1E ). This suggests that while a marked decrease in cell size occurs during the first several rounds of cell division in many organisms, the magnitude of this change is not always similar. We next questioned whether the spindle and/or mitotic centrosome scaled to the longest cell axis (e.g. cell length) in C. elegans and zebrafish embryos. Spindle, mitotic centrosomes, and cell length were measured in C. elegans embryos that stably expressed a centrosome marker (g-tubulin::GFP), cell membrane marker (PH::mCherry) and/or a nuclear marker (H2B::mCherry) ( Figure 1A , 1D, Supplementary Figure 1A Metaphase mitotic spindle length was measured from mitotic centrosome to mitotic centrosome, cell length was measured from cell membrane to cell membrane along the same plane of the metaphase spindle, and metaphase mitotic centrosome area was measured (modeled in Figure 1H ). Cell length decreased with every division over time in C. elegans and zebrafish ( Figure 1C , gray), similar to the trend identified in cell area (Supplementary Figure 1E) . This decrease was not as drastic as the decrease in spindle length ( Figure 1C , orange). When considered as a ratio between spindle and cell length, mitotic spindles occupy a higher percentage of the cell length in later cell divisions compared to earlier divisions in both organisms leading to a significant decrease in the distance from mitotic centrosomes to cell membrane (Supplementary Figure 1F -G). Despite the stark size difference between cells in C. elegans and zebrafish embryos ( Figure 1H) , these data suggest a conserved trend of disproportional changes in cell and spindle dimensions during early cell divisions. When measuring mitotic centrosome size in C. elegans and zebrafish embryos ( Figure 1D -E), a significant decrease in mitotic centrosome area was identified from one round of division to the next ( Figure 1F ). Strikingly, mitotic centrosomes in zebrafish embryos were extremely large with g-tubulin organized into a wheel-like structure (246.44±11.93μm 2 at 8-cell stage, Figure 1E ), compared to g-tubulin in C. elegans (6.75±0.28μm 2 at 1-cell stage, Figure 1D ) or in zebrafish at the 512-cell stage (3.16 ±1.36μm 2 , Supplementary Figure 1H ). The values obtained from cell length, spindle length, and mitotic centrosome area measurements were normalized to determine their relative change. Size values were normalized to the 1-cell stage in C. elegans and to the 8-cell stage in zebrafish. In both organisms, we determined that the change in cell length scaled more closely with the change in mitotic centrosome area than that of spindle length ( Figure 1G ). In both C. elegans and zebrafish embryos, cell length and mitotic centrosome area decreased by 30-40% over time. Spindle length, however, decreased <20% during this time in both organisms ( Figure 1G ). Taken together, these data suggest that decreases in cell size scale more closely with mitotic centrosome size than spindle length ( Figure 1H , percentages represent size decrease). To characterize spindle and centrosome dynamics in the early embryo we focused on the zebrafish embryo due to its uniquely large centrosomes. To do this we employed bactin::EMTB-3xGFP [7] and bactin::centrin-GFP [9] embryos to mark microtubules and centrosomes. Volumetric projections of embryos from these transgenic lines were acquired over time (Figure 2A Figure 2C . At prophase, the mitotic centrosomes are placed on either side of the nucleus ( Figure 2E , 2F) and begin to nucleate a robust microtubule-based spindle for metaphase ( Figure 2D ). During anaphase, the mitotic centrosomes begin to fragment and disperse, and reform during telophase to prepare for immediate cell cycle re-entry ( Figure 2E -F, Supplementary Video 4). Notably, centrin normally marks centrioles [10, 11] , but in this case it marks a uniquely large structure that colocalizes with the PCM protein g-tubulin [12] Figure 3A ). This > 2-fold difference in mitotic centrosome size is maintained from prophase/prometaphase to anaphase (Supplementary Figure 3B , D). This asymmetry in mitotic centrosome size was consistent in cells placed next to the midline or further away from the midline (representative shown in Figure 3A ). When calculating centrosome size at metaphase in the centrosome transgenic line, bactin::centrin-GFP, an approximate 1.5-fold change was calculated at the 8-cell and 16-cell stage ( Figure 3D , Supplementary Figure 3E ). Even though centrin-GFP organization at mitotic centrosomes is asymmetric ( Figure 3D ), a metaphase spindle with the mitotic centrosome organizing the largest area of centrin-GFP was not as consistently positioned towards the center of the embryo (60.1± 4.8 % of the time across n=16 embryos at the 8-and 16-cell stage measured, Figure 3E , ratios between mitotic centrosomes shown for inner to outer in Supplemental Figure 3F ). Taken together, these data suggest a model in which zebrafish mitotic centrosomes present with an asymmetry in PCM components (e.g. γ-tubulin) starting at prophase/prometaphase (Supplementary Figure 3B , 3D) and this asymmetry biases the positioning of the larger centrosome towards the midline at the 8-and 16-cell stage (modeled in Figure 3F ). positioning. As cells progress through the cell cycle, they normally require PLK4 to duplicate their centrosome and PLK1 for robust PCM assembly during bipolar spindle construction [13] . The assembly of PCM components that interact with g-tubulin, such as pericentrin and CEP215, is facilitated by the phosphorylation activity of PLK1 [13, 14] . With PLK4 inhibition, centriole duplication is disrupted, causing spindles to assemble through acentriolar organization of PCM [15] [16] [17] . However, the role of PLK1 and/or PLK4 at mitotic centrosomes in the early zebrafish embryo is unknown. Transcripts for PLK1 and PLK4 have been detected as early as the 1-cell stage in zebrafish embryos, indicating that they are maternally supplied prior to zygotic genome activation, albeit PLK4 transcript levels are significantly lower [18] . Due to this, we tested the hypothesis that PLK1 and/or PLK4 regulate g-tubulin organization at mitotic centrosomes in zebrafish embryos. The PLK1 and PLK4 small molecule inhibitors, BI2536 [10, 19, 20] and centrinone [21] , were injected into 1 cell stage embryos (concentrations used 1µM or 100nM, BI2536 described with zebrafish in [19, 20] ). An injection approach was utilized versus soaking the embryos in drug due to the low permeability of the zebrafish chorion and embryo fragility at this stage. Control embryos were injected with 1% DMSO at the 1-cell stage and analyzed at the 16-cell stage. In 87.14±4.16% of these control embryos, the larger mitotic centrosome was positioned towards the midline (101.65±4.91μm 2 ) whereas the smaller was positioned away (52.28±272μm 2 , Figure 4A -B). This directional positioning of the larger mitotic centrosome towards the midline was significantly decreased when embryos were injected with BI2536 (60.92±5.11% with 100nM and 44.78±7.18% with 1μM), or centrinone (66.13±8.17%, with 100nM and 60.67±6.87% with 1μM, Figure 4A This suggests that not only does PLK1 and PLK4 regulate mitotic centrosome structure and asymmetry, but they regulate the directionality of larger centrosome placement towards the midline of the embryo's grid of cells ( Figure 4E ). We were surprised that PLK1 inhibition caused an increase in the area occupied by g-tubulin in mitotic centrosomes due to its known role in recruiting the pericentrin-CEP215 complex that anchors the γ-TURC at the centrosome [13, 22] . One possible explanation for this is that PLK1 has been proposed to regulate PCM architecture by facilitating its phase separation in C. elegans [23, 24] and that inhibiting PLK1 in zebrafish embryos may change the physical state of the PCM causing it to increase in size. This could explain why mitotic centrosome area significantly increases in a dosage-dependent manner with BI2536 treatment, as losing a PCM architecture regulator may cause the surrounding PCM to lose its tight matrix configuration and expand in size ( Figure 4A , D). Centrinone treatment did not exhibit the same dosagedependent change in mitotic centrosome area ( Figure 4D ). A possible explanation is that PLK4 is present at much lower concentrations in the early zebrafish embryo compared to PLK1 [18] . It is therefore likely that lower drug concentrations are required to target the small pool of PLK4, leading to a similar phenotype with drug concentrations above this small threshold. In order to determine the importance of PLK1/4-dependent asymmetric mitotic centrosome size placement in early zebrafish divisions, we raised embryos after injection of 1% DMSO, 1μM BI2536, or 1μM centrinone ( Figure 4F ). We found that compared to control embryos, PLK1-or PLK4-inhibition resulted in a lower survival rate over the first five days post-fertilization. At five days, we noted heart edema, embryo elongation defects, yolk elongation defects, and small eyes in the small fraction of embryos that survived drug treatment ( Figure 4F , Supplementary Figure 4D ), which are all common defects seen after early development perturbation. Given that the injections of BI2536 or centrinone likely diffuse out when the chorion starts to become more permeable, it is probable that the earliest cell divisions are impacted the most from this treatment. This led us to conclude that PLK1-and PLK4-dependent asymmetric mitotic centrosome placement in early embryos impacts later development. Through these studies, a unique centrosome structure has been characterized that may contribute to a better understanding of how mitotic spindles are able to coordinate cell division in disproportionately large cells. We demonstrate that mitotic centrosome size adapts to the decreasing cell size during the cleavage stage of zebrafish development. During this time in development, we found that zebrafish mitotic centrosomes are asymmetric in size and display a directionality, where the larger mitotic centrosome within a spindle is positioned towards the embryonic midline in a PLK1-and PLK4-dependent manner. Furthermore, we identified an ability for mitotic centrosomes to scale with cell size and maintain a 2-fold asymmetry across the spindle while doing so. When centrosome size scaling and asymmetry is disrupted an increase in embryonic lethality and developmental defects occurs, suggesting that early zebrafish embryonic cell divisions are not only important for early embryogenesis but likely also impact later developmental processes. Animal Lines. Zebrafish lines were maintained using standard procedures approved by the Syracuse University IACUC committee (protocol #18-006). Embryos were staged as described in Kimmel For live cell imaging of C. elegans embryos, a spinning disk confocal system was used. The system is equipped with a Nikon Eclipse and is an inverted microscope with a 60X 1.40NA objective, a CSU-22 spinning disc system and a Photometrics EM-CCD camera from Visitech International. Images were obtained every 2 minutes with a 1micron z-stack step size. Pharmacological treatments. Zebrafish embryos were injected with either 1% DMSO, or BI2536 or centrinone (final concentration 100nM or 1µM) at the 1-2-cell stage. Embryos are incubated at 30°C until they reach the developmental stage of interest, at which time they are fixed with 4% paraformaldehyde in PBS. Immunohistochemistry then proceeds as detailed below. Zebrafish immunohistochemistry. Zebrafish embryos were fixed using 4% PFA containing 0.5% Triton-X 100 overnight at 4°C. Zebrafish were then dechorionated and incubated in PBST (phosphate buffered saline + 0.1% Tween) for 30 minutes. Embryos were blocked using a Fish Wash Buffer (PBS + 1% BSA + 1% DMSO + 0.1% Triton-X 100) for 30 minutes followed by primary antibodies incubation (antibodies diluted 1:200 in Fish Wash Buffer) overnight at 4°C or 3 hours at room temperature. Embryos are then washed five times in Fish Wash Buffer and incubated in secondary antibodies (diluted 1:200 in Fish Wash Buffer) for 3 hours at room temperature. After five more washes, embryos were incubated with 4',6-diamidino-2-phenylindole (NucBlue® Fixed Cell ReadyProbes® Reagent) for 30 minutes. For imaging, embryos were either halved and mounted on slides using Prolong Diamond (Thermo Fisher Scientific cat. # P36971) or whole-mounted in 2% agar (Thermo-Fisher cat. # 16520100). Image and Data Analysis. Images were processed using both FIJI/ImageJ software and Adobe Photoshop. All graphs and statistical analysis were produced using Graphpad Prism software. 3-D images, movies, and surface rendering were performed using Bitplane IMARIS software (Surface, Smoothing, Masking, and Thresholding functions). To calculate two-dimensional area, a boundary was drawn around the structure of interest (cell, spindle pole, etc.) in ImageJ/FIJI and the area within this shape was calculated. To calculate spindle length, cell length, aspect ratio, etc., a line was drawn in ImageJ/FIJI from one end of the structure of interest to the other. This length was then measured and recorded. Phenotypic characterization. Wildtype zebrafish embryos were injected as described in the pharmacological section described above. The embryos were maintained at 30°C and assessed for abnormality in development and the number of deaths every 30 minutes for 9-10 hours post injection then once 24 hours post injection. At 5 days post fertilization, the phenotypes of injected embryos were characterized and the number of embryos with developmental defects were recorded. To generate death curves for the pharmacological treatments, the number of embryos treated with each drug were standardized to the starting number of embryos and were displayed as ratios over time. Statistical analysis. Unpaired, two-tailed Student's t-tests and one-way ANOVA analyses were performed using GraphPad Prism software. **** depicts a p-value <0.0001, *** p-value <0.001, **p-value<0.01, *p-value <0.05. See Methods Tables for detailed information regarding statistics. A comparative analysis of spindle morphometrics across metazoans Scaling, selection, and evolutionary dynamics of the mitotic spindle Microtubule nucleation remote from centrosomes may explain how asters span large cells Re-evaluating centrosome function Cell cycle timing regulation during asynchronous divisions of the early C. elegans embryo Stages of embryonic development of the zebrafish A model for cleavage plane determination in early amphibian and fish embryos Measuring time during early embryonic development Polarization and orientation of retinal ganglion cells in vivo Chromosome misalignment is associated with PLK1 activity at cenexin-positive mitotic centrosomes Centrin-2 is required for centriole duplication in mammalian cells Subdiffraction imaging of centrosomes reveals higher-order organizational features of pericentriolar material Regulating a key mitotic regulator, polo-like kinase 1 (PLK1) The centrosomespecific phosphorylation of Cnn by Polo/Plk1 drives Cnn scaffold assembly and centrosome maturation Increased lateral microtubule contact at the cell cortex is sufficient to drive mammalian spindle elongation SAK/PLK4 is required for centriole duplication and flagella development NuMA assemblies organize microtubule asters to establish spindle bipolarity in acentrosomal human cells The zebrafish transcriptome during early development Cytokinetic bridge triggers de novo lumen formation in vivo Subcellular drug targeting illuminates local kinase action Reversible centriole depletion with an inhibitor of Polo Mitosisspecific Anchoring of γ-Tubulin Complexes by Pericentrin Controls Spindle Organization and Mitotic Entry Regulated changes in material properties underlie centrosome disassembly during mitotic exit Regulated assembly of a supramolecular centrosome scaffold in vitro We thank Lilianna Solnica-Krezel (UW) for sharing the bactin::centrin-GFP zebrafish line. This work was supported by National Institutes of Health Grants no.