Die mitotische Spindel ist ideal für die Erforschung der Selbstorganisation und Plastizität molekularer Kollektive im Zytoplasma. Die Geometrie der Spindel ist entscheidend für die korrekte Chromosomentrennung, muss sich aber an die Zellgröße anpassen. Es ist unbekannt, ob und wie Zellen während ihrer Differenzierung die Spindelarchitektur anpassen, was insbesondere während der Gehirnentwicklung relevant ist. Wir untersuchten dies mit Maus-Embryonalstammzellen, die in frühe neuronale Vorläuferzellen differenziert wurden. Wir entwickelten ein automatisiertes Mikroskopieprotokoll um einen umfassenden Datensatz von mitotischen Zellen zu generieren. Außerdem entwickelten wir Spindle3D, ein Werkzeug zur dreidimensionalen Analyse von Spindeln. Überraschenderweise waren die Spindelvolumina in differenzierenden Zellen bis zu 24% kleiner als in pluripotenten Zellen. Während die Wachstumsgeschwindigkeit der Mikrotubuli unverändert blieb, verschob sich in sich differenzierenden Zellen die Nukleation von Mikrotubuli zugunsten der astralen Population. Diese Veränderung der Spindelarchitektur basierte auf der differenzierungsbedingten Verdünnung des Zytoplasmas. Dies aktivierte CPAP, ein Protein, das die Zentrosomenreifung reguliert, was zur Superskalierung des perizentriolären Materials und verstärkte Rekrutierung von gamma-Tubulin an den Zentrosomen und somit zur Umlagerung von Mikrotubuli innerhalb der Spindel führte. Diese Veränderungen der mitotischen Architektur konnten durch externe osmotische Einwirkung in undifferenzierten Zellen nachgestellt werden. Insgesamt verbinden unsere Ergebnisse zelltypspezifische zytoplasmatische Materialeigenschaften mit der Spindelarchitektur. / The mitotic spindle provides an excellent system in which to study the plasticity of molecular collectives. To segregate chromosomes accurately, the spindle’s geometry must be adaptive to changes in cell size. It is unknown whether and how differentiating cells adjust spindle architecture, specifically during brain development when spindle defects have severe pathological consequences. Using murine embryonic stem cells, we recapitulated the transition from pluripotency to early neural cell identities in vitro. Aiming at a systematic exploration of spindle and cell morphology throughout this process, we developed an automated microscopy protocol and Spindle3D, a morphometric tool for the analysis of spindles in confocal images. Intriguingly, in cells with comparable cell volume, spindle volumes were up to 24% smaller in cells undergoing differentiation. While microtubule growth speed remained equal, we measured increased nucleation of astral microtubules at the expense of the spindle bulk in differentiating cells. The shift in spindle architecture was explained by the differentiation-driven cytoplasmic dilution. This activated the centrosomal regulator CPAP, causing the superscaling of the pericentriolar material and the concomitant increased recruitment of gamma-tubulin to the centrosomes, redistributing microtubule numbers within the spindle. Mimicking the dilution effect by osmotic challenge reproduced the same mitotic architecture in undifferentiated cells. Collectively, our results link cell state-specific cytoplasmic material properties to spindle architecture.

The centrosome is a dynamic organelle found in all animal cells that serves as a microtubule organizing center during cell division. Most of the centrosome components have been identified by genetic screens over the last decade, but little is known about how these components interact with each other to form a functional centrosome. Towards a better understanding of the molecular organization of the centrosome, we investigated the mechanism that regulates the size of the centrosome in the early C. elegans embryo. For this, we monitored fluorescently labeled centrosomes in living embryos and developed a suite of image analysis algorithms to quantify the centrosomes in the resulting 3D time-lapse images. In particular, we developed a novel algorithm involving a two-stage linking process for tracking entrosomes, which is a multi-object tracking task. This fully automated analysis pipeline enabled us to acquire time-resolved data of centrosome growth in a large number of embryos and could detect subtle phenotypes that were missed by previous assays based on manual image analysis. In a first set of experiments, we quantified centrosome size over development in wild-type embryos and made three essential observations. First, centrosome volume scales proportionately with cell volume. Second, beginning at the 4-cell stage, when cells are small, centrosome size plateaus during the cell cycle. Third, the total centrosome volume the embryo gives rise to in any one cell stage is approximately constant. Based on our observations, we propose a ‘limiting component’ model in which centrosome size is limited by the amounts of maternally derived centrosome components. In a second set of experiments, we tested our hypothesis by varying cell size, centrosome number and microtubule-mediated pulling forces. We then manipulated the amounts of several centrosomal proteins and found that the conserved centriolar and pericentriolar material protein SPD-2 is one such component that determines centrosome size.

The centrosome is a dynamic organelle found in all animal cells that serves as a microtubule organizing center during cell division. Most of the centrosome components have been identified by genetic screens over the last decade, but little is known about how these components interact with each other to form a functional centrosome. Towards a better understanding of the molecular organization of the centrosome, we investigated the mechanism that regulates the size of the centrosome in the early C. elegans embryo. For this, we monitored fluorescently labeled centrosomes in living embryos and developed a suite of image analysis algorithms to quantify the centrosomes in the resulting 3D time-lapse images. In particular, we developed a novel algorithm involving a two-stage linking process for tracking entrosomes, which is a multi-object tracking task. This fully automated analysis pipeline enabled us to acquire time-resolved data of centrosome growth in a large number of embryos and could detect subtle phenotypes that were missed by previous assays based on manual image analysis. In a first set of experiments, we quantified centrosome size over development in wild-type embryos and made three essential observations. First, centrosome volume scales proportionately with cell volume. Second, beginning at the 4-cell stage, when cells are small, centrosome size plateaus during the cell cycle. Third, the total centrosome volume the embryo gives rise to in any one cell stage is approximately constant. Based on our observations, we propose a ‘limiting component’ model in which centrosome size is limited by the amounts of maternally derived centrosome components. In a second set of experiments, we tested our hypothesis by varying cell size, centrosome number and microtubule-mediated pulling forces. We then manipulated the amounts of several centrosomal proteins and found that the conserved centriolar and pericentriolar material protein SPD-2 is one such component that determines centrosome size.