Growth, interaction, and positioning of microtubule asters in extremely large vertebrate embryo cells

Abstract

Ray Rappaport spent many years studying microtubule asters, and how they induce cleavage furrows. Here, we review recent progress on aster structure and dynamics in zygotes and early blastomeres of Xenopus laevis and Zebrafish, where cells are extremely large. Mitotic and interphase asters differ markedly in size, and only interphase asters span the cell. Growth of interphase asters occurs by a mechanism that allows microtubule density at the aster periphery to remain approximately constant as radius increases. We discuss models for aster growth, and favor a branching nucleation process. Neighboring asters that grow into each other interact to block further growth at the shared boundary. We compare the morphology of interaction zones formed between pairs of asters that grow out from the poles of the same mitotic spindle (sister asters) and between pairs not related by mitosis (non-sister asters) that meet following polyspermic fertilization. We argue growing asters recognize each other by interaction between antiparallel microtubules at the mutual boundary, and discuss models for molecular organization of interaction zones. Finally, we discuss models for how asters, and the centrosomes within them, are positioned by dynein-mediated pulling forces so as to generate stereotyped cleavage patterns. Studying these problems in extremely large cells is starting to reveal how general principles of cell organization scale with cell size.

Figure 1. Growth and interaction of sister asters in the first two divisions in Xenopus laevis

Fertilized eggs were fixed, stained for tubulin (A, A’ upper, B, C, C’, C”, D, D”) DNA (A’ lower) and γ-tubulin (D’,D’”) as described (Wühr et al 2010). The animal half of the egg was cut off and imaged from the cut surface, so the z axis is parallel to the animal-vegetal axis of the zygote. One letter is used to designate each different embryo. A, A’: metaphase of 1^st mitosis. Note small asters. B: anaphase-telophase of 1^st mitosis. Note aster growth and formation of an interaction zone between sister asters at mid-cell. C, C’, C”: later telophase. Note the dense, bushy appearance of microtubules at the aster periphery, low microtubule density and probable anti-parallel bundles in the interaction zone. D, D’, D”, D’” telophase of 2^nd mitosis, with the 1^st cleavage plane oriented North-South. The presumptive 2^nd cleavage plane will cut each blastomere between sister asters, at ~90° to the 1^st cleavage plane. Note bushy asters and interaction zones. γ-tubulin staining is brightest at points corresponding to centrosomes, but dimmer staining is evident throughout the aster.

Figure 2. Growth and interaction of sister asters in Zebrafish 1^st mitosis

Fertilized eggs from fish stably expressing the microtubule binding domain of ensconsin fused to 3×GFP were imaged live by confocal microscopy (Wühr et al 2010, 2011). The cell is imaged with the animal pole next to the 20× immersion objective lens. 0 min: prophase, note the intact nucleus. 4 min: metaphase, note the small aster radius at this stage. 8-20min: after anaphase onset the paired sister asters rapidly grow, and they meet and interact at the mid-plane of the cell. Note that the density of microtubules at the aster periphery, which is artificially highlighted using the ensconsin probe, remains approximately constant. Note also that microtubules at the aster periphery often appear curved and somewhat disorganized. The nucleus (n) and centrosomes (c) are highlighted in the last panel. Note the centrosomes inside one aster separate on the North-South axis as they move away from the interaction zones. The 2^nd mitotic spindles will later assemble on this North-South axis.

Figure 3. Models for aster growth in large cells

A segment of the spherical aster is shown. Microtubule density at the aster periphery decreases with radius in model A. In the other models, microtubule density at the periphery can remain constant, or even increase, with radius. A: Conventional radial elongation. Minus ends are nucleated and retained at centrosomes (red circle) while plus ends (arrowheads) elongate. B: Nucleation away from the centrosome. As the aster grows, microtubules are nucleated from the side of pre-existing microtubules (yellow circles), or from golgi membranes (blue stacks). C: Release and outward transport. Minus ends are released from centrosomal nucleation sites, and microtubules slide outward (yellow arrows indicate sliding). D: Capture of non-astral microtubules. Microtubules are nucleated away from the aster, e.g. at the cortex (pale green). As astral microtubules (dark green) grow out, they capture and orient the non-astral microtubules, and perhaps transport them inwards (yellow arrows).

Figure 4

Microtubules in zygotes of the nemertean worm Cerebratulus Confocal images of fixed embryos kindly provided by George von Dassow, (Oregon Institute of Marine Biology). Images are projections of 61 sections spaced 0.3 μm apart. The microtubule distribution in metaphase (A) appears radial, and microtubule density decreases rapidly with radius. In late anaphase (B) it appears more bundled and bushy, and microtubule density decreases less with radius. The dark zone in the center in B is presumably caused by a steric block to antibody penetration. A similar block is present at the center of the anaphase midzone and telophase midbody in somatic cells, and may also be present at the center of the aster-aster interaction zone in frog and fish embryos.

Figure 5. Consequences of aster-aster interaction depend on the system

Possible branching nucleation is omitted for simplicity A: In pure tubulin, plus ends simply grow past each other (Brinkley et al 1981). B: In M-phase extract prepared from unfertilized Xenopus eggs, asters (and spindles) that touch each other adhere, move together, and fuse. Movement is driven by dynein, which is thought to cross-bridge anti-parallel microtubules (Gatlin et al 2009). C: In interphase in early embryos, growing asters interact to generate a zone of anti-parallel overlap and low microtubule density at their mutual boundary. The asters then tend to move apart, pulled by cytoplasmic dynein anchored in bulk cytoplasm. Wühr et al (2010), see text for details. A block to microtubule interpenetration during anaphase-telphase was also noted in a classic EM study in echinoderm embryos (Asnes and Schroeder 1979).

Figure 6. Interaction of non-sister asters following polyspermic fertilization in amphybian embryos

A-C; Sequential stages from fertilization to 1^st cleavage following forced polyspermy in the frog Rana fusca. Images drawn from histological sections (Brachet 1910). D-F Tubulin staining in fixed embryos from Xenopus laevis (D,E,E’) and the axolotl Ambostoma mexicanum (F,F’) A: Neighboring sperm asters tend to move apart as they center. B: Asters tend to space out regularly before 1^st mitosis, dividing the cytoplasm into regularly spaced units. C: Cleavage furrows are induced between sister asters at 1^st cleavage in Rana fusca, while zones between non-sisters do not. The same is true in Xenopus laevis (Render and Elinson 1986, Wakabayashi and Shinagawa 2001) D. Forced trispermy in Xenopus laevis, stage between B and C. Note an interaction zone of low microtubule density between two asters (white arrows). The arrow marked with a? indicates an interaction where some microtubule interpenetration may have occurred. E,E’: Forced polyspermy in Xenopus, stage a little earlier than C, after anaphase of 1^st mitosis but before furrow induction. Two pairs of sister asters can be recognized in E by the brighter bundles of microtubules at the center of the interaction zone where the metaphase spindle used to be. Note the sharply defined interaction zones between both sister and non-sister asters. E’ is a higher mag view from E where (s) denotes interaction zones between sister asters, and (ns) between non-sister asters. F,F’: Natural polyspermy in Ambystoma mexicanum, stage similar to B. Two focal planes are shown. This zygote contains at least 6 sperm. The egg is enormous (~3 mm) which makes imaging difficult. Arrows denote likely aster-aster interaction zones as planes between asters where microtubule density is low.

Figure 7. Model for centrosome movement and orientation of the axis between centrosomes

A: Dynein in bulk cytoplasm, presumably anchored to organelles, generates a pulling force on centrosomes that increases with microtubule length (Hamaguchi and Hiramoto 1986, Kimura and Kimura 2011). B: Orientation of centrosome pairs orthogonal to the aster-aster interaction zone. Sister asters are dome shaped at telophase due to the interaction zone. This results in microtubule lengths, and net force on each centrosome, as indicated. In response, centrosome pairs move away from the interaction zone, while the centrosomes within each pair separate and orient on an axis parallel to the interaction zone. The axis between the centrosomes will later become the axis of 2^nd mitotic spindle (Wühr et al 2010). C,C’: Example of centrosome separation within telophase asters following 1^st mitosis in Xenopus. C’ is a high mag. view from C highlighting centrosomes. D: Model for forces on centrosomes in a compressed egg. Compression forces cleavage to orient across the short axis of the egg (Pflüger 1884, Hertwig 1893). See Minc et al (2011) for a mathematical model of this situation. E, E’: Recent repeat of Hertwig’s classic egg compression experiment (Wühr 2010). This compressed egg was fixed in prophase and stained for tubulin (green) and γ-tubulin (red). The centrosome pair has already oriented along the long axis of the cell, presumably in response to compression of the sperm aster as indicated in D.

Figure 8. Preliminary recapitulation of interphase aster growth and interaction in a cell free system

Xenopus egg mitotic extract (Desai 1999) was mixed with fluorescent tubulin and artificial centrosomes (Tsai and Zheng 2005), converted to interphase with Ca⁺⁺, spread between passivated coverslips and imaged by widefield fluorescence microscopy with a 10× objective. Large asters grew with a bushy morphology at their peripheries (e.g. 32min (3×)). When asters grew to touch each they generated interaction zones with locally low microtubule density (e.g. arrows at 42 min, shown at higher mag. in 42min (3×)). These interaction zones blocked aster expansion and were stable for tens of minutes (compare 32, 52 min). When two artificial centrosomes were initially close together, they tended to initiate a single aster, and later split apart within that aster (e.g. the pair indicated by arrowheads at 2, 22 and 53 min). This splitting was reminiscent of centrosome separation within telophase asters in embryos (Fig 6C).