Microtubule Nucleation Remote from Centrosomes May Explain
Total Page:16
File Type:pdf, Size:1020Kb
Microtubule nucleation remote from centrosomes may INAUGURAL ARTICLE explain how asters span large cells Keisuke Ishiharaa,b,1, Phuong A. Nguyena,b, Aaron C. Groena,b, Christine M. Fielda,b, and Timothy J. Mitchisona,b,1 aDepartment of Systems Biology, Harvard Medical School, Boston, MA 02115; and bMarine Biological Laboratory, Woods Hole, MA 02543 This contribution is part of the special series of Inaugural Articles by members of the National Academy of Sciences elected in 2014. Edited by Ronald D. Vale, Howard Hughes Medical Institute and University of California, San Francisco, CA, and approved November 13, 2014 (received for review October 6, 2014) A major challenge in cell biology is to understand how nanometer- aster growth in large cells, such as microtubule sliding, tread- sized molecules can organize micrometer-sized cells in space and milling, or nucleation remote from centrosomes. time. One solution in many animal cells is a radial array of Previously we developed a cell-free system to reconstitute microtubules called an aster, which is nucleated by a central cleavage furrow signaling where growing asters interacted (5, organizing center and spans the entire cytoplasm. Frog (here 12). Here, we combine cell-free reconstitution and quantitative Xenopus laevis) embryos are more than 1 mm in diameter and imaging to identify microtubule nucleation away from the cen- divide with a defined geometry every 30 min. Like smaller cells, trosome as the key biophysical mechanism underlying aster they are organized by asters, which grow, interact, and move to growth. We propose that aster growth in large cells should be precisely position the cleavage planes. It has been unclear understood as a spatial propagation of microtubule-stimulated whether asters grow to fill the enormous egg by the same mech- microtubule nucleation. anismusedinsmallersomaticcells, or whether special mecha- nisms are required. We addressed this question by imaging Results growing asters in a cell-free system derived from eggs, where Reconstitution of Large Asters in a Cell-Free System. To study aster asters grew to hundreds of microns in diameter. By tracking growth, we used extracts made from unfertilized frog eggs (13) marks on the lattice, we found that microtubules could slide out- and added calcium immediately before imaging to initiate the CELL BIOLOGY ward, but this was not essential for rapid aster growth. Polymer interphase cell cycle (12). These extracts are essentially undiluted treadmilling did not occur. By measuring the number and posi- cytoplasm and support the growth and interaction of large asters tions of microtubule ends over time, we found that most micro- that reconstitute spatially organized signaling characteristic of tubules were nucleated away from the centrosome and that cytokinesis in zygotes (Fig. 1A and ref. 5). In most experiments, interphase egg cytoplasm supported spontaneous nucleation asters were nucleated with beads coated with activating antibody after a time lag. We propose that aster growth is initiated to Aurora kinase A (AurkA) (14). This kinase plays a key role by centrosomes but that asters grow by propagating a wave in microtubule nucleation. The beads mimic centrosomes and SCIENCES of microtubule nucleation stimulated by the presence of pre- lead to similar aster growth as the physiological nucleation site, APPLIED PHYSICAL existing microtubules. a pair of centrosomes attached to sperm chromatin (Fig. S1A). However, unlike demembranated sperm, the beads provide a aster | centrosome | microtubule nucleation | embryo | Xenopus point-like center to the aster and facilitate image analysis by avoiding chromatin-mediated nucleation and splitting of the he large cells in early vertebrate embryos are organized by centrosome pair. Tradial arrays of microtubules called asters. This general or- At higher bead densities, asters interact, which mutually ganization was described by early cytologists (1) but is clearly inhibits further growth (Fig. 1A and ref. 5). At low bead density, illustrated by modern fixed immunofluorescence or live imaging. At the end of mitosis, a pair of asters is observed at the spindle Significance poles but remains small in radius, presumably because cyclin- dependent kinase 1 (Cdk1) inhibits aster growth (2). Once the cell How the cell cytoplasm is spatially organized is of fundamental μ enters interphase, the asters grow at rates of 30 m/min in Xenopus interest. In ordinary animal cells the cytoplasm is organized by μ zygotes and 15 m/min in zebrafish, while maintaining a high a radial array of microtubules, called an aster. Aster micro- density of microtubules at their periphery (2–4). Paired asters tubules are nucleated by the centrosome and elongate to the interact at the cell’s midplane to form a specialized zone of mi- periphery. We investigated how asters grow in an extremely crotubule overlaps, which in turn recruit cytokinesis factors to the large cell, the frog egg, using microscopy of an extract system. cell cortex (5, 6). Cell-spanning dimensions are presumably required Asters were initially nucleated at centrosomes, but then addi- so that the microtubules can touch the cortex to accurately po- tional microtubules nucleated far from the centrosome, ap- sition the cleavage furrow according to cell geometry (3, 7, 8). parently stimulated by preexisting microtubules. The resulting In the standard model of aster growth, microtubules are nu- growth process allows asters to scale to the size of huge egg cleated with their minus-ends anchored at the centrosome (9) cells while maintaining a high density of microtubules at the and polymerize outward with plus-ends undergoing dynamic in- periphery. Microtubule-stimulated microtubule nucleation might stability (10). However, there are several issues in applying this be a general principle for organizing large cells. model to a very large cytoplasm (11). Because of the radial ge- ometry, the standard model implies a decrease in microtubule Author contributions: K.I. and T.J.M. designed research; K.I., P.A.N., A.C.G., and C.M.F. performed research; K.I., P.A.N., A.C.G., and C.M.F. contributed new reagents/analytic density with increasing radius. In contrast, microtubule density tools; K.I. analyzed data; and K.I. and T.J.M. wrote the paper. seems to be constant or even increase toward the aster periphery The authors declare no conflict of interest. in frog and fish zygotes (3). Furthermore, this radial elongation 1To whom correspondence may be addressed. Email: [email protected] or model predicts that a subset of microtubules spans the entire [email protected]. aster radius, but it is unknown whether such long microtubules This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10. exist. We wondered whether additional mechanisms promoted 1073/pnas.1418796111/-/DCSupplemental. www.pnas.org/cgi/doi/10.1073/pnas.1418796111 PNAS | December 16, 2014 | vol. 111 | no. 50 | 17715–17722 Downloaded by guest on September 28, 2021 ABC Fig. 1. Reconstitution of large microtubule asters in a cell-free system. (A) High density of beads coated with AurkA antibody in interphase Xenopus egg extract nucleate many asters that mutually inhibit their growth upon contact, leading to limited radii; 34 min after calcium addition. (B) At low bead densities, asters are spatially separated and grow in an uninterrupted manner. Large asters spanning ∼500 μm in radius are observed; 45 min after calcium addition. Wide-field images of fluorescently labeled tubulin. (C) Asters grow and eventually span the entire cytoplasm of Xenopus laevis zygote. Laser scanning confocal images of eggs fixed at 45 min after fertilization and stained for tubulin by immunofluorescence. we observed assembly of asters as large as those observed in vivo, scales of aster growth similar to those in frog zygotes, and we spanning 500 μm or more in radius (Fig. 1 B and C). Aster radius decided to use this system to study aster growth mechanisms. grew continuously for 30–80 min after addition of calcium. At later time points, microtubules often appeared spontaneously in Large Asters Assemble in the Absence of Microtubule Sliding. One the cytoplasm, which interfered with further aster growth (Fig. biophysical process that might contribute to aster growth is S1B). The centers of asters in extract usually did not show a outward microtubule sliding (4, 15). To assess this, we applied drastic loss of microtubules, as seen in fixed frog zygotes or live fluorescent speckle microscopy (16). Speckles result from ran- zebrafish embryos (3). We concluded that our cell-free system dom incorporation of a small fraction of labeled tubulin into was capable of reconstituting large asters at length and time the microtubule lattice, and can be used as fiduciary marks to uncoated surface uncoated surface PLL-PEG surface + dynein inhibition A C E B D F B’ D’ F’ 60 s interior periphery Fig. 2. Large asters assemble in the absence of dynein-mediated microtubule outward sliding. (A) Asters assembled between uncoated coverslips show curved and disorganized microtubules. Image acquired on a wide-field microscope; 15 min after calcium addition. (Scale bar, 50 μm.) (B) First frame of a fluo- rescence speckle microscopy movie (Movie S1) showing labeled tubulin in an aster near its center. Image acquired on a spinning disk microscope. Movies were subjected to particle image velocimetry to determine the outward sliding rate as 6.1 ± 2.8 μm/min (SD, n = 12 asters). (Scale bar, 10 μm.) (B’) Kymographs generated from yellow line in B, where left to right corresponds to interior to periphery of the aster. (Scale bar, 10 μm.) (C–D’)SameasA–B’, in the presence of p150-CC1 protein to inhibit dynein motor activity. Note the decreased curving of aster microtubules and the complete suppression of microtubule outward sliding, −0.16 ± 0.25 μm/min (n = 12 asters) (Movie S2). (E–F’)SameasA–B’, with the coverslip surface passivated by PLL-PEG.