Stimulation of Microtubule Aster Formation and Spindle Assembly by the Small GTPase Ran

See allHide authors and affiliations

Science  21 May 1999:
Vol. 284, Issue 5418, pp. 1359-1362
DOI: 10.1126/science.284.5418.1359


Ran, a small guanosine triphosphatase, is suggested to have additional functions beyond its well-characterized role in nuclear trafficking. Guanosine triphosphate–bound Ran, but not guanosine diphosphate–bound Ran, stimulated polymerization of astral microtubules from centrosomes assembled on Xenopus sperm. Moreover, a Ran allele with a mutation in the effector domain (RanL43E) induced the formation of microtubule asters and spindle assembly, in the absence of sperm nuclei, in a γTuRC (γ-tubulin ring complex)– and XMAP215 (Xenopusmicrotubule associated protein)–dependent manner. Therefore, Ran could be a key signaling molecule regulating microtubule polymerization during mitosis.

In animal cells, the transition from interphase to mitosis is accompanied by pronounced changes in cellular architecture. One of the biggest changes is the conversion of the interphase microtubule array into a highly dynamic mitotic spindle. This requires more than the presence of microtubule nucleating centers (called centrosomes in animal cells) and the conversion of cytosol into a mitotic state (1, 2). Nuclear signals released into the cytoplasm upon nuclear envelope breakdown (NEB) influence the organization of the microtubule arrays. For example, premature rupture of the nuclear envelope in grasshopper spermatocytes during prophase causes the formation of a mitotic spindle next to the chromosomes (3). Chromosomes and nuclei themselves can initiate large microtubule asters from nearby centrosomes (4,5). Even in the absence of centrosomes, for example, during meiosis in female Drosophila, microtubule assembly is promoted near chromosomes (6). Chromosomes can also polarize existing microtubule arrays into a spindle-like structure in mitosis (7). Furthermore, artificial chromosomes tethered to beads stimulate mitotic spindle formation in the absence of centrosomes and kinetochores in Xenopus egg extracts (8). These studies indicate that nuclear signals, including those associated with chromosomes, influence the formation and rearrangement of microtubule structures during mitosis after NEB.

Ran-GTP, the predominant form of Ran in the nucleus (9), is a good candidate for such a signal (10–13). In the yeastSaccharomyces cerevisiae, overexpression of RanGEF1, the guanine nucleotide exchange factor for Ran, specifically suppresses a class of α-tubulin mutations that arrest with excess microtubules as large, budded cells. This suggests a link between the Ran pathway and microtubule polymerization (12). Moreover, cells with mutations in RanGEF1 and the Ran-binding protein, RanBP1, arrest with a misaligned mitotic spindle (11). In mammalian cells, a Ran-binding protein, RanBPM, is localized to the centrosome. Overexpression of RanBPM leads to the formation of ectopic microtubule asters, the centers of which contain γ-tubulin (13). These observations indicate that Ran could influence microtubule polymerization during mitosis.

To study the role of Ran in microtubule assembly during mitosis, we used cytostatic factor (CSF)–arrested Xenopus egg extract, a system used extensively to study centrosome and spindle assembly in vitro (7, 14). When demembranated sperm nuclei are added to the extract, the sperm centrioles rapidly assemble into centrosomes that nucleate microtubule asters, which subsequently organize into spindles (7). This system allowed us to address the effect of Ran on microtubule assembly in the absence of nuclear trafficking.

To determine if Ran affects microtubule assembly from sperm centrosomes, we expressed and purified various glutathione-S-transferase (GST)–Ran alleles (15) from bacteria, then loaded them with guanine nucleotide (16). These proteins were added toXenopus egg extracts containing demembranated sperm nuclei to a final concentration of 25 μM (the endogenous Ran concentration is ∼25 μM) (17). Wild-type Ran-GTP, and an allele of Ran with a mutation in the effector domain, RanL43E, stimulated the formation of large microtubule asters on sperm centrosomes (Fig. 1A). The microtubule polymer mass increased 5- and 10-fold, respectively, above the control (Fig. 1B). In contrast, wild-type Ran-GDP and the Ran alleles RanT24N (a dominant negative allele predominantly bound to GDP) and RanG19V (an activated allele locked in the GTP-bound state) had little effect. The increase in microtubule polymerization activity induced by RanL43E may result from the specific mutation in the effector domain (in which leucine 43 is changed to glutamic acid). This mutation not only keeps the protein in the GTP-bound active form (15), but could also enhance interactions with downstream effectors. RanL43E not only stimulated aster formation on all sperm, but also increased the number of asters not associated with sperm (Fig. 1C).

Figure 1

The effect of Ran on microtubule aster formation by Xenopus sperm centrosomes in Xenopus egg extract. (A) Demembranated sperm nuclei were incubated for 7 min at room temperature in Xenopus egg extract containing rhodamine-labeled tubulin in the presence or absence of 25 μM GST-Ran fusion proteins. Sperm chromatin was visualized with 4′,6-diamidino-2-phenylindole (DAPI) (blue). Images were taken and manipulated (22) such that the microtubule intensities (red) are comparable. (B) Quantitation of the microtubule polymer mass in sperm asters (22). (C) The extent of microtubule aster formation induced by Ran alleles in egg extract. Each of the three classes of sperm alone, sperm with asters, and asters alone were plotted as a percentage of the total. Bar (A), 10 μm.

We next determined whether different GST-Ran alleles could induce microtubule aster formation in Xenopus egg extracts in the absence of sperm nuclei. Only RanL43E induced microtubule aster formation. These asters had a greater variation in size (Fig. 2C) than asters induced by sperm. However, the average microtubule mass of asters induced by RanL43E or sperm was comparable (Fig. 2C).

Figure 2

Induction of microtubule aster and spindle formation in Xenopus egg extract. (A) Addition toXenopus egg extract of GST-RanL43E fusion protein to 25 μM (i to iv), DMSO to 0.5% (v and vi), or demembranated Xenopus sperm (vii and viii) for 7 min (i and v) or 30 min (ii, iii, iv, vi, vii, and viii) at room temperature. (B) Immunofluorescence analysis of microtubule structures formed in Xenopus egg extract. Red, α-tubulin; green, γ-tubulin or NuMA; and blue, DNA (DAPI). Yellow represents the colocalization of α-tubulin and γ-tubulin or α-tubulin and NuMA. Quantification of the total microtubule polymer content of asters (C) or spindles (D), formed by sperm or RanL43E (22). Bar (A and B), 10 μm.

Taxol and dimethyl sulfoxide (DMSO) can induce microtubule aster formation in Xenopus egg extract by stabilizing microtubules directly (18, 19). We tested whether GST-Ran fusion proteins could directly stimulate the polymerization of purified α/β-tubulin into microtubules. At concentrations of up to 65 μM, none of the Ran alleles had an effect on microtubule formation (20). Therefore, RanL43E must activate cellular factors in the egg extract to stimulate microtubule polymerization.

Microtubule asters formed by sperm centrosomes in CSF-arrestedXenopus egg extracts can polarize toward chromosomes to form monopolar and bipolar spindles (7). To test whether microtubule asters induced by RanL43E could rearrange into spindle-like structures, we incubated egg extracts supplemented with RanL43E for 30 min. At this time point RanL43E induced the formation of bipolar and multipolar microtubule structures that resembled mitotic spindles as well as asters (Fig. 2A). After 30 min, sperm nuclei when added to extract formed predominantly half-spindles with a few bipolar spindles (Fig. 2A). In contrast, DMSO induced only the formation of microtubule asters and disorganized arrays.

To examine whether the RanL43E spindle-like structures resembled the sperm spindles, we determined the localization of γ-tubulin and NuMA, known markers of spindle poles, by immunofluorescence (19, 21). Both proteins localized to the poles of spindle-like structures induced by RanL43E- and sperm nuclei–induced spindles (Fig. 2B). Therefore, asters induced by RanL43E have the intrinsic ability to become polarized even in the absence of chromosomes. Most of the microtubule structures induced by RanL43E after 30 min were asters, varying from ∼40 to 75% of the total, whereas bipolar and multipolar spindles made up between 20 to 30% and 5 to 30% of the total, respectively. No half-spindle structures were found in the RanL43E samples.

To compare directly the size and microtubule polymer mass between RanL43E and sperm spindles, we quantitated the fluorescence intensity of rhodamine-labeled tubulin in the microtubule structures produced (22). Bipolar spindles induced by RanL43E varied greatly in the amount of microtubule mass polymerized and contained fewer microtubules than half-spindles induced by sperm (Fig. 2D). Bipolar spindles induced by RanL43E were also smaller in size than the half-, or bipolar spindles, formed in the presence of sperm (Fig. 2). Extending the incubation time to 1 hour did not increase the size of the spindles induced by RanL43E (20). Although RanL43E alone can induce the formation of a bipolar spindle, centrosomes and chromosomes may be required to further stabilize and amplify the microtubule population in a spindle.

We next investigated whether RanL43E-stimulated aster formation was dependent on γTuRC, a microtubule nucleator (23, 24), or XMAP215, a microtubule binding protein that promotes elongation at the plus ends of microtubules (25). We immunodepleted γTuRC or XMAP215 from the egg extract using the antibodies XenC (23,24) and DDL (26). The same amount of normal rabbit immunoglobulin G (IgG) (NR) was used as a control. Depletion of either γTuRC or XMAP215 prevented RanL43E and sperm nuclei from inducing microtubule aster formation after 15 min. However, DMSO still stimulated the production of microtubule arrays (Fig. 3A). After 30 min of incubation, RanL43E and sperm did cause the formation of some asters in γTuRC- and XMAP215-depleted extracts (Fig. 3B), but no spindles were formed. Thus, Ran-activated microtubule polymerization depends on both γTuRC and XMAP215.

Figure 3

Inhibition of Ran-mediated aster formation after immunodepletion of γTuRC or XMAP215. (A) Xenopus egg extract was depleted of γTuRC (with antibody XenC) or XMAP215 (with antibody DDL) or treated with rabbit random IgG (NR) (26), before being incubated for 15 min at room temperature with either 25 μM GST-RanL43E fusion protein, DMSO to 0.5%, or demembranated sperm nuclei in the presence of rhodamine-labeled tubulin. (B) The average number of RanL43E-induced asters per field, or the total number of sperm-induced asters per 100 fields, was determined after incubation of the samples at room temperature for 15 and 30 min. Bar (A), 10 μm.

The dependence of RanL43E-induced asters on γTuRC might indicate that a pericentriolar-like material is assembled. Alternatively, microtubule motors might focus the microtubules induced by RanL43E into an aster. Dynein, a minus end–directed motor, is required to focus DMSO-induced microtubules into asters as well as focused spindle poles. We therefore tested whether dynein had a role in the formation of microtubule asters induced by RanL43E using the antibody 70.1, which binds to the intermediate chain of dynein, thereby inhibiting, its activity (19). Antibody 70.1 was added to Xenopus egg extract to a final concentration of 2 mg/ml, in the presence of either DMSO, RanL43E, or sperm nuclei and its effect was compared with that of a control antibody, 9E10 (27). Antibody 70.1 had little effect on asters induced by sperm nuclei, but completely blocked microtubule aster formation caused by DMSO (Fig. 4A). Although antibody 70.1 prevented the formation of large microtubule asters normally induced by RanL43E, smaller groups of microtubules remained associated at one of their ends (Fig. 4A). These structures could be formed in two ways: Microtubule motors other than dynein could organize the microtubules, or small fragments of pericentriolar-like material could be formed, which then nucleates microtubules.

Figure 4

The role of dynein in microtubule aster formation. (A) Addition of antibody 70.1, which recognizes dynein, or (B) antibody 9E10, which recognizes Myc, to egg extract containing rhodamine-labeled tubulin in the presence of 25 μM GST-RanL43E fusion protein, DMSO to 0.5%, or demembranated sperm nuclei. Bar (A and B), 10 μm.

Ran appears to regulate microtubule structures through known cellular activators of microtubule polymerization. Ran-GTP is thought to be predominantly nuclear, thereby only coming into contact with centrosomes during mitosis after NEB occurs. Therefore, Ran could regulate microtubule structures in mitosis. This may not be surprising, because signals from the nucleus, either associated with the chromosome or closely surrounding it, are thought to be involved in stabilizing and rearranging the microtubule array during mitosis (3–8). Ran-GTP may stimulate microtubule assembly through RanBPM as well as cytosolic factors such as γTuRC and XMAP215. Furthermore, Ran-GTP could act later to initiate or stabilize spindle formation. The Ran-GTPase–activating protein, RanGAP1, is localized to spindles throughout mitosis (28). Tethering RanGAP1 to the spindle could locally activate Ran-GTP hydrolysis to regulate spindle microtubules.

  • * To whom correspondence should be addressed. E-mail: zheng{at}


Stay Connected to Science

Navigate This Article