Asymmetric division in mouse oocytes: with or without Mos

Asymmetric division in mouse oocytes: with or without Mos

Brief Communication 1303 Asymmetric division in mouse oocytes: with or without Mos Marie-Hélène Verlhac, Christophe Lefebvre, Philippe Guillaud, Pas...

1MB Sizes 139 Downloads 79 Views

Brief Communication

1303

Asymmetric division in mouse oocytes: with or without Mos Marie-Hélène Verlhac, Christophe Lefebvre, Philippe Guillaud, Pascale Rassinier and Bernard Maro In both vertebrates and invertebrates, meiotic divisions in oocytes are typically asymmetric, resulting in the formation of a large oocyte and small polar bodies. The size difference between the daughter cells is usually a consequence of asymmetric positioning of the spindle before cytokinesis. Spindle movements are often related to interactions between the cell cortex and the spindle asters [1,2]. The spindles of mammalian oocytes are, however, typically devoid of astral microtubules, which normally connect the spindle to the cortex, suggesting that another mechanism is responsible for the unequal divisions in these oocytes. We observed the formation of the first polar body in wild-type oocytes and oocytes derived from c-Mos knockout mice [3]. In wild-type oocytes, the meiotic spindle formed in the centre of the cell and migrated to the cortex just before polar-body extrusion. The spindle did not elongate during anaphase. In mos–/– oocytes, the spindle formed centrally but did not migrate, although an asymmetric division still took place. In these oocytes, the spindle elongated during anaphase and the pole closest to the cortex moved while the other remained in place. Thus, a compensation mechanism exists in mouse oocytes and formation of the first polar body can be achieved in two ways: either after migration of the spindle to the cortex in wild-type oocytes, or after elongation, without migration, of the first meiotic spindle in mos–/– oocytes. Address: Biologie Cellulaire et Moléculaire du Developpement, UMR 7622, CNRS/Université Pierre et Marie Curie, 9 Quai Saint Bernard, Batiment C-5, 75252 Paris cedex 05, France. Correspondence: Marie-Hélène Verlhac E-mail: [email protected] Received: 16 June 2000 Revised: 25 July 2000 Accepted: 18 August 2000 Published: 6 October 2000 Current Biology 2000, 10:1303–1306 0960-9822/00/$ – see front matter © 2000 Elsevier Science Ltd. All rights reserved.

Results and discussion To understand the mechanisms that establish the asymmetric cell division in mouse oocytes, we set up a system to follow the migration of the spindle in vivo. We microinjected mRNA encoding a fusion protein between tubulin and the green fluorescent protein (tubulin–GFP) into immature oocytes (diameter, 80 µm) at the germinal

Figure 1 (c) 90

(a)

60 α(°) 30 (b)

d α

0

0

10 20 d (µm)

30

Current Biology

Orientation of the first meiotic spindle. (a) Superposition of the phase-contrast image showing the contour of the GV (in blue) with the fluorescent image of the spindle (tubulin–GFP in green). (b) Determination of the axis (in red) corresponding to the shortest distance between the cortex and the GV (d) and the angle (α) between the d axis and the spindle axis (in green). (c) Spindle orientation was not determined by the proximity of the cortex (absence of correlation between d and α).

vesicle (GV) stage, and the distribution of microtubules was followed by video microscopy. As was observed previously, the spindle formed 2 hours after germinal vesicle breakdown (GVBD) [4]. To determine whether the cortex has an influence on the establishment of the spindle axis, we superposed phase-contrast images showing the contour of the GV with the fluorescent image of the spindle (Figure 1a). Using image-analysis software, the axis corresponding to the shortest distance between the cortex and the GV (d; Figure 1b) and the angle between the d axis and the spindle axis (α; Figure 1b) were defined and measured. As shown on Figure 1c, there was no correlation between d and α. Thus, we can conclude that, although the GV is responsible for the localisation of the spindle (the spindle forms in the vicinity of the condensing chromosomes), the position of the GV with respect to the cortex has no influence on the positioning of the first meiotic spindle axis. Once the axis of the first meiotic spindle was established, it defined the path of spindle migration (Figure 2a), as the spindle always migrated along this axis. During prometaphase, the spindle was usually slightly off centre, one pole being located 12.6 ± 7.2 µm nearer to the cortex. The migration started about 2.5 ± 0.3 hours before polar body extrusion. In most cases (> 88%), the spindle migrated towards the nearest part of the cortex and, in contrast to

1304

Current Biology Vol 10 No 20

what happens during the formation of the second polar body [5,6], the extrusion of the first polar body did not require prior rotation of the spindle (Figure 2a). It has been reported that mos–/– oocytes, which lack mitogen-activated protein (MAP) kinase activity, extrude bigger polar bodies [7,8]. The metaphase I spindles in mos–/– and in wild-type oocytes were identical: they had the same size (mos–/– oocytes, 32 ± 2 µm; wild-type oocytes 35 ± 2 µm; p = 0.132) and they both lacked microtubule asters [7] (Figure 2e). Although mos–/– oocytes extruded their first polar bodies at the same time as control oocytes [7], we did not observe a migration of the spindle. Instead, the spindle elongated at anaphase, the pole closest to the cortex moving towards it, bringing one set of chromosomes close to the periphery of the oocyte (Figure 2b). We measured the movements of the poles in oocytes that had their spindles parallel to the observation plane (Figure 2c). In wild-type oocytes, the two poles moved together as a result of spindle migration (Figure 2d), and the length of the spindle remained constant (Figure 2e), even at anaphase. In

mos–/– oocytes, however, the pole opposite to the site of polar-body extrusion (P1) did not move, while the pole closest to the cortex (P2) moved at anaphase (Figure 2d). As shown in Figure 2e, the spindle elongated greatly (about 1.4 times) at anaphase in these oocytes. Polar bodies were bigger in mos–/– oocytes because the central spindle, which defines the position of the cleavage furrow, formed deeper in the oocyte. Microinjection of mos–/– oocytes with mos mRNA rescued spindle migration (data not shown). Thus, our results demonstrate that Mos is involved in the control of spindle migration. Moreover, they also show that, in mos–/– oocytes, the cellular machinery is still able to produce an asymmetric division. The migration of the first meiotic spindle depends on microfilaments and not microtubules. This can be shown by treating oocytes with either cytochalasin D or nocodazole [9,10]. To exclude possible delays in the timing of spindle migration and anaphase in oocytes treated with these drugs, we followed the movements of the chromosomes in vivo. Cytochalasin D did not perturb the timing

Figure 2

(a)

In vivo observation of tubulin–GFP in (a) wild-type and (b) mos–/– oocytes by video microscopy before and at the time of polarbody extrusion. Images were taken every 20 min (left to right and top to bottom). (c) Determination of the position of the spindle poles (P1, P2) and of the spindle length (S). (d) Position of the poles (P1, P2) before and at the time of polar-body extrusion (PBE) in wild-type and mos–/– oocytes. (e) Spindle length before and at the time of polar-body extrusion in wild-type and mos–/– oocytes.

(b)

P2 (d) 80 P1 Pole position (P1, P2; µm)

S

(e) 50 Spindle size (S; µm)

(c)

60

40

40 35 30

20 –3 –2 –1 0 Time (h before PBE)

Current Biology

45

P1 mos+/– P2 mos+/–

P1 mos–/– P2 mos–/–

25

–3

–2 –1 0 Time (h before PBE) mos+/–

mos–/–

Brief Communication

(a)

Figure 4 (b)

(a)

6 (b)

6 Current Biology

Current Biology

In vivo observation of chromosomes (stained with 2 ng/ml vital bisBenzimide, Hoechst 33258, Sigma) in wild-type oocytes treated with (a) 1 µg/ml cytochalasin D or (b) 1 µM nocodazole. Images were taken every 20 min (left to right and top to bottom). Red lines mark the position of the chromosomes at the beginning of the experiment.

of anaphase (10.0 ± 0.4 hours versus 9.9 ± 0.3 hours in controls), yet it prevented the migration of the chromosomes (and of the spindle) to the cortex (Figure 3a). Nocodazole treatment inhibited anaphase, but the chromosomes moved to the cortex (Figure 3b) 1 hour earlier than in nontreated oocytes (6.5 ± 0.5 hours versus 7.5 ± 0.4 hours in controls). Nevertheless, the migration in nocodazole lasted 2.3 ± 0.3 hours, like in control, non-treated oocytes (2.5 ± 0.3 hours), suggesting that the same mechanism is at work. We must point out that in nocodazole-treated oocytes, the chromosomes were hypercondensed and clumped in one mass that tended to be located in a more peripheral position than in control oocytes (Figure 3b). We have demonstrated unambiguously that microfilaments are involved in the migration of the first meiotic spindle, whereas microtubules are not involved in this process. Thus, the presence of the spindle influences the direction of migration but is not required for the movement of the chromosomes to the cortex. These results also establish that the Mos-MAP kinase pathway controls the activity of the actin microfilament network, perhaps through myosin IIA, a molecule involved in spindle migration [11]. Making an asymmetric division during mouse meiosis requires, first, spindle migration to the cortex and, second, a restriction of the cleavage furrow, if the spindle is not perpendicular to the cortex. In oocytes arrested at metaphase II, the cleavage furrow is restricted to a cortical domain that is enriched in microfilaments and is devoid of microvilli [5,6,10]. The chromosomes, which are located under the cortex, play an important role in the formation of this domain. If metaphase II chromosomes are artificially dispersed, by nocodazole treatment, each group of chromosomes will induce the formation of a domain rich in microfilaments and poor in microvilli [5,9]. As the chromosomes

(c) 100

7

7

8

8

Percentage of oocytes

Figure 3

1305

75 50 25 0

5

6 7 8 9 Time (h) mos+/– mos–/–

Setting up the cortical polarity in (a) wild-type and (b) mos–/– oocytes. The microvilli were stained with ConA–FITC (green) and the chromosomes with propidium iodide (red). Oocytes were fixed 6, 7 and 8 h post-GVBD. Projection of five confocal sections (except for the 8 h oocyte in (a), where all sections were used). The scale bar represents 20 µm. (c) The percentage of wild-type (n = 170) and mos–/– (n = 72) oocytes that were still in metaphase I and showed cortical polarity was plotted against time (h) after GVBD.

are almost in the centre of the cell at prometaphase and then under the cortex at the end of metaphase I, we determined whether we could also detect a polarisation of the cortex above the chromosomes in metaphase I. Oocytes synchronised at GVBD were fixed during metaphase I and stained with concanavalin A coupled to fluorescein isothiocyanate (ConA–FITC), which allows the distribution of microvilli to be followed [6,12]. In wild-type oocytes, a microvilli-free cortical domain appeared progressively, concomitantly with spindle migration (Figure 4a,c). In mos–/– oocytes, however, signs of cortical polarisation were present only at anaphase, with less microvilli in the polar body compared with the oocyte (Figure 4b). Thus, mos–/– oocytes can reorganise their cortex in response to the chromosomes, but they do it only when the first meiotic spindle elongates and brings one set of chromosomes close to the cortex. Whatever the nature of the signal associated with the chromosomes, it diffuses at a short distance and induces very rapid changes in the cortex. The spindle migrates along its long axis, but the choice of its direction of migration is a consequence of a slight offcentre positioning of the GV rather than the consequence of a pre-defined cortical site that could influence its migration. First, the cortex did not seem to influence the axis of the spindle: the position of the spindle was determined by the position of the GV and, if a defined position in the cortex were able to influence the positioning of the GV, it would also control the spindle axis; there was, however, no correlation between the spindle axis and the position of the GV relative to the cortex. Second, the spindle formed in a slightly off-centre position and, in 88% of the cases, migrated on the shortest distance. Third, cortical polarisation developed concomitantly with chromosome migration. Therefore, the direction of spindle migration results

1306

Current Biology Vol 10 No 20

from a slight asymmetry in the positioning of the GV and, thus, of the spindle. This asymmetry is then reinforced by the effect of the chromosomes on the cortex. To set up an asymmetric division in the absence of Mos, two mechanisms are at work: elongation of the spindle at anaphase B, and migration of the spindle pole nearest to the cortex. Spindle elongation in mos–/– oocytes was made possible because MAP kinase was not activated. Actually, it has been shown that MAP kinase maintains microtubules in a highly dynamic state in the absence of maturation promoting factor (MPF) activity, both in vitro [13] and in vivo [7]. Thus, in mos–/– oocytes, when MPF activity drops at the metaphase–anaphase transition, the absence of MAP kinase activity allows spindle elongation at anaphase B. The displacement of one pole towards the cortex may require an asymmetry of the forces exerted on both poles, which could be mediated by astral microtubules like in Caenorhabditis elegans [14]. In mouse oocytes, however, astral microtubules are not present, suggesting that another mechanism is at work. One possibility may be the attraction of one set of chromosomes to the cortex (the one nearest to it). In mos–/– oocytes, the cortex is still able to respond to the signal coming from the chromosomes, but at a shorter distance. It is possible that the chromosomes are still able to induce their migration, but again at a shorter distance, and this would happen only at the end of anaphase when the chromosomes are closer to the cortex. This would suggest that Mos behaves like an ‘amplifier’, increasing the range of action of this signal. Long-range effects of the chromosomes on behaviour of cytoskeletal elements have been demonstrated already in the case of microtubules [15]. Taken together, our results present two different modes of achieving an asymmetric division during mouse meiosis. One is dependent on Mos and involves spindle migration concomitant with cortical polarisation. Another is independent of Mos and requires a huge spindle elongation at anaphase followed by cortical polarisation.

Materials and methods Collection and culture of mouse oocytes Immature oocytes arrested at prophase I of meiosis were obtained by removing ovaries from 11 week old OF1 and mos–/– female mice. Oocytes were removed and cultured as described [16]. Immunofluorescence was performed as described [16]. To visualise the microvilli, oocytes were first fixed for 5 min in 1% paraformaldehyde in PBS at 30°C, then incubated for 30 sec in ConA–FITC (0.7 mg/ml in PBS; Molecular Probes), washed in PBS then post-fixed in 3.7% formaldehyde in PBS. They were permeabilised for 20 min in 1% Triton X-100 in PBS, then washed in 0.1% Tween-20/PBS, incubated for 5 min in propidium iodide (5 µg/ml in 0.1% Tween-20/PBS), washed three times in PBS before mounting in Citifluor (Chem. Lab, UCK).

In vitro synthesis and microinjection of mRNA The pRN3-β5-tubulin–GFP plasmid has been described [16]. In vitro synthesis of capped RNAs was performed using linearised plasmid with the mMessage mMachine kit (Ambion). The mRNAs were purified on RNeasy columns (Qiagen) and eluted in injection buffer (10 mM Tris, 0.1 mM EDTA, pH 7.4) at a final concentration of 0.5 µg/µl.

Aliquots of 4 µl were stored at –80°C. The mRNAs were microinjected into the cytoplasm of immature oocytes using an Eppendorf pressure microinjector and sterile pipettes. The oocytes were kept in M2 medium supplemented with dbcAMP during injection, then cultured at 37°C, in an atmosphere of 5% CO2 in air for 5 h in M2 medium supplemented with dbcAMP to allow overexpression of the exogenous proteins. The resumption of meiotic maturation was triggered by removal of the injected oocytes from the dbcAMP-containing medium and transfer into dbcAMP-free medium. Tubulin–GFP was visualised using a cooled charged-coupled device camera (Photometrics PH250) under a computer-controlled video microscope (Zeiss/Oncor).

Acknowledgements We thank W. Colledge and M. Evans for the generous gift of the mos–/– mice, B. Ludin for the gift of the β5-tubulin–GFP clone, and R. Schwartzmann for expert photographic work. This work was supported by grants from CNRS to B.M. and ARC 9913 to M.-H.V. C.L. is the recipient of a MERT fellowship.

References 1. Hyman AA, White J: Determination of cell division axes in the early embryogenesis of Caenorhabditis elegans. J Cell Biol 1987, 105:2123-2135. 2. Hyman AA: Centrosome movement in the early divisions of Caenorhabditis elegans: a cortical site determining centrosome position. J Cell Biol 1989, 109:1185-1193. 3. Colledge WH, Carlton MBL, Udy GB, Evans MJ: Disruption of c-mos causes parthenogenetic development of unfertilized mouse eggs. Nature 1994, 370:65-68. 4. Verlhac M-H, Kubiak JZ, Clarke HJ, Maro B: Microtubule and chromatin behavior follow MAP kinase activity but not MPF activity during meiosis in mouse oocytes. Development 1994, 120:1017-1025. 5. Maro B, Johnson MH, Webb M, Flach G: Mechanism of polar body formation in the mouse oocyte: an interaction between the chromosomes, the cytoskeleton and the plasma membrane. J Embryol Exp Morphol 1986, 92:11-32. 6. Maro B, Johnson MH, Pickering SJ, Flach G: Changes in the actin distribution during fertilisation of the mouse egg. J Embryol Exp Morphol 1984, 81:211-237. 7. Verlhac M-H, Kubiak JZ, Weber M, Géraud G, Colledge WH, Evans MJ, Maro B: Mos is required for MAP kinase activation and is involved in microtubule organisation during mouse meiosis. Development 1996, 122:815-822. 8. Choi TS, Fukasawa K, Zhou RP, Tessarollo L, Borror K, Resau J, Vandewoude GF: The Mos/mitogen-activated protein kinase (MAPK) pathway regulates the size and degradation of the first polar body in maturing mouse oocytes. Proc Natl Acad Sci USA 1996, 93:7032-7035. 9. Van Blerkom J, Bell H: Regulation of development in the fully grown mouse oocyte: chromosome-mediated temporal and spatial differentiation of cytoplasm and plasma membrane. J Embryol Exp Morphol 1986, 93:213-238. 10. Longo FJ, Chen DY: Development of cortical polarity in mouse eggs: involvement of the meiotic apparatus. Dev Biol 1985, 107:382-394. 11. Simerly C, Nowak G, de Lanerolle P, Schatten G: Differential expression and functions of cortical myosin IIA and IIB isotypes during meiotic maturation, fertilization and mitosis in mouse oocytes and embryos. Mol Biol Cell 1998, 9:2509-2525. 12. Johnson MH, Eager D, Muggleton-Harris AL, Graves HM: Mosaicism in the organisation of concanavalin A receptors on surface membrane of mouse eggs. Nature 1975, 257:321-322. 13. Gotoh Y, Nishida E, Matsuda S, Shiina N, Kosako H, Shiokawa K, et al.: In vitro effects on microtubule dynamics of purified Xenopus M phase-activated MAP kinase. Nature 1991, 349:251-254. 14. Gonczy P, Pichler S, Kirkham M, Hyman AA: Cytoplasmic dynein is required for distinct aspects of MTOC positioning, including centrosome separation, in the one cell stage Caenorhabditis elegans embryo. J Cell Biol 1999, 147:135-50. 15. Dogterom M, Felix MA, Guet CC, Leibler S: Influence of M-phase chromatin on the anisotropy of microtubule asters. J Cell Biol 1996, 133:125-140. 16. Brunet S, Polanski Z, Verlhac M-H, Kubiak JZ, Maro B: Bipolar meiotic spindle formation without chromatin. Curr Biol 1998, 8:1231-1234.