Patterns of microtubule polymerization relating to cortical rotation in Xenopus laevis eggs

The unfertilized frog egg is organized with radial symmetry about an animal-vegetal axis. The bilateral symmetry of the tadpole is introduced during the second half of the first cell cycle by a cytoplasmic reorganization originally termed the ‘rotation of symmetrization’ by Ancel and Vintemberger (1948). This process comprises a rotation of the egg cortex with respect to the vegetal yolk mass by 30°, and results in the formation of a ‘grey crescent’ in the eggs of certain species as part of the clear vegetal cortex is brought to overly pigmented animal cytoplasm. The grey crescent marks the future site of the dorsal lip of the blastopore, the ‘organizer’ of the amphibian embryo, and the rotation is necessary for the localization of ‘dorsal information’ to this region (see Elinson and Kao, 1989, for review).

The grey crescent can form at any position around the pigment boundary of the egg, and in normal conditions it forms opposite the entrance point of the sperm (see Clavert, 1962, for review of the early literature). This effect may be mediated in some way by the sperm aster, a giant centriole-nucleated microtubule aster that expands during the first half of the cell cycle, migrates from an eccentric location to fill much of the animal hemisphere (Subtelny and Bradt, 1963; Kubota, 1967; Manes and Barbieri, 1977; Stewart-Savage and Grey, 1982) and causes extensive reorganization of the animal cytoplasm (Ubbels et al. 1983; Ubbels and Vermeulen, 1986). In Xenopus, a true grey crescent does not form but the correspondence between sperm entry site and embryonic axis orientation holds (Gerhart et al. 1981; Danilchik and Black, 1988) and the rotation of symmetrization has been demonstrated by marking dye spots on the vegetal yolk and cortex (Vincent et al. 1986).

More details of the cellular basis for the cortical rotation are now known. Inhibitor studies suggest that microtubules but not microfilaments are actively involved (Manes et al. 1978; Scharf and Gerhart, 1983; Scharf et al. 1986; Vincent et al. 1987). Furthermore, an impressive array of microtubules aligned with the direction of rotation appears beneath the vegetal surface at the time of rotation (Elinson and Rowning, 1988; Zisckind and Elinson, 1990) coincident with the main period of microtubule polymerization in the egg (Elinson, 1985). These observations raise the possibility that the relationship between the site of sperm entry and the direction of rotation could be due to a link between the microtubules of the enlarging sperm aster and those of the vegetal array. In order to address this possibility, we have performed anti-tubulin immunofluorescence on sections of Xenopus laevis eggs fixed at sequential times following fertilization. Eggs activated artificially were also examined since in these a vegetal microtubule array is known to form despite the absence of a sperm aster (Elinson and Rowning, 1988).

Procedures for maintaining Xenopus laevis, induction of ovulation, insemination and dejellying were described in Kao and Elinson (1988). Dejellied unfertilized eggs were maintained for 20 min in 100% Steinberg’s solution prior to fixation or artificial activation to allow damaged or activated eggs to be discarded. Electrical activation was achieved by applying voltage (80 V for 100 mS, 2 pulses) across a plexiglas chamber containing jellied or dejellied eggs in 20% Steinberg’s solution. Prick activation was performed on dejellied eggs with a glass needle either towards the edge of the animal cortex or, when the spindle was to be removed, beside the first polar body at the animal pole, removing a small amount of cortex and cytoplasm from this area (method adapted from Porter, 1939).

In order to compare experiments, all results are expressed on a normalized time (NT) scale (Scharf and Gerhart, 1983). Insemination is designated 0 NT, and first cleavage 1.0 NT. 3 min were added to the activated egg times to account for the time of sperm passage through the jelly (Stewart-Savage and Grey, 1982).

Fixation and microtubule staining of whole eggs was performed exactly as in Elinson and Rowning (1988). Eggs for sectioning were fixed overnight at −20 °C in methanol containing 20% DMSO (Dent and Klymkowsky, 1989) or methanol containing 1 % formaldehyde (from 37 % formaldehyde solution). They were rehydrated through 70% and 30 % methanol/PBS (>15 min each) to PBS, extracted in PBS containing 0.25% Triton X-100 for 30 min, washed in PBS and then dehydrated (30%, then 70%, then 100% ethanol each for >15min, then fresh ethanol overnight). The eggs were transferred to polyester wax (99 % poly(ethylene glycol 400) disterate, 1 % cetyl alcohol, Steedman, 1957) at 40 °C, and left to infiltrate for 1 –3 h with one change of wax. 7 μm thick sections were cut in the plane of the animal-vegetal axis, and of the sperm entry point where appropriate, expanded using 0.1% Triton X-100, dried onto coverslips coated with Gatenby’s glue (0.1% chrom alum, 1.3% gelatin in 6% acetic acid, 25% ethanol), dewaxed with 100% ethanol, rehydrated through 70 % and 30 % ethanol, blocked with 3 % dried milk powder/0.1% Triton X-100 in PBS and rinsed in PBS.

Anti-tubulin staining was achieved by incubating sections overnight at 4 °C with a mouse monoclonal anti-/Ftubulin (N.357: Amersham) diluted 1/500, or a rat monoclonal anti-a-tubulin specific for the tyrosinated form (YL1/2: gift of Dr J. Kilmartin, Kilmartin et al. 1982) diluted 1/1000, washing with several changes of 0.1% Tween 20 in PBS, and then incubating with DTAF-goat anti-mouse or DTAF-rat IgG+IgM (Jackson Immunoresearch) for 2h at room temperature, washing as before, mounting in 50% v/v glycerol/PBS containing 1.5% n-propyl gallate pH 8.3 and viewing on a Leitz Orthoplan microscope. In order to compare staining patterns between different regions of a single section, identical exposure and printing times were used for photographs.

N.357 was used previously for staining frog egg microtubules (Elinson and Rowning, 1988). YL1/2 gave similar staining patterns but with a lower background (compare Fig. 2I and K), probably because it does not recognize all unpolymerized tubulin (some of which is detyrosinated).

Although arrested at metaphase II of meiosis with certain characteristics of an M-phase cell (Karsenti et al. 1984), the unfertilized Xenopus egg has high levels of polymerized tubulin including definitive microtubules (Elinson, 1985; Jessus et al. 1987). Previous studies using immunocytochemistry and electron microscopy demonstrated the presence of microtubules in the unfertilized egg cortex and meiotic spindle but did not reveal their presence or distribution in the cytoplasm (see Elinson and Houliston, 1990 for review). Our method of antibody staining on polyester wax sections of methanol-fixed eggs revealed microtubules distributed evenly through the cytoplasm of the unfertilized egg (Fig. 1A,B). They were present in both animal and vegetal hemispheres but appeared depleted from a small region in the immediate vicinity of the meiotic spindle (Fig. 1C). Although the meiotic spindle has a characteristic barrel shape compared with mitotic spindles and is considered anastral (Karsenti et al. 1984), this staining procedure demonstrated the presence of some long microtubules connecting the spindle poles with the cytoplasmic microtubule network (Fig. 1D,E).

In agreement with the previous biochemical data (Elinson, 1985), eggs fixed shortly after fertilization contained progressively fewer microtubules, so that no microtubules could be detected in most sections of eggs fixed at 0.2 NT. The distinct non-filamentous appearance of the fluorescence in these eggs (as seen for instance in Fig. 2A,C,E,G,I,K compared with 2B,D,F,H,J) gave us confidence that our fixation methods were not promoting microtubule polymerization.

The first microtubules to reappear during the first cell cycle in fertilized eggs were found radially arranged in a group positioned asymmetrically in the animal cytoplasm on the sperm entry side of the egg. This was assumed to be the sperm aster. As expected, it was found to enlarge dramatically to fill most of the animal hemisphere by 0.5 NT. More surprisingly, the microtubules from the astral region also spread through the vegetal cytoplasm, starting at about 0.4 NT (Fig. 2). The extension of astral fibers between the large vegetal yolk platelets differed from their growth in the animal hemisphere in that it was not accompanied by detectable rearrangement of the platelets into a radial pattern. This might explain why the vegetal fibers were not detected previously (Palacek et al. 1978; Ubbels et al. 1983; Ubbels and Vermeulen, 1986). During the period 0.4 –0.5 NT, progressively more microtubules from the region of the sperm aster penetrated the vegetal cytoplasm, creating a distinct asymmetry in the vegetal cytoplasm (in Fig. 2 compare A and B, E and F, H and I, J and K). Upon reaching the vegetal cortex the microtubules appeared to turn into the yolk-free cortical region and run parallel to the egg surface (Fig. 2B,F,H,J, 3C,D). This behaviour was distinct from that at the animal cortex, where the microtubules of the sperm aster ended abruptly and were seen running parallel to the surface only rarely (Fig. 3A,B).

The microtubules running parallel to and just below the vegetal surface of the egg correspond to those identified previously by whole-mount immunocytochemistry (Elinson and Rowning, 1988; Fig. 3D). The initial appearance of microtubules in this region was asymmetric in two ways. First, at times before 0.5 NT astral microtubules were only present on the sperm entry side of the vegetal cytoplasm so a corresponding cortical asymmetry resulted (Fig. 2E,F,J,K). Second, the microtubules approach the cortex at an oblique angle so tended to turn to run through the cortical region in the same direction, away from the sperm entry point and toward the vegetal pole (Fig. 2F,J). At about 0.5 NT, short microtubules began to appear throughout the cytoplasm. In addition, longer microtubules were observed, arranged in a radial fashion but apparently distinct from the sperm aster, being sparser and less ordered. These two populations of non-sperm aster microtubules probably correspond to the randomly arranged microtubules that appear first in the cortex on whole-mount specimens (Elinson and Rowning, 1988). Thus it seems that not all the microtubules of the vegetal cortical array arise from the sperm aster.

At times between 0.6 and 0.8 NT, many radial microtubules were found penetrating the vegetal cytoplasm on the non-sperm side, as well as the sperm side, of the egg and in the cortical region (Fig. 4A-C). Before this period, at around 0.5 NT, the two pronuclei come together and the single astral centre is replaced by two (Stewart-Savage and Grey, 1982). The arrangement of microtubules at the cortex on the non-sperm side was different from that on the sperm side, with microtubules being less ordered, sometimes bending up towards the animal pole. This was probably because the cortex would be undergoing its translocation relative to the cytoplasm by this time (Vincent et al. 1986), moving downward on the sperm side and upward on the nonsperm side of the egg.

The changing pattern of microtubules in the fertilized egg is summarized in Fig. 5A. Note that this indicates the general distribution and orientation of the microtubules. Our data does not allow’ us to decide whether single microtubules extend from the center of the animal cytoplasm to the cortex. At all times following the arrival of astral microtubules at the vegetal cortex, microtubules could be seen connecting the vegetal array and the cytoplasmic network (Fig. 4D,E), suggesting that the vegetal yolk and cortical microtubule array form a coherent mass during the period of the cortical rotation.

In artificially activated eggs, vegetal cortical microtubule arrays form despite the absence of a sperm aster. The microtubules are not always aligned in the same way as in fertilized eggs, sometimes forming spiral or cartwheel patterns (Elinson and Rowning, 1988), with corresponding abnormal cortical movements also occurring (Vincent and Gerhart, 1987). To see how the cortical array arises in activated eggs, we again examined the pattern of microtubule polymerization on sections. The microtubules again disappeared rapidly following activation. Their reappearance was delayed compared with fertilized eggs, not occurring until about 0.45 NT. This corresponds to the delayed polymerization of tubulin in activated versus fertilized eggs detected biochemically (Elinson, 1985). The first microtubules to be found were short microtubules scattered throughout the animal and vegetal cytoplasm. Radially arranged microtubules appeared shortly afterwards, seeming to emanate from the central region of the animal cytoplasm (Fig. 6A-C). The appearance of microtubules running parallel to the egg surface in the vegetal cortex was correlated with their presence in the cytoplasm just below (Fig. 6E,F). This suggests that most of the cortical microtubules in activated eggs derive from the extension of radial ones. It should be noted that a few do form independently before the radial ones arrive there (Fig. 6D). As in fertilized eggs, links were evident between the radial microtubules and the accumulating layer of microtubules in the vegetal cortex throughout the second half of the cell cycle (Fig. 6F-H). The progression of microtubules through the vegetal cytoplasm and into the cortical region in activated eggs was thus similar to that in fertilized eggs, but without the asymmetry provided by the sperm aster (see Fig. 5A,B). It should be noted that any minor asymmetry would probably not be detected since, in the absence of a sperm entry point, the eggs could not be oriented easily for sectioning with respect to the direction of rotation.

Frog eggs are not thought to contain definitive centrioles (see Karsenti et al. 1984). In fertilized eggs, the female pronucleus becomes associated with the sperm aster without forming an aster itself (StewartSavage and Grey, 1982). In parthenogenetically activated eggs, material associated with the female pronucleus is able to nucleate microtubule growth and promote limited pronuclear migration (Subtelney and Bradt, 1963; Manes and Barbieri, 1977; Sambuichi, 1981; Ubbels et al. 1983; Ubbels and Vermeulen, 1986). It also appears to be responsible for promoting abortive furrowing in place of a true first cleavage since, if eggs are enucleated by removing the metaphase spindle at the time of activation, abortive furrowing does not occur (at least not for many hours after control activated eggs, Briggs and King, 1953). These observations suggest that some component associated with the female pronucleus is responsible for organizing microtubules in such a way as to promote furrowing in activated eggs, and may thus be the origin of the radial array of microtubules that we saw m activated eggs. To test this idea, we examined the pattern of microtubule polymerization in enucleate eggs prepared by removal of the spindle at the time of activation.

Dejellied Xenopus eggs were activated by using a glass needle to remove cytoplasm in the spindle region and fixed at times when control prick-activated eggs had both radially arranged microtubules in the vegetal cytoplasm and cortical microtubule arrays. Success of enucleation was assessed by failure to form furrows in the same experimental batch. Only batches with <30 % furrowing were used. Vegetal microtubule arrays formed in 18 out of 19 enucleated eggs examined, with radial microtubules also clearly visible through the vegetal cytoplasm in most of them (Fig. 7A). Furthermore, radial cytoplasmic vegetal microtubules could be detected in unfurrowed enucleated eggs at times when all control prick-activated eggs had undergone furrowing (1.58 NT) (Fig. 7B,C), although the cortical array was hard to distinguish at this time. It thus appears that the egg can organize both a radial cytoplasmic pattern and a vegetal cortical microtubule array in the absence of all sperm and egg nuclear components.

The suspicion that the array of aligned microtubules below the vegetal surface of the egg is involved in the process of cortical rotation is based on the sensitivity of the rotation to microtubule disrupting agents, the timing of formation and orientation of the array and evidence that the force for rotation is located close to the vegetal surface of the egg (see Elinson and Rowning, 1988; Elinson, 1989; Gerhart et al. 1989). The observations described here strengthen this idea and have further implications for the mechanisms by which the array becomes oriented and by which the force for the cortical rotation is generated.

In all activated eggs, we found that microtubules extend in a radial fashion through the animal, then the vegetal halves of the egg, and that microtubules accumulate rapidly beneath the vegetal surface as these microtubules reach the vegetal cortex. The yolk-free region of the vegetal cortex appears to favour microtubule polymerization in comparison to the animal cortex, which is known to have distinct organization, including a thick, pigmented, contractile layer (Merriam et al. 1983; see Elinson and Houliston, 1990 for review). Some of the cortical microtubules appear to be the ends of microtubules that have polymerized outward through the vegetal cytoplasm, while others probably arise de novo as cytoplasmic conditions change to favour microtubule growth (see Elinson, 1985). The rapid polymerization of tubulin during the first cell cycle may be regulated by the phosphorylation of a specific microtubule-associated protein (Gard and Kirschner, 1987; Verde et al. 1990).

In fertilized eggs, a distinct bias is introduced to the pattern of radial polymerization by the sperm aster. It was surprising to see a direct link between the growing sperm aster and the array since previous studies gave the impression that astral rays extended only within the animal cytoplasm (Palacek et al. 1978, Ubbels et al. 1983). Since there is little disturbance to the pattern of yolk platelets in the vegetal half, the fine radial microtubules could easily be missed. The eccentric position of the sperm aster appears to aid the alignment of the vegetal array in two ways. First, the microtubules on the sperm entry side of the egg reach the vegetal cortex before those on the opposite side. Second, the first astral microtubules to arrive at the vegetal cortex do so at an angle favouring their continued growth away from the sperm entry point, toward the vegetal pole. The last microtubules to reach the cortex tend to bend up towards the animal pole (also seen by Ubbels et al. 1983) and probably have little influence on the formation of the array, which is complete by this time. Once the cortex has started to move relative to the cytoplasm, microtubule ends arriving at the vegetal cell surface will tend to become bent in the direction of movement, thus reinforcing the alignment of the array.

It is hard to say how the vegetal cortical array becomes oriented in activated eggs. Many of the radially arranged microtubules in the cytoplasm grow outward toward the vegetal cortex, but we do not know for certain whether the formation of the cortical array depends on their presence. The relationship between the two is suggested by the correlation between the appearance of microtubules in the cytoplasm immediately below the cortex and the formation of the array. If many of the cortical microtubules do indeed derive from outwardly growing cytoplasmic ones, random differences in the cytoplasmic organization may be able to introduce sufficient asymmetry in their growth to mimic the situation in fertilized eggs. It seems that this is not always the case since symmetric spiral or cartwheel formations sometimes arise in the vegetal cortex of activated eggs (Elinson and Rowning, 1988). The orientation of the microtubule array (Zisckind and Elinson, 1990) and of the dorsal-ventral axis (Ancel and Vintemberger, 1948; Black and Gerhart, 1985) can be aligned by centrifuging or tilting the egg prior to the time of formation of the microtubule array. These treatments may either introduce cytoplasmic asymmetries that affect the growth of the radial microtubules or start movement of the cytoplasm relative to cortex and thus align the microtubules as they arrive in the cortical region.

An important finding of this study is that microtubules from the cytoplasm are connected with microtubules of the cortical array throughout the second half of the first cell cycle. The cytoplasm at this time acquires a solid consistency (Elinson, 1983), enabling it to move as a unit with respect to the cortex, and these observations suggest that those vegetal array microtubules connected to the cytoplasmic ones will move with it. Thus the shear force responsible for the rotation would lie between the microtubules of the array and the cortex proper, or possibly between cytoplasmic linked microtubules and other microtubules present in the array. In addition, the occurrence of directed polymerization raises the possibility that some of the force for the cortical rotation could be provided by the polymerization itself. Astral growth has previously been proposed to move pronuclei away from the cortex towards the egg centre (Subtelny and Bradt, 1963; Manes and Barbieri, 1977). In this case, rapid microtubule growth in the subcortical region might create force between the cortex and cytoplasm. Such a phenomenon has been observed in sea urchin eggs, an extensive rotation of cytoplasm relative to the cortex being accompanied similarly by polymerization first of a large microtubule aster and then a spirally arranged cortical array (Harris et al. 1980; Schroeder and Battaglia, 1985). This process is not associated with normal development, the microtubule polymerization being promoted by experimental warming of the egg. In the frog egg, the force of microtubule polymerization may provide an initial shove. It could provide sustained force if localized in some way, since there is net depolymerization of egg microtubules during the second half of the cell cycle (Elinson, 1985).

Another proposal for force generation is that the microtubules of the array may act as tracks along which translocation could occur mediated by a microtubule motor molecule (Vincent et al 1987; Elinson and Rowning, 1988). This would require all or most of the microtubules to be onented with a common polarity The polarity of the cortical microtubules is not yet known. Our data suggest that they may have their fast growing (plus) end towards the future dorsal side of the egg. This conclusion is based on the observation that microtubules growing outward from the cytoplasm appear to contribute to the array in all activated eggs. The plus ends will tend to reach the vegetal cortex first, whether or not the microtubules belong to the sperm aster, because of the presence of a factor (XMAP) that specifically promotes plus end assembly (Gard and Kirschner, 1987). Since the microtubules appear to run in the direction of cortical rotation (see Fig. 5), an anterograde (kinesin-like) motor molecule would be the most likely candidate to carry the cortex towards the dorsal side of the egg. Alternatively, other microtubule-associated force-generating molecules related to dynein or dynamin could promote sliding between different populations of microtubules (see Vallee and Shpetner, 1990 for a review of the characteristics of different microtubule motor molecules).

The presence of non-spindle microtubules in the unfertilized egg (Elinson, 1985; Jessus et al. 1987; Huchon et al. 1988; this study) is unusual for a metaphase cell. Microtubules were found distributed evenly throughout the cytoplasm, except in the immediate vicinity of the meiotic spindle. This may reflect the massive volume of the egg, with the influence of condensed chromatin on microtubule organization (Karsenti et al. 1984) being only local. The rapid loss of microtubules associated with the transition to anaphase is the opposite of the change observed in other cell cycles (see Kirschner and Mitchison, 1986), probably being due to the sharp rise in the level of cytoplasmic calcium associated with activation (see Elinson, 1986). In contrast, microtubule dynamics typical of metaphase and interphase cells are observed in cytoplasmic extracts made by high speed centrifugation of unfertilized and activated eggs, respectively (Verde et al. 1990). Thus the Xenopus egg shows both typical and atypical aspects of microtubule organization in relation to its cell cycle state.

In this study, we found that microtubules polymerized in a radial manner in all activated eggs, with the centre of polymerization lying somewhere in the animal cytoplasm. This organization did not depend on the sperm centriole, or on any nucleating structure derived from the metaphase II spindle. Radially arranged microtubules have been observed previously in the full-grown oocyte, running between the germinal vesicle (oocyte nucleus) and the cell periphery (Palaèek et al 1985). It appears that some components from inside or just next to the germinal vesicle facilitate aster formation in the oocyte cytoplasm (Heidemann and Kirschner, 1978) and that microtubule nucleating material is concentrated around the base of the germinal vesicle, since a large microtubule aster emanates from this region during oocyte maturation (Jessus et al. 1986). The fate of this material is unclear. A yolk-free region thought to derive from the germinal vesicle remains centrally located in the animal cytoplasm after activation (see Ubbels et al. 1983). Perhaps this material provides a focus for microtubule polymerization in activated eggs, either independently or through association with the female pronucleus. It may also play some role in microtubule organization in the fertilized egg, since we observed radial microtubules independent of the sperm aster perhaps corresponding to the ‘wave front’ described by Ubbels and Vermeulen (1986).

This study was supported by a grant to RPE from NSERC, Canada. EH acknowledges the receipt of a SERC/NATO postdoctoral fellowship.