The Drosophila wispy Gene Is Required for RNA Localization and Other Microtubule-Based Events of Meiosis and Early Embryogenesis
A. E. Brent, A. MacQueen, T. Hazelrigg

Abstract

RNAs are localized by microtubule-based pathways to both the anterior and posterior poles of the developing Drosophila oocyte. We describe a new gene, wispy, required for localization of mRNAs to both poles of the egg. Embryos from wispy mothers arrest development after abnormal oocyte meiosis and failure of pronuclei to fuse. Our analysis of spindle and chromosome movements during meiosis reveals defects in spindle structures correlated with very high frequencies of chromosome nondisjunction and loss. Spindle defects include abnormally shaped spindles, spindle spurs, and ectopic spindles associated with lost chromosomes, as well as mispositioning of the meiosis II spindles. The polar body nuclei do not associate with their normal monastral arrays of microtubules, the sperm aster is reduced in size, and the centrosomes often dissociate from a mitotic spindle that forms in association with the male pronucleus. We show that wispy is required to recruit or maintain known centrosomal proteins with two types of microtubule organizing centers (MTOCs): (1) the central MTOC that forms between the meiosis II tandem spindles and (2) the centrosomes of the mitotic spindle. We propose that the wispy gene product functions directly in several microtubule-based events in meiosis and early embryogenesis and speculate about its possible mode of action.

RNAs are localized to subcellular compartments in many different types of cells, including neurons, tissue culture cells, budding yeast, and the oocytes of many species (reviewed in St. Johnston 1995; Bashirullaet al. 1998; Hazelrigg 1998). In Drosophila oogenesis, several mRNAs are localized to the anterior and posterior poles of the oocyte. Localization of these mRNAs in the developing oocyte is a key regulatory event in anterior-posterior patterning of the embryo (reviewed in St. Johnston and Nüsslein-Volhard 1992). bicoid (bcd) mRNA is localized to the anterior pole of the oocyte, where it is subsequently translated, forming an anterior-posterior protein gradient in the young embryo (Berlethet al. 1988; Driever and Nüsslein-Volhard 1988). The bicoid protein is a transcription factor that directs the formation of anterior embryonic structures, in a concentration-dependent fashion. At the posterior pole of the oocyte, localization of oskar (osk) mRNA (Ephrussiet al. 1991; Kim-Haet al. 1991) and nanos (nos) mRNA (Wang and Lehmann 1991; Gavis and Lehmann 1992) is essential for localizing their protein products, which play pivotal roles in abdomen specification and germ cell formation and function (Hulskampet al. 1989; Irishet al. 1989; Ephrussi and Lehmann 1992; Kobayashiet al. 1996; Forbes and Lehmann 1998).

Microtubules are required for localizing RNAs to both the anterior and posterior poles of Drosophila oocytes (Pokrywka and Stephenson 1991, 1995; Clarket al. 1994). The microtubule cytoskeleton in Drosophila egg chambers has sufficient polarity to supply a framework for differential distributions of RNAs within the oocyte, but other localized factors may also be involved in anchoring RNAs at the poles (Theurkaufet al. 1992; Theurkauf and Hazelrigg 1998). Mutations that alter the overall oocyte microtubule architecture also perturb RNA localization (reviewed in St. Johnston 1995). These observations have supported models for localization of RNAs in Drosophila oocytes based on motor-driven transport of RNA-protein complexes along microtubules (Wilhelm and Vale 1993; Clarket al. 1994; St. Johnston 1995; Theurkauf and Hazelrigg 1998).

We report here the characterization of a new gene required for localization of RNAs to both the anterior and posterior poles of the embryo, the wispy gene. Embryos from mutant mothers have less bcd mRNA at their anterior poles, and the mRNA appears diffused compared to wild-type embryos. The effects of wispy mutations are not confined to bcd mRNA, or to one pole of the embryo, as localization of other polar mRNAs is also perturbed. Embryos from mutant mothers arrest development very early, after abnormal oocyte meiosis and failure of pronuclei to fuse. Our analysis reveals specific defects in the structures and functions of meiotic and mitotic spindles and other microtubule structures present in the young embryo. The phenotype of wispy mutants provides strong genetic evidence for a role for microtubules in RNA localization in Drosophila oogenesis and indicates that wispy plays a direct role in both mRNA localization and other microtubule-based cellular events.

MATERIALS AND METHODS

Immunohistochemistry of embryos: Embryos were collected from apple juice agar plates every 5 min or every 3 hr, dechorionated in 75% Clorox bleach for 1–2 min, rinsed with PBT (1% Triton X-100 in PBS), and fixed in a 1:1 mixture of fixative (97% methanol + 3% 0.5 m EGTA, pH 7.4) and heptane, with gentle rocking for 5 min. Subsequently the embryos were shaken vigorously for 30–60 sec to remove vitelline membranes. The heptane layer was removed and replaced by an equal volume of cold methanol fix, and the embryos were rocked at 4° for 2 hr, after which they were gradually rehydrated into PBT and stored at 4° until antibody staining.

α-Tubulin, sperm tail, and DNA labeling: Embryos (5–15 min and 0–3 hr) were labeled with either an anti-α-tubulin antibody or an anti-sperm tail antibody and with a DNA binding dye according to the following procedure. The embryos were rocked overnight in PBT, blocked for 1 hr in 1% bovine serum albumin (BSA; ICN Biomedicals, Inc.) in PBT, and then incubated (with gentle rocking) in a 1:100 dilution in PBT of either mouse monoclonal anti-α-tubulin (Sigma, St. Louis) or a mouse monoclonal anti-sperm tail antibody (gift of T. L. Karr; Karr 1991), for 3 hr. The embryos were then washed in PBT (3 × 3 min) and rocked for 2 hr in the dark in a 1:200 dilution in PBT of preabsorbed lissamine rhodamine (LRSC)-conjugated secondary antibody [AffiniPure Fab Fragment goat anti-mouse IgG (H+L); Jackson ImmunoResearch Laboratories, West Grove, PA], followed by washes (3 × 3 min) with PBT. DNA staining was done with either Sytox green (Molecular Probes, Eugene, OR) or 4′,6-diamidino-2-phenylindole (DAPI; Boehringer Mannheim, Indianapolis). For Sytox green staining, the embryos were placed in a 1:10 dilution of 10 mg/ml RNase in PBT and were incubated at 37° for 2 hr. The RNase solution was removed and replaced with a 1:300 dilution in PBT of Sytox green (30 μm stock in PBS), and the embryos were rocked in the dark for 25–30 min, at room temperature, followed by washes (3 × 3 min) with PBT. For DAPI staining, the embryos were placed in a 1:300 dilution of DAPI (0.1 mg/ml stock in Tris, pH 7.2) in PBT and rocked in the dark for 6 min, followed by washes (3 × 3 min) in PBT. The embryos were mounted in Fluoromount G (Southern Biotechnology Associated) on slides under a glass coverslip sealed with nail polish.

α-Tubulin, γ-tubulin, CP60, CP190, and DNA labeling: Embryos (5–15 min and 0–3 hr) were triply labeled with an anti-α-tubulin antibody, either an anti-γ-tubulin, anti-CP60, or anti-CP190 antibody, and with DAPI, as follows. The embryos were fixed, permeabalized and blocked as described above, and incubated in a 1:100 dilution of rabbit polyclonal anti-γ-tubulin antibody (directed against the C terminus of Drosophila γ-tubulin 37C; gift of Yixian Zheng), rabbit polyclonal anti-CP60, or rabbit polyclonal anti-CP190 (gifts of Karen Oegema). The embryos were rocked in the antibody for 3 hr, followed by washes (3 × 3 min) in PBT. For secondary labeling, the embryos were incubated in a 1:150 dilution in PBT of preabsorbed fluorescein-5-isothiocyanate (FITC)-conjugated secondary antibody [AffiniPure goat anti-rabbit IgG (H+L); Jackson ImmunoResearch Laboratories] for 2 hr, rocking, followed by washes (3 × 3 min) in PBT. This was followed by incubation in a 1:100 dilution in PBT of a mouse monoclonal anti-α-tubulin antibody (Sigma) for 3 hr, rocking. After three more washes in PBT (3 min each), the embryos were incubated in a 1:100 dilution in PBT of preabsorbed LRSC-conjugated secondary antibody [AffiniPure Fab Fragment goat anti-mouse IgG (H+L); Jackson ImmunoResearch Laboratories] for 2 hr, followed by washes (3 × 3 min) in PBT. The DNA was stained with DAPI, as above, and the embryos were mounted as above.

Immunohistochemistry of mature oocytes: Stage 14 oocytes were isolated and fixed using the procedure developed by Theurkauf and Hawley (1992). This method was designed to avoid premature activation of oocytes during fixation. Our procedure had one modification: oocytes were obtained from hand-dissected ovaries rather than mass isolated by blender. After fixation, the oocytes were stored at 4° in PBT until antibody staining.

α-Tubulin and histone labeling: Stage 14 oocytes were rocked in PBT for 3 hr and blocked as above and then incubated in a 1:100 dilution in PBT of mouse monoclonal anti-histone antibody (Chemicon, Temecula, CA) for 3 hr with rocking, followed by washes (3 × 3 min) in PBT. A 1:100 dilution in PBT of preabsorbed LRSC-conjugated secondary antibody [AffiniPure Fab Fragment goat anti-mouse IgG (H+L); Jackson ImmunoResearch Laboratories] was added, and the oocytes were rocked for 2 hr, followed by washes in PBT (2 × 3 min). The oocytes were then incubated for 1 hr while rocking in a 1:100 dilution of normal mouse serum (Sigma) in PBT, washed (2 × 3 min) in PBT, and rocked for 3 hr in a 1:100 dilution in PBT of mouse monoclonal anti-α-tubulin FITC conjugate (Sigma), followed by washing (3 × 3 min) in PBT, and mounted as above.

In situ hybridization: Embryos were collected hourly from apple juice agar plates, dechorionated in 50% Clorox bleach for 1–2 min, rinsed with PBT, and then transferred to a mixture of one part 4% paraformaldehyde in PBS and four parts heptane, and shaken gently for 20 min. The fix was then removed, and a 1:1 mixture of methanol and heptane was added. The embryos were shaken vigorously for 30–60 sec to remove the vitelline membranes, washed (3 × 3 min) in PBT, gradually dehydrated in methanol, and stored at −20° overnight.

Isolation and fixation of ovaries: Ovaries were hand dissected in PBT and were fixed in a mixture of eight parts 4% paraformaldehyde in PBS, one part Clorox bleach, and one part DMSO for 15 min while rocking. The ovaries were then washed in PBT (3 × 3 min), gradually dehydrated in methanol, and stored at −20° overnight.

In situ hybridization: The embryos and ovaries were gradually rehydrated into PBT and washed in PBT (2 × 5 min each). The embryos were postfixed in 4% paraformaldehyde in PBS for 15 min, followed by five washes in PBT (3 min each), and then incubated in 40 μg/ml proteinase K (Boehringer Mannheim) in PBT for 5 min (bcd, hts) or 7 min (osk, nos) followed by two quick washes in 20 mg/ml glycine in PBT and one wash in PBT. This was followed by a further fixation in 4% paraformaldehyde for 20 min and then by washes (5 × 3 min each) in PBT. Next, the embryos were washed in a 1:1 mixture of PBT and hybridization buffer (50% deionized formamide; 5× SSC; 100 μg/ml sonicated, boiled salmon sperm DNA; 100 μg/ml tRNA; 50 μg/ml heparin; 0.1% Tween 20) for 5 min, followed by preincubation in 100% hybridization buffer for at least 1.5 hr. Hybridization was overnight in 10–20 μl of heat-denatured digoxigenin-labeled DNA probe in hybridization buffer. The probe was synthesized from plasmid using a Dig-Nick kit (Boehringer Mannheim); its final concentration was approximately 0.1 μg/ml of probe in 50 μl of hybridization buffer. Following hybridization, the embryos were washed (3 × 20 min) in hybridization buffer. The hybridization buffer was replaced gradually with PBT, and the embryos were washed (5 × 3 min) in PBT, rocked for 1 hr in a 1:2000 dilution in PBT of preabsorbed anti-Dig alkaline phosphatase-conjugated antibody (Boehringer Mannheim), washed in PBT (5 × 3 min), and equilibrated in freshly prepared AP buffer (100 mm NaCl; 50 mm MgCl2; 100 mm Tris pH 9.5; 0.1% Tween 20). Staining was done in 1 ml AP buffer, with the addition of 4.5 μl 4-nitro blue tetrazolium chloride (Boehringer Mannheim) and 3.5 μl 5-bromo-4-chloro-3-indolyl-phosphate (Boehringer Mannheim). The reaction was stopped by washing in PBT. The embryos were gradually transferred to 60% glycerol, allowed to equilibrate in the 60% glycerol, and mounted on slides in a drop of 60% glycerol. The ovaries were treated the same with the exception that the first postfix in 4% paraformaldehyde was skipped, and proteinase K was used at a concentration of 50 μg/ml with 10 min of incubation.

Kinesin-galactosidase staining: Ovaries were dissected in Drosophila Ringer solution (182 mm KCl; 46 mm NaCl; 3 mm CaCl2; 10 mm Tris-HCl, pH 7.2; Ashburner 1989), transferred to glutaraldehyde fixative (1% glutaraldehyde in 50 mm sodium cacodylate buffer, pH 7.3), and rocked for 20 min, followed by washes (3 × 5 min) in PBT. The ovaries were then placed in staining buffer [0.05 m K3(Fe(CN)6), 0.05 m K4(Fe(CN)6) in PBS] with X-gal stock solution (8% in N,N-dimethylformamide) added to a final concentration of 0.2% and stained at 37° for 2 hr in the dark, followed by overnight staining at room temperature. The next day the ovaries were washed (3 × 5 min) in PBT and mounted on slides in 60% glycerol.

RESULTS

The wispy gene was identified in a collection of X-linked female sterile mutations isolated by Mohler (1977) and Mohler and Carroll (1984). We analyzed a subset of their maternal-effect mutations that cause early embryonic arrest of progeny, to determine whether any caused mislocalization of bcd mRNA. We reasoned that some genes required for RNA localization may also be essential for other very early embryonic functions, unlike genes previously identified for bcd RNA localization, such as exuperantia, swallow, and staufen, which when mutated produce embryos that die late in embryogenesis (reviewed in St. Johnston and Nüsslein-Volhard 1992). Our analysis showed that the gene designated fs(1)M19 is required for normal bcd mRNA localization. We have named this gene wispy (wisp) because of its mutant effects on spindle structures, described below.

Genetics: Mohler previously mapped wisp (M19) to polytene interval 10A2-11B1-2 (Mohler 1977; Mohler and Carroll 1984). Using a set of overlapping deficiencies, we refined the map position to 10F1-F7, on the basis of the failure of wisp alleles to complement Df(1)KA6 (10E1;11A8) and Df(1)RA47(10F1;10F10), but not Df(1)N105(10F7;11D10). Three separate wisp alleles were analyzed in this study: wisp12-3147, wisp14-1299, and wisp11-600, and all three produced similar phenotypes. Furthermore, females heterozygous for a given wispy allele and a deficiency that removes the wispy gene [wisp/Df(1)RA47 females] produced phenotypes indistinguishable from homozygous females (wisp/wisp females), suggesting that these alleles may represent the null condition.

Figure 1.

Localized mRNAs are present at reduced levels and are dispersed in wisp embryos. Whole mount in situ hybridizations showing mRNAs localized in embryos from wild-type (WT) and wisp mothers (wisp). (a) bcd mRNA in a wild-type embryo. The mRNA is localized to the anterior, dorsal pole of the embryo. (b) bcd mRNA in an embryo from a wisp11-600 mother. There is less RNA at the anterior pole, and the signal is dispersed compared to wild-type embryos. (c) hts mRNA in a wild-type embryo. The mRNA is anteriorly localized, but extends more posterior than bcd mRNA. (d) hts mRNA in an embryo from a wisp12-3147 mother. The mRNA occupies a broader domain than in wild-type embryos. (e) osk mRNA in a wild-type embryo. The mRNA is localized to the posterior pole and also present at low levels throughout the embryo. (f) osk mRNA in an embryo from a wisp14-1299 mother. There is less apparent osk mRNA at the posterior pole and throughout the embryo. The defect in osk RNA localization is variable; some embryos appeared normal. (g) nos mRNA in a wild-type embryo. Like osk mRNA, nos mRNA is localized to the posterior pole and also present at low levels throughout the embryo. (h) nos mRNA in an embryo from a wisp12-3147 mother. There is less nos mRNA at the posterior pole and less signal throughout the embryo.

RNA localization defects: We analyzed the distribution of bcd mRNA in 0- to 1-hr-old embryos, by whole-mount in situ hybridization. Figure 1, a and b, compares bcd mRNA localization in wild-type embryos and embryos from wisp/Df(1)RA47 mothers (hereafter referred to as wisp embryos). Each of the three wisp alleles had similar effects on bcd mRNA localization. In wild-type embryos, bcd mRNA is highly concentrated at the anterior pole, in a dorsal position (Figure 1a). In contrast, the amount of bcd mRNA at the anterior pole appears reduced in wisp embryos (Figure 1b) and is spread out diffusely over the entire anterior cortex, instead of being dorsally positioned.

View this table:
TABLE 1

Developmental stages of 5- to 15-min-old embryos

To determine whether the effects of wisp mutations on RNA localization are limited to bcd mRNA, we also examined another anteriorly localized mRNA, hu-li tai shao (hts; Yue and Spradling 1992; Zaccai and Lipshitz 1996), and two posteriorly localized mRNAs, oskar (osk) and nanos (nos; Ephrussiet al. 1991; Kim-Haet al. 1991; Wang and Lehmann 1991; Gavis and Lehmann 1992). In the case of hts, there appears to be less mRNA in wisp embryos, and it is more diffused compared to wild-type embryos (Figure 1, c and d). The osk mRNA signal is variable; in some embryos there is less localized osk mRNA (Figure 1, e and f), whereas in others it appears normal. The amount of nos mRNA also appears reduced (Figure 1, g and h). Since localization of nos mRNA requires the action of localized Osk product (Ephrussi and Lehmann 1992), the defects in nos mRNA localization may be indirect, through reduced osk product. In summary, the effects of wisp mutations on RNA localization are not limited to one pole of the embryo.

We next examined bcd and osk mRNA in developing egg chambers to determine when RNA localization defects first occur during oogenesis in wisp mutants. In the oocyte, both RNAs are localized correctly during the stages of oogenesis that we could confidently analyze (stages 1–10; data not shown). In the case of bcd mRNA, localization to discrete apical patches within the nurse cell cytoplasm appeared more variable than that in wild type, but the variability made analysis of the nurse cell distribution less certain. Egg chamber morphology and oocyte determination appeared normal by visual inspection of wisp mutant ovaries. To confirm that microtubule polarity in the oocyte was normal, we examined the localization of a kinesin-β-galactosidase fusion, which is targeted to the posterior of wild-type stage 8–9 oocytes (Clarket al. 1994). We found that the fusion protein is localized normally in wisp egg chambers (not shown). This result indicates that the overall pattern of microtubule polarity is normal in wisp oocytes. However, subtle defects may exist in microtubule organization in the oocyte, or the nurse cells, that would not be detected by this assay. Since polar RNAs are localized during oogenesis in wisp mutants, we conclude that the phenotype, reduced and dispersed RNAs in the young embryo, results from lower efficiency of RNA localization or a failure to maintain the localized state. It is also possible that premature degradation of polar mRNAs occurs in wisp embryos.

Early embryonic arrest in wisp embryos: The reader is referred to Foe et al. (1993) for a detailed description of the early events of Drosophila development. Meiosis is completed shortly after ovulation (Sonnenblick 1950), concomitant with the formation of an extensive sperm aster nucleated by the sperm-derived centrosome. After wild-type embryos complete meiosis II (MII), the female products of meiosis enter a new cell cycle. The female interphase pronucleus, the female haploid nucleus that is closest to the sperm aster, migrates toward the male pronucleus. The male and female pronuclei, once opposed, now assemble on the first mitotic spindle. Final fusion of the parental genomes occurs at telophase of the first mitotic cycle (Callaini and Riparbelli 1996). The three other products of female meiosis, the polar bodies, arrest in a metaphase state (Rabinowitz 1941) and associate with radiating arrays of microtubules. Typically some or all polar bodies fuse, so wild-type embryos do not usually contain three separate polar bodies (Page and Orr-Weaver 1997). Embryos undergo 13 nuclear mitotic cycles without cytoplasmic division, and cellularization ensues during cycle 14.

Eggs were collected every 3 hr from wild-type (+/+) and wisp mothers [wisp/Df(1)RA47], fixed, and fluorescently labeled with an antibody to α-tubulin to visualize the spindles and with Sytox green or DAPI stains to detect the chromosomes. The wild-type embryos (N = 143) were in various stages of development, ranging from completion of meiosis I to early gastrulation. In marked contrast, embryos from wisp mothers were either completing meiosis or had arrested following the completion of meiosis. This was true for embryos from wisp12-3147/Df (N = 119), wisp11-600/Df (N = 102), and wisp14-1299/Df (N = 112) mothers. Meiotic defects occur in wisp mutants (see below); hence, it was possible that some embryos arrested in meiosis I or II. To examine larger numbers of younger embryos, eggs were collected rapidly every 5 min and fixed within the following 10 min (Table 1). Analysis of these embryos showed that meiosis II is usually completed in wisp embryos, since we often observed the four products of meiosis in interphase. In no case was pronuclear fusion observed in wisp embryos. The arrested phenotype resembled that seen in 0- to 3-hr-old embryos.

Figure 2.

Embryos from wisp mothers arrest development after failure of pronuclei to fuse. To study the wisp arrested phenotype, embryos were fixed and labeled with an antibody to α-tubulin to visualize the spindles. (a) A typical wild-type embryo in the fifth mitotic cycle. (The polar bodies cannot be seen in this focal plane.) (b) An embryo from a wisp12-3147 mother. Meiosis II has been completed, and the four products of female meiosis have entered new nuclear division cycles (two have apparently fused) and assembled on three meiotic-like bipolar spindles. The male pronucleus has also entered a new division cycle and is associated with a mitotic spindle, from which the centrosomes are dissociating.

A typical wild-type embryo undergoing cleavage is shown in Figure 2a. The polar bodies lie near the cortex and are not visible in this plane of focus. In contrast, wisp embryos arrest after completion of meiosis II, before pronuclear fusion. Labeling with anti-sperm tail antibody (Karr 1991) confirmed that wisp eggs are fertilized (data not shown). The female meiotic products and the sperm pronucleus enter interphase, but the female pronucleus does not migrate toward the male pronucleus, and the pronuclei are never observed opposed to each other. Figure 2b shows the characteristic arrested phenotype of wisp embryos: one elongated mitotic spindle associated with the male pronucleus, with one or both centrosomes displaced in the nearby cytoplasm, and one, two, three, or four long, wispy, acentriolar spindles associated with the products of female meiosis (like polar bodies, these nuclei often fuse). This phenotype is seen in very young embryos from short egg collections as well as in embryos that were collected and aged for 1 hr (data not shown), indicating that this phenotype occurs rapidly and is the arrested state of wisp embryos.

A comparison of the wild-type first-division mitotic spindle and the wisp mitotic spindle associated with the male pronucleus revealed differences in size and shape, microtubule density, number and distribution of chromosomes, and position of the centrosomes (Figure 3, a–d). Unlike a normal first-division mitotic spindle (Figure 3a), the wisp mitotic spindle has certain meiotic-like characteristics (Theurkauf and Hawley 1992): the density of microtubules is highest in the center of the spindle, and it is long and tapered like a meiotic spindle (Figure 3c). Unlike the chromosomes associated with the wild-type metaphase spindle (Figure 3b), the chromosomes of the wisp mitotic spindle are usually spread out unevenly across the long axis (Figure 3d), without separation of individual chromatids. The chromosomal array on the wisp mitotic spindle does not resemble either metaphase or anaphase and may signal a defect in either stage of the nuclear division cycle.

Normally, polar bodies in wild-type embryos are associated with a monastral microtubule array, and the chromosomes are arranged in a radial pattern with centromeres oriented toward the center of the structure. The female meiotic products often fuse, so that typically only one polyploid polar body is present in an embryo (Page and Orr-Weaver 1997). The wild-type polar body shown in Figure 3, e and f, illustrates the typical pattern of chromosomes and associated microtubules. In contrast, in wisp embryos the four products of female meiosis assemble on one, two, three, or four meiotic-like, acentriolar bipolar spindles. These spindles have the microtubule density characteristics of meiotic spindles and also have structural abnormalities such as bent, twisted ends and occasional spurs associated with migrant chromosomes. Figure 3, g and h, shows a wisp spindle associated with three products of female meiosis. The chromatids have separated and are spread out atypically across the spindle, and spindle spurs extend to displaced chromosomes.

In wisp embryos, one or both centrosomes of the mitotic spindle are often detached from the spindle poles and mispositioned nearby in the cytoplasm (Figures 2b and 3c). In wild-type mitotic spindles, astral microtubules radiate outward from the centrosomes, becoming especially pronounced at anaphase and telophase. In the case of wisp embryos, all of the centrosomes, those attached and those displaced in the cytoplasm, either lack astral microtubules or are associated with abnormally short astral microtubules. Normally the sperm-derived centrosome duplicates once during the growth of the sperm aster, and centrosomes duplicate again in late anaphase of the first mitotic division (Callaini and Riparbelli 1996). Only one or two centrosomes are ever observed in a single wisp embryo, indicating that the centrosome duplicates just once.

Figure 3.

Abnormal spindle structures in wisp embryos. Embryos were fixed and labeled with an antibody to α-tubulin to visualize the spindles and Sytox green to detect the chromosomes. Spindles are shown on the left, and their associated chromosomes are shown on the right. a–d compare a normal first division mitotic spindle from a wild-type embryo with the male pronuclear spindle and its associated chromosomes. (a) First-division mitotic spindle at metaphase in a wild-type embryo. (b) The chromosomes of this spindle are aligned at the metaphase plate, with the male and female complements forming two distinct groups. (c) The mitotic spindle that forms in association with the male pronucleus in an embryo from a wisp12-3147 mother. Compared to the wild-type spindle, this spindle is much smaller and has a microtubule density more similar to a meiotic spindle than a mitotic spindle. One centrosome has dissociated from the spindle and is displaced in the nearby cytoplasm. (d) The nucleus is haploid, and the four chromosomes are spread out irregularly along the length of the spindle, unlike metaphase or anaphase. e–h compare a typical polar body and its associated monastral microtubule array, with the bipolar, meiotic-like spindles that form in association with the polar bodies in wisp embryos. (e) Polar body monastral microtubule array in a wild-type embryo. (f) The chromosomes are arranged with their centrosomes oriented toward the center of the structure. (g) A meiotic-like, acentriolar bipolar spindle associated with the fused polar bodies in an embryo from a wisp11-600 mother. (h) The chromosomes are spread out irregularly along the length of the spindle.

To examine the nature of the centrosomes present in the wisp mitotic spindle, we determined whether they contain three proteins known to be associated with centrosomes: γ-tubulin, CP60, and CP190. γ-Tubulin is an intrinsic component of centrosomes, present at all stages of the cell cycle (Stearnset al. 1991; Zheng et al. 1991, 1995). CP190 and CP60 localize to the centrosome in a cell-cycle-dependent fashion. In syncytial embryos, CP190 is localized to centrosomes from the time of nuclear breakdown through telophase (Oegemaet al. 1997). Maximum amounts of CP60 are detected at the centrosomes during anaphase and telophase (Kelloget al. 1995; Oegemaet al. 1997). The spindles shown in Figure 4 were simultaneously labeled with antibodies to α-tubulin (which labels both the spindle arms and the centrosomes) and γ-tubulin, CP190, or CP60 antibodies. In wisp spindles, γ-tubulin (Figure 4h) and CP190 (Figure 4j) are present in the centrosomes, but appear reduced compared to wild type. Often these two proteins were barely detectable in wisp centrosomes. CP60 was not detected in the centrosomes of wisp spindles (Figure 4l).

Meiosis I defects: To examine meiosis I (MI), we analyzed mature stage 14 oocytes, labeled with antibodies to histone and α-tubulin, from ovaries of wild-type and wisp/Df(1)RA47 females. Stage 14 oocytes are normally arrested in metaphase of meiosis I (Sonnenblick 1950). MI spindles (Figure 5, a–d) lack centrosomes, are anastral, and are highly tapered (Hatsumi and Endow 1992; Theurkauf and Hawley 1992; Matthieset al. 1996; Endow and Komma 1997). The chromosomes are arranged either in a tight karyosome (Figure 5b) or with the fourth chromosomes separated from the karyosome, moving toward the poles (Figure 5d). Occasionally other nonexchange chromosomes separate precociously from the main karyosome mass (Theurkauf and Hawley 1992). Table 2 shows the proportions of wild-type (+/+) stage 14 oocytes with each of these configurations.

In the majority of wisp oocytes, metaphase I arrest is not maintained, and the chromosomes separate aberrantly (Figure 5, e–j). With very high frequencies, the chromosomes are dispersed across the length of the spindle, resembling neither metaphase I arrest nor anaphase I (Table 2). Consistent with the accepted model that chromosomes nucleate the MI spindle (Theurkauf and Hawley 1992; McKim and Hawley 1995), the shapes of the wisp spindles are often distorted due to unequal numbers of chromosomes on each spindle arm (Figure 5, g and h). The wisp spindles frequently appear ragged, with microtubule spurs extending into the cytoplasm, associated with lost chromosomes. Often the fourth chromosomes are separated from the main mass, either both present at one pole indicating nondisjunction or displaced in the cytoplasm (Figure 5, f and j). These lost chromosomes were frequently associated with their own small spindles (Figure 5i). While there was no single pattern to the chromosome figures, loss and nondisjunction of the fourth chromosomes was common (Table 3). Separation of homologous chromosomes was evident in some cases (e.g., Figure 5h), suggesting that the dispersal of chromosomes could reflect, or accompany, a reduction in meiotic recombination, as has been seen in mutants that are known to be recombination defective (McKim and Hawley 1995).

Figure 4.

Centrosomal proteins are reduced or absent in the centrosomes of wisp embryos. To examine the centrosomes, spindles of young embryos were labeled with an α-tubulin antibody and with antibodies to γ-tubulin, CP60, or CP190. α-Tubulin is shown above, and the centrosomal protein in the same spindle is shown below, in each set of images. (a) A wild-type mitotic spindle in metaphase. (b) γ-Tubulin is localized to the centrosomes. (c) A wild-type mitotic spindle in metaphase. (d) CP190 is localized to the centrosomes and is also found on the arms of the spindle. (e) A wild-type mitotic spindle in anaphase. (f) CP60 in the centrosomes increases to maximum levels during anaphase and telophase and can also be found on the arms of the spindle. (g) Mitotic spindle associated with the male pronucleus in an embryo from a wisp11-600 mother. One centrosome has dissociated. (h) The level of γ-tubulin in wisp centrosomes is less than in wild type. (i) Mitotic spindle in an embryo from a wisp12-3147 mother, with one dissociated centrosome. (j) There is less CP190 in wisp centrosomes than in wild type, and no CP190 is detected along the body of the spindle. (k) Mitotic spindle in an embryo from a wisp12-3147 mother, with one dissociated centrosome. (l) No CP60 is detected in wisp centrosomes or along the spindle arms.

As a control for possible dominant effects of Df(1)RA47, we also included +/Df(1)RA47 oocytes in our analysis. Compared to +/+ oocytes, +/Df(1)RA47 oocytes have a higher frequency of MI figures in which the chromosomes are initiating separation, possibly reflecting haplo-insufficiency for the wisp locus. However, the behavior of separated chromosomes in +/Df(1)RA47 oocytes is usually normal: the chromosomes separate as two groups, as normally occurs in meiosis I. Only 2% of the oocytes in our +/Df(1)RA47 set had configurations in which the bulk of the chromosomes were abnormally dispersed (last column, Table 2), and only 6% exhibited loss or nondisjunction of the fourth chromosomes (Table 3, line 2). In contrast, in the majority of wisp/Df(1)RA47 oocytes the chromosomes separated and dispersed abnormally on the MI spindle (73–84%; Table 2, last column), and nondisjunction or loss of fourth chromosomes occurred with very high frequencies (28–44%; Table 3, last column).

Figure 5.

Meiosis I is aberrant in wisp oocytes. Mature oocytes were isolated from hand-dissected ovaries and double-labeled with antibodies to α-tubulin and histone. In each pair of images, the spindle is shown on the left, and the chromosomes are on the right. (a) A wild-type meiosis I spindle. (b) The chromosomes are massed in a tight karyosome in the center of the spindle. (c) A wild-type meiosis I spindle. (d) The fourth chromosomes are moving toward the spindle poles. (e) wisp12-3147 meiosis I spindle. (f) The fourth chromosomes are separated from the main spindle and are attached by spurs. The rest of the chromosomes are dispersed in an irregular fashion, resembling neither metaphase I arrest nor anaphase I. (g) wisp14-1299 meiosis I spindle. The irregular shape of the spindle reflects the distribution of the chromosomes. (h) The chromosomes are separating in an irregular fashion that is unlike normal anaphase I, in which chromosomes move as two compact groups to the spindle poles, and the chromosomes appear to be undergoing nondisjunction. (i) wisp14-1299 meiosis I spindle. The small ectopic spindle on the bottom is associated with a lost fourth chromosome. (j) Both fourth chromosomes have separated abnormally far from the main mass, and one is lost from the spindle altogether.

Meiosis II defects: Oocytes are naturally activated to complete meiosis I and enter meiosis II (MII) by passage through the oviduct (Doane 1960; Mahowaldet al. 1983). Meiosis is normally completed during the first few minutes following egg ovulation (Sonnenblick 1950; Riparbelli and Callaini 1996a; Page and Orr-Weaver 1997). We observed meiosis II in our young 5- to 15-min-old embryos (see above and Table 1). The wild-type MII spindles are tandemly arranged, positioned perpendicular to the anterior-posterior axis of the embryo and radiating inward from the cortex (Figure 6a). In wisp mutant embryos the position of the MII tandem spindles was variable, often near the center of the embryo aligned parallel (instead of perpendicular) to the anterior-posterior axis (Figure 6b). Mispositioning of the MII spindles occurred in 37–56% of observed meioses (Table 4).

Chromosome positioning defects occur on wisp MII spindles, similar to those described above for MI and postmeiotic spindles. Normally the chromosomes are aligned in tight bundles in the center of each tandem MII metaphase spindle, and during anaphase II the chromosomes move to opposite poles as two compact groups. In the wild-type example shown in Figure 7, a and b, the left spindle is in metaphase, and the right spindle is in anaphase; often the two MII spindles are slightly out of synchrony, as in this example. In contrast, chromosomes are often dispersed along the entire length of wisp MII spindles. In the example shown in Figure 7, c and d, the two MII spindles contain unequal numbers of chromosomes, probably as a result of nondisjunction that occurred in MI. The chromosomes associated with the top spindle are abnormally dispersed along its length. Displaced chromosomes, usually fourth chromosomes, are often found near wisp MII spindles, associated with small ectopic spindles (not shown).

Between the two MII spindles lies a microtubule organizing center (MTOC) that nucleates an extensive monastral array of radiating microtubules (Puro 1991; Riparbelli and Callaini 1996a, 1998; Endow and Komma 1998). γ-Tubulin (Endow and Komma 1998; Riparbelli and Callaini 1998), the centrosomal protein CP190 (Riparbelli and Callaini 1996a), and the kinesin-like protein klp3A (Williamset al. 1997) are all present in the MII central MTOC. We show here that the centrosomal protein CP60 is also present in this structure (Figure 8). The MII central MTOC is defective in eggs from wisp mothers. Often the wisp MII spindles lack a detectable central MTOC (see examples in Figures 6b and 7c). The examples shown in Figure 8 illustrate the appearance of wisp MII spindles that do contain central MTOCs; in all cases this structure is abnormal. The spindles have been simultaneously labeled with antibodies to α-tubulin and γ-tubulin (a–d), CP190 (e–h), or CP60 (i–l). Although γ-tubulin, CP190, and CP60 are present in the wisp MII central MTOC, their amounts appear greatly reduced, and the MTOC lacks the normal extensive array of microtubules that radiate to the egg cortex.

TABLE 2

Meiosis I chromosome configurations

Defects in the sperm aster: The sperm aster in wild-type embryos develops while oocyte meiosis is being completed, reaching its maximum size during anaphase II and telophase II, when its radiating microtubules reach to the embryo cortex (Riparbelli and Callaini 1996b; Figure 6a). In contrast, the sperm aster in wisp mutant embryos is dramatically reduced (Figure 6b).

DISCUSSION

We show here that the Drosophila wispy gene is required for female meiosis and the earliest events of embryogenesis. The first defects we detected in wisp mutants occurred in meiosis I, late in oogenesis. Normally, mature oocytes are arrested at metaphase of MI, with the chromosomes condensed in a compact karyosome at the midzone of an acentriolar, bipolar spindle. Metaphase I arrest is disrupted in wisp oocytes: the chromosomes are usually dispersed along the length of the spindle, frequently undergoing chromosome nondisjunction and loss (Tables 2 and 3). The shape of the mutant spindles is abnormal, reflecting the dispersal of chromosomes, and spindle spurs and small ectopic spindles are associated with lost chromosomes (Figure 5). Following completion of female meiosis in wild-type embryos, the polar bodies fuse (sometimes all together) and arrest at metaphase in association with a unique monastral array of microtubules. The pronuclei fuse, leading to formation of the first mitotic spindle. In wisp embryos, meiotic-like bipolar spindles form on the products of female meiosis, and pronuclear fusion never occurs (Figures 2 and 3). The haploid male pronucleus enters a mitotic nuclear division cycle and arrests with the chromosomes dispersed aberrantly along the length of its spindle (Figure 3).

View this table:
TABLE 3

Frequencies of fourth chromosome loss and nondisjunction

Figure 6.

The female meiosis II spindles are mispositioned in the wisp embryo, and the sperm aster is reduced in size. The completion of meiosis was observed in 5- to 15-min-old embryos that had been fixed and labeled with an antibody to α-tubulin. (a) In wild-type embryos, the MII tandem spindles are always positioned perpendicular to and in association with the egg cortex. A sperm aster forms (arrow), from which an extensive array of microtubules radiate toward the egg cortex. (b) In embryos from wisp mothers the MII spindles appear detached from the cortex and are frequently aligned in parallel with the long axis of the egg. The sperm aster appears small, with fewer and shorter radiating microtubules.

A prominent sperm aster forms shortly after fertilization in wild-type eggs, nucleated by the sperm-derived centrosome (Huettner 1924; Sonnenblick 1950; Riparbelli and Callaini 1996b). In wisp embryos, the sperm aster is reduced (Figure 6). The subsequent failure of male and female pronuclei to fuse is consistent with the hypothesis that the sperm aster provides tracts of microtubules for movement of the female pronucleus toward the male pronucleus (see Foeet al. 1993). Alternatively, failure of pronuclear fusion could signal a defect in the function of a microtubule-associated motor required for movement of the female pronucleus along the sperm aster microtubules (see below).

Failure of pronuclear fusion could also be a consequence of the abnormal orientation of wisp meiosis II spindles. The normal perpendicular orientation of the MII spindles, relative to the egg cortex, may serve to position the female pronucleus close to the male pronucleus, in preparation for their fusion (Riparbelli and Callaini 1996a). In wild-type embryos, a microtubule organizing center forms between the two tandem MII spindles, extending an extensive array of radiating microtubules toward the egg cortex. In contrast, the central MTOC is reduced or absent in wisp embryos (Figures 6, 7 and 8). Riparbelli and Callaini (1996a) suggested that the MII central MTOC microtubules tether the MII spindles in their normal perpendicular position relative to the egg cortex. Our results are consistent with this idea, since wisp MII spindles, which lack the normal array of monastral microtubules, are mispositioned in the embryo.

Another type of microtubule organizing center that functions abnormally in wisp embryos are the centrosomes of the haploid mitotic spindle associated with the male pronucleus. The centrosomes in mutant embryos do not nucleate a normal array of astral microtubules and frequently dissociate from the spindle. We examined both the meiosis II central MTOC and the spindle centrosomes for the presence of centrosomal proteins, including γ-tubulin, CP60, and CP190. All three proteins were present, but appeared reduced, in the MII central MTOC (Figure 8). In the spindle centrosomes, both γ-tubulin and CP190 were reduced and CP60 could not be detected (Figure 4). Our failure to detect CP60 in the centrosomes could be a consequence of a wisp-induced block in the mitotic cell cycle, since CP60 normally increases during anaphase and telophase (Kelloget al. 1995; Oegemaet al. 1997). However, the fact that all three proteins are reduced in both types of MTOCs suggests that the wisp gene product may be required for recruiting these proteins or for stabilizing their association with these structures.

In addition to meiotic and mitotic defects, wisp mutations alter the localization of RNAs to the poles of the egg (Figure 1). While both bcd and osk mRNAs were localized during oogenesis in mutant females, the phenotype, reduced and dispersed polar mRNAs in young embryos from mutant mothers, is consistent with the wisp gene being required for the efficiency of RNA localization or maintenance of the localized state. The reduced RNA signals could also indicate that wisp mutations cause premature degradation of localized mRNAs.

The microtubule cytoskeleton is required for localizing RNAs in diverse types of cells (reviewed in Wilhelm and Vale 1993; St. Johnston 1995; Bassell and Singer 1997; Gavis 1997; Bashirullahet al. 1998; Hazelrigg 1998; Oleynikov and Singer 1998). Examples include mRNAs localized in the processes of nerve cells and oligodendrocytes, mRNAs localized to the animal or vegetal poles of Xenopus oocytes, and several RNAs localized to the anterior and posterior poles of the Drosophila oocyte. In the case of RNAs localized during Drosophila oogenesis, microtubule inhibitors disrupt localization (Pokrywka and Stephenson 1991, 1995). The Drosophila oocyte microtubule cytoskeleton has sufficient polarity to suggest that RNAs may be targeted to the anterior or posterior poles by minus- or plus-end directed microtubule motor proteins (Theurkaufet al. 1992; Wilhelm and Vale 1993; Clarket al. 1994; St. Johnston 1995; Theurkauf and Hazelrigg 1998). Despite the attractiveness of such microtubule-based RNA transport models, mutations in known motor proteins that affect RNA localization in Drosophila oocytes have not been identified. This could reflect redundancy in the motors that mediate RNA transport (Theurkauf and Hazelrigg 1998). Alternatively, genes encoding the relevant motors could have been missed in previous genetic screens due to their requirements in other cell types and stages of development.

View this table:
TABLE 4

Position of the meiosis II spindles in embryos from wild-type and wisp mothers

Microtubule motor proteins, including both kinesins and dyneins, play essential roles in meiosis and mitosis (reviewed in Barton and Goldstein 1996; Endow 1999). Kinesins exist as a large family of proteins (Moore and Endow 1996). In Drosophila, the loss of maternal expression of genes encoding kinesins is associated with meiotic and early embryonic defects. These include the genes non-claret disjunctional, ncd (Davis 1969; Hatsumi and Endow 1992; Endowet al. 1994; Endow and Komma 1996, 1998; Matthieset al. 1996), no distributive disjunction, nod (Carpenter 1973; Zhang and Hawley 1990; Zhang et al. 1990a,b; Hatsumi and Endow 1992; Theurkauf and Hawley 1992), Klp3A (Williamset al. 1997), and Klp38B (Rudenet al. 1997). Interestingly, wisp mutants exhibit defects strikingly similar to kinesin mutants, including: chromosome nondisjunction and loss (ncd and nod); alterations in the meiosis II central MTOC (ncd and Klp3A); failure to recruit or maintain γ-tubulin in the MII central MTOC (ncd); failure to form monopolar spindles in association with the polar bodies (ncd); centrosome loss from mitotic spindles (ncd and Klp3A); and failure of male and female pronuclei to fuse (Klp3A). These similarities in mutant phenotypes suggest the possibility that wisp may encode a kinesin or a regulator of kinesin activity.

Figure 7.

Chromosomes separate abnormally on wisp meiosis II spindles. Meiosis II was observed in 5- to 15-min-old embryos that had been fixed and double-labeled with antibodies to α-tubulin and histone. The spindles are shown on the left, and the associated chromosomes are shown on the right. (a) Meiosis II spindles in a wild-type embryo. Between the two tandem spindles a central MTOC (arrow) that extends radiating microtubules forms. (b) Often the chromosomes on one spindle arm begin to separate before those on the other, as in this example. The chromosomes separate as two compact groups to the poles of the spindle. (c) wisp11-600 meiosis II spindles. The central MTOC cannot be detected. (d) The chromosomes on the top spindle are dispersed abnormally along the length of the spindle. The numbers of chromosomes on each spindle are unequal, probably as a result of meiosis I nondisjunction.

Figure 8.

The wisp meiosis II central MTOC is abnormal. The MTOC that forms between the two tandem MII spindles was detected in fixed embryos by labeling with antibodies to α-tubulin and the centrosomal proteins γ-tubulin, CP60, or CP190. In each set of images, α-tubulin is shown on the left, and the centrosomal protein is shown in the smaller image on the right. (a) Wild-type meiosis II spindles. (b) γ-Tubulin in the central MTOC. (c) wisp12-3147 meiosis II spindles. (d) γ-Tubulin is reduced in the wisp central MTOC. (e) Wild-type meiosis II spindles. (f) CP190 in the central MTOC. (g) wisp11-600 meiosis II spindles. (h) There is less CP190 in the central MTOC. (i) Wild-type meiosis II spindles. (j) CP60 in the central MTOC. As with the centrosomes, the level of CP60 is maximum during anaphase and telophase. (k) wisp14-1299 meiosis II spindles. (l) CP60 is reduced in the wisp central MTOC. The level of all three centrosomal proteins was always less in the wisp meiosis II central MTOCs than in wild type; the signal shown in d, h, and l was the maximum observed. When no central MTOC was detected in wisp meiosis II spindles by α-tubulin labeling, no γ-tubulin, CP190, or CP60 was observed.

Unlike mutations in previously identified genes for mRNA localization in Drosophila oogenesis, including exuperantia (exu), swallow (swa), and staufen (stau), which cause embryos to arrest late in embryogenesis with characteristic head and abdominal defects (reviewed in St. Johnston and Nüsslein-Volhard 1992), wisp maternal-effect mutations lead to very early embryonic arrest, before anterior-posterior patterning defects can be detected. Our analysis has shown that early arrest is associated with defects in several microtubule-based events, including meiosis, sperm aster formation, pronuclear fusion, and mitotic spindle function. These defects suggest that the wisp gene product's essential function lies in regulating microtubule structure or microtubule-based motility, in a component shared with RNA localization pathways. Further study of RNA localization in other maternal-effect mutations causing very early embryonic arrest may in the future identify additional genes in this category and help to elucidate the role of microtubules in RNA localization.

Acknowledgments

We thank T. Karr for antibody to the Drosophila sperm tail, Y. Zheng for antibody to γ-tubulin, K. Oegema for CP60 and CP190 antibodies, and R. Nagoshi for providing M19 stocks. This work was supported by National Institutes of Health grant GM-48060 to T.H. and by a Columbia University Undergraduate I. I. Rabi Fellowship to A.E.B.

Footnotes

  • Communicating editor: R. S. Hawley

  • Received August 30, 1999.
  • Accepted December 13, 1999.

LITERATURE CITED

View Abstract