We have examined cytoskeletal requirements for bicoid (bed) RNA localization during Drosophila oogenesis, bed is an anterior morphogen whose proper function relies on the localization of its messenger RNA to the anterior cortex of the egg. Drugs that depolymerize microtubules perturb all aspects of bed RNA localization. During recovery from drug treatment, bed RNA relocalizes to the oocyte cortex, suggesting that the localization machinery is a component of the cortical cytoskeleton.

Taxol, a drug that stabilizes microtubules, also effectively disrupts bed RNA localization, and the effects of taxol treatments on exuperantia and swallow mutants suggest general roles for these gene products in the multi-step bed RNA localization process.

In Drosophila, the organization of the unfertilized egg determines the spatial pattern of the embryo. During oogenesis, developmental morphogens are deposited asymmetrically in the oocyte, and these determinants direct the proper development of the embryo. Three classes of maternal gene products are required to establish the anteroposterior body plan (reviewed in French, 1988). The anterior group includes bicoid (bed), exuperantia (exu) and swallow (sww), and determines the development of head and thoracic structures. The posterior group directs development of the abdomen and comprises at least 8 genes, including the posterior determinant nanos. The terminal group genes specify both extreme ends of the embryo.

The development of anterior structures is the best characterized patterning activity. Products of the bed gene are responsible for the initiation of head and thoracic development (Driever and Niisslein-Volhard, 1988b; Driever et al. 1990). During oogenesis, bed RNA is transcribed in nurse cells and transported to the oocyte, where it is localized to the anterior pole (Berleth et al. 1988). bed RNA is translated after fertilization, and the protein diffuses away from the anterior end, forming a steep concentration gradient (Driever and Niisslein-Volhard, 1988a). There is a direct correlation between concentrations of bed protein and transcriptional activation of downstream genes such as hunchback, which in turn direct the formation of anterior structures (Struhl et al. 1989).

Thus, the establishment of a protein gradient directs the polarity and extent of anterior development. It follows that the initial anterior localization of bed RNA during oogenesis is essential for head and thoracic development.

Two genes, exuperantia (exu) (Schüpbach and Weis-chaus, 1986) and swallow (.ww) (Frohnhôfer and Niisslein-Volhard, 1987; Stephenson et al. 1988), have been identified as essential for bed mRNA localization. Mutations at either locus cause a loss of bed RNA localization, with the result that the bed protein gradient is shallow or absent (Driever and Niisslein-Volhard, 1988b). The loss of this gradient results in the loss of head structures, since bed protein concentrations in the anterior embryo are no longer high enough to activate transcription of anterior-specific genes. In exu mutants, bed RNA localization is lost early in oogenesis (Berleth et al. 1988; St. Johnston et al. 1989; Pokrywka and Stephenson, unpublished), while in sww mutants, early localization events appear normal, but localization is lost at later stages (Berleth et al. 1988; Stephenson et al. 1988). A third locus, staufen (Schiipbach and Wieschaus, 1986), may play some role in maintaining bed RNA distribution in mature eggs (St. Johnston et al. 1989).

The products of these loci may play an accessory role in anchoring bed RNA to a stable framework such as the cytoskeleton, but little is known about cytoskeletal components that might be involved in the process. Cytoskeletal requirements for the localization of molecules or organelles thought to be determinants have been investigated in other systems (for review, see Jeffery, 1989). In C. elegans, P-granules are segregated into the posterior cell at the first cleavage by a mechanism that requires microfilament function (Strome and Wood, 1983; Hill and Strome, 1988). Both microtubules and microfilaments are required for Vg1 RNA localization in Xenopus oocytes (Yisraeli et al. 1990). The egg cortex, a distinctive cytoplasmic domain rich in cytoskeletal elements, has long been suspected as the likely location of determinants, but the organization of the cortical cytoplasm is poorly understood.

Here we describe experiments that reveal a role for the cytoskeleton in bed RNA localization. Intact microtubules are required at all stages of bed RNA localization. We also present an initial characterization of the localization machinery and define steps in the process by which bed RNA localization in the oocyte is achieved.

Drug treatments

Wild-type females were from the Oregon R strain, exu egg chambers were from females homozygous for exuPJ42 (Schüpbach and Weischaus, 1986), and sww egg chambers were from females homozygous for sww’497(Gans et al. 1975). Ovaries were dissected from 1- to 2-day-old females into R-14 (Robb, 1969), made with fresh linoleic acid. Individual stage 10 egg chambers were teased out and incubated in R-14 containing the appropriate drugs at 25°C with gentle rocking to insure adequate aeration. Stock solutions were made as follows: 20 mg ml-1 colchicine (Sigma) in absolute ethanol, 2 mg ml-1 nocodazole (Sigma) in DMSO, 10 HIM taxol (gift of National Cancer Institute) in DMSO, 10 mg ml-1 tubulozole-T or -C (Janssen) in absolute ethanol, and 2 mg ml-1 cytochalasin D (Sigma), in absolute ethanol. In vitro working concentrations were: colchicine, 20μml-1; nocodazole, 20/.;gmrl; taxol, 20 UM or 100μ M; tubulozole-T or -C, 10μ ml-1; cytochalasin D, 2μ gml-1Control egg chambers were incubated in R-14 for up to 4 h without adverse effects.

In vivo incubations were performed using day old females. They were injected with 100 nl of either 200 pg ml-1 colchicine or 20μ gml- cytochalasin D in PBS. Injections were made directly into the abdomen, and injected flies were allowed to recover at 25 °C for 6 to 24 h. Control injections were performed with 1 % ethanol in PBS to simulate the solution injected with the drugs, and had no effect on bed RNA localization.

In situ hybridizations

Individual egg chamber populations and whole ovaries were fixed as follows. Tissue was washed briefly in PBS and then fixed for 30– 45 min in [10% methanol, 0.1M sodium acetate, pH7.0, 3.7% formaldehyde], saturated with heptane. After fixation, tissue was rinsed in [80mM NaCl, 20mM KC1, 2mM EDTA, 2 HIM EGTA, lOmM Pipes, pH7.2], The tissue was then dehydrated by 30 min washes in 30 %, 50 %, 70 %, 95 %, and 100 % ethanol. The 95 % ethanol contained 0.2 % eosin Y to aid in detecting the egg chambers in following steps. Tissue was cleared in 1:1 100% ethanokbenzene and pure benzene for 8min each, and then transferred to paraffin (TissuePrep, Fisher). After several changes of paraffin, tissue was embedded.

35S-labelled riboprobes were made by in vitro transcription, utilizing linearized templates, 35S-UTP and the Bluescript T7 promoter, as described in Stephenson et al. 1988. Other in situ procedures were as described in Stephenson et al. (1988).

Microtubule staining

Ovaries were dissected in PBS, and fixed, dehydrated, cleared, embedded and sectioned as above, except that eosin was omitted. Sections were deparaffinized in 2 changes of xylene and rehydrated through an ethanol series into PBS containing 0.3 % Triton-X 100. Sections were blocked for 3h in [PBS, 0.3% Triton-X 100, 5% NGS, and 0.1% BSA], at 4°C. Antibody incubations were carried out overnight in a humid chamber at 4°C. Slides were rinsed in at least four changes of PBS with 0.3% Triton-X 100 for 2h at 4°C, and then incubated with goat anti-rabbit rhodaminated secondary antibodies. The secondary antibodies were also diluted in blocking solution (dilutions ranging from 1:200 to 1:1000 were used), and incubations lasted 2h. Slides were then washed in several changes of PBS with 0.3 % Triton-X 100 and mounted in Gelvatol. Polyclonal antibodies to tubulin were a generous gift of J. Olmsted, and were diluted 1:10 in blocking solution

Microtubules are required for establishing and maintaining the position of bed mRNA in nurse cells and the oocyte

To explore the possibility that components of the cytoskeleton participate in bed RNA localization, we inhibited specific cytoskeletal elements in Drosophila egg chambers and determined the position of bed RNA by in situ hybridization. Control egg chambers were cultured in R-14 without drugs. In vitro development proceeds normally and is complete after 8 to 10 h, consistent with the time course of development in vivo (Mahowald and Kambysellis, 1980). In agreement with the findings of Petri et al. (1979), the majority of cultured egg chambers develop to stage 14 (stages according to King, 1970). Although a fraction of the egg chambers do not complete development, we analyzed only egg chambers that had undergone appropriate development. Egg chambers incubated in R-14 for four hours exhibit a completely normal patten of bed mRNA (Fig. 1A). At stage 10, bed mRNA distribution in nurse cells is patchy, and usually concentrated on one side of each nucleus. At this stage, bed message is also localized to the anterior margin of the oocyte (Berleth et al. 1988; Stephenson et al. 1988; St. Johnston et al. 1989; Stephenson and Pokrywka, 1991). In oocytes that have matured in vitro, bed RNA is distributed as an anterior cap, indicating that the localization process is not perturbed by these culture conditions (Fig. 1B).

Fig. 1.

bed RNA localization in egg chambers cultured in vitro, bed RNA was detected by in situ hybridization to paraffin sections. Bright-field and dark-field micrographs are shown. In this and all in situ figures anterior is left and posterior right. The scale bar in all figures is 100μm, unless noted, (n) nurse cell complex, (o) oocyte, (f) follicle cells. (A) A wild-type stage 10 egg chamber incubated in R-14 for 4h. bed RNA localization in nurse cells and the oocyte is completely normal. (B) A wild-type stage 14 oocyte which completed development in vitro, bed RNA is localized to the anterior cortex. (C) A stage 10 egg chamber cultured for 1 h in R-14 containing 20 μg ml-1 nocodazole. bed RNA localization is lost in nurse cells and the oocyte.

Fig. 1.

bed RNA localization in egg chambers cultured in vitro, bed RNA was detected by in situ hybridization to paraffin sections. Bright-field and dark-field micrographs are shown. In this and all in situ figures anterior is left and posterior right. The scale bar in all figures is 100μm, unless noted, (n) nurse cell complex, (o) oocyte, (f) follicle cells. (A) A wild-type stage 10 egg chamber incubated in R-14 for 4h. bed RNA localization in nurse cells and the oocyte is completely normal. (B) A wild-type stage 14 oocyte which completed development in vitro, bed RNA is localized to the anterior cortex. (C) A stage 10 egg chamber cultured for 1 h in R-14 containing 20 μg ml-1 nocodazole. bed RNA localization is lost in nurse cells and the oocyte.

We have tested the effects of the microtubule destabilizers nocodazole, colchicine and tubulozole-C on bed RNA localization. Individual stage 10 egg chambers were dissected from young wild-type females and incubated in Robb’s R-14 (Robb, 1969) containing one of these drugs for one to two hours, processed for paraffin sectioning, and sections analyzed by in situ hybridization to determine the distribution of bed RNA. All three drugs gave similar results, an example of which is shown in Fig. 1C. Nocodazole inhibits the microtubule-based cytoplasmic streaming in late oocytes (Gutzeit, 1989b): as a result the contents of treated oocytes are stratified into dark-staining yolky ooplasm and light-staining cytoplasm that has just entered from nurse cells. In addition, inhibition of microtubule function results in the release of bed RNA from its anterior location. In many cases, the loss of localization is complete, and no bed message is found at the anterior margin.

At stage 10, there are two populations of bed RNA in the egg chamber: message already localized at the anterior cortex of the oocyte and message still in nurse cells, awaiting import into the oocyte. During a one hour nocodazole treatment, much of the latter population enters the egg and is concentrated in the lighterstaining nurse cell cytoplasm. Thus, the unlocalized bed RNA deep within the oocyte is likely to be composed of message that was delocalized by nocodazole treatment as well as newly entering bed message which has never become localized. These results suggest that microtubules are necessary both for localizing bed RNA in the oocyte, and for maintaining its localization while development proceeds.

The loss of bed mRNA localization in response to microtubule disruption also applies to the nurse cell pattern. When wild-type egg chambers are cultured in R-14 alone, bed RNA displays a characteristic patchy nurse cell distribution, as shown in Fig. IA. However, in oocytes treated with microtubule inhibitors bed mRNA has a uniform distribution in nurse cells (see Fig. 1C). This implies that bed message is anchored to some framework in nurse cells as well as in the oocyte, and that microtubule networks in nurse cells are directly or indirectly responsible for the observed distribution of bed RNA.

We believe the effects of these drugs on bed RNA localization result from their ability to alter microtubule stability and not from a non-specific poisoning effect. The three drugs used differ in their chemical structure and interactions with microtubules, yet all have the same effects on bed message localization. Egg chambers treated in vitro with colchicine or nocodazole for up to 5 h continue to develop to an extent comparable to controls, suggesting the physiology of the egg chamber is normal apart from microtubule disruption (data not shown). We also incubated egg chambers with tubulozole-C, a microtubule destabilizer, or its inactive isomer tubulozole-T (Geuens et al. 1985). While tubulozole-C disrupts the localization of bed RNA efficiently, tubulozole-T-treated egg chambers appear normal in all respects (data not shown).

We also tested the effect of cytochalasin D treatments on bed message localization. Cytochalasin D inhibits the polymerization of actin filaments (Cooper, 1987), and has been shown to interfere with the actin-based bulk flow of cytoplasm from nurse cells into the oocyte at stage 10b (Gutzeit, 1986a). Egg chambers were treated with 2μgml-1 cytochalasin D in R-14forone to two hours; oocyte size increased slowly in treated egg chambers, indicating that the rapid bulk flow of nurse cell cytoplasm into the oocyte was blocked as expected. However, the localization of bed message in nurse cells and the oocyte is normal (results not shown). Repeated experiments of this nature revealed no effect of cytochalalsin D on bed RNA distribution. These experiments imply that the localization system tolerates perturbation of microfilaments. However, since cytochalasin D does not completely abolish microfilament networks (Cooper, 1987), it is possible that bed RNA localization utilizes actin filaments not affected by these experimental conditions.

Microtubules are required for all stages of bed RNA localization

The results obtained above demonstrate a requirement for microtubules in bed RNA localization during later stages of oogenesis. Young egg chambers do not develop in vitro, and so the effects of cytoskeletal drugs on early stages could not be tested in this way. To examine the effect of cytoskeletal inhibitors on localization events early in oogenesis, we injected concentrated solutions of either cytochalasin D or colchicine directly into the abdomens of female flies, and allowed the flies to recover for periods ranging from 6 to 24 h. Over 90% of injected flies recovered and were active. After various in vivo incubation times, the position of bed message was analyzed by in situ hybridization.

Injection of colchicine results in the loss of bed message localization in all egg .chambers examined, as shown in Fig. 2. The Drosophila ovary contains a continuum of developmental stages, from the germarium to stage 14 oocytes, which were mature at the time of injection, bed RNA normally begins accumulating in the oocyte of stage 5 egg chambers. However, bed message is uniformly distributed in stage 5-8 egg chambers treated with colchicine in vivo (Fig. 2A). At later stages, bed RNA is distributed uniformly in the oocyte, but higher levels of message are detected in nurse cells (see Fig. 2B). Mature oocytes also exhibit a uniform pattern of bed RNA when treated with colchicine in vivo. Thus, the presence of microtubules appears to be required for bed RNA localization during all stages of oogenesis. We were unable to test the effects of other microtubule destabilizers due to their poor solubility in physiological buffers. Mock injections were performed with PBS solutions containing 1 % ethanol; these injections have no effect on oogenesis or bed mRNA distribution. In agreement with in vitro experiments, injections of cytochalasin D and B into females has no effect on bed RNA localization (data not shown).

Fig. 2.

Egg chambers treated with colchicine in vivo. Egg chambers were dissected from females injected with colchicine 20 h earlier. (A) Early egg chambers, stages 5-7. No bed RNA localization is observed. (B) A stage 10 egg chamber in which localization of bed message is lost. (C) A mature oocyte, displaying a uniform distribution of bed RNA.

Fig. 2.

Egg chambers treated with colchicine in vivo. Egg chambers were dissected from females injected with colchicine 20 h earlier. (A) Early egg chambers, stages 5-7. No bed RNA localization is observed. (B) A stage 10 egg chamber in which localization of bed message is lost. (C) A mature oocyte, displaying a uniform distribution of bed RNA.

bed mRNA is relocalized when microtubule inhibitors are removed from the egg chamber

To explore the relationship between microtubules and bed RNA localization machinery, we tested the ability of egg chambers to localize bed RNA after recovery from treatment with the microtubule destabilizing drug nocodazole. Nocodazole can be efficiently rinsed out of cells, and at least one microtubule-based process, ooplasmic streaming, resumes in a few minutes (Gutzeit, 1986b). Stage 10 oocytes were incubated for 1–2 h in R-14 containing 20lugml-1 nocodazole. Half of the egg chambers were removed and fixed immediately. The remaining egg chambers were transferred to media free of nocodazole and allowed to develop for several hours before fixation.

The results of this experiment are shown in Fig. 3. After nocodazole treatment, 70–80 % of egg chambers show some disruption of bed RNA localization, with bed message commonly found dispersed deep within the oocyte cytoplasm. In contrast, 80–90% of egg chambers allowed to recover in the absence of nocodazole show relocalization of bed mRNA to the oocyte cortex, and little or no bed message is detectable in the interior of the oocyte. The relocalization of bed messge to the oocyte cortex is consistent and reproducible; however, the distribution of bed RNA in recovered egg chambers is variable. In many egg chambers, we see a nearly wild-type distribution of bed RNA. In others, bed message is localized to the anterior cortex, but is present in more posterior regions of the cortex as well. A few oocytes appear to have bed message extending almost completely around the periphery of the oocyte. The pattern of relocalization appears to depend on the effectiveness of the nocodazole treatment. A short (one hour) nocodazole treatment results in bed RNA that is released from the anterior cortex but remains in the anterior portion of the oocyte (Fig. 3A). Egg chambers allowed to recover from these short drug treatments relocalize bed RNA primarily to the anterior oocyte cortex. Longer nocodazole treatments (1.5 to 2.0 h) yield egg chambers in which dispersed bed message is present deeper in the oocyte. The corresponding recovered egg chambers display bed RNA distributed all along the cortex, extending well into the posterior half of the oocyte (Fig. 3D,E). However, higher concentrations of bed mRNA are always present at the anterior tip of recovered oocytes, even those in which the extent of bed message mislocalization was presumably extensive.

Fig. 3.

Recovery of egg chambers from nocodazole treatment. (A) An egg chamber incubated in 20μgml-1 nocodazole for 1 h to disrupt bed RNA localization. (B) A corresponding egg chamber allowed to recover from nocodazole treatment for 3h in fresh media. Most bed RNA is localized to the anterior cortex. (C) An egg chamber treated with 20μgml-1nocodazole for 1.5 h. bed message is more extensively dispersed with longer nocodazole treatments. (D,E) Egg chambers treated as in C, then allowed to recover in fresh R-14 for 2h. bed RNA is distributed along the lateral and posterior oocyte cortex, as well as anteriorly.

Fig. 3.

Recovery of egg chambers from nocodazole treatment. (A) An egg chamber incubated in 20μgml-1 nocodazole for 1 h to disrupt bed RNA localization. (B) A corresponding egg chamber allowed to recover from nocodazole treatment for 3h in fresh media. Most bed RNA is localized to the anterior cortex. (C) An egg chamber treated with 20μgml-1nocodazole for 1.5 h. bed message is more extensively dispersed with longer nocodazole treatments. (D,E) Egg chambers treated as in C, then allowed to recover in fresh R-14 for 2h. bed RNA is distributed along the lateral and posterior oocyte cortex, as well as anteriorly.

Taxol treatment results in the localization of bed mRNA at ectopic positions in the oocyte

To determine whether excessive microtubule polymerization alters bed message localization, we also examined the effects of taxol, a microtubule stabilizer that promotes the polymerization of tubulin monomers and induces the formation of novel microtubule networks independent of microtubule organizing centers (Dustin, 1984). Taxol treatment of egg chambers also results in the mislocalization of bed message. The effect of taxol treatment at either 20 μM or 100 μM is reproducible, although the patterns of bed mRNA mislocalization are highly variable. A fraction of treated egg chambers appear similar to those treated with colchicine or nocodazole; bed RNA is released from its localized position and appears in the interior of the oocyte (Fig. 4A). In addition, we observe novel patterns of bed message localization. In most treated egg chambers, bed RNA is concentrated in aggregates within the oocyte, and boundaries between regions containing and lacking bed message are distinct, bed message is often clumped around the oocyte nucleus; indeed this organelle can commonly be identified as a patch of nonhybridization surrounded by strong signal. We also detect irregular patches of bed message at the cortex, although these regions are seldom positioned anteriorly. Approximately 75 % of egg chambers show some ectopic localization, almost always to the cortex or oocyte nucleus (Fig. 4B). These regions can be considered as the preferred sites of ectopic localization. The effects of taxol on nurse cell distribution of bed RNA are less certain: taxol treatment does not appear to affect the localization of bed message in nurse cells, since bed message is clumped around nurse cell nuclei in a pattern similar to that in untreated egg chambers.

Fig. 4.

Effects of taxol on bed RNA distribution. Wild-type egg chambers were cultured in R-14 in the presence of 100MM taxol for 2h. (A) An egg chamber in which bed message has become dispersed in the oocyte. (B) An egg chamber with ectopic localization of bed RNA to the lateral oocyte cortex.

Fig. 4.

Effects of taxol on bed RNA distribution. Wild-type egg chambers were cultured in R-14 in the presence of 100MM taxol for 2h. (A) An egg chamber in which bed message has become dispersed in the oocyte. (B) An egg chamber with ectopic localization of bed RNA to the lateral oocyte cortex.

The microtubule staining patterns of wild-type and taxol-treated egg chambers were examined by immunocytochemistry. In untreated egg chambers, microtubule staining is largely diffuse, presumably because most tubulin is not polymerized. Follicle cells exhibit bright, highly organized staining, while nurse cells and the oocyte contain few distinct structures. Fibrous staining is seen around nurse cell nuclei; within the oocyte, staining is brightest around the cortex and the oocyte nucleus (Fig. 5). In taxol-treated egg chambers, microtubule staining is strong and consistent, as shown in Fig. 5C,D,E. Microtubule staining highlights spikes and bundles radiating from nurse cell nuclei. Within the oocyte, bright, discontinous staining is observed around the cortex. Heavy staining is occasionally observed around the oocyte nucleus (data not shown), and bundles and spikes are stained in isolated areas of the cortical cytoplasm. These microtubule staining patterns are consistent with the distribution of bed message in taxol-treated egg chambers..

Fig. 5.

Microtubule distribution in untreated and taxol-treated egg chambers. 5μm paraffin sections were stained with polyclonal antibodies to tubulin, and rhodamine-conjugated secondaries. Bar, is 10μm. (ncn) nurse cell nuclei, (o) oocyte, (f) follicle cells. (A) The nurse cell complex of a stage 10 egg chamber. Fibers surrounding nurse cell nuclei are stained. (B) Part of the oocyte and overlying follicle cells; follicle cells stain brightly, as does the oocyte cortex. (C,D,E) Microtubule staining of taxol-treated egg chambers. Bundles and spikes radiate from nurse cell nuclei, and are found in the ooplasm. Cortical staining in the oocyte is unorganized and discontinous. (F,G) exu egg chambers incubated in taxol. Abnormal microtubular structures are similar to those induced in wild-type egg chambers.

Fig. 5.

Microtubule distribution in untreated and taxol-treated egg chambers. 5μm paraffin sections were stained with polyclonal antibodies to tubulin, and rhodamine-conjugated secondaries. Bar, is 10μm. (ncn) nurse cell nuclei, (o) oocyte, (f) follicle cells. (A) The nurse cell complex of a stage 10 egg chamber. Fibers surrounding nurse cell nuclei are stained. (B) Part of the oocyte and overlying follicle cells; follicle cells stain brightly, as does the oocyte cortex. (C,D,E) Microtubule staining of taxol-treated egg chambers. Bundles and spikes radiate from nurse cell nuclei, and are found in the ooplasm. Cortical staining in the oocyte is unorganized and discontinous. (F,G) exu egg chambers incubated in taxol. Abnormal microtubular structures are similar to those induced in wild-type egg chambers.

Taxol has differential effects on bed RNA localization in sww and exu mutants

We also examined the effects of taxol on bed message localization in exu and sww egg chambers. Untreated exu egg chambers show a loss of bed message localization starting at early stages. In these mutants, there is a complete loss of bed RNA localization around nurse cell nuclei and at the anterior oocyte cortex (Fig. 6C). Untreated sww egg chambers appear normal in early oogenesis, but the localization of bed message deteriorates in a stepwise manner beginning at stages 10–11. Initially bed message is mislocalized to lateral and posterior regions of the oocyte cortex (Fig. 6A), and at later stages bed RNA is dispersed in internal regions of egg cytoplasm as well (Berleth, et al. 1988; Stephenson, et al. 1988). A third mutant, staufen, causes defects in bed RNA localization after fertilization. The phenotypes of exu and sww mutants are consistent with the idea that the exu gene product is required for the association of bed message with the localization machinery and that sww is required for maintenance of localization within the oocyte. We were thus interested in learning whether the bed RNA pattern in these mutants could be altered by taxol treatment. Since staufen does not participate in localization during oogenesis, analysis of this mutant was not performed.

Fig. 6.

The effects of taxol treatment on bed RNA localization in and exu egg chambers. (A) An untreated JIW egg chamber, stage 11. bed RNA localization in the oocyte has slipped to the lateral cortex and unlocalized bed RNA can be seen within the oocyte. The deterioration of bed RNA localization continues in later stages of oogenesis. (B) A stage 11 sww egg chamber treated with taxol. and wild-type egg chambers respond similarly to taxol treatment, bed RNA is localized in discontinous patches at random points of the cortex (arrows), and aggregates can also be seen in the interior of the oocyte. (C) An untreated exu egg chamber, in which bed RNA is unlocalized. (D) An exu egg chamber treated with taxol. Taxol has no effect on the distribution of bed RNA in exu egg chambers.

Fig. 6.

The effects of taxol treatment on bed RNA localization in and exu egg chambers. (A) An untreated JIW egg chamber, stage 11. bed RNA localization in the oocyte has slipped to the lateral cortex and unlocalized bed RNA can be seen within the oocyte. The deterioration of bed RNA localization continues in later stages of oogenesis. (B) A stage 11 sww egg chamber treated with taxol. and wild-type egg chambers respond similarly to taxol treatment, bed RNA is localized in discontinous patches at random points of the cortex (arrows), and aggregates can also be seen in the interior of the oocyte. (C) An untreated exu egg chamber, in which bed RNA is unlocalized. (D) An exu egg chamber treated with taxol. Taxol has no effect on the distribution of bed RNA in exu egg chambers.

The effects of taxol on sww and exu egg chambers support their proposed roles in bed RNA localization during oogenesis. When sww egg chambers are treated with 100pM concentrations of taxol, bed message becomes clumped and localized to ectopic positions within the oocyte (Fig. 6B). The sww egg chambers are identical to wild-type oocytes treated in the same way. Taxol-treated sww egg chambers are distinct from untreated sww egg chambers in showing localization of bed RNA around the oocyte nucleus and discontinous patches of bed RNA at the oocyte cortex. It appears that the position of bed message in sww egg chambers is sensitive to changes in microtubule equilibrium in the same manner as normal egg chambers. In contrast, exu egg chambers incubated with taxol are indistinguishable from untreated counterparts (Fig. 6D). That is, taxol treatment does not alter the distribution of bed RNA in exu egg chambers, suggesting that changes in microtubule stability have no effect on bed RNA distribution in exu egg chambers. We stained exu egg chambers with tubulin antibodies to determine the microtubule distribution. Untreated exu egg chambers look identical to untreated wild-type egg chambers with respect to microtubule staining (data not shown). Taxol treatment of exu egg chambers produces the same aberrant microtubule bundles and spikes seen in taxol-treated wild-type egg chambers (Fig. 5F,G). Thus, it appears that microtubules in wild-type and mutant egg chambers respond similarly to taxol treatment.

A role for microtubules in bed RNA localization

In many developing systems, cytoskeletal components play a role in the localization of developmental determinants. In this report, we show that microtubules are involved in the localization of bed RNA to the anterior pole of Drosophila oocytes. When egg chambers are treated with microtubule inhibitors, bed RNA becomes unlocalized in nurse cells and the oocyte. These effects occur at all stages of oogenesis and in response to a variety of microtubule inhibitors and experimental conditions. In contrast, the microfilament inhibitor cytochalasin D does not have any effect on bed message localization. Thus, microtubules play an important role in all aspects of the localization of bed message, from initial localization events in nurse cells to the stabilization of localized message in mature oocytes. These results appear to be specific for microtubules, and are not a side effect of drug treatment, since several microtubule drugs have similar effects on localization. In addition, the ability of egg chambers to resume bed RNA localization once microtubule inhibition is lifted argues strongly that microtubule destabilizers do not have long-lasting, irreversible effects on egg chamber viability.

It seems likely that correct regulation of microtubule dynamics is crucial to the proper localization of bed RNA. Microtubule destabilizing drugs like colchicine, nocodazole and tubulozole-C function by preventing the addition of tubulin subunits to dynamic microtubules. The sensitivity of bed message localization to these drugs implies that the population of microtubules necessary for localization is in dynamic equilibrium with the tubulin monomer pool. Further, the sensitivity of bed message localization to taxol, which stabilizes and promotes the growth of existing microtubules, suggests that microtubule dynamic equilibrium itself may be an important aspect of localization. Perhaps stabilizing microtubules involved in bed message localization interferes with their function in this process. Alternately, the novel microtubules induced by taxol may compete with organized microtubule networks for bed RNA particles, or may disrupt microtubule networks that hold bed RNA at the oocyte cortex.

Several observations are consistent with a relatively direct role for microtubules in bed RNA localization. First, the sites of accumulation of bed message in untreated wild-type egg chambers, around nurse cell nuclei and at the oocyte cortex, are the sites of highest microtubule density (Gutzeit, 19866; data herein). Second, taxol induces novel clumps of bed message that are similar in occurrence and position to the unusual microtubule structures induced by taxol. However, the experiments reported here do not distinguish between a very direct interaction in which bed mRNP itself is attached to microtubules, and a much more indirect interaction, in which bed RNA is attached to or organized into an organelle whose position is maintained by microtubules. We intend the phrase ‘associated with microtubules’ to encompass the range of direct and indirect interactions suggested by the microtubuledependent localization process described here.

Localization machinery and the question of oocyte polarity

We have shown that bed RNA populations in the oocyte are displaced by drug treatment, and subsequent removal of microtubule inhibitors results in relocalization to the cortex, although relocalization is not specific to the anterior cortex. Our results are consistent with the idea that bed message mislocalized in the presence of nocodazole is relocalized to the closest cortex when microtubule function is restored. Short periods of nocodazole treatment result in mild mislocalization and nearly normal relocalization, while longer nocodazole treatments result in severe mislocalization and relocalization to even posterior cortical regions.

The relocalization of bed message to the cortex suggests that the cortex is the ‘natural’ location of the bed message localization machinery. However, these experiments do not tell us whether the anterior cortex is specially modified relative to other parts of the egg periphery. It is possible that the capacity to localize bed RNA is present in every part of the oocyte cortex, but bed message is restricted to the anterior margin because of the proximity of nurse cells. As an alternative, the bed message localization machinery may be naturally restricted to the anterior oocyte cortex. In the latter case one would posit that the normally anterior localization machinery is reassembled in an abnormal fashion at any cortical region in the nocodazole recovery experiments.

There are two additional points to be made from these recovery experiments. First, in recovered egg chambers, all of the bed message is cortical. This suggests that bed RNA released into the interior of the oocyte by drug treatment remains competent to be relocalized. Second, in egg chambers in which bed RNA was probably dispersed almost completely, recovery always results in higher bed RNA concentrations at the anterior oocyte margin. We believe this represents the restoration of the normal localization machinery as well as the return of mislocalized bed RNA to the cortex. That is, bed message that enters the oocyte for the first time during the recovery period appears to be properly localized to the anterior end. It further suggests that bed RNA released by drug treatment is relocalized improperly, while bed message entering the oocyte during recovery is correctly localized. At the present time, we cannot be certain that the heavier anterior localization represents newly entering bed RNA; however, it represents the simplest interpretation of available data. This observation suggests that the inherent polarity of the oocyte-nurse cell complex may be sufficient to account for at least some of the anterior localization observed in normal egg chambers.

Steps in bed RNA localization

Results presented here and elsewhere (St. Johnston et al. 1989) define steps in bed RNA localization (Fig. 7). The earliest step in the process is localization of newly synthesized bed message around nurse cell nuclei. We propose that the exu gene product and microtubules are required for perinuclear localization. Since taxol fails to induce novel bed RNA localization in exu oocytes, we believe the defect in exu mutants is the failure of bed RNA to become associated with the microtubule-based localization machinery. These results suggest that exu gene product mediates an association with microtubules around nurse cell nuclei.

Fig. 7.

A model for bed RNA localization, bed RNA is synthesized in nurse cells and becomes associated with microtubules in a process that requires the exu gene product. This association results in the nurse cell perinuclear localization observed in mid-oogenesis, and represents the earliest event requiring exu. bed message is transported to the oocyte by an unknown mechanism, where it is localized at the anterior cortex. This event also requires microtubules. Localization at late stages of oogenesis and the maintenance of localization require both microtubules and the sww gene product. It is likely that exu participates in oocyte localization as well; for clarity, only its initial role in localization is diagrammed here.

Fig. 7.

A model for bed RNA localization, bed RNA is synthesized in nurse cells and becomes associated with microtubules in a process that requires the exu gene product. This association results in the nurse cell perinuclear localization observed in mid-oogenesis, and represents the earliest event requiring exu. bed message is transported to the oocyte by an unknown mechanism, where it is localized at the anterior cortex. This event also requires microtubules. Localization at late stages of oogenesis and the maintenance of localization require both microtubules and the sww gene product. It is likely that exu participates in oocyte localization as well; for clarity, only its initial role in localization is diagrammed here.

The significance of perinuclear localization is not clear. We have suggested (Stephenson and Pokrywka, 1991) that bed message localization around nurse cell nuclei is a prelude to oocyte localization in postvitellogenic stages. During this time, large amounts of nurse cell cytoplasm flow into the oocyte. Retention of bed RNA in the anterior oocyte during the bulk flow of nurse cell cytoplasm may require that bed RNA become associated with the localization machinery in nurse cells. The microtubule dependence of both nurse cell localization and oocyte localization is consistent with this idea. Perhaps bed RNA becomes associated with microtubules around nurse cell nuclei and these microtubules are assembled into networks in the anterior oocyte. Alternatively bed RNA may become associated in nurse cells with a microtubule-based transport system which is continuous with the anterior oocyte cortex.

Once an initial association with microtubules is established, bed RNA moves into the oocyte through an unknown mechanism, and is localized at the anterior cortex. Data presented elsewhere (Stephenson and Pokrywka, 1991) suggest that bed RNA is restricted to a layer of the peripheral cortex less than 5 fim in thickness. The relocalization of bed message after nocodazole treatment suggests that microtubules are required for maintaining bed RNA in this cortical location.

The maintenance of localization in the oocyte also requires the SHW gene product, which carries out an essential role only at later stages of oogenesis and only in the oocyte. In sww mutants, localization in the oocyte is normal until late stage 10, when localization becomes unstable and is eventually lost completely. The late and gradual loss of bed RNA localization during oogenesis, and abnormalities in nuclear cleavages during embryogenesis (Zalokar et al. 1975; Pokrywka and Stephenson, in preparation) have been interpreted as suggesting a role for sww in cytoskeletal regulation during oogenesis and early embryogenesis. The results of taxol treatments of swiv mutant egg chambers are consistent with the idea that the asssociation of bed RNA with microtubules is normal in these egg chambers. We predict that the microtubule networks to which bed RNA is attached are less stable or incorrectly regulated in sww oocytes, so that localization is not maintained.

In this report, we demonstrate that microtubules are involved in the localization of bed RNA during Drosophila oogenesis. The localization of developmentally important molecules in other systems has also been shown to rely on cytoskeletal elements. In C. elegans, P-granules are segregated in a process requiring microfilaments (Hill and Strome, 1988). Vgl RNA localization in Xenopus has been shown to require microtubules and microfilaments (Yisraeli et al. 1990). Like bed RNA, Vgl RNA is localized to a distinct cortical domain, suggesting that the cortical cytoskeleton may be enriched in the elements necessary for localization. In Drosophila, cyclin B is localized to the posterior pole during oogenesis and early embryogenesis (Raff et al. 1990), an event that is also thought to be dependent on microtubules. Further research will be needed to determine whether other similarities exist between these and other systems of determinant localization.

We thank Joanna Olmsted for providing anti-tubulin antibodies and for assistance with the immunocytochemistry. Thanks also to Vijay Hegde for immunocytochemistry protocols. We are grateful to Drs Bob Angerer, David Goldfarb and Joanna Olmsted for helpful comments on the manuscript. This work was supported by grants from the National Science Foundation (DCB 8702160) and National Institutes of Health (GM 41513). N.J.P. was supported by an NIH Predoctoral Training Grant in Genetics and Regulation awarded to the University of Rochester.

Berleth
,
T.
,
Burri
,
M.
,
Thoma
,
G.
,
Bopp
,
D.
,
Richstein
,
S.
,
Frigerio
,
G.
,
Noll
,
M.
and
Nosslein-Volhard
,
C.
(
1988
).
The role of localization of bicoid RNA in organizing the anterior pattern of the Drosophila embryo
.
EMBO J
.
7
,
1749
1756
.
Cooper
,
J. A.
(
1987
).
Effects of cytochalasin and phalloidin on actin
.
J. Cell Biol
.
105
,
1473
1478
.
Driever
,
W.
and
Nosslein-Volhard
,
C.
(
1988a
).
A gradient of bicoid protein in Drosophila embryos
.
Cell
54
,
83
93
.
Driever
,
W.
and
Nosslein-Volhard
,
C.
(
1988b
).
The bicoid protein determines position in the Drosophila embryo in a concentration-dependent manner
.
Cell
54
,
95
104
.
Driever
,
W.
,
Siegel
,
V.
and
Nosslein-Volhard
,
C.
(
1990
).
Autonomous determination of anterior structures in the early Drosophila embryo by the bicoid morphogen
.
Development
109
,
811
820
.
Dustin
,
P.
(
1984
).
Taxol: An agent that favorizes microtubule assembly
.
In Microtubules
, pp.
213
217
.
Berlin
:
Springer-Verlag
.
French
,
V.
(
1988
).
Gradients and insect segmentation
.
Development
103
Supplement
, pp.
3
16
.
Frohnhôfer
,
H. G.
and
Nosslein-Volhard
,
C.
(
1987
).
Maternal genes required for the anterior localization of bicoid activity in the embryo of Drosophila
.
Genes Dev
.
1
,
880
890
.
Gans
,
M.
,
Audit
,
C.
and
Masson
,
M.
(
1975
).
Isolation and characterization of sex-linked female-sterile mutants in Drosophila melanogaster
.
Genetics
81
,
683
704
.
Geuens
,
G. M. A.
,
Nuydens
,
R. M.
,
Willebrords
,
R. E.
,
Van De Veire
,
R. M. L.
,
Goossens
,
F.
,
Dragonetti
,
C. H.
,
Mareel
,
M. M. K.
and
Debrabander
,
M. J.
(
1985
).
Effects of tubulozole on the microtubule system of cells in culture and in vivo
.
Cancer Res
.
45
,
733
752
.
Gutzeit
,
H. O.
(
1986a
).
The role of microfilaments in cytoplasmic streaming in Drosophila follicles
.
J. Cell Sci
.
80
,
159
169
.
Gutzeit
,
H. O.
(
1986b
).
The role of microtubules in the differentiation of ovarian follicles during vitellogenesis in Drosophila
.
Roux’s Arch, devl Biol
.
195
,
173
181
.
Hill
,
D. P.
and
Strome
,
S.
(
1988
).
An analysis of the role of microfilaments in the establishment and maintenence of asymmetry in Caenorhabditis elegans zygotes
.
Devl Biol
.
125
,
75
84
.
Jeffery
,
W. R.
(
1989
).
Localized mRNA and the egg cytoskeleton
.
Inti Rev. Cytol
.
119
,
151
195
.
King
,
R. C.
(
1970
).
Ovarian Development in Drosophila melanogaster
.
New York
:
Academic Press
.
Lasko
,
P. F.
and
Ashburner
,
M.
(
1990
).
Posterior localization of vasa protein correlates with, but is not sufficient for, pole cell development
.
Genes Dev
.
4
,
905
921
.
Mahowald
,
A. P.
and
Kambysellis
,
M. P.
(
1980
).
Oogenesis
.
In Genetics and Biology of Drosophila
, vol.
2c
, (ed.
M.
Ashburner
and
T. F. R.
Wright
), pp.
141
224
.
New York
:
Academic Press
.
Petri
,
W. H.
,
Mindrinos
,
M. N.
,
Lombard
,
M. F.
and
Margaritis
,
L. H.
(
1979
).
In vitro development of the Drosophila chorion.in a chemically defined organ culture medium
.
Wilhelm Roux’s Arch, devl Biol
.
186
,
351
.
Raff
,
J. W.
,
Whitefield
,
W. G. F.
and
Glover
,
D. M.
(
1990
).
Two distinct mechanisms localize cyclin B transcripts in syncytial Drosophila embryos
.
Development
110
,
1249
1261
.
Robb
,
J. A.
(
1969
).
Maintenance of imaginai discs of Drosophila melanogaster in chemically defined media
.
J. Cell Biol
.
41
,
876
885
.
Schüpbach
,
T.
ànd Wieschaus
,
E.
(
1986
).
Maternal-effect mutations altering anterior posterior pattern in the Drosophila embryo
.
Wilhelm Roux’s Arch’, devl Biol
.
195
,
302
317
.
Stephenson
,
E. C.
,
Chao
,
Y. C.
and
Fackenthal
,
J. D.
(
1988
).
Molecular analysis of the swallow gene of Drosophila melanogaster
.
Genes and Development
2
,
1655
1665
.
Stephenson
,
E. C.
and
Pokrywka
,
N. J.
(
1991
).
Localization of bicoid message during Drosophila oogenesis
.
In The Cytoskeleton in Development
(ed.
E.
Bearer
).
Orlando
:
Academic Press, Inc. (In press
).
St Johnston
,
D.
,
Driever
,
W.
,
Berleth
,
T.
,
Richstein
,
S.
and
Nüsslein-Volhard
,
C.
(
1989
).
Multiple steps in the localization of bicoid RNA to the anterior pole of Drosophila embryos
.
Development
107
Supplement
,
13
19
.
Strome
,
S.
and
Wood
,
W. B.
(
1983
).
Generation of asymmetry and segregation of germ line granules in early C. elegans embryos
.
Cell
35
,
15
25
.
Struhl
,
G.
,
Struhl
,
K.
and
Macdonald
,
P. M.
(
1989
).
The gradient morphogen bicoid is a concentration-dependent transcriptional activator
.
Cell
57
,
1259
1273
.
Yisraeli
,
J. K.
,
Sokol
,
S.
and
Melton
,
D. A.
(
1990
).
A two-step model for the localization of maternal mRNA in Xenopus oocytes: Involvement of microtubules and microfilaments in the translocation and anchoring of Vgl mRNA
.
Development
108
,
289
298
.
Zalokar
,
M.
,
Audit
,
C.
and
Erk
,
I.
(
1975
).
Developmental defects of female-sterile mutants of Drosophila melanogaster
.
Devl Biol
.
47
,
419
432
.