During Drosophila oogenesis, the posterior determinant, Oskar, is tightly localized at the posterior pole of the oocyte. The exclusive accumulation of Oskar at this site is ensured by localization-dependent translation of oskar mRNA: translation of oskar mRNA is repressed during transport and activated upon localization at the posterior cortex. Previous studies have suggested that oskar translation is poly(A)-independent. We show that a long poly(A) tail is required for efficient oskar translation, both in vivo and in vitro, but is not sufficient to overcome BRE-mediated repression. Moreover, we show that accumulation of Oskar activity requires the Drosophila homolog of Cytoplasmic Polyadenylation Element Binding protein (CPEB), Orb. As posterior localization of oskar mRNA is an essential prerequisite for its translation, it was critical to identify an allele of orb that does localize oskar mRNA to the posterior pole of the oocyte. We show that flies bearing the weak mutation orbmel localizeoskar transcripts with a shortened poly(A) that fails to enhanceoskar translation, resulting in reduced Oskar levels and posterior patterning defects. We conclude that Orb-mediated cytoplasmic polyadenylation stimulates oskar translation to achieve the high levels of Oskar protein necessary for posterior patterning and germline differentiation.

The definition of spatial cues is essential to many biological processes,from cell movement to embryonic development. The ability of cells to sense space and initiate and maintain polarity relies on a set of processes that culminates with the creation of specialized cytoplasmic and cortical compartments, and the establishment of axes of symmetry.

During embryogenesis, the establishment of the embryonic axes is a prerequisite to proper patterning of the embryo. In Drosophila melanogaster, both the anteroposterior (AP) and dorsoventral (DV) axes are established during oogenesis, through the prior asymmetric localization of cytoplasmic determinants in the oocyte. To ensure the tight and exclusive presence of these determinants at the sites where their activity is required,they are often synthesized in loco after prelocalization of their transcripts,which are then subject to temporal and spatial translational regulation(St Johnston, 1995).

oskar (osk) mRNA, which encodes the posterior determinant, is localized at the posterior pole of the oocyte and embryo,where Osk protein is required to assemble the germ plasm(Beams and Kessel, 1974;Ephrussi et al., 1991;Kim-Ha et al., 1991), a cytoplasmic sub-compartment containing the abdominal and germline determinants of the fly (Lehmann and Nüsslein-Volhard, 1986). Females bearing mutations inosk produce embryos lacking both abdomen and germline. Conversely,mislocalization of Osk activity leads to formation of ectopic abdominal structures at the expense of the anterior structures(Ephrussi et al., 1991;Ephrussi and Lehmann, 1992;Smith et al., 1992). Hence,restriction of Osk activity to the posterior pole is also essential for normal development to occur.

Coupling of osk mRNA translational activation to its localization at the posterior pole of the oocyte ensures specific and exclusive accumulation of Osk protein at the posterior pole of the oocyte(Kim-Ha et al., 1995;Markussen et al., 1995;Rongo et al., 1995). The oocyte develops in a 16-cell cyst, consisting of 15 transcriptionally active nurse cells and the oocyte itself, all interconnected by cytoplasmic bridges(Spradling, 1993). From the early stages of oogenesis onwards, osk mRNA is transcribed by the nurse cells and accumulates in the oocyte. At mid-oogenesis (stage 8)osk mRNA becomes expressed at high levels at the posterior pole,where it is translated (Ephrussi et al.,1991; Kim-Ha et al.,1991).

Prior to its localization, osk mRNA is translationally repressed through the binding of Bruno (Bru) repressor protein to discrete elements(Bruno Response Elements or BRE) present in the osk 3′untranslated region (3′UTR) (Kim-Ha et al., 1995; Webster et al.,1997). Upon posterior localization, osk translation is derepressed (Gunkel et al.,1998). The mechanisms that underlie repression and derepression ofosk remain elusive. In many species, however, it appears that the translational status of a regulated transcript is determined by the length of its poly(A) tail, and the switch from a silenced to a translationally active state is controlled by cytoplasmic polyadenylation (reviewed byRichter, 1999). Accordingly,an increase in translation is often associated with poly(A) tail elongation,whereas translational silencing correlates with poly(A) tail shortening(Lieberfarb et al., 1996;Sallés et al., 1994;Sheets et al., 1995). InXenopus, where this phenomenon has been most extensively studied, the cis-acting elements involved in cytoplasmic polyadenylation have been identified and include the AAUAAA hexanucleotide (also required for nuclear polyadenylation) and the U-rich cytoplasmic polyadenylation element (CPE)(Fox, 1989; McGrew and Richter,1990). The CPE is specifically bound by the polyadenylation-inducing protein CPEB(Hake and Richter, 1994;Stebbins-Boaz et al., 1996),whose presumed function is to recruit and stabilize the cytoplasmic polyadenylation machinery (Mendez et al.,2000).

In Drosophila, no discrete cis-acting elements involved in cytoplasmic polyadenylation have been identified so far. This has made it difficult to assess directly the involvement of cytoplasmic polyadenylation in translational control of regulated transcripts. A putative CPEB is encoded by the oo18 RNA binding (orb) locus. Strong orballeles affect osk mRNA localization(Christerson and McKearin,1994; Lantz et al.,1994), preventing an analysis of their effect on osktranslation. Weaker orb alleles, for which the pattern ofosk mRNA localization has not yet been analyzed, are available and might provide a useful tool with which to dissect the role of Orb in modulating osk translation. Among those, the hypomorphic orballele, orbmel, completes oogenesis and has been previously shown to affect the establishment of the AP axis, by interfering with Osk protein accumulation (Christerson and McKearin, 1994). The observation by Chang et al.(Chang et al., 1999) that in this mutant osk poly(A) tail length is shortened supports the idea that cytoplasmic polyadenylation might be involved at least in some aspects ofosk translational regulation. However, as posterior localization is required for osk mRNA translation, characterization of the effect of the orbmel mutation on osk mRNA localization is an essential prerequisite to any assessment of a possible role of Orb in regulation of osk polyadenylation and Osk protein accumulation.

At least two non-mutually exclusive scenarios may be hypothesized. Polyadenylation of osk transcript might be required to overcome translational repression and to activate translation upon posterior localization, a mechanism widely used to repress/derepress translation of several developmentally regulated transcripts. This would predict that, even upon posterior localization, no Osk activity is synthesized in the absence of polyadenylation. Alternatively, poly(A) tail elongation might not be a prerequisite for translation per se, but rather be required to enhance translation efficiency, allowing for Osk protein accumulation. In fact, it is already established that the poly(A) tail of an mRNA can synergize with its 5′ cap structure to enhance translation initiation. This synergistic effect of the cap and the poly(A) tail arises from the simultaneous binding of the poly(A)-binding protein PABP and the cap-binding protein eIF4E, to the translation initiation factor eIF4G, promoting reinitiation of translation by terminating ribosomes on the same transcript(Gallie, 1991;Tarun and Sachs, 1996;Wells et al., 1998). This second hypothesis predicts that upon posterior localization of oskmRNA, translation would initiate, but with limited efficiency, unless enhanced by polyadenylation. In this light it is interesting that posterior patterning and germline differentiation require different thresholds of Osk activity(Markussen et al., 1995;Rongo et al., 1995). Posterior patterning is robust and requires only low levels of Osk, whereas pole cell formation appears to be very sensitive to any reduction in Osk protein levels. To guarantee production of sufficient amounts of Osk for fulfilment of both functions, translational derepression and enhancement might be involved.

We address the role of cytoplasmic polyadenylation in control ofosk translation. We have measured osk poly(A) tail length in vivo and correlated its polyadenylation status to its translation status both in vivo and in vitro. We show that within RNA corresponding to all stages of oogenesis, a discrete population of osk mRNA bears a long poly(A)tail. A poly(A) tail of the maximum length observed for osk mRNA in vivo is necessary and sufficient to enhance translation of a chimericosk transcript in vitro, in the absence of Bruno repressor protein. However, addition of a poly(A) tail of any length does not suffice to overcome BRE-mediated repression, at least in vitro. We also show that maintenance of a long poly(A) tail on osk transcript requires Orb and is essential for Osk protein accumulation. In orbmel mutant egg chambers,the osk poly(A) tail is shorter than in wild type. Shortening of theosk poly(A) tail correlates with a reduction in Osk protein accumulation despite posterior localization of the transcript. This leads to complete sterility and, in extreme cases, to loss of posterior embryonic patterning. However, osk translation in orbmelappears to be only attenuated but not abolished. We therefore suggest a role for Orb in posterior patterning by enhancing osk translation through the addition or maintenance of a long poly(A) tail.

### Poly(A) tail measurement

Drosophila melanogaster females were kept at 25°C for 2-3 days on yeast before dissection. Ovaries were recovered in PBT (PBS/0.1% Tween-20)and frozen in liquid N2 without any buffer. Total RNA was extracted following the RNA clean protocol (Hybaid).

osk poly(A) tail was measured mainly using an RNaseH-based method,as described by Zangar et al. (Zangar et al., 1995), including some modification as described by Lie and Macdonald (Lie and Macdonald,1999). After denaturation at 85°C for 5 minutes, 20 μg of total ovary RNA were hybridized to 5 μg of a complementary DNA oligo specific for osk mRNA (5′-CGC CAG AAT TCT ACA CTG TG-3′)with and without 5 μg of dT16 oligo, at 42°C for 10 minutes. The RNA-DNA hybrid was than specifically digested with RNase H (1 U from Gibco) at 30°C for 30 minutes in 20 μl of RNase H buffer (40 mM Tris-HCl pH 8, 4 mM MgCl2, 1 mM DTT, 30 ng/ml BSA). Reactions were stopped by phenol-chloroform extraction and ethanol precipitation. The RNA fragments were resuspended in Ambion RNA loading buffer, separated on 5%polyacrylamide denaturing gel and transferred by electro-blotting for 3-4 hours at 20 V to NEN Genescreen hybridization membrane. osk mRNA was detected by northern hybridization. Quantification of the RNA populations bearing different poly(A) tail length was performed on a Macintosh computer using the public domain NIH Image program (developed at the US National Institutes of Health and available on the internet athttp://rsb.info.nih.gov/nih-image/).

Alternatively, poly(A) tail length was measured using the poly(A) test(PAT). cDNAs were synthesized from 500 ng of total ovary RNA, using the BRL Superscript retrotranscriptase. The protocol used was modified to maintain the poly(A) tract in the cDNA. Prior to first strand synthesis, the poly(A) tail was hybridized to saturation with dT16 oligo and the oligo dT16 subunits ligated before adding the oligodT-anchor, as described by Sallés et al.(Sallés et al., 1994). Specific osk cDNAs were amplified by PCR in the presence of[α32P]dCTP using the oligodT-anchor and a specificosk oligo (5′-AAG CGC TTG TTT GTA GCA CA-3′). PCR products were separated on 5% denaturing polyacrylamide gels.

### DNA constructs

m1m2lacWT was previously described(Gunkel et al., 1998). m1m2lacΔABC was derived from m1m2lacWT by deletion of anEcoRI/DraI fragment containing the AB repressor region, and additional deletion of the C region by PCR-directed mutagenesis, using the PCR primers ΔCbclIIup (5′-ACT GTC TAG AAC GTT TTT TTT GTC C-3′)and T3XL (5′-CGA AAT TAA CCC TCA CTA AAG GGA-3′). This construct lacks nucleotides 3660-3778 and 4416-4487 of the m1m2lacWT chimericosk-lacZ transcript.

Cassettes of 36, 73, 98 and 150 adenosines (As) were cloned downstream of the various 3′UTRs using the unique NotI and KpnI sites in the plasmids. The derived constructs were named after the original plasmids and differ only in the length of the poly(A) cassette. Cassettes encoding a poly(A) tract longer than 150 nucleotides could not be cloned because of instability in bacteria. Poly(A) tails longer than 150A were added by enzymatic polyadenylation, using recombinant yeast Poly(A) Polymerase(yPAP) from Amersham. Transcript (1 μg) was polyadenylated at 30°C for 20 minutes in the presence of 10 mM ATP, 1× PAP buffer and 600 U of yPAP.

### In vitro transcription

Capped chimeric osk-lacZ mRNAs were synthesized using the SP6 mMessage mMachine kit from Ambion. After a 2 hour reaction, the template DNA was eliminated by digestion with DNAseI, and the RNA purified using RNeasy columns from QIAGEN. Prior to transcription, the template was linearized with an appropriate restriction enzyme, cutting at a unique site downstream of the 3′UTR. C36luca mRNA was synthesized as described(Gray and Hentze, 1994). The RNAs were trace labeled with [α32P]UTP to facilitate assessment of their concentration and integrity. All RNAs used in the same experiment were synthesized in parallel.

### In vitro translation assay

Translation assays were performed as described previously(Castagnetti et al., 2000). Briefly, 50 ng of template osk mRNA were translated in a 12.5 μl reaction containing 60 μM amino acids, 16.8 mM creatine phosphate, 80 ng/μl creatine kinase, 24 mM HEPES (pH 7.4), 0.6 mM Mg(OAc)2, 60 mM KOAc, 0.1 mM Spermidine, 1.2 mM DTT, 100 ng/μl calf liver tRNA and 40%ovary or embryo extract. Luciferase mRNA (20 ng) was co-translated as an internal control. The reactions were incubated for 90 minutes at 25°C. The translation efficiency of the osk chimeric mRNAs was quantified using the chemiluminescent β-Gal Reporter Gene Assay (Roche), following the protocol provided by the manufacturer. Luciferase activity was measured according to Brasier et al. (Brasier et al., 1989).

### In situ hybridization and immunostaining

Two- to three-day-old females were dissected in PBS and ovaries were fixed for 20 minutes in 4% paraformaldehyde in PBS. After washing twice in PBT (0.1%Triton X-100), ovaries were extracted for 1.5 hours in PBS 1% Triton X-100,blocked for 1 hour in PBS 0.3% Triton X-100, 0.5% BSA and then incubated overnight with primary antibodies (α-Stau at 1/250, α-Osk at 1/3000). After washing, samples were incubated with FITC/rhodhamine-coupled secondary antibodies (1/500 in PBT) for detection. Microscopy was carried out using a Leica TCS SP confocal microscope.

For in situ hybridization, ovaries were fixed in 4% paraformaldehyde and processed according to Glotzer and Ephrussi(Glotzer and Ephrussi, 1999)using osk digoxigenin-labeled probes.

### Immunoprecipitation

Total ovarian extract, prepared as for the in vitro translation assay, was incubated with 3 μl of specific antibody in 100 μl of hybridization buffer [20 mM HEPES (pH 7.5), 150 mM NaCl, 2.5 mM MgCl2, 250 mM sucrose, 0.05% NP-40, 0.5% Triton X-100, 1X EDTA-free protease inhibitors cocktail from Roche]. After an overnight incubation, sepharose-protein A beads were added for 50 minutes to allow binding. Beads were washed three times with hybridization buffer without MgCl2. Protein were denatured in Laemmli buffer, run on a 10% SDS polyacrylamide gel and detected using monoclonal Orb antibodies (1:500). The monoclonal antibodies orb6H4 and orb4H8 developed by P. Schedl were obtained from the Developmental Studies Hybridoma Bank developed under the auspices of the NICHD and maintained by The University of Iowa, Department of Biological Sciences, Iowa City, IA 52242.

### An Orb-dependent long poly(A) tail on osk mRNA in vivo

Modulation of poly(A) tail length is a powerful and widely employed mechanism in translational control. In several species, translational silencing is correlated with a deadenylated status of the mRNA, and activation of translation parallels poly(A) tail elongation. We decided to address directly the role of the poly(A) tail in osk translational control,measuring the length of the osk poly(A) tail in vivo, and assessing the influence of such a tail on translation both in vivo and in vitro.

Previous evaluations of the length of the osk poly(A) tail involved different methods of measurement, yielding differing results(Chang et al., 1999;Lie and Macdonald, 1999;Sallés et al., 1994). We therefore decided to perform the two assays typically used for this purpose on the same samples: the PAT assay, involving PCR amplification; and the more direct, RNaseH/northern blot-based assay. For the PAT assay, ovaries were divided into early (stage 1-5) and late stages (stage 7-14). During the early stages of oogenesis, osk mRNA is unlocalized and translationally silent. During the late stages, at least a portion of osk mRNA is localized and translated. In both pools, osk poly(A) tails ranging from 100 to 230 A were present (Fig. 1A). Using the RNaseH-based assay, poly(A) tails 200 in length were also detected in both samples (Fig. 1B and data not shown). Hence, a significant fraction ofosk mRNA molecules bears remarkably long poly(A) tails, consistent with some of the previous reports (Chang et al., 1999; Sallés et al., 1994).

Fig. 1.

osk mRNA has a long poly(A) tail in vivo. (A) osk mRNA poly(A) tail measurement using the PAT assay previously described by Sallés et al. (Sallés et al., 1994) and Chang et al.(Chang et al., 1999). The longest detected osk poly(A) tail is about 200 A long, both in early(up to stage 5) and late (5-14) stages of oogenesis. The length of the poly(A)tail is equal to the difference between the top of the smear and the fragment of osk mRNA amplified, both indicated by arrows. (B) Measurement of the osk poly(A) tail length using an RNAseH based assay. Total ovarian RNA was hybridized to a DNA oligonucleotide complementary to the 3′-most region of the osk 3′UTR, in the presence (+) or in the absence (-) of excess oligo dT16. The maximum length present in the osk mRNA population corresponds to the difference between the top of the smear in the -' lane and the baseline given by the +' lane. In wild-type ovaries, the osk poly(A) tail reaches a length of 200 A,whereas in orbmel ovaries the osk poly(A) tail is shortened to about 130-150 A. Quantitation using the NIH image program shows that in wild-type ovaries 36% of osk mRNA has a tail length of 150-200 A and that this population is reduced to 4.5% inorbmel homozygous ovaries. The same amount of total RNA was processed in each sample. The result obtained for the wild-type RNA was confirmed using a second osk oligo. (C) Translation efficiency of chimeric osk-lacZ mRNAs bearing poly(A) tails of different lengths in embryo extract. Efficient translation activation was observed when a poly(A)tail longer than 200 A was added to the transcript. Tails of 36, 53, 73 and 150A in length did not activate translation. The difference in translation between A0, A36, A53, A73 and A150 can be explained by a comparable increase in stability of the transcripts. The difference in half-life between the A0 and An transcript does not explain the difference in their translation efficiency. Previous reports had suggested that osk translation was poly(A) independent (Castagnetti et al.,2000; Lie and Macdonald,1999). In both cases the poly(A) tails used in the assay were shorter than 200 A and therefore not competent to activate translation according to our present analysis.

Fig. 1.

osk mRNA has a long poly(A) tail in vivo. (A) osk mRNA poly(A) tail measurement using the PAT assay previously described by Sallés et al. (Sallés et al., 1994) and Chang et al.(Chang et al., 1999). The longest detected osk poly(A) tail is about 200 A long, both in early(up to stage 5) and late (5-14) stages of oogenesis. The length of the poly(A)tail is equal to the difference between the top of the smear and the fragment of osk mRNA amplified, both indicated by arrows. (B) Measurement of the osk poly(A) tail length using an RNAseH based assay. Total ovarian RNA was hybridized to a DNA oligonucleotide complementary to the 3′-most region of the osk 3′UTR, in the presence (+) or in the absence (-) of excess oligo dT16. The maximum length present in the osk mRNA population corresponds to the difference between the top of the smear in the -' lane and the baseline given by the +' lane. In wild-type ovaries, the osk poly(A) tail reaches a length of 200 A,whereas in orbmel ovaries the osk poly(A) tail is shortened to about 130-150 A. Quantitation using the NIH image program shows that in wild-type ovaries 36% of osk mRNA has a tail length of 150-200 A and that this population is reduced to 4.5% inorbmel homozygous ovaries. The same amount of total RNA was processed in each sample. The result obtained for the wild-type RNA was confirmed using a second osk oligo. (C) Translation efficiency of chimeric osk-lacZ mRNAs bearing poly(A) tails of different lengths in embryo extract. Efficient translation activation was observed when a poly(A)tail longer than 200 A was added to the transcript. Tails of 36, 53, 73 and 150A in length did not activate translation. The difference in translation between A0, A36, A53, A73 and A150 can be explained by a comparable increase in stability of the transcripts. The difference in half-life between the A0 and An transcript does not explain the difference in their translation efficiency. Previous reports had suggested that osk translation was poly(A) independent (Castagnetti et al.,2000; Lie and Macdonald,1999). In both cases the poly(A) tails used in the assay were shorter than 200 A and therefore not competent to activate translation according to our present analysis.

As a first step towards investigating a potential role of cytoplasmic polyadenylation and the poly(A) tail in osk translational control, we measured the length of the osk poly(A) tail in ovaries of the hypomorphic orb mutant, orbmel. orbmelmutant females complete oogenesis and produce embryos out of which 25% show posterior patterning defects, reflecting the reported reduction in Osk protein accumulation (Christerson and McKearin,1994; Markussen et al.,1995). As measured by the RNaseH-based assay, the oskpoly(A) tail is ∼50 A shorter in orbmel mutant females than in the wild type (Fig. 1B), extending to only 150 A. Thus, the presence of a long poly(A)tail on osk mRNA requires orb function. This confirms the observation that osk mRNA is subject to cytoplasmic polyadenylation in vivo (Chang et al.,1999).

The fact that the osk poly(A) tail is somewhat shorter inorbmel than in wild type suggested that this difference might be the basis of the observed defect in Osk protein accumulation in the mutant. However, despite their shorter poly(A) tails inorbmel ovaries, the osk transcripts still bear relatively long tails that should in principle be able to stimulate translation. To determine whether the shorter poly(A) tail of oskmRNA in orbmel could affect Osk accumulation, we tested whether the length of the poly(A) tail is critical for translational stimulation of osk mRNA expression. We compared the in vitro translation efficiency of chimeric osk-lacZ transcripts bearing poly(A) tails of different lengths. As investigation of the role of the poly(A) tail in osk translational stimulation requires an environment in which osk translation is not otherwise silenced, we used the cell-free translation system obtained from Drosophila embryos, which contains no Bru repressor protein. As shown inFig. 1C, a poly(A) tail as long as 150 A does not stimulate translation above that of transcripts bearing no poly(A) tail. This result is consistent with previous reports from Lie and Macdonald (Lie and Macdonald,1999) and Castagnetti et al.(Castagnetti et al., 2000) in which poly(A) tails of 36 A and 73 A failed to stimulate translation ofosk reporter transcripts in vitro. Remarkably, however, addition of a tail in excess of 200 A enhances translation 20- to 100-fold above that of a transcript bearing no poly(A) tail (Fig. 1C). This increase in translation following polyadenylation is due to a substantial increase in the rate of translation initiation, rather than to stabilization of the transcripts, whose half-lives are comparable(Fig. 1C; data not shown). These results suggest that a tail longer than 150 A may also be necessary for efficient Osk accumulation in vivo, consistent with a role of cytoplasmic polyadenylation and Orb protein in osk mRNA translational regulation.

### A poly(A) tail longer than 150A is necessary for Osk accumulation in vivo

Previous reports have shown the involvement of Orb in the establishment of anteroposterior polarity (Christerson and McKearin, 1994). To study the potential dependence of osktranslation on the poly(A) status of the mRNA in vivo, we first had to determine whether the mRNA is correctly localized in the weakorbmel mutants, as analysis of strong orb alleles has highlighted a role of Orb in osk mRNA posterior localization. In situ hybridization revealed that osk mRNA is correctly localized in 67% and partially localized in 10% of orbmel oocytes(Fig. 2, upper panel).osk mRNA is not detected at the posterior pole of the remaining 23%of orbmel oocytes. Staufen (Stau) protein, whose distribution has been shown to mirror that of osk mRNA during oogenesis (St Johnston et al.,1991), shows a similar degree of localization inorbmel egg chambers(Fig. 2 lower panel). As shown in Fig. 2, 100% of wild-type and 81% of orbmel oocytes at stage 9/10 accumulate Stau at the posterior pole. In 80% of orbmel oocytes that show posterior Stau accumulation, the protein is fully localized at the posterior pole, and in 20% Stau is also detected in the cytoplasm. Thus, in contrast to the strong orb alleles(Christerson and McKearin,1994; Lantz et al.,1994), orbmel only mildly affects oskmRNA localization, rendering it suitable for an analysis of the role of cytoplasmic polyadenylation in osk translation.

Fig. 2.

osk mRNA localization is only mildly affected inorbmel mutants. Whole-mount in situ hybridization (upper panel) was performed on ovaries from wild-type and orbmelhomozygous flies, using a DIG-labeled osk probe. The result obtained by RNA in situ hybridization was confirmed by immunostaining with α-Stau antibody (lower panel).

Fig. 2.

osk mRNA localization is only mildly affected inorbmel mutants. Whole-mount in situ hybridization (upper panel) was performed on ovaries from wild-type and orbmelhomozygous flies, using a DIG-labeled osk probe. The result obtained by RNA in situ hybridization was confirmed by immunostaining with α-Stau antibody (lower panel).

An indication that Orb is required for osk translation is that, as shown in Fig. 3A and previously reported (Markussen et al.,1995), the amount of both Osk isoforms is dramatically reduced inorbmel when compared with wild type. This reduction could in principle be a consequence of the osk mRNA localization defect. However, the amount of Osk protein in orbmel is at best 25% of the wild type, in spite of the fact that 65% oforbmel oocytes show normal posterior osk mRNA localization. This suggests an involvement of Orb in osk translation. Consistent with this, antibody staining of ovaries reveals that only 67% oforbmel oocytes that localize Stau also accumulate Osk at the posterior (Fig. 3B). In the remaining 33%, no Osk protein is detected. However, trace amounts of Osk are presumably also produced in these oocytes, as maintenance of Stau andosk mRNA at the posterior pole requires Osk protein itself(Rongo et al., 1995;Vanzo and Ephrussi, 2002). It therefore appears that, in orbmel ovaries, osktranslation is impaired but not abolished. In these egg chambers, oskmRNA bears a poly(A) tail that is insufficient to support efficient translation in vitro. Taken together these observations suggest that Orb-mediated cytoplasmic polyadenylation is required to enhance translation,but not for translational derepression of osk mRNA.

Fig. 3.

Osk accumulation in orbmel homozygous egg chambers. (A)Western blot analysis was performed on total ovarian protein extracts of similarly aged wild-type and orbmel flies. The reduction in Osk phosphorylation is presumably a consequence of the reduction in Osk accumulation, as Osk protein has been shown to be required for its own phosphorylation and stabilization(Markussen et al., 1997;Riechmann et al., 2002). (B)Whole-mount antibody staining of orbmel egg chambers with Stau, and Stau and Osk antibodies.

Fig. 3.

Osk accumulation in orbmel homozygous egg chambers. (A)Western blot analysis was performed on total ovarian protein extracts of similarly aged wild-type and orbmel flies. The reduction in Osk phosphorylation is presumably a consequence of the reduction in Osk accumulation, as Osk protein has been shown to be required for its own phosphorylation and stabilization(Markussen et al., 1997;Riechmann et al., 2002). (B)Whole-mount antibody staining of orbmel egg chambers with Stau, and Stau and Osk antibodies.

To confirm this hypothesis, we checked if the addition of a long poly(A)tail could overcome BRE-mediated repression in vitro. For this purpose, we used the ovarian extract, which contains Bru and recapitulates BRE-mediated repression. As shown in Fig. 4,addition of a >250 A tail does not overcome Bru-mediated repression, as BRE-containing transcripts are less efficiently translated than their BRE-deleted counterparts, whether or not they bear a poly(A) tail. As in the embryo extract, the polyadenylated BRE+ and BRE- transcripts are translated more efficiently than the corresponding transcripts lacking a poly(A) tail. Taken together, these results suggest that addition of a long poly(A) tail is not sufficient to overcome repression, but has a stimulatory effect on translation of both BRE+ and BRE- transcripts. The physiological relevance of cytoplasmic polyadenylation in osk regulation is strengthen by the observation that, in orbmel egg-chambers, even when levels of Osk sufficient to support abdomen formation are produced, those embryos that develop into adult females are sterile(Table 1).

Fig. 4.

A long poly(A) tail is not sufficient to overcome BRE-mediated repression. Translation efficiency of chimeric osk-lacZ mRNAs bearing either no poly(A) tail or a tail of >200A in ovary extract. The transcripts used in these experiments bear the 5′ region of osk mRNA up to the second AUG codon, and either the wild-type osk 3′UTR (WT) or anosk 3′UTR from which the BREs were deleted (ΔABC). The data shown are representative of three independent experiments. The values were normalized for those of a luciferase RNA co-translated as an internal control and the experimental error varies between 3% and 10%.

Fig. 4.

A long poly(A) tail is not sufficient to overcome BRE-mediated repression. Translation efficiency of chimeric osk-lacZ mRNAs bearing either no poly(A) tail or a tail of >200A in ovary extract. The transcripts used in these experiments bear the 5′ region of osk mRNA up to the second AUG codon, and either the wild-type osk 3′UTR (WT) or anosk 3′UTR from which the BREs were deleted (ΔABC). The data shown are representative of three independent experiments. The values were normalized for those of a luciferase RNA co-translated as an internal control and the experimental error varies between 3% and 10%.

Table 1.

Phenotypes and fertility of embryos produced byorbmel/orbmel females

Embryonic phenotypeSterility*
Undeveloped (51%)
Posterior patterning defect (27%)
Wild type (22%) 100%
Embryonic phenotypeSterility*
Undeveloped (51%)
Posterior patterning defect (27%)
Wild type (22%) 100%
*

Sterility was determined by dissection of adult flies and confirmed by Vasa staining of pole cells in embryos.

### Orb interacts with Bic-C and Bru

Cytoplasmic polyadenylation of mRNA requires CPEB to recruit the enzyme poly(A) polymerase (PAP) on the regulated mRNA. Until recently, only one family of PAP, containing both a catalytic domain and an RRM-like domain, was known. Wang et al. have now identified a novel family of PAP that differs from the canonical PAP for the absence of the RRM-like domain. The prototype of this family is represented by C. elegans GLD-2 whose binding to the RNA is mediated by GLD-3, a KH domain containing protein of the BicC family(Wang et al., 2002).

Interestingly, we found that Drosophila BicC interacts physically with Orb in co-immunoprecipitation experiments(Fig. 5A). As the phenotype ofBicC mutants implicates BicC protein as a negative regulator ofosk translation (Saffman et al.,1998), we tested whether Orb interacts with the translational repressor Bru. Indeed, we could detect a physical interaction between Bru and Orb (Fig. 5B), as revealed by the co-immunoprecipitation of Bru with Orb. By contrast, we detected no direct interaction between Bru and BicC (data not shown) in our assay.

Fig. 5.

BicC, Orb and Bru interact in vitro. (A) Immunoprecipitation experiments were performed on total ovarian extract using α-Orb (lane 1),α-BicC (lane 2) or α-Hrp48 (lane 3) antibodies, or beads alone(lane 4). After precipitation, proteins were detected using α-Orb antibodies. Equal amounts of extract were used in each sample. No Orb protein is retained on the beads alone (lane 4) or with α-Hrp48 (lane 3)antibody. Orb is readily precipitated by α-Orb (lane 1) and α-BicC(lane 2) antibodies. (B) Bru protein is pulled down from total ovarian extract by the α-Orb antibody. The same blot was probed with α-Orb (1) andα-Bru antibodies (2).

Fig. 5.

BicC, Orb and Bru interact in vitro. (A) Immunoprecipitation experiments were performed on total ovarian extract using α-Orb (lane 1),α-BicC (lane 2) or α-Hrp48 (lane 3) antibodies, or beads alone(lane 4). After precipitation, proteins were detected using α-Orb antibodies. Equal amounts of extract were used in each sample. No Orb protein is retained on the beads alone (lane 4) or with α-Hrp48 (lane 3)antibody. Orb is readily precipitated by α-Orb (lane 1) and α-BicC(lane 2) antibodies. (B) Bru protein is pulled down from total ovarian extract by the α-Orb antibody. The same blot was probed with α-Orb (1) andα-Bru antibodies (2).

The relevance of these interactions in vivo is further confirmed by the genetic interactions between the BicC locus and the orb andaret loci — the latter encoding Bru protein. Females heterozygous for BicC show a number of AP patterning defects, ranging from head defects to bicaudal embryos(Mahone et al., 1995). TheBicC phenotype is suppressed when the mutation is combined with anorb allele or the strong aret allele,aretQB72. Table 2 shows that 85% of embryos produced byBic-CYC33/+ females fail to hatch and of those 60% are bicaudal. The null allele orbF343 efficiently suppresses the Bic-C phenotype and only 13% of the embryos produced byBicCYC33/+; orbF343/+ fail to hatch,none of which shows a bicaudal phenotype. As shown inTable 2, the strength of the phenotype and the extent of the suppression depends on the orballele. Embryonic viability is also improved, up to 72%, in embryos produced by BicCYC33/ aretQB72 females.

Table 2.

Hatching rate and phenotype of embryos

Phenotype of non-hatchers
Number*Hatching rate (%)Bicaudal (%)Head defects (%)
BicCYC33/cyO 624 15 60 14
aretQB72/cyO 267 74
orbmel/TM3 236 65
orbF303/TM3 276 75
orbF343/TM3 536 85
BicCYC33/aretQB72;+/+ 604 72
BicCYC33/cyO; orbmel/TM3 766 65 39 52
BicCYC33/cyO; orbF303/TM3 746 30 62 30
BicCYC33/cyO; orbF343/TM3 662 87
Phenotype of non-hatchers
Number*Hatching rate (%)Bicaudal (%)Head defects (%)
BicCYC33/cyO 624 15 60 14
aretQB72/cyO 267 74
orbmel/TM3 236 65
orbF303/TM3 276 75
orbF343/TM3 536 85
BicCYC33/aretQB72;+/+ 604 72
BicCYC33/cyO; orbmel/TM3 766 65 39 52
BicCYC33/cyO; orbF303/TM3 746 30 62 30
BicCYC33/cyO; orbF343/TM3 662 87
*

Number of eggs scored.

Embryos to 100% either undeveloped or unfertilized.

Accumulation of Osk activity at the posterior pole of the oocyte is the first step in the assembly of a specialized cytoplasmic sub-compartment, the pole plasm, which contains determinants that direct abdominal patterning and germline differentiation (Ephrussi et al.,1991; Ephrussi and Lehmann,1992; Kim-Ha et al.,1991; Smith et al.,1992). As mislocalization of Osk activity is detrimental to embryonic development, localization-coupled translational activation ensures specific accumulation of Osk at the posterior pole of the oocyte. Previous studies have shown that Bru is required to silence osk translation during transport, to prevent ectopic Osk accumulation(Kim-Ha et al., 1995). BRE-mediated repression is alleviated upon posterior localization of the transcript, allowing its translational activation(Gunkel et al., 1998). Combining measurement of osk poly(A) tail length in vivo with quantification of the translation activity of the corresponding mRNAs in vivo and in vitro, we have been able to correlate polyadenylation with the translational status of the mRNA. A mutation in Orb, the DrosophilaCPEB, leads to shortening of the osk poly(A) tail and to a reduction in Osk accumulation. The fact that, at least in vitro, a long poly(A) tail neither overcomes BRE-mediated repression nor is necessary for repression ofosk reporter transcripts, suggests that cytoplasmic polyadenylation is not the decisive event in translational activation of osk at the posterior pole. Rather, it appears that the presence of a 200 A long tail onosk mRNA promotes its efficient translation, allowing accumulation of Osk to levels sufficient for both abdomen and germline formation to proceed.

The prevailing model, which is based on studies of translational control in the Xenopus oocyte, suggests that the polyadenylation status of a transcript correlates with its translational status: a short poly(A) tail corresponding to a silenced mRNA and poly(A) tail elongation triggering translational activation. In Xenopus, upon progesterone treatment a wave of cytoplasmic polyadenylation activates translation of deadenylated and silenced maternally derived mRNAs(Richter, 1999). InDrosophila, translation of bicoid mRNA, which encodes the anterior determinant of the embryo, is repressed until egg activation when poly(A) tail elongation triggers translation initiation(Sallés et al., 1994). Although the correlation between adenylation and translation still holds for several transcripts, a growing body of evidence suggests that the two events may be coincidental but not directly connected. Interestingly, deadenylation and translational repression of Drosophila hunchback (hb)(Chagnovich and Lehmann, 2001)and mouse tPA (Stutz et al.,1998) mRNAs can occur independently of each other. The transcripts are deadenylated concomitant with translational repression, yet repression can occur in the absence of ongoing deadenylation. In arrested primary mouse oocytes, polyadenylation of the tPA mRNA is necessary to counteract the default deadenylation that affects most other oocyte mRNAs, thus preventing its degradation (Stutz et al.,1998).

Our observations suggest that silencing and awakening of osk mRNA translation can occur in the absence of changes in poly(A) tail length and, in fact, osk mRNA bears a long poly(A) tail at all stages of oogenesis,including when it is unlocalized and translationally silent. However, it is still formally possible that at intermediate stages of oogenesis oskmRNA undergoes a deadenylation that goes undetected in our measurements on bulk RNA, and that elongation of the poly(A) tail causes displacement of the repressor complex, leading to translational derepression. This hypothesis is supported by the fact that the repressor protein Bru shares a 50% sequence identity with the Xenopus deadenylation promoting factor EDEN-BP(Kim-Ha et al., 1995). However, we did not detect any obvious pattern of deadenylation in vitro when Bruno was added to the embryonic extract, nor could we observe a shortening of the poly(A) tail of translationally silenced osk transcript recovered form ovarian extract (data not shown). Nevertheless, our results show that BRE-mediated repression is effective independently of the length of the poly(A) tail on osk transcripts, and that a silenced mRNP can be assembled on a naked osk transcript, whether or not it bears a poly(A) tail. These results suggest that polyadenylation is not the sole determining event leading to translational derepression of osk mRNA at the posterior pole, but that the maintenance of a long poly(A) tail, by cytoplasmic polyadenylation, accounts for the enhancement of osktranslation and is required for efficient osk translation, to ensure sufficient accumulation of Osk at the posterior pole of theDrosophila oocyte to promote abdominal patterning and germline differentiation.

Furthermore, the physical interaction detected between Orb and Bru, and Orb and BicC suggests the existence of a multi-protein complex containing both positive and negative regulators of osk translation. In this scenario, translational silencing and polyadenylation are linked through Bru protein, offering a possible explanation as to how CPEB might be recruited to mRNAs in Drosophila, where no canonical CPE has so far been identified. Transcripts properly repressed by Bru, upon localization, could be adenylated by the recruitment of Orb by Bru itself. Loss of Bru repression would, therefore, result in loss of Orb binding with consequent deadenylation and translational silencing. In this model, modulation of the poly(A) tail would be part of the mechanism that regulates translation, ensuring a second level of control over ectopic expression while localizing all the components necessary for efficient translation. Remarkably, mutations in the BRE sites do not result in ectopic osk translation(Kim-Ha et al., 1995),suggesting the existence of a second layer of translational control. Moreover,during embryonic development when osk translation is no longer required, both Orb and Bru proteins are depleted in the embryo andosk mRNA undergoes complete deadenylation(Sallés et al.,1994).

We thank Don Rio for providing the α-Hrp48 antibody and Paul Lasko for the α-BicC antibody. We are very thankful to all members of the Ephrussi laboratory, particularly M. Chekulaeva and V. Riechmann, for their interest and input into the work; to A. Jenny for discussion and advice; and to A. Cyrklaff for technical help. We also thank F. Gebauer and M. Hentze for their comments and advice during the course of this work. S. C. was supported by an EMBL predoctoral fellowship and a grant from the HFSPO to A. E.

Beams, H. W. and Kessel, R. G. (
1974
). The problem of germ cell determinants.
Int. Rev. Cytol.
39
,
413
-479.
Brasier, A. R., Tate, J. E. and Habener, J. F.(
1989
). Optimized use of the firefly luciferase assay as a reporter gene in mammalian cell lines.
Bio Techniques
7
,
1116
-1122.
Castagnetti, S., Hentze, M. H., Ephrussi, A. and Gebauer, F.(
2000
). Control of oskar mRNA translation by Bruno in a novel cell-free system from Drosophila ovaries.
Development
127
,
1063
-1068.
Chagnovich, D. and Lehmann, R. (
2001
). Poly(A)-independent regulation of maternal hunchback translation in the Drosophila embryo.
98
,
11359
-11364.
Chang, J. S., Tan, L. and Schedl, P. (
1999
). The Drosophila CPEB homolog, orb, is required for oskar protein expression in oocytes.
Dev. Biol.
215
,
91
-106.
Christerson, L. B. and McKearin, D. M. (
1994
).orb is required for anteroposterior and dorsoventral patterning during Drosophila oogenesis.
Genes Dev.
8
,
614
-628.
Ephrussi, A., Dickinson, L. K. and Lehmann, R.(
1991
). oskar organizes the germ plasm and directs localization of the posterior determinant nanos.
Cell
66
,
37
-50.
Ephrussi, A. and Lehmann, R. (
1992
). Induction of germ cell formation by oskar.
Nature
358
,
387
-392.
Fox, C. A., Sheets, M. D. and Wickens, M. P.(
1989
). Poly(A) addition during maturation of frog oocytes:distinct nuclear and cytoplasmic activities and regulation by the sequence UUUUUAU.
Genes Dev.
3
,
2151
-2162.
Gallie, D. R. (
1991
). The CAP and poly(A) tail function synergistically to regulate mRNA translational efficiency.
Genes Dev.
5
,
2108
.
Glotzer, J. B. and Ephrussi, A. (
1999
). Analysis of RNA distribution during Drosophila oogenesis using fluorescent in situ hybridization. In
A Comparative Methods Approach to the Study of Oocytes and Embryos
(ed. J. D. Richter), Part III. New York: Oxford University Press.
Gray, N. and Hentze, M. (
1994
). Iron regulatory protein prevents binding of the 43S translation pre-initiation complex to ferritin and eALAS mRNAs.
EMBO J.
13
,
3882
-3891.
Gunkel, N., Yano, T., Markussen, F.-H., Olsen, L. C. and Ephrussi, A. (
1998
). Localization-dependent translation requires a functional interaction between the 5′ and 3′ ends ofoskar mRNA.
Genes Dev.
12
,
1652
-1664.
Hake, L. E. and Richter, J. D. (
1994
). CPEB is a specificity factor that mediates cytoplasmic polyadenylation during Xenopus oocyte maturation.
Cell
79
,
617
-627.
Kim-Ha, J., Smith, J. L. and Macdonald, P. M.(
1991
). oskar mRNA is localized to the posterior pole of the Drosophila oocyte.
Cell
66
,
23
-35.
Kim-Ha, J., Kerr, K. and Macdonald, P. M.(
1995
). Translational regulation of oskar mRNA by bruno,an ovarian RNA-binding protein, is essential.
Cell
81
,
403
-412.
Lantz, V., Chang, J. S., Horabin, J. I., Bopp, D. and Schedl,P. (
1994
). The Drosophila orb RNA-binding protein is required for the formation of the egg chamber and establishment of polarity.
Genes Dev.
8
,
598
-613.
Lehmann, R. and Nüsslein-Volhard, C.(
1986
). Abdominal segmentation, pole cell formation, and embryonic polarity require the localized activity of oskar, a maternal gene inDrosophila.
Cell
47
,
141
-152.
Lie, Y. S. and Macdonald, P. M. (
1999
). Translational regulation of oskar mRNA occurs independent of the cap and poly(A) tail in Drosophila ovarian extracts.
Development
126
,
4989
-4996.
Lieberfarb, M. E., Chu, T., Wreden, C., Theurkauf, W., Gergen,J. P. and Strickland, S. (
1996
). Mutations that perturb poly(A)-dependent maternal mRNA activation block the initiation of development.
Development
122
,
579
-588.
Mahone, M., Saffman, E. E. and Lasko, P. F.(
1995
). Localized Bicaudal-C RNA encodes a protein containing a KH domain, the RNA binding motif of FMR1.
EMBO J.
14
,
2043
-2055.
Markussen, F.-H., Michon, A.-M., Breitwieser, W. and Ephrussi,A. (
1995
). Translational control of oskar generates Short OSK, the isoform that induces pole plasm assembly.
Development
121
,
3723
-3732.
Markussen, F. H., Breitwieser, W. and Ephrussi, A.(
1997
). Efficient translation and phosphorylation of Oskar require Oskar protein and the RNA helicase Vasa.
Cold Spring Harb. Symp. Quant. Biol.
62
,
13
-17.
McGrew, L. L. and Richter, J. D. (
1990
). Translational control by cytoplasmic polyadenylation during Xenopus oocyte maturation: characterization of cis and trans elements and regulation by cyclin/MPF.
EMBO J.
9
,
3743
-3751.
Mendez, R., Hake, L. E., Andresson, T., Littlepage, L. E.,Ruderman, J. V. and Richter, J. D. (
2000
). Phosphorylation of CPE binding factor by Eg2 regulates translation of c-mos mRNA.
Nature
404
,
302
-307.
Paillard, L., Omilli, F., Legagneux, V., Bassez, T., Maniey, D. and Osborne, H. B. (
1998
). EDEN and EDEN-BP, a ciselement and an associated factor that mediate sequence-specific mRNA deadenylation in Xenopus embryos.
EMBO J.
17
,
276
-287.
Richter, J. D. (
1999
). Cytoplasmic polyadenylation in development and beyond.
Microbiol. Mol. Biol. Rev.
63
,
446
-456.
Riechmann, V., Gutierrez, J. G., Filardo, P., Nebreda, A. R. and Ephrussi, A. (
2002
). Par-1 regulates stability of the posterior determinant Oskar by phosphorylation.
Nat. Cell Biol.
5
,
337
-342.
Rongo, C., Gavis, E. R. and Lehmann, R. (
1995
). Localization of oskar RNA regulates oskar translation and requires Oskar protein.
Development
121
,
2737
-2746.
Saffman, E. E., Styhler, S., Rother, K., Li, W., Richard, S. and Lasko, P. (
1998
). Premature translation of oskar in oocytes lacking the RNA-binding protein bicaudal-C.
Mol. Cell. Biol.
18
,
4855
-4862
Sallés, F. J., Lieberfarb, M. E., Wreden, C., Gergen, J. P. and Strickland, S. (
1994
). Coordinate initiation ofDrosophila development by regulated polyadenylation of maternal messenger RNAs.
Science
266
,
1996
-1999.
Sheets, M. D., Wu, M. and Wickens, M. (
1995
). Polyadenylation of c-mos mRNA as a control point in Xenopusmeiotic maturation.
Nature
374
,
511
-516.
Smith, J. L., Wilson, J. E. and Macdonald, P. M.(
1992
). Overexpression of oskar directs ectopic activation of nanos and presumptive pole cell formation in Drosophila embryos.
Cell
70
,
849
-859.
1993
). Developmental genetics of oogenesis. In
The Development of Drosophila melanogaster
, Vol.
1
(ed. M. Bate and A. Martinez-Arias), pp.
1
-70. Cold Spring Harbor, NY:Cold Spring Harbor Laboratory Press.
St Johnston, D., Beuchle, D. and Nüsslein-Vorhard, C.(
1991
). staufen, a gene required to localize maternal RNAs in Drosophila eggs.
Cell
66
,
51
-63.
St Johnston, D. (
1995
). The intracellular localisation of messenger RNAs.
Cell
81
,
161
-170.
Stebbins-Boaz, B., Hake, L. E. and Richter, J. D.(
1996
). CPEB controls the cytoplasmic polyadenylation of cyclin,Cdk2 and c-mos mRNAs and is necessary for oocyte maturation in Xenopus.
EMBO J.
15
,
2582
-2592.
Stutz, A., Conne, B., Huarte, J., Gubler, P., Volkel, V.,Flandin, P. and Vassalli, J. D. (
1998
). Masking, unmasking,and regulated polyadenylation cooperate in the translational control of a dormant mRNA in mouse oocytes.
Genes Dev.
12
,
2535
-2548.
Tarun, S. and Sachs, A. (
1996
). Association of the yeast poly(A) tail binding protein with translation initiation factor eIF-4G.
EMBO J.
15
,
7168
-7177.
Vanzo, N. F. and Ephrussi, A. (
2002
). Oskar anchoring restricts pole plasm formation to the posterior of the Drosophila oocyte.
Development
129
,
3705
-3714.
Wang, L., Eckmann, C. R., Kadyk, L. C., Wickens, M. and Kimble,J. (
2002
). A regulatory cytoplasmic poly(A) polymerase inCaenorhabditis elegans.
Nature
419
,
312
-316.
Webster, P. J., Liang, L., Berg, C. A., Lasko, P. and Macdonald,P. M. (
1997
). Translational repressor bruno plays multiple roles in development and is widely conserved.
Genes Dev.
11
,
2510
-2521.
Wells, S. E., Hillner, P. E., Vale, R. D. and Sachs, A. B.(
1998
). Circularization of mRNA by eukaryotic translation initiation factors.
Mol. Cell
2
,
135
-140.
Zangar, R. C., Hernandez, M., Kocarek, T. A. and Novak, R. F. (
1995
). Determination of the poly(A) tail lengths of a single mRNA species in total hepatic RNA.
BioTechniques
18
,
465
-469.