p70^S6K Controls Selective mRNA Translation during Oocyte Maturation and Early Embryogenesis in Xenopus laevis

Abstract

In mammalian cells, p70^S6K plays a key role in translational control of cell proliferation in response to growth factors. Because of the reliance on translational control in early vertebrate development, we cloned a Xenopus homolog of p70^S6K and investigated the activity profile of p70^S6K during Xenopus oocyte maturation and early embryogenesis. p70^S6K activity is high in resting oocytes and decreases to background levels upon stimulation of maturation with progesterone. During embryonic development, three peaks of activity were observed: immediately after fertilization, shortly before the midblastula transition, and during gastrulation. Rapamycin, an inhibitor of p70^S6K activation, caused oocytes to undergo germinal vesicle breakdown earlier than control oocytes, and sensitivity to progesterone was increased. Injection of a rapamycin-insensitive, constitutively active mutant of p70^S6K reversed the effects of rapamycin. However, increases in S6 phosphorylation were not significantly affected by rapamycin during maturation. mos mRNA, which does not contain a 5′-terminal oligopyrimidine tract (5′-TOP), was translated earlier, and a larger amount of Mos protein was produced in rapamycin-treated oocytes. In fertilized eggs rapamycin treatment increased the translation of the Cdc25A phosphatase, which lacks a 5′-TOP. Translation assays in vivo using both DNA and RNA reporter constructs with the 5′-TOP from elongation factor 2 showed decreased translational activity with rapamycin, whereas constructs without a 5′-TOP or with an internal ribosome entry site were translated more efficiently upon rapamycin treatment. These results suggest that changes in p70^S6K activity during oocyte maturation and early embryogenesis selectively alter the translational capacity available for mRNAs lacking a 5′-TOP region.

In mammalian cells, the p70 and p85 isoforms of the 70-kDa ribosomal protein S6 kinase (p70^S6K) are both rapidly activated upon stimulation of cells with virtually all mitogenic factors (31, 46, 60). The two isoforms are identical except that p85 has an amino-terminal 23-amino-acid extension containing a nuclear localization signal. The two isoforms are derived from the same transcript by alternative translation initiation start sites (50a). p70^S6K is activated by a complex pattern of phosphorylation on several sites by various upstream kinases (48, 60). The first event is phosphorylation of (Ser/Thr)-Pro sites Ser₄₁₁, Ser₄₁₈, Thr₄₂₁, and Ser₄₂₄ in the carboxy-terminal autoinhibitory domain, facilitating phosphorylation of Thr₃₈₉ and Ser₄₀₄ in the linker region, which in turn leads to disruption of the interaction between the carboxy and amino termini of the protein. Finally, phosphorylation of Thr₂₂₉ leads to full activation of p70^S6K. The kinase responsible for Thr₂₂₉ phosphorylation is the constitutively active 3-phosphoinositide-dependent protein kinase 1 (1a, 49a), whereas the kinases involved in the previous phosphorylation events, also required for activation, are not clearly identified. However, phosphorylation of Thr₃₈₉ in mammalian cells is dependent on the kinase activity of target of rapamycin (TOR), also termed FRAP, RAFT, or RAPT (9, 48, 60). The macrolide antibiotic rapamycin is a potent inhibitor of the p70^S6K pathway. It forms a complex with FKBP12, which specifically blocks activity of mammalian TOR, thereby leading to rapid deactivation of p70^S6K (1, 9, 16, 31, 58). Rapamycin has been shown to down-regulate the translation of mRNAs containing a 5′-terminal oligopyrimidine tract (5′-TOP), which include those for ribosomal proteins and other proteins of the translational machinery (4, 28, 29, 41, 59). All known mRNAs for vertebrate ribosomal proteins and protein synthesis elongation factors contain a 5′-TOP, and their translation is regulated in response to mitogens (2–5, 30, 31, 41) via p70^S6K activity (28).

The downstream target of p70^S6K is ribosomal protein S6, which is present in a single copy per 40S subunit. S6 becomes rapidly phosphorylated on five serine residues in its carboxy-terminal region upon stimulation of cells with growth factors. Moreover, phosphorylation occurs in an ordered manner in vivo and in vitro (6, 63). S6 phosphorylation is closely correlated with increased rates of protein synthesis when quiescent cells reenter the cell cycle (17, 58). Conversely, a decrease in protein synthesis is paralleled by lower S6 phosphorylation (58). The concept that the function of p70^S6K is linked to regulation of protein synthesis is also suggested by studies in murine embryonic stem cells with a disrupted p70^S6K gene (32).

Other studies have suggested that p70^S6K is also linked to pathways controlling cell cycle progression. In particular, microinjection of neutralizing p70^S6K antibodies into mammalian cells inhibits G₁ progression (35). Moreover, mice deficient for p70^S6K are significantly smaller, an effect which is most dramatic during embryogenesis. Unexpectedly, mouse embryo fibroblasts derived from p70^S6K-deficient mice were as sensitive to rapamycin as mouse embryo fibroblasts derived from wild-type animals, and there was no effect on the S6 phosphorylation response (57). These studies led to the identification of a new highly homologous S6 kinase which is rapamycin sensitive and whose transcript is up-regulated in all tissues examined (57). The importance of p70^S6K in both protein synthesis and cell cycle progression awaits studies in mice deficient for both S6 kinase genes.

Xenopus oocytes are an interesting system for the study of S6 phosphorylation because of the amplification of ribosomal genes and ribosomes during oogenesis. It has been estimated that in oocytes only 1% of the ribosomes are present on polysomes, with the remainder being gradually utilized after fertilization (64). During progesterone-induced oocyte maturation in Xenopus laevis, phosphorylation of S6 in the total ribosome population changes dramatically. It is low in resting oocytes, is increased greatly when 50% of the oocytes have undergone germinal vesicle breakdown (GVBD), and is maximal in unfertilized eggs (44). In parallel, overall protein synthesis is up-regulated by a factor of approximately 2 (51). Biochemical and molecular cloning studies have indicated the protein kinase responsible for S6 phosphorylation during maturation is p90^Rsk (19–21). Like p70^S6K, Rsk phosphorylates all five sites in S6 in an ordered fashion. Despite the general increase in protein synthesis and S6 phosphorylation during oocyte maturation, production of ribosomal proteins ceases (27), suggesting that translation of ribosomal protein mRNAs (rp-mRNAs) that contain the 5′-TOP is uncoupled from that of non-5′-TOP mRNAs during oocyte maturation and may be regulated by different mechanisms (2–5). After fertilization, translation of S3, L17, and L31 begins, and L5 synthesis is evident from stage 7 onward (49). In both oocytes and embryos, total translational capacity is constant such that new mRNAs compete for translation with existing mRNAs (37).

The p70^S6K has been reported to become rapidly deactivated after induction of Xenopus oocyte maturation, suggesting a role in resting oocytes that is terminated in the initial phase of oocyte maturation (36). p70^S6K function has not previously been investigated in embryos. Therefore, we cloned a full-length Xenopus homolog of p70^S6K and investigated the function of this enzyme during oocyte maturation and early development in X. laevis. p70^S6K was not responsible for the up-regulation of S6 phosphorylation during maturation. Indeed, administration of rapamycin accelerated oocyte maturation that was correlated with reduced translation of mRNAs with a 5′-TOP region and enhanced translation of mos. In embryos, p70^S6K was rapidly activated after fertilization and may contribute to the enhanced translation of several ribosomal proteins after fertilization.

MATERIALS AND METHODS

Cloning of X. laevis p70^S6K cDNA.

A λgt10 cDNA library generated by oligo(dT) priming of RNA from defolliculated X. laevis oocytes was obtained from D. Melton (50). Screening of 4× 10⁵ PFU by hybridization with probes corresponding to the 231 bp of the 5′ coding sequence of rat p70^S6K cDNA and to 665 bp of a partial X. laevis cDNA clone (36), respectively, revealed 20 clones containing identical 1.7-kb inserts. Two phage inserts were subcloned in pBluescript and analyzed by dideoxy sequencing of both strands, using a Sequenase 2.0 sequencing kit (U.S. Biotechnology, Lake Placid, N.Y.) with oligonucleotide priming from the T3 and T7 sequences in the vector polylinker.

Oocytes, eggs, and embryos.

Female X. laevis frogs were injected with 75 IU of pregnant mare’s serum gonadotropin (PMSG) 2 to 7 days prior to dissection of the ovary and manual isolation of oocytes. Isolated oocytes were incubated in 1× modified Barth’s solution [88 mM NaCl, 1 mM KCl, 0.41 mM CaCl₂, 0.33 mM Ca(NO₃)₂, 0.82 mM MgSO₄, 2.4 mM NaHCO₃, 10 mM HEPES (pH 7.4)], and maturation was induced by addition of progesterone as indicated in the figure legends. Rapamycin (Sigma, St. Louis, Mo.), dissolved in dimethyl sulfoxide, was added at a final concentration of 2 μg/ml 1 to 2 h prior to induction of maturation. Controls were exposed to dimethyl sulfoxide alone. When eggs or embryos were needed, frogs were injected with 550 IU of human chorionic gonadotropin to induce egg laying 12 to 14 h later. To obtain activated eggs, freshly laid eggs were dejellied with 2% cysteine (Sigma), pH 8.0, in 1× Maller’s modified Ringer (MMR; 100 mM NaCl, 2 mM KCl, 1 mM CaCl₂, 1 mM MgCl₂, 5 mM HEPES [pH 7.4]), treated with the calcium ionophore A23187 (Calbiochem, La Jolla, Calif.) at a final concentration of 5 μg/ml in 1× MMR for 1.5 to 2 min, and rinsed 8 to 10 times with 0.2× MMR. Preincubation with rapamycin (2 μg/ml in 1× MMR) was performed for 15 to 30 min before addition of the calcium ionophore. Embryos were obtained by in vitro fertilization of freshly laid eggs, dejellied, and cultivated in 0.1× MMR as described previously (26). Embryos were staged as specified by Nieuwkoop and Faber (45). Oocytes, eggs, and embryos were frozen in dry ice at the desired time or stage.

Injection of mRNA and DNA.

Constructs encoding Myc-tagged active and inactive mutants of rat p70^S6K were produced by PCR amplification of plasmids pRK5-myc-p70^S6KD3E-E₃₈₉ and pRK5-myc-p70^S6KQ₁₀₀ (42, 48), using the primers 5′CTTGAATTCGGCAGGAGTGTTTGACATAG3′ and 5′GCGCTCTAGATCATAGATTCATACGCAGGT3′. PCR products were cloned into pGEM T-easy vector (Promega, Madison, Wis.); the resulting plasmids were digested with EcoRI, and the fragment was ligated into EcoRI-digested pCS2+MT (53, 62). Capped mRNA was produced in vitro by using a mMessage mMachine kit (Ambion, Austin, Tex.) after linearization of the plasmid with NotI. Oocytes were injected with 20 ng of mRNA and incubated for 2 h prior to induction of maturation by progesterone. The 5′-flanking region of the hamster elongation factor 2 (EF2) gene (with [−272 to +47] or without [−272 to +1] [+1, transcription initiation site] a 5′-TOP) was inserted upstream of luciferase cDNA in the pGL2-Enhancer vector (Promega) to obtain the EF2+TOP or EF2−TOP luciferase reporter. The hamster EF2 gene has a typical 5′-TOP sequence from its transcription initiation site (+1 CCTCTTCCGCCGCAGCCGCCGCCATCGTCGGCGCCCCTCGCTCTTCT +47) (43). Initiation of transcription of the chimeric mRNAs at the native EF2 transcription site was confirmed by S1 mapping in a mammalian cell system. Plasmid DNA (1.5 ng) encoding luciferase reporters with or without a 5′-TOP, driven by the genomic promoter of hamster EF2, was injected into the oocyte germinal vesicle. After incubation at room temperature for 10 h, oocytes were frozen in dry ice.

The SP6-EF2+TOP reporter construct, which was used for production of RNA, was made by PCR amplification of EF2+TOP by using the primers GATTTAGGTGACACTATAGCTCTTCCGCCCCAGC and CATCGCTGAATACAGTTAC and blunt-end cloning of the product into pT7Blue-3 vector (Novagen, Madison, Wis.). A PstI/BglII fragment from pOTV corresponding to the 3′ untranslated region (3′ UTR) of Xenopus β-globin mRNA was inserted into BamHI/PstI sites of pT7Blue-3. The integrity of the constructs was confirmed by sequencing. RNA was produced in vitro with a mMessage mMachine kit (Ambion) from the SP6 promoter, whose sequence was integrated in the upstream primer by using a PstI-linearized template. The β-galactosidase reporter construct containing the internal ribosome entry site (IRES) of encephalomyocarditis virus was produced by ligation-independent cloning of the β-galactosidase open reading frame (Novagen) into pCITE-5 LIC vector (Novagen).

Gel electrophoresis and Western blotting.

Oocytes, eggs, or embryos were homogenized in extraction buffer (50 mM Tris [pH 7.4], 80 mM β-glycerophosphate, 20 mM EDTA, 20 mM NaF, 0.1 mM sodium vanadate, 1 mM dithiothreitol [DTT], 0.3 μM microcystin, 0.3 mM phenylmethylsulfonyl fluoride, leupeptin [10 μg/ml], pepstatin [10 μg/ml], chymostatin [10 μg/ml]) and centrifuged for 5 min at 4°C. The cytosolic phase equivalent to one oocyte, egg, or embryo was loaded onto Laemmli sodium dodecyl sulfate (SDS)–10% polyacrylamide gels for immunoblotting with rabbit anti-Mos (Santa Cruz Biotechnology, Santa Cruz, Calif.) or rabbit anti-Cdc25A antibodies or onto 12.5% Anderson gels for immunoblotting with rabbit anti-p70^S6K antibody (Santa Cruz Biotechnology). Proteins were transferred to nitrocellulose membranes by using a semidry blotting technique (Pharmacia-LKB, Piscataway, N.J.). Membranes were blocked with 10% nonfat dry milk in phosphate-buffered saline–0.05% Tween 20 and probed with antibodies in phosphate-buffered saline–10% milk–0.05% Tween (anti-Mos and anti-p70^S6K) or with Tris-buffered saline–0.05% Tween (anti-Cdc25A). Bands were visualized by the enhanced chemiluminescence procedure (Amersham, Arlington Heights, Ill.). Isolation of ribosomes and two-dimensional gel analysis of ribosomal proteins were carried out as described previously (44, 47).

Immunoprecipitation and p70^S6K assay.

Extracts, prepared as described above and corresponding to one oocyte or embryo, were incubated with 400 ng of anti-p70^S6K antibody for 2 h on ice. Antibody-antigen complexes were collected onto 10 μl of protein A-Sepharose beads (Pierce Chemical, Rockford, Ill.). Beads were then washed twice in low-salt buffer (50 mM Tris [pH 7.4], 80 mM β-glycerophosphate, 20 mM EDTA, 20 mM NaF, 0.1 mM sodium vanadate, 1 mM DTT, 100 mM NaCl, 0.2 mg of bovine serum albumin per ml, 1% Nonidet P-40 or IGEPAL CA-630), twice in high-salt buffer (500 mM NaCl instead of 100 mM), and then twice with kinase buffer (50 mM morpholinepropanesulfonic acid [pH 7.4], 10 mM MgCl₂, 0.2 mg of bovine serum albumin per ml, 1 mM DTT). The immunoprecipitates were incubated in kinase buffer containing 50 μM ATP, 5 μCi of [γ-³²P]ATP, and 22 μg of ribosomal 40S subunits in a volume of 20 μl at 30°C for 30 min, and the reaction was stopped by addition of SDS-polyacrylamide gel electrophoresis (PAGE) sample buffer. Ribosomal proteins were separated by SDS-PAGE (10% gel). After exposure of the dried gel to X-Omat RP film (Kodak, Rochester, N.Y.), the band corresponding to S6 was excised and counted by liquid scintillation spectrometry. Xenopus ovary 40S ribosomal subunits were obtained as described by Erikson et al. (22).

Reporter assays.

Oocytes injected with luciferase or β-galactosidase reporter constructs were lysed (20 μl per oocyte) with 1× reporter lysis buffer (Promega) or lysis solution (with 0.5 mM DTT; Tropix, Bedford, Mass.) respectively. Routinely, extract corresponding to 0.05 or 0.5 oocyte was assayed for luciferase activity by the injection of 100 μl of luciferase substrate (Promega) into a Mono Light luminometer (Analytical Luminescence Laboratory, Ann Arbor, Mich.) according to the manufacturer’s protocol. For detection of β-galactosidase activity, 10 μl of extract corresponding to 0.5 oocyte was incubated with 70 μl of β-galactosidase reaction buffer (Tropix) for 2 h at room temperature before injection of 100 μl of light emission accelerator into the luminometer.

Nucleotide sequence accession number.

The sequence shown in Fig. 1A has been deposited in the GenBank under accession no. AJ131521.

FIG. 1.

Schematic diagram and cDNA sequence of X. laevis p70^S6K. (A) Amino acid sequence alignment of X. laevis (Xen. l.) p70^S6K and rat p70^S6K. Identical amino acids are indicated by bars; similar amino acids are indicated by dots. (B) Schematic diagram of X. laevis p70^S6K, showing the N-terminal (cross-hatched), catalytic (open), linker (hatched), autoinhibitory (filled), and C-terminal (dotted) domains. Phosphorylation sites conserved with mammalian p70^S6K, and known to be involved in kinase activation, are indicated.

RESULTS

Cloning of X. laevis p70^S6K cDNA.

Earlier studies led to the identification of a PCR product encoding a maternal form of p70^S6K from X. laevis (36). Hybridization screening of an X. laevis cDNA library with rat and Xenopus probes identified multiple clones with a 1.7-kb insert. The insert sequence of 1,717 nucleotides contains a large open reading frame. The first ATG codon, nucleotides 97 to 99, is surrounded by a strong translation initiation start site consensus sequence (33), and the following 1,503 nucleotides encode an amino acid sequence with 93% identity to the sequence of mammalian p70^S6K (Fig. 1). Longer clones containing sequences homologous to the p85^S6K amino terminus were not found; however, no stop codon was found in frame in the 5′ sequence preceding the ATG translation initiation codon. In addition, none of the other 20 phages isolated from the cDNA library contained a longer 5′ sequence.

p70^S6K activity during oocyte maturation and embryonic development.

In the earlier study described above, Lane et al. (36) showed with antibodies to mammalian p70^S6K that p70^S6K activity in oocytes declined after induction of maturation by progesterone. The high sequence identity of xp70^S6K with the mammalian enzyme (Fig. 1) plus conservation of all the phosphorylation sites and regulatory motifs present in the mammalian enzyme supports the use of reagents based on mammalian p70^S6K for study of the Xenopus enzyme. In the experiments reported here, we determined the activity profile of p70^S6K during oocyte maturation by assaying the kinase activity of immunoprecipitated p70^S6K, using 40S ribosomal subunits as the substrate. p70^S6K activity was high in resting oocytes, decreased 6- to 10-fold within the first 1 to 2 h after induction of maturation by progesterone, and stayed low until oocytes reached germinal vesicle breakdown (GVBD) (Fig. 2). Similar results were obtained with oocytes from PMSG-primed or unprimed frogs, and kinase activity in resting oocytes was in the same range in both primed and unprimed oocytes. Usually no increased activity was evident after progesterone treatment, but in one experiment a 15% increase in p70^S6K activity was seen upon stimulation of oocytes with progesterone (Fig. 2A). Western blotting confirmed that electrophoretic shifts mirror the p70^S6K activity changes observed with immunocomplex-kinase assays (Fig. 2B). The anti-p70^S6K antibody recognized two clusters of bands of equal intensity at ∼70 and ∼85 kDa, representing the p70 and p85 isoforms. These blots indicate that the p85 isoform described in other species is also present in X. laevis, and both isoforms undergo changes in activity together during maturation. In immunoblots, both isoforms could be blocked by preincubation of the antibody with the immunogenic peptide (data not shown). The bands in each cluster have previously been shown to differ in phosphorylation state (48). The samples with the highest enzyme activity showed multiple bands for each isoform (lanes 1 and 7). With the gradual decrease of activity after induction of maturation, the most retarded isoforms disappeared and the abundance of the isoforms with higher mobility increased (lanes 2 to 6 and 8 to 12). Thus, the antibody recognized both isoforms on the Western blot, and it also precipitated both forms from extracts (data not shown). Therefore, in kinase assays both the p70 and p85 isoforms appear to contribute to total S6 kinase activity.

FIG. 2.

Activity of p70^S6K during oocyte maturation and early embryogenesis. (A) Oocyte maturation was induced with progesterone (10 μg/ml), and samples were collected every 30 min until 100% GVBD was reached. p70^S6K immunoprecipitates of oocytes from primed and unprimed frogs were assayed for phosphorylation of ribosomal protein S6, as indicated. (B) In the upper panel, extracts corresponding to one oocyte per lane were separated by SDS-PAGE using 12.5% Anderson gels. Proteins were transferred to a nitrocellulose membrane and probed with an anti-p70^S6K antibody. The antibody recognized two clusters of bands of approximately 70 and 85 kDa, which represent the p70 and p85 isoforms of the enzyme. The time after progesterone addition is indicated at the top of each lane. The lower panel shows an autoradiograph of p70^S6K kinase activity of unprimed (lanes 1 to 6) and primed (lanes 7 to 12) oocytes. (C) Immunoprecipitates of p70^S6K from the indicated stages of embryonic development were assayed for phosphorylation of S6. High activity was observed immediately after fertilization, at around stage 7 shortly before the MBT, and during late gastrula stages. (D) Confirmation of the results of the kinase assays in panel C by monitoring the abundance of slower-migrating isoforms of the protein on Western blots (upper panel) and autoradiograph of immunoprecipitated p70^S6K activity from the same embryo samples (lower panel).

After fertilization of X. laevis eggs, all new protein synthesis is translationally controlled prior to the midblastula transition (MBT) (18). Therefore we also investigated the activity of p70^S6K during early embryonic development (Fig. 2C). After fertilization, p70^S6K activity increased about 30-fold during the first cell cycle and then decreased over the next two to three cell cycles. Activity increased again two- to threefold at stage 8, shortly before the MBT, and a third peak was observed during gastrulation at stage 12. The results of the kinase assays were confirmed by changes in the abundance of slower-migrating bands of both isoforms on Western blots (Fig. 2D).

The first cell cycle is the only pre-MBT cell cycle to contain a G phase. Therefore, the detailed kinetics of the increase of p70 activity in the first cell cycle after fertilization was studied. Since embryos require a 30-min dejellying procedure before extracts can be made, we measured p70^S6K activity in dejellied unfertilized eggs activated with the calcium ionophore A23187, which mimics the events of fertilization. p70^S6K activity increased to a high level within the first 30 min after ionophore treatment, with significant activity by 20 min after treatment (Fig. 3A). The deadenylation of mos mRNA after fertilization and the degradation of Mos protein after activation of eggs paralleled increases in p70^S6K activity (Fig. 3B). These results indicate that in X. laevis eggs, p70^S6K can be activated in the absence of growth factors solely by elevation of cellular calcium levels, as has been shown in other systems (14, 25).

FIG. 3.

p70^S6K activity increases within 30 min after activation of eggs. (A) Freshly laid eggs were dejellied and incubated in rapamycin (2 μg/ml) in 1× MMR for 30 min. Eggs were activated with the Ca²⁺ ionophore A23187 (5 μg/ml) at time zero and collected at the indicated time points; S6 kinase was measured in p70^S6K immunoprecipitates. (B) Extracts corresponding to one egg were separated on 10% polyacrylamide gels, the proteins were transferred to a nitrocellulose membrane, and the blot was probed with anti-Mos antibody. The extracts used were the same as those used for panel A.

Rapamycin accelerates oocyte maturation and decreases the threshold level of progesterone required for maturation.

The activity profile of p70^S6K after progesterone treatment suggests that an even earlier down-regulation of its activity might play a role in oocyte maturation. Rapamycin has been shown to be a potent inhibitor of the p70^S6K pathway in different systems, specifically inhibiting TOR, a mediator of p70^S6K function (1, 16). Thus, to investigate the function of p70^S6K during oocyte maturation and embryogenesis, rapamycin was used to block activation of the enzyme in oocytes and embryos. Incubation of oocytes in rapamycin (2 μg/ml) decreased the activity of p70^S6K to the background level (Fig. 4A). In embryos treated with rapamycin after fertilization, p70^S6K activity was also reduced to the background level. At 30 min after fertilization, corresponding to 10 min after addition of rapamycin, activity was lower than in control embryos (Fig. 4B). At all later time points, 45 to 105 min postfertilization, p70^S6K activity was undetectable. This indicates that it takes between 10 and 30 min to fully inhibit p70^S6K activity by incubation of oocytes and embryos in rapamycin.

FIG. 4.

Rapamycin inhibits the activity of p70^S6K in oocytes and embryos. (A) Oocytes were pretreated with rapamycin (2 μg/ml) for 1 to 2 h, and maturation was induced by addition of progesterone (10 μg/ml). Samples were collected, and S6 kinase activity was measured from immunoprecipitated p70^S6K from an extract corresponding to one oocyte per sample. Results are expressed relative to p70^S6K activity in stage VI oocytes. (B) Freshly laid eggs were fertilized in vitro, dejellied, and treated with rapamycin (2 μg/ml) 20 min after fertilization. Extracts were prepared at the indicated times, and kinase activity was measured from immunoprecipitated p70^S6K.

Surprisingly, oocytes treated with rapamycin underwent GVBD faster than untreated oocytes (Fig. 5). Fifty percent GVBD (GVBD₅₀) occurred between 1 and 2 h earlier at the minimum concentration of progesterone that led to 100% GVBD (Fig. 5A), which is different for oocytes from different frogs. This effect of rapamycin could not be detected in oocytes treated with a high dose of progesterone, indicating that high concentrations of progesterone can override the molecular effects of rapamycin. Although rapamycin alone, without addition of progesterone, was not able to induce oocyte maturation, it caused a higher percentage of GVBD in treated oocytes at suboptimal concentrations of progesterone that cause less than 100% GVBD (Fig. 5B). Therefore, rapamycin not only is able to accelerate maturation but also increases the sensitivity of oocytes to progesterone, suggesting that the progesterone-dependent down-regulation of p70^S6K activity is important for normal maturation kinetics.

FIG. 5.

Rapamycin accelerates GVBD and decreases the threshold concentration of progesterone required for maturation. (A) Oocytes were pretreated with rapamycin (2 μg/ml) for 2 h before induction of maturation with a threshold level of progesterone. The percentage of GVBD was scored by occurrence of a well-defined white spot in the animal pole indicative of GVBD. At the lowest concentration of progesterone that leads to 100% GVBD in untreated oocytes (60 ng/ml), GVBD₅₀ occurred ∼1 h earlier in rapamycin-treated oocytes than in controls. (B) At a subthreshold concentration of progesterone (40 ng/ml), at which only 30% of oocytes underwent GVBD, the final percentage of GVBD was increased ∼2-fold by rapamycin treatment.

Rapamycin does not inhibit p70^S6K directly but instead inhibits the upstream kinase TOR, which leads to rapid deactivation of p70^S6K (1, 16). To exclude the possibility that the effect of rapamycin on oocyte maturation was due to inhibition of another downstream target of TOR, such as 4E-BP1/PHAS I (9, 38, 60), oocytes were injected with 20 ng of in vitro-transcribed mRNA encoding a constitutively active and rapamycin-insensitive mutant of rat p70^S6K (p70^S6KD3E-E₃₈₉). The protein level of the expressed kinase was about 5- to 10-fold above that of the endogenous p70^S6K (data not shown). In extracts from oocytes injected with the constitutively active p70^S6K, total p70^S6K activity was about 30-fold higher than in uninjected samples at GVBD (Fig. 6A, lanes 1 and 2). Kinase activity in oocytes injected with an inactive mutant of rat p70^S6K (p70^S6KQ₁₀₀) was the same as in uninjected control oocytes (Fig. 6A, lanes 1 and 3). Importantly, at a subthreshold concentration of progesterone, injection of the active form of p70^S6K reversed the accelerating effect of rapamycin, whereas oocytes injected with the inactive form of p70^S6K behaved like uninjected oocytes incubated with rapamycin (Fig. 6B). This result suggests that the acceleration of oocyte maturation by rapamycin is indeed mediated by specific blocking of the p70^S6K pathway and not by an alternative pathway that is also blocked by rapamycin.

FIG. 6.

Constitutively active p70^S6K reverses the effect of rapamycin. (A) Oocytes were pretreated with rapamycin (2 μg/ml) for 1 to 2 h and then microinjected with 20 ng of mRNA encoding either p70^S6KD3E-E₃₈₉ (active form) or p70^S6KQ₁₀₀ (kinase-dead form). After incubation for 2 h at room temperature, S6 kinase activity of immunoprecipitated p70^S6K was measured in extracts of resting or matured oocytes. Kinase activity was about 30-fold higher at GVBD in oocytes injected with the active form of rp70^S6K than in control GVBD oocytes. (B) Oocytes were treated with rapamycin (rap.), and some were then injected with the constitutively active (p70^S6KD3E-E₃₈₉) or inactive (p70^S6KQ₁₀₀) form of rat p70^S6K. Maturation induced with progesterone (40 ng/ml) occurred a lower percentage in oocytes injected with the active isoform of p70^S6K in the presence of rapamycin, whereas inactive p70^S6K did not affect the kinetics of maturation.

Rapamycin does not block increased S6 phosphorylation during maturation.

If inhibiting p70^S6K accelerates maturation, one might have predicted that expression of a constitutively active p70^S6K would not only reverse the effects of rapamycin on acceleration of GVBD but also retard the kinetics of GVBD with low-dose progesterone. As shown in Fig. 6, elevated p70^S6K activity does not retard maturation. A possible explanation for this result is that the normal function of p70^S6K to phosphorylate S6 is being performed in progesterone-treated oocytes by p90^Rsk. In support of this idea, p90^Rsk was originally purified as the only S6 kinase activity present in fully mature eggs (19–22), and p70^S6K activity is undetectable by GVBD (Fig. 2 and reference 36). Moreover, p90^Rsk phosphorylates all five sites in S6 in the same ordered fashion as observed with p70^S6K (63). Thus, it is likely that in p70^S6K-injected oocytes at GVBD, S6 phosphorylation is already maximal due to the activity of p90^Rsk. To test this hypothesis further, we determined the levels of S6 phosphorylation in the presence and absence of rapamycin (44). The results show that S6 phosphorylation was greater in progesterone-treated than in untreated oocytes, with the majority of the protein migrating in derivatives b and c, containing 2 and 3 mol, respectively, of phosphate (Fig. 7A and B, respectively), as reported previously (44). In the presence of rapamycin the progesterone response was largely unaffected, although slightly less phosphorylated S6 derivative d was evident (compare Fig. 7B and C). The insensitivity of S6 phosphorylation to rapamycin (Fig. 7B) supports the identification of p90^Rsk as the enzyme responsible for increased S6 phosphorylation in response to progesterone treatment. Recently, a second immunologically distinct form of p70^S6K, termed p70^S6K2, was identified in p70^S6K1-deficient mice (57). However, this enzyme is also rapamycin sensitive and therefore unlikely to account for S6 phosphorylation in oocytes when p70^S6K1 is down-regulated (Fig. 2). Despite the fact that p90^Rsk accounts for S6 phosphorylation during maturation, it cannot be excluded that p70^S6K regulates translation of mRNAs on the minor (1%) fraction of ribosomes that are on polysomes in resting oocytes, since the constitutively active p70 overcomes rapamycin effects that are evident before Rsk activation (Fig. 6).

FIG. 7.

Rapamycin does not block increased S6 phosphorylation during maturation. Groups of 300 oocytes each were incubated in MMR (A), MMR containing progesterone (10 μg/ml) (B), or MMR containing progesterone and rapamycin (2 μg/ml) (C). After progesterone-treated oocytes reached a time equivalent to 2.0 GVBD₅₀, oocytes were frozen and subsequently ribosomal proteins were analyzed by two-dimensional PAGE and silver staining as indicated in Materials and Methods. The derivatives of S6 labeled a through d represent forms with 1 to 4 mol of phosphate in S6.

Translational up-regulation of non-5′-TOP mRNAs and down-regulation of 5′-TOP mRNAs.

The first new protein synthesis during oocyte maturation that is regulated at the translational level is the synthesis of Mos (34, 54–56). Mos protein concentration needs to reach a threshold level in order to further activate downstream events (12, 13). Therefore, we compared the expression of Mos in control and rapamycin-treated oocytes during maturation (Fig. 8A). In rapamycin-treated oocytes immunoblotting detected earlier expression of Mos, and the amount of protein was increased as well. Since Mos protein is sufficient to induce maturation (52, 65), this provides an explanation for how rapamycin facilitates oocyte maturation. It has been shown in other systems that rapamycin can specifically inhibit initiation of translation of mRNAs containing a 5′-TOP (28, 29, 59). The mouse mos mRNA, and presumably its X. laevis homolog, does not contain a 5′-TOP in the 5′ UTR (23). As 5′-TOP mRNAs can represent up to 20% of the total mRNA in the cell (2, 5), translation of mRNAs lacking this sequence may be up-regulated in rapamycin-treated oocytes by a mechanism involving the release of 5′-TOP mRNAs from polyribosomes. Since oocytes have no spare translational capacity (37), this may enable mRNAs without a 5′-TOP, like mos mRNA, to be translated more efficiently, as shown in Fig. 8A.

FIG. 8.

Mos and Cdc25A protein levels are increased after rapamycin treatment. (A) Extracts equivalent to one oocyte from untreated or rapamycin-treated oocytes were separated by SDS-PAGE (10% gel), transferred to a nitrocellulose membrane, and probed with anti-Mos antibody. Time after addition of progesterone is indicated at the top. (B) Dejellied eggs were treated with rapamycin (2 μg/ml) for 30 min. After activation with the Ca²⁺ ionophore A23187, samples were collected, and extracts were separated by SDS–10% PAGE, transferred to a nitrocellulose membrane, and probed with anti-Cdc25A antibody. Time after activation of the eggs is indicated at the top.

A similar situation may occur after fertilization. An increase in p70^S6K activity (Fig. 2) and protein synthesis is evident after fertilization in X. laevis, and one protein identified in this increase is Cdc25A (Fig. 8B). In ionophore-activated eggs, we observed earlier and increased translation of Cdc25A after rapamycin treatment (Fig. 8B), further supporting the ability of p70^S6K to control the timing and extent of translation of specific mRNAs.

To ascertain directly whether this effect of p70^S6K in oocytes involves changes in 5′-TOP translation, we performed luciferase reporter assays with two different constructs, one with and one without the 5′-TOP of EF2 at the transcriptional start site, both driven by the genomic hamster EF2 promoter. Plasmid DNA encoding these constructs was injected into the nuclei of untreated or rapamycin-treated stage VI oocytes. As a control for possible differences in transcription due to rapamycin, another reporter construct encoding a β-galactosidase was coinjected (see Materials and Methods). After incubation for 10 h at room temperature, luciferase and β-galactosidase activity was measured in extracts equivalent to 0.5 or 1 oocyte. After normalization for transcription based on β-galactosidase assays, the luciferase activity of oocytes not treated with rapamycin was set to 100% and compared to the activity of treated oocytes (Fig. 9). In rapamycin-treated oocytes, luciferase activity was consistently 30 to 40% lower with the construct containing the 5′-TOP, whereas rapamycin treatment led to 30 to 40% higher luciferase activity in extracts from oocytes injected with the construct that did not contain a 5′-TOP (Fig. 9A). This result supports the hypothesis that rapamycin increases translational capacity for non-5′-TOP mRNAs by inhibiting translation of mRNAs with a 5′-TOP region. These results were confirmed in studies using direct injection of mRNAs. The 5′-TOP construct described above was transcribed in vitro and injected into oocytes in the presence and absence of rapamycin. Rapamycin treatment caused a 30 to 40% decrease in translation (Fig. 9B), similar to the level of inhibition after cDNA injection. To evaluate whether this decrease led to a commensurate increase in available translational capacity for other mRNAs, we used injection of an mRNA with an internal IRES, which should be independent of many complex 5′ UTR controls. Indeed, recent studies have shown that in virus-infected cells treated with rapamycin, the viral IRES-containing transcripts are more efficiently translated (8). As shown in Fig. 9B, translation of an IRES mRNA in oocytes was increased nearly 40% by rapamycin treatment.

FIG. 9.

Rapamycin decreases expression of 5′-TOP RNA and increases expression of non 5′-TOP RNA. (A) Plasmid DNA (1.5 ng) of luciferase reporter constructs containing the genomic promoter of hamster EF2 with (+ TOP) or without (− TOP) a 5′-TOP region was injected into the germinal vesicle of either rapamycin-treated (rap) or untreated stage VI oocytes. A β-galactosidase reporter was coinjected to control for any effect of rapamycin on transcription. Extracts were prepared, and luciferase and β-galactosidase activities were measured in an aliquot equivalent to 0.5 oocyte. Luciferase activity was normalized to β-galactosidase activity, and the luciferase activity of untreated oocytes was set to 100%. Similar results were obtained in three independent experiments. (B) In vitro-transcribed mRNA (10 ng) encoding a luciferase reporter construct with 5′-TOP or a β-galactosidase reporter with an IRES was injected into stage VI oocytes in the presence or absence of rapamycin (rap). Luciferase and β-galactosidase activities were measured in extracts corresponding to 0.5 or 0.05 oocyte after incubation of the oocytes for 5 h. Similar results were obtained in five independent experiments.

DISCUSSION

Fully grown stage VI oocytes are arrested in prophase of meiosis I, which corresponds to late G₂ phase in the cell cycle. During maturation a hormonal stimulus releases the oocytes from their arrest, inducing completion of meiosis I and progression to metaphase of meiosis II, where they arrest again awaiting fertilization (24, 40, 54). p70^S6K activity is present in unmatured stage VI oocytes (reference 36 and Fig. 2A, 2B, and 4A) and in earlier stages (data not shown). Upon induction of maturation by progesterone, activity of p70^S6K decreases, suggesting that low kinase activity facilitates progression through the cell cycle in the oocyte. This notion is supported by the observation that incubation of oocytes in rapamycin, leading to a complete loss of p70^S6K activity, accelerates oocyte maturation and increases sensitivity to progesterone (Fig. 4 to 6).

During oogenesis, each oocyte produces a huge stockpile of 10¹² ribosomes that supports embryonic development until the swimming tadpole stage (2–5). Recruitment of rp-mRNAs onto polysomes, a measure of translational activity, increases throughout oogenesis, reaching its maximum in stage VI oocytes (7, 10). This elevated accumulation of ribosomal proteins is due largely to the preferential translation of rp-mRNAs in mid to late stages of oogenesis (15). At that time, translation of rp-mRNAs comprises ∼20% of total protein synthesis, and ribosomal proteins are the major class of protein being produced. p70^S6K activity is present during this period (data not shown) and decreases (Fig. 2) concomitant with the cessation of ribosomal protein synthesis during oocyte maturation (27). This finding suggests that decreased p70^S6K activity may be responsible for the down-regulation of translation of rp-mRNAs, which contain 5′-TOP regions. In our experiments, synthesis of Mos protein, whose mRNA does not contain a 5′-TOP, starts earlier and reaches a higher amount in rapamycin-treated oocytes (Fig. 8A), resulting in accelerated GVBD (Fig. 5). Also, an IRES reporter construct without a 5′-TOP was translated more efficiently in rapamycin-treated oocytes, whereas expression of a construct containing a 5′-TOP was decreased (Fig. 9). This result suggests a model in which preferential translation of rp-mRNAs with 5′-TOP regions occurs in oocytes when p70^S6K activity is high. A reduction of p70^S6K activity reduces translation of rp-mRNAs and releases translational capacity for mRNAs without a 5′-TOP region, like Mos, which are required for oocyte maturation.

p70^S6K activity changes may also affect protein synthesis after fertilization. In the embryo, translation of ribosomal proteins S3, L17, and L31 starts from stage 1 onward, and protein L5 begins to be synthesized around stage 7 (49), corresponding to one of the peaks of activity of p70^S6K (Fig. 2C). It is possible the high p70^S6K activity in early embryos contributes to specific translation of these selected rp-mRNAs.

Phosphorylation of ribosomal protein S6 correlates with translation of 5′-TOP RNA, as shown for EF1α in cultured cells after mitogenic stimulation (30). This finding suggests that in cultured cells the rapamycin effects on translation are mediated via p70^S6K-dependent S6 phosphorylation. In mitogen-treated cells, the highly phosphorylated derivatives of S6 are selectively found in polysomes (61). In contrast, in oocytes despite increased S6 phosphorylation prior to GVBD (reference 44 and Fig. 7), only 1% of the ribosomes are in polysomes. Moreover, the kinase responsible for this increased rapamycin-insensitive S6 phosphorylation does not appear to be p70^S6K, whose activity is undetectable at GVBD, but rather p90^Rsk (19, 20) (Fig. 7). An analogous situation occurs after fertilization, when p90^RSK activity and S6 phosphorylation rapidly decrease (reference 19 and unpublished data) despite an increase in p70^S6K activity (Fig. 2). Although there is a twofold increase in total protein synthesis and maximal S6 phosphorylation at GVBD, ribosomal protein 5′-TOP mRNAs are no longer being translated. Indeed, since the p90^Rsk-dependent S6 phosphorylation in this system occurs long after p70^S6K is inactivated (Fig. 2), it is possible that the minor fraction of ribosomal 40S subunits on polysomes is phosphorylated by p70^S6K before and shortly after induction of maturation with progesterone. Similarly, although most S6 is dephosphorylated after fertilization in concert with deactivation of p90^Rsk (19, 44), it cannot be excluded that a minor fraction of S6 is phosphorylated by p70^S6K after fertilization. Alternatively, these results could indicate that other targets of p70^S6K besides S6, such as trans-acting factors (11, 39), account for rapamycin effects on 5′-TOP mRNA translation.

The fact that progesterone down-regulates p70^S6K activity and rapamycin affects the progesterone response implies that the rapamycin-sensitive pathway is under hormonal control in oocytes. Although it is evident that p70^S6K deactivation in progesterone-treated oocytes is due to dephosphorylation (Fig. 2), at present we cannot distinguish between inhibition of upstream kinases such as mammalian TOR versus activation of phosphatases. Since the effect of rapamycin could be observed only at threshold levels of progesterone, it is likely that other mechanisms activated by progesterone also contribute to the switch to preferential translation of non-5′-TOP RNAs during oocyte maturation. These might include changes in both the 3′ and 5′ UTRs of mRNAs such as mos that occur following progesterone treatment (34). Translation of such mRNAs, however, is still constrained by the need for additional translational capacity afforded by down regulation of p70^S6K activity and reduced translation of 5′-TOP mRNAs.