The active form of Xp54 RNA helicase in translational repression is an RNA-mediated oligomer

Cloning of Xenopus CPEB

The CPEB expression plasmid was constructed by inserting the open reading frame (ORF) of CPEB, as an NdeI–XhoI fragment, into pET21b (Novagen). The cDNA was produced by PCR from a Xenopus cDNA library (46) using oligodeoxynucleotides 5′ CCA GGA GCC TGC CAT ATG GCC TTC CCA CTG 3′ and 5′ GTT GTT CCA ATG CTC GAG GCT GGA GTC ACG 3′; NdeI and XhoI sites are underlined. The amino acid sequence of the bacterially expressed protein differed from that of the endogenous protein by the addition of a C-terminal extension with the sequence LEHHHHHH. Haemagglutinin (HA)-tagged CPEB was produced by insertion of the CPEB ORF as a EcoRI–NotI fragment into a construct containing a 3× HA tag, β-globin 5′ and 3′UTRs and a 3′ A₂₃C₃₀ tail. The cDNA was produced by PCR from pET21b-CPEB using oligodeoxynucleotides 5′ GAG GAA TTC ATA TGG CCT TCC CAC TGA AAG 3′ and 5′ GTG GGC GGC CGC TTA GCT GGA GTC ACG ACT 3′ (sites underlined) and cloned into an EcoRI–NotI-digested vector derived from pS664TEN (a gift from S. Morley); the vector backbone was substituted as an HindIII–PstI fragment into pGEM2 (Promega) with a deleted EcoRI site, retaining the β-globin UTRs and the A₂₃ tail, but allowing subsequent digestion with SmaI and transcription with T7 RNA polymerase. Fused in-frame at the N-terminus are three copies of an influenza HA epitope tag.

Xenopus oocyte preparation

Ovarian lobes were removed from Xenopus laevis females. Stage VI oocytes were isolated and stored at all times in modified Barth’s solution [8.8 mM NaCl, 1 mM KCl, 330 µM Ca(NO₃)₂, 410 µM CaCl₂, 820 µM MgSO₄, 2.4 mM NaHCO₃, 10 mM HEPES-NaOH pH 7.4]. Oocytes were microinjected as required. Maturation was performed, if necessary, by treatment of oocytes with 10 µg/ml progesterone at 20°C for 16 h. Extracts were made either crude in minimal modified Barth’s solution or diluted in an appropriate buffer, and clarified twice by centrifugation at 9300 g for 10 min at 4°C. When ³⁵S labelling was required, 5–10 oocytes were transferred, 3 h post-injection, into 50 µl of modified Barth’s solution containing [³⁵S]promix (Amersham) at 1 µCi/µl and incubated for 16 h at 20°C. After removing excess label, oocytes were washed in modified Barth’s solution and processed as normal.

Tethering

The MSP vector and the Luc-MS2 reporter were supplied by N.K. Gray (47). cDNAs encoding MS2–Xp54 and its mutated forms have been described previously (21): plasmids encoding fusion proteins were linearized with HindIII and Luc-MS2 with BglII prior to transcription with T7 RNA polymerase. Luc-MS2 reporter constructs requiring a poly(A) tail were modified using Escherichia coli poly(A) polymerase from a poly(A) tailing kit (Ambion), following the manufacturer’s instructions to add a tail of about 150 A residues.

Oocyte micromanipulation was performed essentially as described above. A 50 nl aliquot (at ∼500 ng/µl) of a solution of mRNAs encoding fusion proteins was injected into stage VI oocytes. After 3 h, oocytes were treated with progesterone and/or [³⁵S]promix, as required and, after a further 3 h, injected with 10 nl of a solution containing Luc-MS2 reporter mRNA (10 ng/µl) and Renilla luciferase control mRNA (0.35 ng/µl). Incubation was continued overnight before harvesting.

Three to five pools each containing 3–5 oocytes were assayed per experimental point. Oocytes/eggs were homogenized in lysis buffer (50 µl/oocyte; Promega) and 10 µl samples assayed for firefly and Renilla luciferases using a Dual-Luciferase assay system (Promega). Relative light determinations were measured in a TD20/20 Turner Designs luminometer.

Immunoprecipitation

Immunoprecipitation experiments were performed from oocytes injected with MS2 fusion proteins, using mouse monoclonal anti-MS2 antibodies (a generous gift of Mike Kiledjian; Nelson Biological Laboratories, New Jersey). Where oocytes were co-injected with HA-tagged CPEB, 50 nl of an RNA mixture containing 500 ng/µl MS2–Xp54 (or MS2) and 500 ng/µl HA-CPEB was injected and oocytes incubated for 16 h. Oocytes co-injected with reporter constructs were manipulated as described above. S10 lysates were prepared from 20 oocytes/eggs in 1 ml of NET buffer [50 mM Tris–HCl pH 7.5, 150 mM NaCl, 0.5% (v/v) NP-40, 1 mM EDTA pH 8.0, 0.25% (w/v) gelatin, 0.02% (w/v) NaN₃] + 2% (w/v) bovine serum albumin (BSA). Lysates were incubated, with gentle mixing, with 0.1 µl of anti-MS2 antibody for 2 h at 4°C, prior to addition of 10 µl of protein G–Sepharose beads for 2 h at 4°C. Beads were washed in 3 × 1 ml of NET buffer and bound proteins eluted in 20 µl of protein sample buffer. A 10–20 µl aliquot of each sample was separated by SDS–PAGE and proteins detected by silver staining, autoradiography or western blotting. For immunoprecipitations involving RNase A-treated extracts, Xenopus extracts were supplemented with RNase A to 20 pg/µl extract and incubated at 20°C for 10 min, prior to clarification by centrifugation at 9300 g for 10 min at 4°C. This treatment was sufficient to degrade mRNA, as judged by in vitro translation of phenol-extracted RNA. Immunoprecipitation reactions for the isolation of proteins for peptide sequencing were performed essentially as described above by scaling up reaction volumes 25-fold; volumes of protein G–Sepharose, however, were scaled up 10-fold.

Mass spectrometry analysis

Proteins within the gel-excised bands were first reduced, carboxyamido methylated, and then digested to peptides using trypsin on a MassPrepStation (Micromass, Manchester, UK). The resulting peptides were subjected to LC–MS/MS. The liquid chromatographic separation was achieved with a PepMap C18, 180 µm i.d., 15 cm column (LC Packings, Amsterdam). The mass spectrometer was a QTof (Micromass). Fragmentation data were used to search the National Center for Biotechnology Information database using the MASCOT search engine (http://www.matrixscience.com). Probability-based MASCOT scores were used to evaluate identifications. Only matches with P < 0.05 for random occurrence were considered significant (further explanation of MASCOT scores can be found at http://www.matrixscience.com).

Preparation of anti-Xp54 antibody

Anti-Xp54 antibody was obtained from rabbits immunized with acrylamide gel fragments containing recombinant Xp54 (CovalAb). The Xp54 expression plasmid was constructed by inserting the complete ORF of Xp54 as an NheI–SalI fragment into NheI–XhoI-digested pGIT [the 5′UTR and 3′UTR-A₃₃C₁₈ of Xenopus globin were introduced into the XbaI and BlpI sites of pET21d (Novagen), in order to create a single vector suitable for protein expression in bacteria and in vitro synthesis of RNA for microinjections into Xenopus oocytes]. The cDNA was produced by PCR from Xp54 clone B2 (33), using oligodeoxynucleotides 5′ CGC GCT AGC ATG AGC ACC GCC AGA A 3′ and 5′ ACG GTC GAC TTA AGG TTT GTC TTC C 3′. A culture of transformed BL21 star cells was grown to an A₆₀₀ of approximately 0.3 and induced for 4 h at 37°C by the addition of 1 mM isopropyl-β-d-thiogalactopyranoside (IPTG). Cells were harvested by centrifugation at 5000 g for 10 min and resuspended in 0.01–0.02 volumes of NTA buffer [300 mM NaCl, 1% (v/v) Triton X-100, 50 mM sodium phosphate buffer, pH 7.8]. Cells were lysed using a French pressure cell, and cell debris, including the insoluble Xp54 helicase, pelleted at 8000 g for 20 min. Pellets were resuspended in NTA buffer and aliquots run on SDS–polyacrylamide preparative gels. Coomassie-stained Xp54-containing gel slices were excised and used for immunization.

Purification of anti-CPEB peptide antibody

Rabbit anti-peptide Xenopus CPEB antibody was described previously (32). His-tagged CPEB was expressed in BL21 star cells, as described above for Xp54, and 10 µg purified from the post-lysis insoluble fraction by SDS–PAGE and western blotting. The membrane was stained briefly with Ponceau S [0.2% (w/v) dye in 0.3% (w/v) trichloroacetic acid] prior to excision of the CPEB-containing band. Membrane fragments were washed in 3 × 500 µl of phosphate-buffered saline (PBS), prior to incubation in 5% (w/v) BSA in PBS at 4°C for 1 h, with agitation. A 200 µl aliquot of antiserum was incubated with the blocked membrane for 16 h at 4°C, depleted serum removed and membranes washed in 3 × 1 ml of PBS. Bound antibody was eluted in 100 µl of 0.1 M glycine, pH 2.5 for 10 min at 20°C, neutralized with 10 µl of 1 M Tris, pH 8.0, and diluted to 200 µl in PBS.

Immunoblotting

Protein samples were separated by SDS–PAGE on 15% (w/v) acrylamide gels and transferred onto an Immobilon P membrane (Millipore) using a semidry blotting apparatus. Rabbit anti-Xp54 antibody was used at 1:10 000, affinity-purified rabbit anti-CPEB antibody at 1:10 000 and eIF4E antibody (gift of Dr Simon Morley) at 1:30 000. Western blots were subsequently detected by enhanced chemiluminescence (ECL).

FPLC gel filtration

A 24 ml Superose 6 HR 10/30 column was washed with ddH₂O at 0.2 ml/min for 70 min, followed by 0.5 ml/min for 60 min, and equilibrated in eluent buffer (34.2 mM Na₂HPO₄, 15.8 mM NaH₂PO₄, 150 mM NaCl). The void volume was determined by running 1 mg/ml dextran blue and the column calibrated with thyroglobulin (669 kDa), ferritin (440 kDa), catalase (232 kDa), aldolase (158 kDa), ovalbumin (43 kDa), chymotrypsinogen A (25 kDa) and RNase A (13 kDa) (all Amersham Biosciences) in eluent buffer at a flow rate of 0.5 ml/min. Elution volumes for each protein were used to produce a calibration curve from which further approximate molecular weights were determined. A 200 µl aliquot of stage V/VI S10 extract was fractionated through the column at a flow rate of 0.5 ml/min, and 0.3 ml fractions collected. Alternate fractions were separated by SDS–PAGE and visualized by western blotting. Where required, lysate was treated with RNase A, 167 µg/ml for 10 min at 20°C prior to loading on the column.

Xenopus CPEB and Xp54 interact in vivo

We previously showed that clam p47 helicase and clam p82/CPEB co-precipitate in oocyte lysates, using both anti-CPEB and anti-helicase antibodies (21). Here we examined whether the Xenopus forms of the RNA helicase and CPEB interact in oocytes. As Xp54 and XCPEB proteins have very similar sizes on SDS–PAGE, and these are close to the size of the heavy chain of IgG, and both available antibodies were raised in rabbits, we could not examine such an interaction with the endogenous proteins. Thus, oocytes were co-injected with HA-CPEB and MS2–Xp54 mRNA or control MS2 mRNA and labelled with [³⁵S]promix. Oocytes were harvested after 22 h and lysates used in immunoprecipitation experiments with MS2 antibodies and protein G–Sepharose. In vitro translated proteins, labelled in the rabbit reticulocyte lysate cell-free system, were analysed alongside the precipitated proteins to show protein size (Fig. 1). The results indicate that HA-CPEB co-precipitates with MS2–Xp54, but not with MS2 control protein. Furthermore, experiments reported below show that MS2–Xp54 associates with endogenous CPEB, and that the native proteins co-purify in gel filtration (see Figs 5 and 6). We conclude that as in Spisula, Xenopus CPEB and Xp54 interact in vivo.

Tethered Xp54 represses translation in oocytes and in eggs

To examine the role of Xp54 in translation, we used the tethered function assay (48). As Xp54 associates with CPEB, a specific RNA-binding protein that binds 3′UTR CPEs, Xp54 was fused with MS2 coat protein and tethered to the 3′UTR of a reporter RNA via MS2 coat protein-binding sites. First, mRNA encoding MS2–Xp54 was microinjected into oocytes, some of which were matured with progesterone (see Materials and Methods). After several hours of incubation to allow the synthesis of fusion protein, firefly luciferase mRNA containing MS2-binding sites in its 3′UTR was injected, along with an internal control Renilla luciferase mRNA lacking any regulatory elements in its 3′UTR (Fig. 2). Variants of Xp54, containing single amino acid changes (DEAD→DQAD) and (HRIGR→HRIGQ) in the conserved helicase motifs II and VI, which reduce or abolish ATPase hydrolysis and ATP-binding, respectively [reviewed in Linder (41)], were also examined (Fig. 2). We monitored the effects of MS2–Xp54, relative to MS2 alone, on the ratio of firefly to Renilla luciferase activities. As the levels of MS2, and all Xp54 fusion proteins were approximately equal (see Figs 3B, 4 and 5), and firefly reporter RNA levels were not affected by ectopic Xp54 expression [not shown, see Minshall et al. (21)], the effects were exerted at the level of translation. We estimate that ∼1–5 ng of Xp54 is expressed ectopically in each oocyte in our conditions, in contrast to the 50–100 ng levels of the endogenous helicase (not shown); precluding any possible effects of wild-type or dominant-negative forms of the helicase on total protein synthesis.

First, we observed that Xp54 tethered to the 3′UTR represses translation of firefly luciferase reporter RNA in both oocytes and eggs to very similar extents (approximately 3–4 times) relative to MS2 (Fig. 3A). In contrast, mutant helicases activated translation, in both cell types, by ∼3- to 5-fold, relative to MS2 (Fig. 3A). Thus the effects of ectopic Xp54, all variants, on reporter mRNA were the same irrespective of cell type, implying that the helicase is not functionally modified during maturation. During meiotic maturation, some maternal mRNAs are rapidly mobilized into polysomes by a process involving cytoplasmic polyadenylation (see Introduction). In this experiment, the firefly reporter RNA was not adenylated prior to injection and, due to lack of CPE and hexanucleotide elements in its 3′UTR, was not adenylated during maturation. When firefly reporter mRNA was adenylated prior to microinjection (by an estimated 150 nt, see Materials and Methods), its translation was stimulated ∼10-fold in oocytes (Fig. 3A), as expected from previous studies showing the stimulatory effect of a poly(A) tail. This stimulation of translation was not affected by co-expression of either wild-type or mutant helicases [Fig. 3A and Minshall et al. (21)]. Thus Xp54, while able to repress translation in eggs when the tethering mRNA is non-adenylated, does not repress the translation of polyadenylated RNA.

Xp54 represses translation in association with p50

Next, we examined the possible co-purification of endogenous proteins with wild-type and mutant (DQAD/HRIGQ) Xp54 proteins fused with MS2. Oocytes or eggs, co-injected with firefly luciferase reporter mRNA and MS2–Xp54 mRNAs or MS2 control mRNA, were incubated in Barth’s solution containing [³⁵S]promix, to allow sensitive detection of protein ligands. Cells were harvested after 22 h and lysates used in immunoprecipitation experiments with anti-MS2 antibodies and protein G–Sepharose. Several proteins co-precipitated with Xp54 proteins specifically, as they were absent from the MS2 control samples (first lane in each panel). Strikingly, one protein, of ∼50 kDa, consistently co-precipitated with wild-type MS2–Xp54, but not with either of the two mutant helicases (Fig. 3B). Moreover, MS2–Xp54 does not co-purify with endogenous p50 in cells injected with the polyadenylated reporter RNA (Fig. 3B). Altogether, the data shown in Figure 3 indicate that recruitment of newly synthesized p50 by MS2–Xp54 correlates with the ability of the tethered helicase to repress translation. In parallel experiments in which the lysates were treated with RNase A prior to immunoprecipitation, we noted that the association of p50 with MS2–Xp54 was RNA dependent (Fig. 4).

Identification of p50 as Xp54

Large-scale immunoprecipitations were performed using lysates from unlabelled oocytes microinjected with MS2–Xp54 mRNA. The protein G–Sepharose MS2 antibody-bound proteins were stained with Coomassie colloidal blue, and the band migrating at ∼50 kDa was submitted for mass spectrometer peptide sequencing following trypsin digestion (see Materials and Methods). The highest probability-based Mowse score, 236, in a MASCOT search of NCBInr (MATRIX Science; http://www.matrixscience.com), was shown by vertebrate p54 helicase proteins with eight matching peptides (GNEFEDYCLK, SGAYLIPLLER, VMATTGGTNLR, LDDTVHVVIATPGR, GVTQYYAYV TER, VFHDFR, NLVCTDLFTR and LAETYLHR).

To confirm that the 50 kDa polypeptide interacting with the wild-type helicase was Xp54, control MS2 and MS2–Xp54 mRNAs were injected into oocytes, some of which were matured into eggs, and the MS2 immunoprecipitation reactions were analysed by western blot with Xp54 antibody. As shown in Figure 5, this antibody recognizes the MS2–Xp54 helicase and a protein of ∼50 kDa which co-precipitates with MS2–Xp54, in an RNA-dependent manner, in both oocytes and eggs. (A very low level of non-specific binding of Xp54 to MS2, also RNase-sensitive, was also noted.) On the basis of both the protein sequence and the western data, we conclude that Xp54 interacts with itself, as a dimer or higher oligomer, via RNA. From the higher ECL signal ratio of endogenous Xp54 to MS2–Xp54 (Fig. 5), it seems likely that at least a portion of the immunoprecipitated Xp54 is in an oligomeric form. Since using antibodies in immunoprecipitation and westerns assays may not yield quantitative data, we have not sought to determine the precise oligomerization number.

Changes in Xp54–CPEB–eIF4E interactions during meiotic maturation

To confirm and extend the nature of the interactions seen previously between Xp54 and CPEB-tagged proteins in oocytes (Fig. 1), the MS2 western blot was overprobed with CPEB antibody (Fig. 5). Endogenous CPEB interacts specifically with MS2–Xp54 in oocytes (Fig. 5), but, in contrast to the MS2–Xp54:Xp54 interaction, which is sensitive to RNase, the co-precipitation of CPEB with MS2–Xp54 does not require RNA (Fig. 5; see Discussion). In eggs, due to CPEB degradation during meiotic maturation (30–32), the total levels of CPEB are very much reduced and CPEB is not detectable in MS2–Xp54 immunoprecipitates (Fig. 5).

To examine possible interactions between Xp54 and components of the translational machinery, the MS2 western blot was also overprobed with eIF4E antibody (Fig. 5). Interestingly, the cap-binding initiation factor interacts with MS2–Xp54 in an RNA-independent manner in oocytes. More unexpectedly, the interaction between Xp54 and eIF4E changes from being independent of RNA in oocytes to being mediated by RNA in eggs (Fig. 5). Altogether, the MS2–Xp54 pull-down experiment confirms the identification of p50 as Xp54, as well as the interaction between Xp54 and CPEB. Moreover, evidence was obtained of dynamic changes in Xp54–CPEB–eIF4E interactions during meiotic maturation.

Xp54 is present in large and intermediate sized complexes

To extend our tethering and pull-down experiments, we sought an independent way to assess the protein ligands of endogenous Xp54. Lysates, prepared from stage V/VI oocytes, either in an untreated form or after treatment with RNase A, were analysed by gel filtration on a Superose 6 HR 10/30 column by FPLC (Fig. 6). Previous workers had shown that in stage V and VI oocytes, Xp54 is entirely cytoplasmic (42). Column fractions were analysed by western blotting using Xp54, CPEB and eIF4E antibodies. In untreated lysates, Xp54 fractionated with a broad profile throughout the column, in fractions ranging from ∼50 kDa to >2 MDa, representing monomer/dimer and larger complexes. Most of the large and intermediate sized complexes are sensitive to RNase, since in the nuclease-treated lysate the majority of Xp54 is found in fractions that correspond approximately to a monomer species. A minor fraction of Xp54, in the large complexes, is resistant to RNase (Fig. 6). In contrast, CPEB is found only in large, RNase-resistant complexes, in association with an Xp54 monomer. These large complexes also contain a fraction of eIF4E. These results support those obtained in tethering and immunoprecipitation studies. First, it appears that both native and tagged Xp54 is found in RNA-dependent dimer (and higher oligomer) forms, and secondly, both associate, directly or indirectly, with CPEB in oocytes in very large complexes which contain additional protein partners, including eIF4E.

Oligomerization of Xp54 and other RNA/DNA helicases

Using the tethered function approach and gel filtration analysis, we have shown that the DEAD-box helicase Xp54 oligomerizes in an RNA-dependent manner, and that this self-association correlates with the ability of the helicase to repress translation of reporter RNA (Figs 3, 4 and 6). Self-association is not observed when the tethered Xp54 is inactivated by mutations in the DEAD and HRIGR motifs, nor when the reporter RNA is pre-adenylated and not subject to Xp54 control (Fig. 3). Moreover, our gel filtration data support the existence of a substantial fraction of Xp54 helicase in RNase-sensitive oligomeric forms. [Interestingly, similar biochemical fractionation studies also indicate that Dhh1 is present in very large complexes in yeast cells (44,49)]. Thus, our data imply that Xp54 helicase is active as an oligomer during the translational repression in vivo. One possible explanation of these results is that the inactivating mutations abrogate RNA binding, resulting in loss of oligomerization. The DEAD-box (motif II), the Walker B motif of ATP-binding proteins, contains the highly conserved glutamic residue required for ATP hydrolysis. DQAD or DAAD mutants of eIF4A, Dbp5 and Ded1 bind ATP and RNA with wild-type (or near wild-type) affinities, but lack ATPase activity (50–52), while the principal role of motif VI (HRIGR) is now understood to be to bind ATP [reviewed in Linder (41)]. Based on these examples, it appears unlikely that both the Xp54 mutations resulted in the loss of RNA binding; however, until we can produce significant amounts of soluble proteins, and identify the RNA target(s), we cannot be entirely certain of this conclusion.

Xp54 joins a growing list of helicases that function as dimer/oligomers. A well-characterized example is the hepatitis C NS3 protein, the non-structural viral protein with protease, NTPase and RNA/DNA helicase activities. The crystal structure report of Cho et al. (53) suggests that NS3 helicase, a member of the DEXH family of helicases, forms a dimer, with a cleft between the components of the dimer through which a single-stranded nucleotide can pass as the helicase unwinds the RNA. Dimerization or oligomerization of NS3 helicase requires the presence of single-stranded DNA and is necessary for optimal helicase activity (54,55). A dimer of E.coli UvrD, a DNA helicase, is also the active form of the helicase in vitro. Two UvrD monomers must bind the DNA substrate, and both must be able to hydrolyse ATP, in order to form a functional unwinding complex (56). The highly related DEAD box RNA helicases p68 and p72, involved in both transcription and mRNA processing, exist as homodimers and, more abundantly, as heterodimers, as well as monomers, in cells (57). Mutation of DEAD to NEAD, rendering them ATPase and helicase inactive, weakens their interactions and prevents association with fibrillarin in the yeast two-hybrid system. Dimerization is mediated by a large part of the conserved core of the helicases, but in this case does not require RNA (57). While exact details vary, in the case of those helicases that are known to dimerize, directly or indirectly via nucleic acid substrate, self-association is required for helicase function.

CPEB–Xp54 association

We also showed that Xp54 interacts with CPEB, in both immunoprecipitation and gel filtration analyses, thus extending our previous report that Spisula p82/CPEB binds the p47 homologue in oocytes (21). This interaction may also be conserved in Drosophila. Thus Me31B forms a complex with Exuperantia (Exu), a protein involved in RNA localization, and Ypsilon Schachtel (Yps), the Y-box FRGY2 homologue, in an RNase-sensitive manner (34). Exu binds directly to Yps in oocytes (58). Furthermore, yps interacts genetically with orb, and, Orb is physically associated with both the Exu and Yps proteins, in an RNase-sensitive manner (59). While these findings, taken altogether, are suggestive of a conserved CPEB/RCK association, it remains to be established whether Orb and Me31B themselves interact, and whether such an interaction is mediated by RNA. Previously we reported that in clam oocytes, p82/CPEB bound p47 helicase via RNA (21). This conclusion was based on the observation that RNase treatment of lysates prior to immunoprecipitation with p82 antibodies prevented co-immunoprecipitation of p47. We also showed that RNase treatment of lysates prior to immunoprecipitation with p47 antibodies only partially reduced the levels of co-immunoprecipitating p82. Then, we ascribed these differences to the varying abilities of the antibodies to bind RNP complexes and protect their contents from nuclease (21). However, the data shown here, using essentially the same protocols but in Xenopus oocytes, clearly show that CPEB and Xp54 associate in the absence of RNA (Fig. 5). We believe the two sets of data can be reconciled in light of the observation that Xp54 oligomerizes in an RNA-dependent manner (Figs 4 and 5). Thus, most of the Xp54 polypeptides are lost upon nuclease treatment from CPEB precipitates, and the one attached monomer may fall below the level of detection. In the reverse reaction, all helicase polypetides are recovered, with the associated CPEB. Nevertheless, it seems likely that CPEB and Xp54 do not interact directly, but do so via another protein(s), as we have been unable to detect binding beween the two proteins synthesized in the reticulocyte lysate (data not shown). We are currently determining the domains in each protein that are required for their association in oocytes.

Role of Xp54 in translational repression

On the basis of our data, we propose the following model for Xp54 function in repression (Fig. 7). When tethered to the 3′UTR of non-adenylated mRNA, Xp54 oligomerizes and represses translation. Tethering to this location may be artificially assisted as with the MS2 system, or provided by CPEB which specifically binds the U-rich CPEs in translationally regulated RNAs. Since tethered, oligomeric Xp54 is able to inhibit translation in both oocytes (MS2 and CPEB as tethers) and eggs (MS2 tether), i.e. in the presence or absence of CPEB, we conclude that an Xp54–CPEB interaction is not required for repression. Tethered Xp54 which represses translation in oocytes interacts with eIF4E, possibly via CPEB and maskin (19), and may sequester it from the translational machinery. In progesterone-matured eggs, in the absence of CPEB, Xp54 remains capable of RNA-dependent self-association; however, it is presumably no longer anchored to the CPE motifs owing to the lack of CPEB. This release may render it capable of reorganizing protein complexes and allowing polyadenylation in vivo. When reporter RNA is injected in a polyadenylated form, the 3′ extension appears to prevent recruitment of additional Xp54 molecules, implying its function in the change between repressed and activated states. The interaction with eIF4E also changes dramatically, from being RNA independent in oocytes to requiring RNA in eggs, and may reflect a functional relocation from 3′UTR-based protein ligands to the 5′ cap structure. Interestingly, a dynamic eIF4E–cap interaction may also underlie the decapping enhancer function of Dhh1 in yeast (36).

We also showed that mutant tethered Xp54 helicases that fail to oligomerize do not repress translation but in fact strongly activate translation [see also Minshall et al. (21)]. We are currently investigating the basis of this stimulation; it is possible that a further unidentified component of the mRNP complex is responsible for this dramatic activation. We speculate that the oligomeric form of wild-type Xp54 prevents this protein from binding the complex or from carrying out its activation role.