Release of eukaryotic initiation factor 4E (eIF4E) from its translational repressor eIF4E-binding protein (4E-BP) is a crucial event for the first mitotic division following fertilization of sea urchin eggs. Finding partners of eIF4E following fertilization is crucial to understand how eIF4E functions during this physiological process. The isolation and characterization of cDNA encoding Sphaerechinus granularis eIF4G (SgIF4G) are reported. mRNA of SgIF4G is present as a single 8.5-kb transcript in unfertilized eggs, suggesting that only one ortholog exists in echinoderms. The longest open reading frame predicts a sequence of 5235 nucleotides encoding a deduced polypeptide of 1745 amino acids with a predicted molecular mass of 192 kDa. Among highly conserved domains, SgIF4G protein possesses motifs that correspond to the poly(A) binding protein and eIF4E protein-binding sites. A specific polyclonal antibody was produced and used to characterize the SgIF4G protein in unfertilized and fertilized eggs by SDS-PAGE and western blotting. Multiple differentially migrating bands representing isoforms of sea urchin eIF4G are present in unfertilized eggs. Fertilization triggers modifications of the SgIF4G isoforms and rapid formation of the SgIF4G-eIF4E complex. Whereas rapamycin inhibits the formation of the SgIF4G-eIF4E complex, modification of these SgIF4G isoforms occurs independently from the rapamycin-sensitive pathway. Microinjection of a peptide corresponding to the eIF4E-binding site derived from the sequence of SgIF4G into unfertilized eggs affects the first mitotic division of sea urchin embryos. Association of SgIF4G with eIF4E is a crucial event for the onset of the first mitotic division following fertilization, suggesting that cap-dependent translation is highly regulated during this process. This hypothesis is strengthened by the evidence that microinjection of the cap analog m7GDP into unfertilized eggs inhibits the first mitotic division.
Unfertilized sea urchin eggs are haploid cells, that are arrested after completion of their meiotic divisions at the G1 stage. Fertilization triggers entry into S phase and completion of the first mitotic division of the embryonic development. The overall rate of protein synthesis is low in unfertilized sea urchin eggs and is stimulated rapidly following fertilization. This dramatic rise in protein synthesis is independent of mRNA transcription and ribosome biogenesis (Epel, 1967; Brandhorst, 1976). De novo protein synthesis is dispensable for progression through the S phase but is required for the onset of the M phase and subsequent embryonic cell cycles (Wagenaar, 1983; Dube, 1988). The unfertilized egg apparently contains all the necessary components for translation (Davidson et al., 1982) but protein synthesis is repressed until fertilization occurs. The exact mechanisms by which translation is increased following fertilization and how those controls contribute to proper development are not yet well understood.
Among the translational machinery components, the eukaryotic initiation factor 4E (eIF4E) is a crucial target for regulation of translation initiation (Raught et al., 2000a; Richter and Sonenberg, 2005) and plays an important role in the control of the cell cycle (Cormier et al., 2003; Clemens, 2004; Mamane et al., 2004). In different organisms, eIF4E is also thought to play a role in embryogenesis including gametogenesis (Amiri et al., 2001), fertilization (Cormier et al., 2001) and establishment of embryonic axes (Niessing et al., 2002; Cho et al., 2005). In mammalian cells, the function of eIF4E is to recognize the 7-methyl guanosine triphosphate (m7GTP) `cap' at the 5′ terminus of the mRNA, thereby allowing the latter to bind to the 43S complex composed of the ribosomal 40S subunit, eukaryotic initiation factor 3 (eIF3) and eukaryotic initiation factor 2 (eIF2). For protein synthesis to take place, eIF4E must function in conjunction with a large scaffolding protein, eIF4G, that interacts with several proteins, including eIF4E, eIF4A (an ATP-dependent RNA helicase) and eIF3. eIF4G provides a physical link between the 5′ end of capped mRNA and the ribosome (Morley et al., 1997; Raught et al., 2000a). eIF4G also associates with the poly(A)-binding protein (PABP) to stimulate translation of polyadenylated RNAs, and with the MAP-kinase-interacting kinases Mnk1 and Mnk2, which phosphorylate eIF4E at serine 209 when associated with eIF4G (Gingras et al., 1999; Raught et al., 2000a). In addition to its important function in cap-dependent translation, eIF4G is also required for cap-independent translation initiation of picornavirus RNAs to recruit the 40S ribosomal subunit to mRNA (Lomakin et al., 2000). In this case, eIF4G binds independently of eIF4E to a large secondary structure element within the RNA called the internal ribosome entry site (IRES). Although only one eIF4G ortholog has been found in insects (Hernández et al., 1998), two independent isoforms of eIF4G (eIF4GI and eIF4GII) have been described in humans (Yan et al., 1992; Gradi et al., 1998). The longest possible open reading frame (ORF) of human eIF4GI has been identified only recently and it has been suggested that five isoforms of eIF4GI protein exist in cells, generated by alternative translation initiation (Bradley et al., 2002; Byrd et al., 2002). Little is known about the specific role of these different isoforms in the cell. Recent data indicate a distinctive but complex pattern of cellular localization of different fragments that result from cleavage of eIF4G during apoptosis or viral infection (Coldwell et al., 2004).
eIF4E-binding protein (4E-BP) inhibits cap-dependent translation (Haghighat et al., 1995) by competing with eIF4G for a common site on eIF4E (Mader et al., 1995). Only one 4E-BP ortholog has been described in invertebrates (Bernal and Kimbrell, 2000; Cormier et al., 2001; Miron et al., 2001); whereas three 4E-BP homologs (4E-BP1, 4E-BP2 and 4E-BP3) exist in mammals (Pause et al., 1994; Rousseau et al., 1996; Poulin et al., 1998). In mammals, the phosphorylation state of 4E-BPs regulates their interaction with eIF4E. The underphosphorylated forms of 4E-BPs interact strongly with eIF4E and thereby inhibit cap-dependent translation, whereas the hyperphosphorylated forms do not (Pause et al., 1994). 4E-BP is a downstream effector of phosphoinositide 3-kinase (PI 3-kinase) and Akt/protein kinase B (PKB) (Gingras et al., 1998) and 4E-BP hyperphosphorylation is dependent on FRAP/mTOR (FKBP12 and rapamycin-associated protein/the mammalian target of rapamycin) (Gingras et al., 2001). In sea urchin, rapid 4E-BP degradation following fertilization represents an additional mechanism of 4E-BP regulation that is controlled by a rapamycin-sensitive mTOR pathway (Salaün et al., 2003). Furthermore, the protein 4E-BP provides a molecular link between hypoxia cellular response and prolonged activation of the cell-cycle–DNA-damage checkpoints in sea urchin embryos (Le Bouffant et al., 2006).
Recently, the release of eIF4E from its translational repressor 4E-BP has been shown to be a crucial event for the first mitotic division following fertilization in sea urchin (Salaün et al., 2005). Finding partners of eIF4E following fertilization is thus important to understand how eIF4E functions during early development of sea urchin embryos. Here, we report the isolation and characterization of a full-length cDNA encoding for eIF4G from the sea urchin Sphaerechinus granularis (SgIF4G). We show that several protein isoforms of SgIF4G exist in unfertilized eggs and that these isoforms are modified following fertilization. The protein SgIF4G associates rapidly with eIF4E in response to fertilization and we provide evidence that SgIF4G interaction with eIF4E is functionally important for the completion of the first mitotic division following fertilization, implicating cap-dependent translation in this process. Taken together, our results provide a basis for the future analysis of the role of the cap-dependent translational regulation in early embryonic development.
Isolation and characterization of the cDNA encoding the sea urchin eIF4G
The release of eIF4E from 4E-BP is a major event correlated with the rise in protein synthesis triggered by fertilization of sea urchin eggs. It is then conceivable that eIF4E associates with partners such as eIF4G following fertilization. In order to analyse a potential role of eIF4G after fertilization and during early embryonic development, we decided to isolate and sequence the cDNA encoding for this protein in sea urchin eggs. Full-length S. granularis eIF4G (SgIF4G) was cloned (see Materials and Methods) and we obtained a final composite cDNA sequence that consisted of 7012 bp (EMBL accession number AJ634049). The first ATG in the 5′ extended sequence, which is located 103 nucleotides from the 5′ end of the cDNA, begins an open reading frame of 5235 nt encoding a deduced polypeptide of 1745 amino acids (Fig. 1) with a predicted molecular mass of 192 kDa. Since three in-frame upstream stop codons exist, at position 25, 52 and 88, in the 5′ untranslated region (5′ UTR), it is very likely that the first ATG in the 5′ UTR is the authentic start codon. The 3′ UTR region of the cDNA consists of 1672 nt and contains a canonical polyadenylation signal at position 6973. The cDNA was terminated with a 30-base poly(A) tail.
When the sequence of the peptide encoded by the ORF (Fig. 1A) was compared with the amino acid sequences of human eIF4GI (accession number NP-937884) and eIF4GII (accession number NP-001409), an overall identity of approximately 26% was found. However, this value increased when the central and the C-terminal regions of SgIF4G were aligned with human eIF4G proteins. The last 1005 amino acids of SgIF4G share 40% identity and 55% similarity with the C-terminal part of human eIF4GI. When compared with C-terminal human eIF4GII, only the last 820 amino acids of SgIF4G share identity (24%) and similarity (43%). The highest homology between SgIF4G and eIF4GI is found in some conserved domains such as MIF4G (the middle domain of eIF4G, a predicted eIF4A- and eIF3-binding domain), MA3 (a domain present in DAP-5, eIF4G and MA-3), and eIF5C (domain at the C-termini of GCD6, eIF-2Bϵ, eIF5 and eIF4G). The N-terminal region of SgIF4G was examined for the presence of eIF4E- and PABP-binding sites (Mader et al., 1995; Gradi et al., 1998). The sea urchin eIF4G contains sequence motifs that have significant similarity to the known eIF4E-binding motif (Fig. 1B) and to the PABP-binding domain previously reported (Fig. 1C).
A complete genome of Strongylocentrotus pupuratus (Cameron et al., 2000) is available at the National Human Genome Research Institute of Baylor College of Medicine-Houston, TX. The analysis in silico using the human genome sequencing resource suggested existence of a unique gene encoding for eIF4G in sea urchin (Morales et al., 2006). Indeed, when a 1-kb cDNA within the eIF4G coding region obtained by PCR amplification was used to probe a northern blot of polyadenylated RNAs isolated from unfertilized eggs (see Materials and Methods), a unique 8.5-kb transcript (Fig. 1D) was detected. These data suggest that a single ortholog of eIF4G is expressed in unfertilized sea urchin eggs.
To provide further evidence for the notion that cDNA encodes sea urchin eIF4G, we tested the ability of a sea urchin eIF4G polypeptide to interact with eIF4E. A GST-tagged SgIF4G fusion protein (from residue 591 to residue 887, see Fig. 1A) containing the eIF4E-binding site of the sea urchin eIF4G (GST-tagged SgIF4G) was produced and its ability to associate with a mouse recombinant GST-tagged eIF4E (GST-tagged mIF4E) was analysed (Fig. 1E). As expected, the GST-tagged mIF4E was retained on an m7GTP-affinity column (cap column) but did not co-purify with a pre-incubated GST protein (lane 1) used as a control. When pre-incubated with the GST-tagged mIF4E, the recombinant GST-tagged SgIF4G was retained on the cap column (lane 2), whereas GST-tagged SgIF4G alone did not (lane 3). Taken together, these data demonstrate that the sea urchin recombinant eIF4G polypeptide interacts and co-purifies on a cap column with a heterologous eIF4E.
Multiple isoforms of eIF4G protein exist in unfertilized and post-fertilized eggs of sea urchin
To analyse native eIF4G protein in sea urchin eggs and embryos, a polyclonal antibody directed against the GST-tagged SgIF4G was produced (see Materials and Methods). Using western analysis, we first checked that the anti-SgIF4G antibody recognized the recombinant GST-tagged SgIF4G protein; pre-immune serum did not yield any signal (data not shown). Using this antibody, we then performed western blot analysis of extracts prepared prior to fertilization (Fig. 2A, lanes 1, 3, 5 and 7) or 30 minutes post-fertilization (Fig. 2A, lanes 2, 4, 6 and 8). Immunoblotting of proteins from unfertilized eggs using the anti-SgIF4G antibody revealed several polypeptides (lane 1): a set of upper bands (approximately four) of higher molecular mass (between 190 kDa and 250 kDa, with a higher doublet corresponding to bands a/b and a lower doublet corresponding to bands c/d), one main band of 150-175 kDa (band e), a lower and weak doublet (bands f/g) and a lower band of 125 kDa (band h). Immunoblotting of proteins from eggs 30 minutes post fertilization revealed a dramatic pattern of modification of the polypeptides (Fig. 2A, lane 2). Whereas band h disappeared and the level of the polypeptide corresponding to band e significantly decreased, a new set of polypeptides of more than 250 kDa (set of bands corresponding to i) appeared. Interestingly, emetine treatment of the eggs did not affect the modifications of the isoforms pattern triggered by fertilization (data not shown), demonstrating that SgIF4G isoforms were targeted by post-translational modifications after fertilization. As a control of the specificity of the different bands recognized by the anti-SgIF4G antibodies used, immunoblotting using the pre-immune serum showed no aspecific-reacting proteins (lanes 3 and 4). Batches of anti-SgIF4G antibody were pre-incubated with 1 μg of purified recombinant GST-SgIF4G or with 1 μg of purified GST protein (Fig. 2A, lanes 5 and 6, or 7 and 8, respectively) and then used for immunoblotting protein extracts. Pre-incubation with purified GST-SgIF4G abolished the antibody reactivity towards the different polypeptides recognized by the anti-SgIF4G antibody (compare lanes 1 and 2 with lanes 5 and 6 of Fig. 2A). Conversely, pre-incubation with the purified GST protein did not hinder the antibody–antigen reaction (compare lanes 1 and 2 with lanes 7 and 8 of Fig. 2A). Therefore, the anti-SgIF4G antibody specifically and efficiently recognizes different native SgIF4G forms, suggesting that multiple isoforms of eIF4G exist in extracts prepared from eggs prior to or 30 minutes after fertilization of eggs. We then performed a pull-down assay by incubating the GST-tagged mIF4E with total protein extracts prepared prior to or post-fertilization of eggs (Fig. 2B). The multiple isoforms present in unfertilized or post-fertilized eggs interacted with GST-tagged mIF4E.
The multiple isoforms of SgIF4G associate in vivo with eIF4E
To analyse the association between the different isoforms of sea urchin eIF4G and the endogenous eIF4E, proteins in total extracts and proteins purified pre- or post-fertilization of eggs by using m7GTP columns were analysed by western blotting (Fig. 2C). The multiple isoforms of SgIF4G that were revealed in total extracts co-purified with endogenous eIF4E after m7GTP-column purification (compare left and right upper panels of Fig. 2C). Interestingly, the level of SgIF4G isoforms that co-purified with eIF4E increased dramatically following fertilization. We then asked whether the different isoforms of SgIF4G could be immunoprecipitated using anti-SgIF4G from extracts prepared pre- or post-fertilization (Fig. 2D). Whereas the multiple isoforms of SgIF4G were efficiently immunoprecipitated by the antibody from extracts prior to and following fertilization of eggs (Fig. 2D, top panel), eIF4E was only co-immunoprecipitated from post-fertilization extracts (Fig. 2D, bottom panel). Taken together these data strongly suggest that the different SgIF4G isoforms can associate with eIF4E but that SgIF4G-eIF4E complex formation increases significantly following fertilization of sea urchin eggs.
Fertilization induces rapid modification of SgIF4G isoforms and SgIF4G-eIF4E complex formation
We have shown previously that eIF4E association with 4E-BP decreases rapidly following fertilization (Cormier et al., 2001) making eIF4E available for other partners. It is then conceivable that SgIF4G association with eIF4E correlates with eIF4E–4E-BP complex dissociation following sea urchin fertilization. The proteins 4E-BP or SgIF4G bound to eIF4E were analysed after affinity purification of eIF4E on m7GTP columns using extracts from unfertilized and fertilized eggs collected every 2 minutes following fertilization (Fig. 3A). Resolution of the proteins was performed on a 15% SDS-PAGE to analyse the three different proteins in the same gel. SgIF4G associated with eIF4E approximately 10 minutes after fertilization, concurrent with eIF4E–4E-BP dissociation.
We then analysed the multiple isoforms of SgIF4G following fertilization. Total proteins extracted from eggs prior to and every 2 minutes after fertilization were resolved on 7.5% SDS-PAGE and were analysed by western blot (Fig. 3B). Interestingly, the set of higher-mass polypeptides (set of bands i) appeared and the abundance of the peptide corresponding to band h decreased as quickly as 2 minutes after fertilization. Therefore, these data demonstrate that sea urchin eIF4G modifications occur rapidly after fertilization and that these modifications precede the release of eIF4E from its repressor 4E-BP and consequently also precede the association of SgIF4G with eIF4E.
Rapamycin inhibits SgIF4G-eIF4E association but does not affect SgIF4G isoform modifications induced by fertilization
We have previously shown that dissociation of eIF4E–4E-BP complex is mediated by a rapamycin-sensitive mTOR pathway induced by fertilization (Salaün et al., 2003). We therefore tested whether rapamycin affects the formation of the SgIF4G-eIF4E complex and the modifications of the SgIF4G isoform (Fig. 4). First, we tested the effect of rapamycin on SgIF4G-eIF4E complex formation (Fig. 4A). Rapamycin prevented formation of the SgIF4G-eIF4E complex in treated embryos; SgIF4G associated with eIF4E in controls (Fig. 4A, top panels). The effect of rapamycin was also evaluated for its ability to enhance association of the eIF4E–4E-BP complex (Fig. 4A, bottom panels). Inhibition of formation of the SgIF4G-eIF4E complex correlates with the maintenance of the eIF4E–4E-BP complex in treated embryos (bottom and right panel); association of SgIF4G-eIF4E complex appeared after eIF4E–4E-BP disruption in control embryos (bottom and left panel). We then tested whether rapamycin alters modifications of the SgIF4G isoform that occur after fertilization (Fig. 4B). Rapamycin was entirely ineffective on SgIF4G isoform modifications induced by fertilization. Taken altogether, these data demonstrate that fertilization triggers rapid SgIF4G isoform modifications independently of a rapamycin-sensitive TOR pathway; this signaling pathway is involved in eIF4E–4E-BP dissociation required for formation of the SgIF4G-eIF4E complex.
An eIF4E-binding-motif peptide of sea urchin eIF4G inhibits formation of the SgIF4G-eIF4E complex and affects the onset of the first mitotic division of sea urchin embryos
Since SgIF4G associates with eIF4E and is rapidly modified following fertilization, it was important to test the effect of the inhibition of SgIF4G-eIF4E complex formation on the onset of the first mitotic division induced by fertilization. We produced a peptide based on the eIF4E-binding motif within the sea urchin eIF4G protein, deduced from cDNA and a scrambled peptide as a control. We first tested the ability of the peptide to alter the association between the recombinant protein GST-mIF4E and the recombinant GST-SgIF4G (Fig. 5A). The two recombinant proteins were incubated with 20 μM or 50 μM of eIF4E-binding peptide (KKQYDRDFLLQFQKGCT), corresponding to the eIF4E-binding sequence of the sea urchin eIF4G or a scrambled peptide (DLCQKQYRFLFKQDTGK). After incubation, GST-SgIF4G was analysed after affinity purification of GST-mIF4E on m7GTP columns (Fig. 5A). The eIF4E-binding peptide from sea urchin eIF4G inhibited the formation of the SgIF4G-eIF4E complex formation, whereas the scrambled peptide did not affect the association between the two recombinant proteins. We then tested the effects of the peptides microinjected into unfertilized eggs on the first mitotic division following fertilization. Peptides were microinjected at a final concentration of 20 μM. Fertilization per se was not affected by microinjection, because the fertilization membranes were raised normally after microinjection (data not shown). The peptide based on the eIF4E-binding of sea urchin eIF4G microinjected at this concentration delayed the first cleavage significantly compared with the control, whereas the scrambled peptide did not (Fig. 5B). These data suggest that availability of eIF4E for its partner SgIF4G is required for the completion of the first mitotic division of sea urchin eggs following fertilization.
Microinjection of m7GDP into unfertilized eggs affects the first mitotic division of sea urchin embryos
Since fertilization triggers SgIF4G-eIF4E complex formation, we sought direct evidence of a role of the cap dependent translation in the onset of the first mitotic division of sea urchin embryos. We investigated the effect of the cap analog, m7GDP (1 mM final concentration), microinjected into eggs on the first mitotic division of sea urchin embryos (Fig. 6). m7GDP administered in this way efficiently inhibits the first cleavage compared with the controls. These data suggest that the cap-dependent translation is implicated in the first mitotic division following fertilization and reinforce our hypothesis that SgIF4G interaction with eIF4E is functionally important for the early development of sea urchin embryos.
In this report we have provided experimental evidence supporting the notion that eIF4G is encoded as a single gene in sea urchins but that multiple protein isoforms of SgIF4G are present in eggs and early embryos. The existence of several isoforms of eIF4G in vivo has been described in mammals (Etchison et al., 1982; Grifo et al., 1983), but the nature of these different polypeptides has long remained unclear. Multiple isoforms of sea urchin eIF4G could be the result of a complex translation strategy involving internal initiation (Gan and Rhoads, 1996; Gan et al., 1998; Johannes and Sarnow, 1998; Byrd et al., 2005), upstream ORF codons, alternative splicing or alternate translation initiation sites (Byrd et al., 2002; Byrd et al., 2005). Since it is known that human eIF4G mRNA can direct translation by internal initiation, some of the isoforms present in unfertilized eggs could also arise from this cap-independent translational process and an investigation of SgIF4G mRNA will be required to address this question. In mammals, an upstream ORF in the 5′ UTR has been suggested to serve to downregulate cap-dependent initiation of eIF4GI translation and the cDNA encoding Drosophila eIF4G contains seven short ORFs upstream of the initiator AUG (Hernández et al., 1998). The absence of an upstream ORF in SgIF4G mRNA suggests that its translational regulation occurs independently of this mechanism in sea urchin. Since only one mRNA is detected by northern blot analysis, this also rules out the possibility that the different isoforms present in unfertilized eggs arise from alternative splicing or alternative promoter usage. In animals, the optimal consensus sequence for recognition of the AUG start codon is RCCAUGG, where the purine (R) at position –3 is the most conserved and functionally the most important position (Kozak, 2002). The G at position +4 is also conserved and important, especially in the absence of A at position –3 (Kozak, 1997). Therefore, a `strong' initiation codon is considered to contain the purine at –3, the G at +4 or both. Following the first AUG start codon, the cDNA sequence encoding the first half of the N-terminal part of SgIF4G contains 11 other AUG start codons corresponding to potential functional initiation codons. Consequently, these internal AUG codons could be involved in the production of the multiple isoforms. This hypothesis is supported by two recent studies, which demonstrate that multiple isoforms of human eIF4G are generated by use of alternate translation initiation codons (Bradley et al., 2002; Byrd et al., 2002). Little is known about the function of these different isoforms in the cell. Since eIF4G is a scaffolding protein that interacts with multiple partners, it is then conceivable that production of different eIF4G isoforms can have consequences on mRNA translation.
We found a high level of identity (40%) and similarity (60%) in the C-terminal region (in the last 370 amino acids) between sea urchin eIF4G and human eIF4GI proteins. Since the C-terminal third part of eIF4GI binds to the protein kinase Mnk1 (Pyronnet et al., 1999; Waskiewicz et al., 1999) it is reasonable to assume that SgIF4G associates with a homolog of Mnk1. Indeed, the existence of this kinase in sea urchin eggs is supported by the identification of Mnk1 in the sea urchin genome (our unpublished data). Therefore, we hypothesize that eIF4E phosphorylation occurs following fertilization (Waltz and Lopo, 1987) only when eIF4G brings Mnk1 in the vicinity of eIF4E. We also found that the sequence in the N-terminal region of sea urchin eIF4G contains the PABP recognition site and that this motif is relatively well-conserved with respect to the motifs found formerly in human eIF4GI and eIF4GII. Since PABP is now considered as a canonical initiation factor (Kahvejian et al., 2005), it would be of great interest to analyse the functional interaction between PABP and SgIF4G following sea urchin fertilization and during the embryonic cell cycle.
In agreement with the presence of the eIF4E-binding site and using convergent approaches, we have demonstrated that sea urchin eIF4G associates with eIF4E following fertilization of sea urchin eggs. Our study shows that eIF4G is rapidly involved in this early event of life, because association of SgIF4G with eIF4E increases within minutes following fertilization. Our data show perfect synchronization between the release of eIF4E from its repressor 4E-BP and the formation of the SgIF4G-eIF4E complex. By microinjecting eIF4E-binding peptide derived from SgIF4G into sea urchin eggs, we have also provided evidence that eIF4E availability for SgIF4G association is required for the onset of the first mitosis triggered by fertilization of the eggs. Taken together, these data demonstrate that, dissociation of eIF4E–4E-BP followed by formation of the SgIF4G-eIF4E complex represent important events that correlate with the rise in mRNA translation which occurs after fertilization, and these events are required for the onset of the first mitotic division of sea urchin embryonic development. Furthermore, our experiment showing that m7GDP microinjected into unfertilized eggs efficiently inhibits the first mitotic division triggered by fertilization reinforces the hypothesis that cap-dependent translation is an important way to regulate gene expression following fertilization. However, we observed that a low level of SgIF4G can be associated with eIF4E in unfertilized eggs (see Fig. 2C), suggesting that low amounts of SgIF4G-eIF4E complex exists before fertilization. This observation is in agreement with the fact that in quiescent or starved HEK 293 cells, a similarly low amount eIF4G protein is recovered after cap-column purification of eIF4E (Raught et al., 2000b). Consequently, it would be interesting to know whether pre-existing SgIF4G-eIF4E complex is involved in the low translation rate in unfertilized eggs (Epel, 1967).
The most surprising aspect of this work is the discovery that fertilization triggers a dramatic pattern modification of the SgIF4G isoforms due to post-translational processes. These modifications precede the association between eIF4E and SgIF4G (Fig. 3). Moreover, pull-down assays demonstrate that the multiple isoforms of SgIF4G present in unfertilized eggs can associate with eIF4E (Fig. 2B). These data suggest that SgIF4G-eIF4E complex formation occurs independently of the SgIF4G modifications. Therefore the functional significance of the various eIF4G isoforms remains to be elucidated. On the one hand, a new set of SgIF4G polypeptides of high molecular mass appears and, on the other hand, the level of some isoforms decreases following fertilization. Human eIF4Gs are phosphoproteins (Duncan et al., 1987; Morley and Traugh, 1989; Morley and Traugh, 1990; Donaldson et al., 1991; Raught et al., 2000b). In HEK 293 cells, several serum-stimulated phosphorylation sites have been identified by mass spectrometry and mutational analyses, and three phosphorylation sites (serines 1108, 1148 and 1192) of eIF4GI have been shown to be sensitive to rapamycin (Raught et al., 2000b). In sea urchin, rapamycin does not affect the SgIF4G modifications we observed following fertilization. Even if this result does not rule out the possibility that SgIF4G is phosphorylated via the FRAP/mTOR pathway, it demonstrates that this pathway is not involved in the appearance of the high molecular mass eIF4G isoforms. Furthermore, in our hands, treatment of extracts 30 minutes post-fertilization with different type of phosphatases does not affect the pattern of the SgIF4G isoforms (data not shown), suggesting that it is not due to phosphorylation. Rather than the result of protein phosphorylation, the SgIF4G isoforms that appear after fertilization might be the result of other post-translational modifications such as ubiquitylation. Ubiquitylation modification is conceivable because our data demonstrate that a set of SgIF4G isoforms disappears or decreases (Fig. 2A, band h or e, respectively) suggesting that fertilization may promote the instability of these isoforms. Ubiquitylation of SgIF4G has been investigated by western blotting analyses using different commercially available antibodies directed against mammalian ubiquitin but, unfortunately, none of the antibodies crossreacted with the sea urchin proteins (data not shown). As eIF4G plays a central role in the assembly of the pre-initiation complex, such modifications of SgIF4G can have drastic consequences on translation initiation induced by fertilization and during early embryonic development of the sea urchin. An intriguing question is why some SgIF4G isoforms decrease while formation of the SgIF4G-eIF4E complex increases (our results) in correlation with the stimulation of protein synthesis induced by fertilization (Epel, 1967). One possibility is that some isoforms of SgIF4G present in unfertilized eggs participate in the inhibition of the cap-dependent translation in unfertilized eggs. Indeed, distinct eIF4G fragments cleaved proteolytically following picornaviral infection or apoptosis in mammals have been shown to inhibit cap-dependent translation (Lamphear et al., 1995; Ohlmann et al., 1996; Clemens et al., 1998). However, the translation of some cellular mRNAs possessing an IRES is maintained (Henis-Korenblit et al., 2002; Stoneley and Willis, 2004). Thus, it will be of great interest to study whether SgIF4G isoforms are involved in some cap-independent translation at a low rate in unfertilized eggs and whether the removal of these isoforms is required to stimulate protein synthesis triggered by fertilization. Therefore, investigation of the molecular causes, as well as the molecular consequences of the different isoforms of sea urchin eIF4G should provide new insights in how this important player regulates translation during physiological processes such as fertilization and embryonic cell cycle progression.
Materials and Methods
Acetylcholin, 4-(2-aminoethyl)-benzenesulfonylfluoride hydrochloride (AEBSF) and glycine were purchased from Interchim. Sodium orthovanadate, EDTA, β-glycerophosphate, dithiothreitol (DTT), N-2-hydroxyethylpiperazine-N′-2-ethanesulfonic acid (HEPES), leupeptin, aprotinin, Tween 20, protamine sulfate, rapamycin, m7GDP were obtained from Sigma. Alexa Fluor 488-dextran was purchased from Molecular Probes. ECL detection reagents, 7-methyl-GTP Sepharose 4B beads were obtained from Amersham Pharmacia Biotech. Mouse monoclonal antibody directed against rabbit eIF4E was purchased from Transduction Laboratories (Lexington, KY). Rabbit anti-guinea pig, goat anti-mouse and swine anti-rabbit IgG (horseradish peroxidase-coupled) were obtained from Dako SA. Polyclonal antibodies directed against 4E-BP2 (Rousseau et al., 1996) were a generous gift from Nahum Sonenberg (McGill University, Montreal, Quebec, Canada). The amino acid sequence of the peptide corresponding to the sea urchin eIF4G (wild type: KKQYDRDFLLQFQKGCT) and the amino acid sequence of the scrambled peptide (DLCQKQYRFLFKQDTGK) were synthesized by Eurogentec (Seraing, Belgium). For the generation of the mouse eIF4E recombinant protein (GST-mIF4E), the construct was a generous gift from Simon Morley (University of Sussex, Brighton, UK).
Preparation of gametes and determination of cleavage rates
Sphaerechinus granularis sea urchins collected in the Brest area (France), were kept in running seawater and used within 5 days. Spawning of gametes, in vitro fertilization and culturing of eggs and embryos were performed as described by Marc et al. (Marc et al., 2002). Stock solutions of 20 mM rapamycin was prepared in ethanol and stored at –20°C. A final concentration of 20 μM rapamycin was added to the eggs 10 minutes before fertilization. In microinjection experiments, de-jellied unfertilized eggs were placed and stuck in a line on a dish coated with 1% protamine sulfate. Peptides and m7GDP were diluted at 2 mM and 100 mM, respectively, in microinjection buffer (10 mM HEPES pH 7.0, 200 mM KCl, 550 mM mannitol) containing 1 mM Alexa Fluor 488-dextran to allow visualization of injected eggs. The microinjection system we used resulted in the injection of approximately 1% of the volume of the egg (De Nadai et al., 1998). Cleavage was scored by observation under a light microscope.
Cloning of S. granularis eIF4G
Primers designed from the Strongylocentrotus purpuratus EST database (5′-GAT GGG CTT CCA AAC ATT CC-3′ and 5′-GTT AAA CAT CTG AGG AGT CAG C-3′) were used to amplify a 376-bp fragment by PCR using a S. granularis cDNA library (Delalande et al., 1998). Further screening of the cDNA library using the PCR fragment as a probe led to the isolation of a partial 3-kb clone. 5′ and 3′ ends were extended by RACE PCR (Smart RACE cDNA amplification kit, Clontech) using primers 5′-GAG GTA CAA GCC ACT GCT GTC ATG CCT C-3′ and 5′-GCC GCT CCT CTG TAT TAA TGG GAA GCT G-3′ to obtain a final 7-kb cDNA, containing a 5235 nt ORF. Sequencing was performed on an Applied biosystem AB3100 automatic sequencer at the Génopole Ouest sequencing facility in Roscoff (France). The EMBL accession number for S. granularis eIF4G (SgIF4G) cDNA is AJ634049.
RNA preparation and northern blot analysis
Total RNA was purified from embryos by the guanidinium thiocyanate method (Chomczynski and Sacchi, 1987) and poly(A)+ RNAs were isolated using the Oligotex mRNA isolation kit (Quiagen). One and 8 μg of poly(A)+ RNAs were separated by electrophoresis on a 1% agarose gel containing 2.2 M formaldehyde. After electrophoresis, RNAs were partially hydrolyzed in 0.05 N NaOH for 20 minutes and transferred to a hybond-N+ nitrocellulose membrane (Amersham Biosciences) in 20×SSC overnight. After transfer, membranes were washed in 5×SSC and UV-crosslinked in a GS gene linker™ UV chamber (Bio-Rad). Blots were prehybridized in a solution containing 5×SSC, 5×Denhardt, 0.1% SDS and 100 μg/ml denatured salmon sperm at 65°C for 1–2 hours. A 1-kb DNA fragment within the SgIF4G coding region (nucleotides 1928-3001) was amplified by PCR using primers 5′-CTT CAA CCC TTC CAG AAA GTA AAG ACG G-3′ and 5′-GCC GCT CCT CTG TAT TAA TGG GAA GCT G-3′, and purified twice on GeneElute Agarose™ spin columns (Sigma). The purified PCR product (25 ng) was labeled in the presence of [α-32P]dCTP with Megaprime™ DNA labeling systems (Amersham Biosciences) according to the manufacturer's instructions. The prehybridization solution was replaced with fresh solution containing SgIF4G cDNA probe, and the filters were incubated overnight at 65°C. Filters were washed at 65°C in 0.1×SSC, 0.1% SDS. Radioactivity on blots was revealed by autoradiography on X-OMAT AR film (Kodak).
GST-recombinant proteins and antibodies directed against SgIF4G protein
A 900-bp fragment within the SgIF4G-coding region (amino acids 591-887), containing the eIF4E-binding site, was amplified using primers 5′-CGC GGA TCC GTG CAA CAG TCA ATG CCA G-3′ and 5′-CCG GAA TTC TCT GGA CGA TCT CCT ACC-3′. The fragment was inserted into pGEX-4T-1 vector digested with BamHI and EcoRI. The GST and the fused GST-SgIF4G and GST-mIF4E proteins were overexpressed in E. coli (strain BL21) and purified on a glutathione Sepharose 4B column according to the manufacturer's instructions (Amersham Pharmacia Biotech). Polyclonal antibodies against the recombinant GST-SgIF4G protein were obtained in guinea pigs by standard immunization protocol (Eurogentec). The pre-immune and immune sera were kept in aliquots at –20°C.
Analysis of in vitro interaction between SgIF4G and mIF4E
After production and purification, recombinant proteins were dialysed overnight in buffer A (50 mM HEPES pH 7.7, 150 mM KCl, 1 mM EDTA, 5% glycerol). One microgram of GST-mIF4E or GST alone was incubated for 1 hour with m7GTP-Sepharose beads in buffer A. After washing, the beads were incubated for 1 hour in buffer A containing 1 mg/ml of BSA, 0.5% IGEPAL (Sigma Aldrich) and peptides as indicated. GST-SgIF4G or GST as controls were then added and incubation was prolonged for a further 1 hour. After extensive washing, the beads were boiled in Laemmli buffer and analysed by western blotting using anti-GST antibodies (Santa Cruz Biotechnology) and a chemifluorescence detection system (ECF, Amersham Pharmacia Biotech).
Preparation of cell lysates
After fertilization, at the times indicated, cells were collected by centrifugation in a Heraeus Labofuge 4000 centrifuge for 2 minutes at 2000 g. The cell pellet was frozen in liquid nitrogen and stored at –80°C. Cells were lysed by passage through a 25G syringe in one cell volume of 2× binding buffer (40 mM HEPES pH 7.4, 100 mM sodium fluoride, 10 mM ATP, 20 mM tetrasodium pyrophosphate (PPi), 100 mM NaCl, 0.4 mM EDTA, 2 mM dithiothreitol, 1 mM AEBSF, and 20 μg/ml of aprotinin and leupeptin). Cell lysates were centrifuged for 15 minutes at 16,000 g at 4°C in an Eppendorf centrifuge 5415R and the supernatants were stored at –20°C before use. Protein quantification was performed in duplicate by the Bradford assay.
Isolation of eIF4E and associated proteins
Isolation of eIF4E from egg extracts and its partners was performed using m7GTP beads as described previously (Salaün et al., 2004). Briefly, extracts were mixed with 25 μl of m7GTP-Sepharose beads (m7GTP column). After 60 minutes incubation at 4°C, the columns were washed three times with 1 ml of 1× binding buffer containing 100 mM NaCl. Laemmli sample buffer was added directly to the beads.
Immunoprecipitation of sea urchin eIF4G
For the affinity purification procedure, 500 μl of supernatant resulting from the 16,000 g centrifugation were incubated in batches with 2% BSA-presaturated immobilized protein A beads (50 μl packed beads) covalently coupled to eIF4G antibodies, in 1 ml of an IP buffer (50 mM Tris–HCl pH 7.6, 500 mM NaCl and 1% Nonidet P40) for 2 hours at 4°C. The beads were then washed three times in the IP buffer and twice in 500 mM NaCl and 50 mM Tris–HCl (pH 7.6). After washing, bound proteins were eluted with Laemmli buffer and resolved on a 10% acrylamide gel and analysed by western blot.
One microgram of GST-tagged mIF4E or GST alone were mixed and pre-incubated with 25 μl of gluthathione-Sepharose in a final volume of 200 μl of 1× binding buffer for 1 hour at 4°C. The beads were washed three times with 1× binding buffer. Five hundred microliters (2 mg of proteins) of supernatant resulting from the 16,000 g centrifugation were mixed and incubated end-over-end for 60 minutes at 4°C. Then the samples were washed three times with 1 ml of 1× binding buffer containing 100 mM NaCl. Laemmli sample buffer was added directly to the beads and the proteins were resolved by SDS-PAGE and analysed by western blotting.
Western blot analyses
Western blot analyses were performed following electrophoretic transfer of proteins from SDS-PAGE onto 0.22-μm nitrocellulose membranes (Towbin et al., 1979). 4E-BP was analysed using rabbit polyclonal antibodies directed against human 4E-BP2 (Rousseau et al., 1996). eIF4E was analysed using mouse monoclonal antibody directed against rabbit eIF4E. Sea urchin eIF4G was analysed using guinea pig polyclonal antibodies directed against SgIF4G.
Membranes were incubated with antibodies directed against 4E-BP2 (1:2000), eIF4E (1:2000), SgIF4G (1:2000) or actin (1:2000) in 20 mM Tris-HCl (pH 7.6), 5% skimmed milk and 0.1% Tween 20 at room temperature. The antigen-antibody complex was measured by chemiluminescence using horseradish peroxidase-coupled secondary antibodies according to the manufacturer's instructions (ECL; Amersham Pharmacia Biotech). Depending on the experiments, a chemifluorescence detection system (ECF, Amersham Pharmacia Biotech) was used for detection of antibody complexes.
We thank Nahum Sonenberg for the gift of 4E-BP2 antibody. We thank Simon Morley for the gift of the construct encoding the recombinant GST-eIF4E protein. We thank Mark Levingston for helpful comments and for manuscript corrections. We are grateful to the members of the cell cycle and development group for helpful discussions and comments on this manuscript. This work was supported by Association pour la Recherche contre la Cancer (ARC, France; grants 4247 and 3507) to P.C., Ligue Nationale contre le Cancer (Délégations Départementales Finistère, Morbihan, Vendée et Côte d'Armor) to P.C., Conseil Régional de Bretagne and Conseil Général du Finistère.