INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

During the first hours of embryonic development in most if not all animal species, transcription is effectively silent. Accordingly, the developing embryos are dependent on the stock of maternal gene products for control of their growth and subsequent differentiation. In Xenopus, Drosophila, and mouse oocytes and embryos the recruitment or release of maternal mRNAs into or from polysomes is correlated, respectively, with the lengthening or shortening of the poly(A) tail (see reference 22 for a review).

Several cis-acting RNA elements that control poly(A) tail length have been identified. The first such element to be characterized was the cytoplasmic polyadenylation element (CPE). This is a uridine-rich element present in the 3' untranslated regions (UTRs) of maternal Xenopus mRNAs that are polyadenylated during oocyte maturation (5, 12). mRNAs devoid of this element are deadenylated by a process called default deadenylation (6, 23).

Sequence elements similar to the Xenopus CPE have been identified in the 3' UTRs of maternal mouse mRNAs (8). In the case of mice it was shown that this CPE-like element directed not only the cytoplasmic polyadenylation during oocyte maturation but also the deadenylation of these same mRNAs in primary oocytes. Hence, this element was named an adenylation control element (ACE). The cytoplasmic polyadenylation directed by both CPEs and ACEs requires the AAUAAA nuclear polyadenylation signal (5, 8, 12). In addition, CPE-driven polyadenylation can be affected by auxiliary cis elements. For instance, the 3' UTRs of Cl1 and Cl2 Xenopus mRNAs contain silencing elements that repress the function of the CPE during oocyte maturation but not after fertilization (19, 20).

The Eg family of maternal mRNAs were originally identified as mRNAs that are poly(A)⁺ in the egg and deadenylated after fertilization (16). c-mos mRNA was later shown to present the same adenylation-deadenylation behavior (17). This postfertilization deadenylation of Xenopus Eg mRNAs is a transcript-specific process that requires sequence information present in the 3' UTRs of these mRNAs (3). The sequence elements that direct the postfertilization deadenylation of Xenopus Eg5 and c-mos mRNAs (embryo deadenylation elements, [EDENs]) have been identified (14). In addition, the 53-kDa protein (EDEN-BP) that binds to these elements has been shown to be a necessary component of the EDEN-dependent deadenylation process in embryos (14).

Independently, Stebbins-Boaz and Richter (21) have studied cis elements implicated in the control of Eg1-cdk2 mRNA poly(A) tail length in Xenopus oocytes and embryos. They localized two cis elements (5'cdk2 and 3'cdk2) in the 3' UTR of this mRNA that appear to act synergistically to cause the deadenylation of this mRNA in embryos but not in maturing oocytes. Although binding of EDEN-BP was detected for the full-length Eg1-cdk2 3' UTR (10, 21), no binding of EDEN-BP was detected for either the 5'cdk2 or the 3'cdk2 element (12a, 21). This implies that the deadenylation activity targeted by these elements is independent of EDEN-BP. However, previous unpublished data from this laboratory (12a) suggested that the (AUU)₃ motif present in the 3'cdk2 element could modulate the deadenylation kinetics of EDEN-containing RNAs. The present study was therefore performed to characterize this possible interplay between the two cis elements. We show here that this (AUU)₃ motif acts as an enhancer of EDEN-dependent deadenylation. This enhancer activity was observed for two distinct EDEN sequences, implying that it is a general effect.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Plasmid constructions. The generic structure of the different chimeric genes used is given in Fig. 1A, as is the sequence of the 3' UTR common to all of these chimeric genes.

View larger version (21K): [in this window] [in a new window]   FIG. 1.   Structure of chimeric genes and composition of RNAs. (A) Generic structure and sequence of the 3' UTR of the chimeric genes used in this study as transcription templates. Transcription was driven by the T7 promoter. The reporter sequence containing the globin open reading frame is indicated by a thick line; the position of the c-mos EDEN or of the n-fold UGUA repeats [(UGUA)n] is indicated by a boxed "E." The position of the (AUU)3 motif or mutations of this motif is represented by a boxed "U." The position of the AAUAAA nuclear polyadenylation signal is indicated by a narrow open box. In the 3' UTR sequence the globin translational stop codon and the nuclear polyadenylation signals are in bold type. The positions of the EcoRV (Ev) and BamHI (B) restriction sites used to make the transcription templates for RNAs with or without the (AUU)3 motif are also shown and underlined in the sequence. p(A), poly(A) tail added to the RNA by E. coli PAP after transcription. (B) For each chimeric gene, the restriction enzymes used and the corresponding RNAs are indicated. The nature of the sequence at positions E and U is also indicated. The sequence of the c-mos EDEN is 5'TATATGTATGTGTTGTTTTATGTGTGTGTGTGTGCT3'.

RNA synthesis. ³²P-labeled and capped transcripts were made in vitro with the Promega Riboprobe transcription kit as previously described (10). The transcription templates were linearized by restriction with the enzymes indicated in Fig. 1B. In vitro polyadenylation was achieved by incubating (5 min at 37°C) 1 to 5 µg of in vitro-synthesized RNA with 3 U of Escherichia coli poly(A) polymerase (PAP; Pharmacia) in a 50-µl reaction volume (50 mM Tris-HCl, pH 8.0; 250 mM NaCl; 10 mM MgCl₂; 1 mM MnCl₂; 2 mM dithiothreitol; 50 ng of bovine serum albumin per µl; 0.5 mM ATP; and 1.6 U of RNasin per µl). The reaction was stopped by adding 50 µl of 100 mM Tris-HCl (pH 7.5)-300 mM NaCl-10 mM EDTA-2% sodium dodecyl sulfate (SDS). The RNA was phenol-chloroform extracted twice and ethanol precipitated. The polyadenylated radiolabeled RNA was purified by electrophoresis on 5% denaturing polyacrylamide gels as described previously (1). RNA size markers were synthesized as described by Audic et al. (2).

Biological and analytical methods. Embryos were obtained by standard procedures and incubated in F1 solution (15). Embryos at the two-cell stage were injected with 18.4 nl of in vitro transcript (about 1,000 cpm; 0.25 to 1 fmol of RNA). The embryos were allowed to develop at 22°C. At various times, five-embryo lots were taken, and the total RNA was extracted as described by Harland and Misher (7). When required for digestion with RNase H, the RNA samples were resuspended in water and treated as described by Osborne et al. (13). The samples of total RNA or the RNase H-treated samples were dissolved in 50% formamide and analyzed by electrophoresis on 4% polyacrylamide-urea gels as described by Audic et al. (1). The proportion of deadenylated RNA was quantified by the ImageQuant program (Molecular Dynamics) to treat numerical data produced by scanning multiple autogradiograms. The amount of poly(A) RNA, the final deadenylation product, was taken as the signal within a window between A₀ and A₂₀. The total amount of RNA was considered the summation, in each lane, between A₀ and the top of the signal in the 0-h lane.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

To analyze the effect of the (AUU)₃ motif on EDEN-dependent deadenylation a chimeric gene was constructed with the Xenopus laevis -globin coding sequence and a 3' UTR that contained the c-mos EDEN followed by the nuclear polyadenylation signal (AATAAA) and the (ATT)₃ motif. An EcoRV restriction site was placed just 3' to the (ATT)₃ motif, and a BamHI site was placed 5' to this motif. Templates for in vitro transcription that did or did not contain this motif were generated by restriction of this plasmid with either EcoRV or BamHI, respectively (Fig. 1B).

The capped ³²P-labeled RNAs transcribed in vitro from these templates were polyadenylated by E. coli PAP and injected into two-cell embryos. Samples of five embryos were taken immediately after injection (0 h) and then at 1, 2, and 4 h after injection. Total RNA was extracted from these samples and analyzed by electrophoresis on polyacrylamide-urea gels followed by autoradiography. The results of such an experiment are shown in Fig. 2. The Gb/EDENm/AUU mRNA was deadenylated after injection (Fig. 2A, lanes 1 to 4) with kinetics very similar to that previously observed for the GbORF/mosEDEN chimeric mRNA that contains the same 3' UTR (14). The injected RNA molecules were deadenylated in an asynchronous manner, with a coexistence at the 1-h point of the final deadenylated product and of molecules that migrated at the position of the injected poly(A)⁺ RNAs. For the RNAs devoid of the (AUU)₃ motif (Fig. 2A, lanes 5 to 8) the deadenylation appeared to be less efficient. At the 1-h point a smaller proportion of these injected RNAs had been shortened to the size of the final deadenylation product (Fig. 2A; compare lanes 2 and 6), and at the 2-h point the signal from RNAs of intermediate size (partially deadenylated molecules) was greater for the RNAs devoid of the (AUU)₃ motif (compare lanes 3 and 7). This difference between the rates of deadenylation of the RNAs containing and devoid of the (AUU)₃ motif was confirmed by quantification of these data (Fig. 2C). The presence of the (AUU)₃ motif increased the rate at which an RNA containing the c-mos EDEN was deadenylated about 2.5-fold.

View larger version (34K): [in this window] [in a new window]   FIG. 2.   Effect of the (AUU)3 motif on c-mos EDEN-dependent deadenylation. (A) Between 0.15 and 0.3 fmol of the 32P-labeled and capped Gb/EDENm/AUU and Gb/EDENm RNAs were injected as polyadenylated transcripts into two-cell embryos as described in Materials and Methods. Samples of five embryos were taken for RNA extraction at the indicated times after injection, and aliquots of 2.5-embryo equivalents were analyzed in a 5% polyacrylamide-urea gel. After being dried the gel was subjected to autoradiography. Lane A0, nonadenylated RNAs before injection; lane M, RNA molecular weight markers; sizes are indicated on the left. (B) The chimeric RNAs devoid of an EDEN sequence (Gb/EDEN and Gb/EDEN/AUU) were synthesized as 32P-labeled, capped transcripts and polyadenylated in vitro prior to injection into two-cell embryos. As a control the polyadenylated Gb/EDENm/AUU RNA was similarly prepared and injected into two-cell embryos. At the indicated times after injection aliquots of five embryos were taken and further processed as described for panel A. Lanes A0 and M, same as for panel A. (C) Quantification of the proportion of poly(A) RNA produced 4 h after injection into two-cell embryos. For the chimeric transcripts used for panels A and B, the amount of poly(A) RNA, the final deadenylation product, relative to total RNA in each lane was calculated as described in Materials and Methods. (D) Effect of the (AUU)3 motif on the UV cross-linking of EDEN-BP to the c-mos EDEN. RNAs containing (lanes 2 and 4) or devoid of (lanes 1 and 3) the c-mos EDEN and with (lanes 3 and 4) or without (lanes 1 and 2) the (AUU)3 motif were incubated in activated egg extracts. After irradiation with UV light and treatment with RNase A, the proteins were resolved on a 10% polyacrylamide gel in the presence of SDS. Radiolabeled proteins were revealed by autoradiography of the dried gel.

A close examination of the data in Fig. 2A indicates that the size of the in vivo-deadenylated RNAs is slightly greater than that of the in vitro-synthesized A₀ transcripts, which implies that in vivo the final deadenylated product retains a short oligo(A) tail. This is similar to the situation described for several yeast mRNAs (see reference 9 for a review) and for mRNAs containing c-fos destabilizing elements expressed in NIH 3T3 cells (18). In these cases the poly(A) tail is shortened to an oligo(A) tail before the body of the mRNA is degraded. In primary mouse oocytes the deadenylation of maternal mRNAs driven by the ACE also leaves an oligo(A) tail (8). With the RNA size markers and the nonadenylated in vitro transcript, it was estimated that, for the RNAs deadenylated via the EDEN-dependent pathway in Xenopus embryos, the oligo(A) tail contained 15 to 20 A's (an average of five independent experiments). Hereafter we refer to the final product of in vivo deadenylation as poly(A) RNA and the nonadenylated in vitro transcript as A₀ RNA.

To ensure that the (AUU)₃ motif stimulated EDEN-dependent deadenylation, RNAs with or without the (AUU)₃ motif and from which the EDENm element had been deleted (Gb/EDEN/AUU and Gb/EDEN, respectively) were injected as ³²P-labeled, capped transcripts into two-cell embryos after in vitro polyadenylation. These RNAs did not bind EDEN-BP (Fig. 2D, lanes 1 and 3). As a positive control of EDEN-dependent deadenylation in this experiment the in vitro-polyadenylated Gb/EDENm/AUU RNA was also injected. Total RNA was extracted from these injected embryos and analyzed (Fig. 2B) as described above.

The Gb/EDEN RNA (Fig. 2B, lanes 1 to 4) and the Gb/EDEN/AUU RNA (lanes 5 to 8) were deadenylated synchronously, which is indicative of default deadenylation (14). In comparison, the Gb/EDENm/AUU RNA (Fig. 2B, lanes 9 to 12) showed the expected asynchronous deadenylation with the coexistence of the initial poly(A)⁺ RNA with the final deadenylated product. Quantification of the amounts of poly(A) Gb/EDEN and Gb/EDEN/AUU RNAs at the 4-h point, relative to the total amount of these RNAs at this time (Fig. 2C), showed that the (AUU)₃ motif did not increase the rate of deadenylation in the absence of an EDEN.

One explanation for the observed increase in EDEN-dependent deadenylation when the (AUU)₃ motif was present in the RNA is that this motif causes an increase in EDEN-BP binding. To test this possibility, UV-cross-linking experiments were performed with the RNAs containing or devoid of an EDEN and with or without the (AUU)₃ motif. The data (Fig. 2D) show that the EDEN-BP cross-linking signal, observed only with the EDEN-containing RNAs (lanes 2 and 4), was not significantly affected by the presence of the (AUU)₃ motif.

The natural EDENs characterized to date are not strictly conserved sequence motifs; the only identified consensus is a repetition of UR dinucleotides. Hence, the effect of the (AUU)₃ motif described above could eventually be attributed to a particular interaction or synergy between the c-mos EDEN and the (AUU)₃ motif. Therefore, to further characterize the effect of the (AUU)₃ motif on EDEN-dependent deadenylation we decided to make a synthetic EDEN based on a repetition of the simple UGUA motif. These synthetic EDENs were embedded into the 3' UTR of -globin mRNA.

Cross-linking of EDEN-BP to UGUA repeats. EDEN-BP is a necessary component of EDEN-dependent deadenylation (14). Therefore, binding of EDEN-BP is a primary prerequisite of a synthetic EDEN. Initial UV-cross-linking experiments using radiolabeled RNAs [Gb/(UGUA)_n], whose 3' UTRs contained two, four, six, or nine repetitions of the UGUA motif, showed that a ³²P-labeled protein of about 53 kDa was detected only with RNAs containing four or more UGUA repeats (data not shown).

View larger version (52K): [in this window] [in a new window]   FIG. 3.   Detection by UV cross-linking of EDEN-BP binding to RNAs containing UGUA repeats. 32P-labeled Gb/(UGUA)n RNAs were incubated in activated egg extracts that had been treated with preimmune serum (mock depleted) or anti-EDEN-BP serum (immunodepleted). After irradiation with UV light the samples were processed as described in the legend to Fig. 2D. The numbers (n) of UGUA repeats in the different chimeric RNAs are indicated above the lanes. Lane M, molecular mass markers; sizes (in kilodaltons) are indicated between the two panels.

UGUA repeats constitute a functional EDEN whose efficiency is enhanced by the (AUU)₃-containing element. The putative EDEN (UGUA)₉ is the same length as the c-mos EDEN and binds EDEN-BP proficiently. Therefore, RNAs containing this motif in the 3' UTR and with [Gb/(UGUA)₉/AUU] or without [Gb/(UGUA)₉] the (AUU)₃ motif (Fig. 1) were synthesized as ³²P-labeled, capped molecules. A poly(A) tail was added to these RNAs by in vitro polyadenylation with PAP prior to injection into two-cell embryos. Total RNA was extracted from these embryos at 0, 1, 2, and 4 h postinjection and analyzed by electrophoresis on polyacrylamide-urea gels followed by autoradiography. In the experiment whose results are shown in Fig. 4A, the Gb/(UGUA)₉ RNA (lanes 1 to 4) was deadenylated efficiently; poly(A) RNA was already detected at the 1-h point (lane 2). In the presence of the (AUU)₃ motif (Fig. 4A, lanes 5 to 8) the initial rate of deadenylation was increased; however, at the 4-h point the amounts of deadenylated product for these two RNAs appeared to be similar. Quantification of these data showed that for both of these RNAs the proportion of the deadenylated product, relative to the total amount of RNA, at the 4-h point was 35% ± 5% (mean ± standard deviation).

View larger version (34K): [in this window] [in a new window]   FIG. 4.   The (AUU)3 motif can stimulate EDEN-dependent deadenylation. (A) The Gb/(UGUA)9 and Gb/(UGUA)9/AUU RNAs were synthesized as 32P-labeled, capped transcripts and then polyadenylated in vitro with E. coli PAP. Between 0.15 and 0.3 fmol of these RNAs was injected into two-cell embryos. At the indicated times after injection, aliquots of five embryos were taken for RNA extraction and analysis as described in the legend to Fig. 2. Lanes A0, nonadenylated RNAs before injection. (B) Samples of five embryos injected with the radiolabeled, capped polyadenylated Gb/(UGUA)9 RNA were taken for RNA extraction at the indicated times after injection. Half of each sample was treated with RNase H in the presence (+) or absence () of oligo(dT). The samples were then analyzed as described in the legend to Fig. 2. (C) The three different chimeric RNAs, Gb/(UGUA)6, Gb/(UGUA)6/AUU, and Gb/(UGUA)6/N9, that all contained six UGUA repeats in the 3' UTR were synthesized and polyadenylated in vitro. Two-cell embryos injected with these 32P-labeled RNAs were further processed as described for panel A. In Gb/(UGUA)6/AUU the (AUU)3 motif was present between the AAUAAA signal and the poly(A) tail. In Gb/(UGUA)6/N9 the (AUU)3 motif was replaced by an unrelated sequence, and in Gb/(UGUA)6 the motif was absent. Lanes A0, nonadenylated RNAs before injection.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

In the work presented here we have shown that an (AUU)₃ motif can increase the rate of deadenylation [formation of poly(A) RNA] targeted by an EDEN. This effect was observed for both a natural EDEN (c-mos EDEN) and two synthetic EDENs, composed of UGUA repeats, which have graded efficiencies.

The (AUU)₃ motif was originally part of the 3'cdk2 element described by Stebbins-Boaz and Richter (21). In the context of the chimeric RNA used by these authors (5'cdk2/B4/3'ckd2) the 5'cdk2 and 3'cdk2 elements act synergistically to confer deadenylation. These elements are not related to the EDEN either in sequence or in protein binding capacity. This suggests that the (AUU)₃ motif may act as a general enhancer of sequence-specific deadenylation in embryos.

Although the 5'cdk2 and 3'cdk2 elements do not bind EDEN-BP, the complete Eg1-cdk2 3' UTR does contain an EDEN-BP binding site (see the introduction). Hence, this mRNA is a potential substrate for EDEN-dependent deadenylation. The data presented here show that if this were the case then the 3'cdk2 element would work in synergy not only with the 5'cdk2 element but also with the EDEN. In essence this implies that the Eg1-cdk2 mRNA would be a substrate for two distinct deadenylation pathways. Since the Eg1-cdk2 mRNA encodes an important cell cycle-regulatory protein (for a review, see reference 4), multiple deadenylation pathways may have evolved to ensure its translational silencing after fertilization.

Maternal mRNAs that are deadenylated in embryos contain a CPE that ensures their poly(A)⁺ status in the laid egg. Therefore, one possible way to enhance the rate of deadenylation in embryos would be to inhibit CPE-dependent polyadenylation. However, when a chimeric RNA that contained the Eg5 3' UTR and the cloned 65-adenosine tail with the associated (AUU)₃ motif but with its EDEN deleted, was injected into two-cell Xenopus embryos it was further polyadenylated (14). Hence, in this context at least, the (AUU)₃ motif does not prevent polyadenylation. It is not clear at present how the (AUU)₃ motif exerts its effect on EDEN-dependent deadenylation. There is no complementary sequence in the -globin 3' UTR that could base-pair with the (AUU)₃ motif and thereby form a secondary structure. In addition, no factor that associates with this motif could be detected by UV cross-linking (unpublished data). Therefore, the mechanism by which the (AUU)₃ motif enhances deadenylation remains to be determined.

Of the synthetic EDENs used during this work, the one containing nine UGUA repeats was the most efficient in directing deadenylation. The element with six UGUA repeats had an intermediate deadenylation activity. This effect of the number of UGUA repeats on EDEN-dependent deadenylation activity is suggestive of a correlation between the affinity of EDEN-BP binding and the size (length) of the target sequence. Increased affinity would probably lead to an increased temporal stability of the deadenylation complex, which in turn could cause an increase in deadenylation activity.

No binding of EDEN-BP was detected when only two UGUA repeats (four contiguous UR dinucleotides) were present (Fig. 3, lane 2). For the Eg5 EDEN, mutants that retained six contiguous UR dinucleotides bound EDEN-BP (14). From this it could be implied that an EDEN-BP binding site should contain at least five contiguous UR dinucleotides. However, in at least one natural EDEN this does not appear to be the case. The 3' UTR of Eg2 mRNA does not contain more than four contiguous UR dinucleotides. Hence, it may be the density of UR dinucleotides rather than the number of contiguous repeats that is the important parameter in defining an EDEN. The recent cloning of EDEN-BP will allow us to use a SELEX approach to better define the constraints for EDEN-BP binding.

Finally, the data presented here show that the deadenylation of maternal mRNAs in Xenopus embryos can be controlled by multiple cis elements. The Eg mRNAs contain both a CPE and an EDEN, and although the EDEN does not act simply as an inhibitor of CPE function (11), it is conceivable that there could be a competition between these two antagonistic functions. Hence, the sequential changes in poly(A) tail length observed for the maternal Eg mRNAs in vivo will probably be defined by the net result of a cross talk between the various cis elements present in a given mRNA.