| Nucleic Acids Research | Pages |
Dissociation of mRNA cytoplasmic polyadenylation from translational activation by structural modification of the 5[prime]-UTR
Introduction
Materials And Methods
Fly strains
DNA templates for in vitro transcription
In vitro transcription
In vitro translation and western blot analysis
Embryo RNA injections, recovery and analysis
Results
Creating structural modification of the 5[prime]-UTR of bcd mRNA
bcd-AS RNA is poly(A) elongated to wild-type levels
bcd-AS RNA is efficiently translated in vitro but not in vivo
Discussion
Acknowledgements
References
Dissociation of mRNA cytoplasmic polyadenylation from translational activation by structural modification of the 5[prime]-UTR
Received June 11, 1999; Accepted July 8, 1999
ABSTRACT During early metazoan development, certain maternal mRNAs are translationally activated by elongation of their poly(A) tails. Bicoid (bcd) mRNA is a Drosophila maternal mRNA that is translationally activated by cytoplasmic polyadenylation during the first hour after egg deposition. The sequences necessary and sufficient to promote its poly(A) elongation, and hence translation, are contained within its 3[prime]-untranslated region (UTR). The mechanism by which poly(A) elongation at the 3[prime]-end affects translational initiation at the 5[prime]-end remains unknown. To investigate this question, we have analyzed a bicoid mRNA whose 5[prime]-UTR contains a short antisense sequence directed against a portion of the coding region. This mutated RNA is efficiently translated in vitro. After injection into Drosophila embryos, this RNA is stable and polyadenylated, but inefficiently translated. These experiments show that structural modification of the 5[prime]-end of an mRNA can perturb the translational activation normally conferred by polyadenylation in vivo.
INTRODUCTION
A general principle of animal development is that late oocytes and early embryos are transcriptionally quiescent. Therefore, changes in gene expression during this crucial time must rely on post-transcriptional mechanisms. A prevalent form of such regulation is translational control of maternal mRNAs; this control includes both activation as well as silencing of mRNAs at precise times and locations in development (1,2).
One form of translational regulation in oocytes and embryos involves changes in mRNA poly(A) tail length. The general scheme established from work on many species is as follows (3-5): (i) transcripts destined to become dormant are synthesized in the nucleus and acquire a normal length poly(A) tail; (ii) once the transcripts are transported to the cytoplasm, they undergo cytoplasmic deadenylation that shortens their poly(A) tails to 20-40 nt and silences their translation; (iii) during oocyte maturation or early development (depending on the species and specific mRNA), the dormant mRNAs are cytoplasmically polyadenylated and become translationally active; (iv) in many cases, the mRNAs become unstable after polyadenylation and are non-specifically deadenylated and degraded. Not all of these features have been established with each species tested, but in broad terms this scheme has been implicated in translational regulation in Caenorhabditis elegans, Dictyostelium, Drosophila, mouse, sea urchin, Spisula, starfish and Xenopus.
The effect of polyadenylation on translation has been studied in various systems (6). In Xenopus and mouse oocytes, dormant messages are monosomic and become associated with polysomes after polyadenylation (7), and in yeast, the presence of poly(A) binding protein (Pab1p) on an mRNA is required for translational initiation (8). These results raise the question of how a modification of the 3[prime]-end influences initiation at the 5[prime]-end (9-11). One mechanism for stimulation is that the Pab1p-poly(A) tail complex recruits the 40S ribosomal subunit to the mRNA (12). In addition, the Pab1p-poly(A) tail complex can interact with the 5[prime] cap structure (m7GpppN) by associating with the translation initiation factor eIF4G, a component of the 40S ribosomal subunit (13). Another intriguing result linking polyadenylation and the 5[prime]-end of mRNAs is that poly(A) addition induces 5[prime] cap ribose methylation; it is possible that this modification plays a major role in increasing translational efficiency (14). In Drosophila, localization-dependent translation of oskar (osk) mRNA requires functional communication between specific elements in its 5[prime]- and 3[prime]-untranslated regions (UTRs) (15). However, translational activation of osk mRNA does not depend upon lengthening of the poly(A) tail (16).
Many of these experiments are consistent with a closed-loop model of translational stimulation by poly(A) based on a circular mRNA that facilitates interactions between the 5[prime]- and 3[prime]-ends (17). Strong evidence supporting this model is the recent visualization, using atomic force microscopy, of circularized, capped, polyadenylated RNA in the presence of an eIF4E (the cap binding protein)/eIF4G/Pab1p complex (18).
Bicoid mRNA is localized to the anterior end of Drosophila embryos and gives rise to an anterior to posterior bicoid protein gradient that directs formation of anterior structures (19,20). During the first hour after egg deposition, bcd mRNA is elongated by about 150 A residues and this poly(A) elongation results in translational activation (16,21). In order to better define the relationship between polyadenylation/translation and RNA structure, we have introduced into the 5[prime]-UTR of bcd mRNA a short antisense sequence directed against a C-terminal portion of the coding region. This RNA is efficiently translated in vitro and is stable and fully polyadenylated upon injection into Drosophila embryos. However, its in vivo translation is inefficient. This mutated RNA uncouples poly(A) elongation from in vivo translational activation, making it a valuable tool for studying the relationship between polyadenylation/translation and RNA structure.
MATERIALS AND METHODS
Fly strains
Canton-S (named wild-type for this paper) and nosBN (22) fly strains were used. Egg collection chambers contained Canton-S or nosBN homozygous females and Canton-S males. The nosBN chromosome was balanced over TM3. Canton-S flies were obtained from Peter Gergen and nosBN flies from Ruth Lehmann and Robin Wharton.
DNA templates for in vitro transcription
The bcd wild-type (bcd-WT) transcript was prepared from pSK-bcdWT. This plasmid was generated by subcloning an ~2.5 kb XhoI-NotI (New England Biolabs/Boehringer Mannheim) fragment, containing the full-length bcd mRNA, excised from pBCDWT (16), into pBluescript II SK (Stratagene).
The bcd antisense (bcd-AS) transcript was derived from pSK-bcdAS. This plasmid was generated from pGEM4-bcd (23) by PCR amplifying a fragment of ~550 nt using bcd-XhoI (5[prime]-CCCTCGAGTTGCCCGCGGCGTTCCG) and bcd6 (5[prime]-GTT-TCGCTGACAGATCCGC) primers. This PCR fragment was digested with XhoI and PstI (New England Biolabs/Boehringer Mannheim) and cloned into pSK-bcdWT to replace the 5[prime]-UTR and initial portion of the bcd coding sequence. The above PCR fragment contained a 63 nt antisense sequence (5[prime]-TT-GCCCGCGGCGTTCCGATGGGGATTATACGCTTGCATTATCGTATCCGTCGTGCATTGATAT) directed against a C-terminal portion of the bcd coding region (nt 1549-1611 of the published cDNA sequence; 24).
One nanogram of pGEM4-bcd was amplified in a 50 µl reaction in the presence of 5 µl of 10× buffer from Perkin-Elmer Cetus, 0.3 mM dNTPs, 25 pmol of each primer and 2.5 U of Taq DNA polymerase. The PCR conditions were as follows: 5 min at 93°C; 30 cycles of 1 min at 93°C, 1 min at 60°C, 1 min 30 s at 72°C; followed by a 7 min extension at 72°C.
pSK-bcdAS was sequenced as described (25), using Sequenase from US Biochemical and primers bcd5 (5[prime]-GGAGTGTTTGGGGAAAATGG), bcd6, bcd8 (5[prime]-GCTCCCGTTCGCTCATCG) and bcd10 (5[prime]-CGCCGCGTTATCTTTACG), to confirm fidelity of PCR amplification.
In vitro transcription
pSK-bcdWT and pSK-bcdAS were linearized using NotI (New England Biolabs/Boehringer Mannheim), phenol/chloroform extracted and ethanol precipitated in the presence of 0.3 M sodium acetate. Approximately 1 µg of each linearized template was used in a 50 µl reaction containing 1× Epicentre Technologies transcription buffer, 10 mM DTT, 0.5 mM ATP, 0.5 mM CTP, 0.05 mM GTP, 0.012 mM UTP, 0.5 mM CAP analog (m7GpppG; Ambion), 80 U RNasin (Promega), 100 µCi [[alpha]-32P]UTP (3000 mCi/mmol; Du Pont NEN) and T7 RNA polymerase (Epicentre Technologies). The RNAs were used at a specific activity of 2 × 107-4 × 107 c.p.m./µg (~200 ng/µl).
A T7 RNA polymerase transcript generated from linearized pSK-bcdWT contained 31 nt of polylinker, bcd mRNA (nt 1-2456), a 16 nt poly(A) tail and NotI sequence. A T7 RNA polymerase transcript generated from linearized pSK-bcdAS contained 31 nt of polylinker, the 63 nt antisense region, bcd mRNA (nt 11-2456), a 16 nt poly(A) tail and NotI sequence.
In vitro translation and western blot analysis
Cold in vitro translation of bcd-WT and bcd-AS RNAs was performed using the TNT Coupled Wheat Germ Extract and the TNT Coupled Reticulocyte Lysate (data not shown) Systems (Promega) in the presence of 1 µg of plasmid template and TNT T7 RNA polymerase. The reactions were performed as described by the manufacturer. Fifty percent of each reaction was diluted 1:1 into 2× SDS sample buffer and boiled for 3 min before loading on a 10% SDS-PAGE gel. After overnight transfer at 4°C, 100 mA, western blot analysis was performed as described (25) using an anti-bcd rabbit polyclonal antibody (a gift from Gary Struhl).
Embryo RNA injections, recovery and analysis
For the polyadenylation assay (Fig. 2), wild-type embryos (0-0.5 h old) were collected, injected anteriorly at room temperature, allowed to develop for 1 h and harvested to recover the injected RNA as described (26). RNAs were analyzed by electrophoresis through a 0.8% agarose/6% formaldehyde gel, blotted and visualized by autoradiography. The equivalent of 20 embryos was loaded on a single lane.
For the translational assay (Fig. 4), embryos (0-0.5 h old) derived from nosBN/nosBN females were collected, injected posteriorly at room temperature, allowed to develop for 24 h, fixed with Hoyer's medium/lactic acid (1:1), baked at 60°C for 4 h and visualized using dark-field microscopy.
For immunohistochemical detection of protein and whole mount in situ hybridization (Fig. 5), embryos (0-0.5 h old) derived from homozygous nosBN females were collected, injected posteriorly, allowed to develop at 25°C for 1 h, fixed and manually devitellinized (27). Immunostaining was performed (21) using the same anti-bcd rabbit polyclonal antibody utilized for the western blot analysis. Whole mount in situ hybridization was performed (21) using a digoxigenin-labeled (Boehringer Mannheim Biochemicals) bcd probe. As template, a PstI-BglII bcd fragment (nt 503-1353 of the published cDNA sequence; 24) was used. Embryos were visualized using Nomarski optics.
RESULTS
Creating structural modification of the 5[prime]-UTR of bcd mRNA
In order to study the relationship between RNA structure and cytoplasmic polyadenylation/translation, we created a mutated version of the maternal mRNA bicoid, referred to as bcd-AS. This RNA has a wild-type 3[prime]-UTR, known to contain the cis-acting sequences for poly(A) elongation (16,28), and coding sequence, but the 5[prime]-UTR is altered by the insertion of a 63 nt antisense sequence (Materials and Methods) directed against a portion of the bcd open reading frame (nt 1549-1611 of the published cDNA sequence; 24). It is probable that the RNA secondary structure is modified by the presence of a sense-antisense double-stranded RNA structure, as suggested by computer models using the mfold program (29,30; an earlier version of this program has been previously used to fold the 3[prime]-UTR of bcd mRNA; 31). Figure 1 shows the predicted secondary structures of bcd-WT and bcd-AS RNAs. The overall foldings are very similar with the exception that bcd-AS RNA has a double-stranded structure due to annealing between the sense and the antisense 63 nt sequences. In contrast, bcd-WT RNA has several stem-loops in the corresponding region. To further investigate the ability of this mutated RNA to form a stable sense-antisense structure we have performed RNase protection assays using high specific activity, in vitro transcribed bcd-AS RNA (data not shown). The results obtained suggested that the RNA contained an RNase-resistant short double-stranded sequence.
Figure 1. Predicted secondary structures of bcd-WT and bcd-AS RNAs. The potential secondary structures of (A) bcd-WT and (B) bcd-AS RNAs were determined using the mfold version 3.0 program for RNA folding by Zuker and Turner (available through the internet at http://www.ibc.wustl.edu/~zuker/rna/ ). A temperature of 25°C was set as a parameter, because all embryo injections were performed at room temperature. The structures with the lowest free energy are shown. The formation of a double-stranded structure, due to the presence of the 63 nt antisense sequence, is pinpointed. Note also the overall structural similarities, including folding of the 3[prime]-UTRs.
Figure 2. Polyadenylation status of injected bcd-WT and bcd-AS RNAs. The RNAs were injected into the anterior end of 0-0.5 h old wild-type embryos. The transcripts were recovered 1 h after injection (lanes I) and subjected to electrophoresis using uninjected RNAs as size markers (lanes U) to determine levels of polyadenylation in vivo. Both RNAs are poly(A) elongated with ~150 A residues, a level of polyadenylation previously reported for endogenous bcd mRNA (17).
Due to the structural modification at the 5[prime]-end of the RNA, coupled with the wild-type 3[prime]-UTR, we reasoned that analysis of the behavior of bcd-AS RNA might be revealing as to the relationship between the structure of the RNA and polyadenylation/translational activation. RNA injection experiments into Drosophila embryos have allowed us to test these biological RNA features in vivo.
bcd-AS RNA is poly(A) elongated to wild-type levels
bcd-WT mRNA is elongated with ~150 A residues upon injection into the anterior end of Drosophila wild-type embryos (16,26). This level of polyadenylation promotes efficient translation. To determine if the 5[prime] antisense sequence would disrupt polyadenylation or stability of the mRNA, we injected either bcd-WT or bcd-AS RNA into the anterior end of 0-0.5 h old wild-type embryos. As shown in Figure 2, bcd-AS RNA is polyadenylated to wild-type levels and stable. This experiment demonstrates that the short antisense sequence is not interfering with proper cytoplasmic polyadenylation or stability of the injected RNA. This result was expected because the 3[prime]-UTR sequences that normally direct poly(A) elongation (28) or stability (32) of bcd mRNA have not been deleted. In addition, these injections demonstrate that endonucleolytic cleavage of the mutated RNA, due to the possible sense-antisense double-stranded structure, does not occur in vivo.
bcd-AS RNA is efficiently translated in vitro but not in vivo
The presence of different types of structural modification in the 5[prime]-UTR of an RNA can alter its translational capabilities by impeding RNA accessibility to the ribosomal subunits (33). In addition, intramolecular duplex structures involving the 5[prime]-UTR can inhibit translation by preventing the assembly of a functional initiation complex (34). The basic translatability of an RNA can be tested using the wheat germ extract or rabbit reticulocyte lysate systems.
In vitro translation systems do not require polyadenylation of the 3[prime]-end or capping of the 5[prime]-UTR of the RNA; however, the presence of the cap structure or a poly(A) tail can improve translational efficiency. In contrast, in vivo both the cap structure at the 5[prime]-end and the poly(A) tail at the 3[prime]-end are, in some cases, necessary, acting synergistically as enhancers of translation (35).
The novel secondary/tertiary structure of bcd-AS RNA could impede the RNA accessibility to the translational machinery or prevent interaction of the poly(A) tail with the 5[prime]-end. In order to investigate these possibilities, we checked the level of translation of bcd-AS RNA compared to bcd-WT RNA both in vitro and in vivo.
First, we performed in vitro translations under non-competitive conditions (nuclease-treated extracts) and verified the specific production of bcd protein by western blot, using an anti-bcd polyclonal antibody (Fig. 3). bcd-AS and bcd-WT RNAs were translated in vitro to comparable levels (lanes 3 and 4), demonstrating that both RNAs were accessible to the ribosomes and the translational machinery. The same experiment could not be performed under competitive conditions (non-nuclease-treated extracts), as previously done for yeast mRNAs, because in this case an immunoprecipitation reaction is required (36) and the available anti-bcd antibody does not work in such experiments (data not shown).
Figure 3. In vitro translations/western blot analysis. Cold in vitro translations were performed using the TNT Coupled Wheat Germ Extract System, followed by western blot analysis using an anti-bcd polyclonal antibody. Similar results were obtained using the TNT Coupled Reticulocyte Lysate System (data not shown). Lane 1, the reaction was performed in the absence of DNA template; lane 2, a luciferase-containing plasmid was used as template; lanes 3 and 4, pSK-bcdAS and pSK-bcdWT, respectively, were used. Similar levels of bcd protein (indicated by an arrow) are detected in lanes 3 and 4 of the western blot.
Figure 4. In vivo translational assay by RNA injection into the posterior end of embryos derived from nosBN homozygous females. (A) The graphics report the number of injected embryos showing formation of anterior structures (levels of translation ranged from suppression of the posterior filzkörper to induction of a complete head skeleton). (B) Cuticle preparations of embryos derived from nosBN/nosBN females and injected into the posterior end with bcd-WT or bcd-AS RNA. (Left) The arrow indicates the presence of head skeleton structures in the posterior end of an embryo injected with bcd-WT RNA, demonstrating translation of the injected RNA. (Right) The arrow indicates the presence of filzkörper in the posterior end of an embryo injected with bcd-AS RNA, demonstrating inefficient translation of the injected RNA. The morphology of this embryo is identical to an uninjected embryo of the same genotype. Dark-field microscopy was used for these pictures. These embryos are oriented with the anterior to the left.
Figure 5. Immunohistochemistry and whole mount in situ hybridization of injected embryos derived from nosBN/nosBN females. (Top) Upon injection at the posterior end, bcd-WT RNA, but not bcd-AS RNA, is efficiently translated, as determined by immunostaining. Anteriorly, similar levels of endogenous protein are detected. (Bottom) Both bcd-WT and bcd-AS RNAs are localized to the site of injection at the posterior end, as determined by whole mount in situ hybridization. Anteriorly, comparable levels of endogenous bcd RNA are detected. Nomarski optics were used for these pictures. These embryos are oriented with the anterior to the left.
Second, we injected the same two RNAs into the posterior end of nanos-deficient (nos-) embryos, which can be used as a sensitive in vivo translational assay (26). For our purposes, the posterior end of nos- embryos is equivalent to the anterior end of bcd-deficient (bcd-) embryos, since in the absence of nos protein polyadenylated bcd-WT RNA is translated upon injection (26). bcd-WT and bcd-AS RNAs were injected into the posterior end of 0-0.5 h old nos- embryos and translation was assayed by formation of anterior structures. bcd-WT RNA promoted formation of anterior structures as expected, but bcd-AS did not (Fig. 4A), indicating that its translation in vivo was inefficient. The morphologies of bcd-WT-injected (anterior structures formed in the posterior) and bcd-AS-injected (no anterior structures formed in the posterior) embryos are shown in Figure 4B.
Third, we directly assayed translation in vivo by immunostaining nos-deficient embryos injected posteriorly with either bcd-WT or bcd-AS RNA. Upon injection, only bcd-WT RNA induced specific staining of the posterior end of embryos, demonstrating that bcd-AS RNA was not efficiently translated in vivo (Fig. 5, top). Furthermore, lack of induction of anterior structures at the posterior by bcd-AS RNA was not due to mislocalization of the injected RNA, because whole mount in situ hybridization experiments showed that both bcd-WT and bcd-AS RNAs were properly localized to the posterior end site of injection (Fig. 5, bottom).
Since bcd-AS RNA is stable, polyadenylated and localized to the site of injection to the same extent as bcd-WT RNA, these experiments demonstrate that poly(A) elongation alone is insufficient to drive efficient translation in vivo and that a perturbation of the structure of the RNA can interfere with proper translational activation.
DISCUSSION
The relevance of translational control by cytoplasmic polyadenylation during early development has been shown with key molecules such as c-mos in Xenopus (37) and mouse (38), and bicoid (16) and Toll (39) in Drosophila. The cis-acting sequences responsible for elongation are contained within the 3[prime]-UTR. A truncation of ~450 nt of the bicoid 3[prime]-UTR (bcd-Sma) abolishes polyadenylation and inhibits translation of the corresponding RNA. A long poly(A) (~150 A residues) added to the truncated RNA re-establishes translation (16). Taken together, these observations show a requirement for cytoplasmic polyadenylation during Drosophila embryogenesis and also suggest that a long poly(A) tail is sufficient to promote translation.
In this report, we have uncoupled cytoplasmic polyadenylation from translational activation in vivo. Our experiments demonstrate that bcd-AS RNA is translatable in vitro, but the poly(A) elongation that occurs in vivo is not sufficient to promote efficient translation during early embryogenesis. This is the first time that an mRNA that is normally activated by cytoplasmic polyadenylation and maintains its ability to be elongated has been translationally impaired by a mechanism different from deadenylation (26,40-42). In Xenopus, studies with the FGF receptor-1 (XFGFR) maternal mRNA demonstrate that polyadenylation and translation are separable events in this mRNA (43). This mRNA is cytoplasmically polyadenylated during early development, but this elongation is not sufficient to promote translation. However, XFGFR mRNA is not translationally activated by cytoplasmic polyadenylation during early development and the two events appear regulated and timed independently.
Our work suggests that the secondary/tertiary structure of an RNA is important for proper translational activation during early development. In the particular case of cytoplasmic polyadenylation, the folding of the RNA could be critical to properly orient the long poly(A) tail. We envision a scenario where the poly(A) tail folds back to interact with the 5[prime]-UTR in order to stimulate the beginning of translation. A perturbation of the RNA folding could prevent 5[prime]- and 3[prime]-ends `cross-talk', because now the physical distance between them is too large to be covered by a poly(A) tail of ~150 A residues. Therefore, a long poly(A) tail can be defined as still necessary but no longer sufficient to promote efficient translation. It is possible that RNA polyadenylation and folding collaborate to correctly coordinate translation.
The importance of folding to determine function has been studied extensively in proteins. Each protein likely possesses a predetermined and reproducible structure that is essential for many molecular interactions, including those with other proteins, RNA and DNA. The protein folding process is complex and requires the orchestrated action of a series of molecules. Mutations within the protein that modify its structure are often accompanied by loss-of-function. It is possible to predict a similar situation for RNA molecules.
This study suggests that the secondary/tertiary structure of a eukaryotic RNA is important for its proper translation. The analysis of different mechanisms of translational activation, besides cytoplasmic polyadenylation, and their relationship with the folding of the RNA could clarify the role of RNA structure in translation.
ACKNOWLEDGEMENTS
We would like to thank Lihsia Chen, Anupama Dahanukar, Wolfgang Driever, Peter Gergen, Ann Jacobson, Ruth Lehmann, Alan Sachs, Fernando Sallés, Gary Struhl, Tim Wadkins, Robin Wharton and Michael Zuker for fly stocks, reagents, helpful discussions and/or critical reading of the manuscript. A.C.V. thanks Robin Wharton for helpful suggestions and for providing space and reagents during the completion of this paper. This work was supported by a National Institutes of Health Grant to S.S. (GM-51584).
REFERENCES
*To whom correspondence should be addressed at present address: Department of Genetics, Howard Hughes Medical Institute, Duke University Medical Center, Box 3657, 327 CARL Building, Research Drive, Durham, NC 27710, USA. Tel: +1 919 681 2264; Fax: +1 919 681 8984; Email: verro001{at}mercutio.mc.duke.edu Present address: Christopher Wreden, Department of Neurology, UCSF School of Medicine, San Francisco, CA 93143-0435, USA
This page is run by Oxford University Press, Great Clarendon Street, Oxford OX2 6DP, as part of the OUP Journals
Comments and feedback: jnl.info{at}oup.co.uk
Last modification:
Copyright© Oxford University Press, 1999.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||