Coupled amplification and degradation of exogenous RNA injected in amp
http://www.100md.com
《核酸研究医学期刊》
Laboratoire de Biologie du Développement, Université P. et M. Curie, CNRS, UMR 7622, 9 Quai St Bernard, 75251 Paris Cedex 05 France and 1 Laboratoire des régulations post-transcriptionnelles, Institut André Lwoff, CNRS, UPR 1983, 7 rue Guy Moquet, 94801 Villejuif Cedex France
*To whom correspondence should be addressed. Tel: +33 1 44 27 27 82; Fax: +33 1 44 27 35 78; Email: andeol@ccr.jussieu.fr
ABSTRACT
The early development of amphibians takes place in the absence of significant transcription and is controlled at the post-transcriptional level. We have reported that in vitro synthesized transcripts injected into axolotl fertilized eggs or oocytes were not continuously degraded as their abundance apparently fluctuated over time, with detected amounts sometimes higher than initial injected amounts. To further characterize this phenomenon, we have co-injected RNA chain terminators to prevent RNA synthesis. This led to the suppression of fluctuations and to a regular decrease in the amount of transcripts that appeared to be more stable in the presence of inhibitors. These observations indicate a coupling between RNA synthesis and an accelerated degradation. Throughout the time course, cRNA molecules could be detected, and their abundance increased in the early phase of the kinetics, supporting the implication of an RNA-dependent RNA polymerase in an asymmetric amplification process. Finally, when the fate of the injected transcripts was investigated in individual oocytes, we observed an absolute increase in abundance in some but not all oocytes, supporting the existence of a limiting step in the initiation of the RNA amplification stochastic process.
INTRODUCTION
In eukaryotes, the spatial and temporal uncoupling of transcription, translation and RNA degradation provides multiple opportunities for post-transcriptional regulation of gene expression . These different regulations are critical for the early development of amphibians since this takes place in the absence of detectable transcription until the mid-blastula transition and is consequently controlled by post-transcriptional regulation acting on the pool of maternally inherited mRNA . We previously set up an original in vivo system allowing the investigation of the relationship between maternal mRNA stability/degradation before the appearance of zygotic transcription at the MBT (6,7). Our approach is based on the use of two amphibian species which differ in their developmental timing, Xenopus laevis and axolotl (Ambystoma mexicanum). In these two species, MBT occurs after 6 h (12 divisions) (8) and 21 h (10 divisions) (4), respectively. Following injection of Xenopus transcripts into axolotl eggs, early development was allowed to proceed and the fate of the injected RNA was monitored for up to 24 h. Unexpectedly, when using a probe specific for the injected synthetic RNA, the detected molecules did not continuously decrease in abundance following injection but exhibited an irregular pattern of decreases and increases (9). These fluctuations were observed in almost all experiments , and in most cases several increases were detected during a 24 h time course. Fluctuations in the level of the detected molecules were also observed in stage VI oocytes or unfertilized eggs (UFEs), indicating that this phenomenon was not restricted to early development but could also be observed in late oogenesis stages and throughout oocyte maturation. In these studies, fluctuations could be observed with a wide variety of in vitro synthesized transcripts, suggesting that they were not due to specific sequence elements (10). In addition, injection of axolotl transcripts into Xenopus oocytes also led to fluctuations. Nevertheless, whatever the animal model used, no intrinsic periodicity could be evidenced even when the injected cells were obtained from a single female. Because these kinetics were derived from independent groups of cells and did not represent the repeated observations of one set of cells, we proposed that a stochastic event might be involved in the generation of the fluctuations from the late stages of oogenesis and in the period preceding MBT, during early development.
Since RNA analysis is a destructive process, the above-mentioned observations could have simply reflected stochastic cell to cell differences in the RNA degradation rates. However, some of our observations led us to conclude that this hypothesis was too simple. First, labelling the in vitro synthesized RNA indicated that the injected transcripts accounted for only part of the RNA detected later. Secondly, cordycepin (3' deoxyadenosine), an RNA elongation inhibitor, prevented the appearance of fluctuations when co-injected with the RNA. Thirdly, in a few instances, the amount of detected RNA slightly exceeded what had been injected. Taken together, these data suggested that, in addition to RNA degradation, a hitherto unknown amplification process contributed to these fluctuations. Within this framework, the underlying stochastic event could be the initiation of an in vivo synthesis of RNA molecules contributing to the detected RNA molecules. In this respect, it is worth noting that a stochastic event has previously been invoked in the initiation step of co-suppression in plants and attributed to the genesis of double-stranded RNA by an RNA-dependent RNA polymerase (RdRp) (11).
Although the possibility of an RNA amplification independent of the presence of the gene has been considered for a long time, it is only through the study of RNA interference (RNAi) that its importance in cell biology has been established. RNAi is the process by which double-stranded RNA can induce a sequence-specific RNA degradation in almost all eukaryotes (12). In several organisms, including plants, fungi, Schizosaccharomyces pombe and Caenorhabditis elegans, an RdRp has been involved in the initiation of RNAi by generating double-stranded RNA (11). This enzymatic activity has also been implicated in the amplification of the silencing signal (13,14) and the extension of the silencing to adjacent sequences (transitive interference) (15). While the presence of a family of genes closely related to the tomato RdRp has been observed in the genome of plants (16), fungi and nematode, no homologue has been identified in the Drosophila or mammalian genomes. Thus it is unclear whether RdRp activities exist in vertebrates and what their contribution to post-transcriptional regulations could be (17,18).
In the present study, we further investigate the origin of the fluctuations of exogenous RNA in axolotl oocytes. First, as cordycepin (3'dATP) may also interfere with the poly(A) tail metabolism, we co-injected another inhibitor of RNA synthesis, 3'dUTP, with in vitro synthesized transcripts. We confirm the existence of an in vivo coupling between RNA amplification and accelerated degradation of the injected RNA molecules, and we establish that this phenomenon depends on an in vivo RNA synthesis activity. Secondly, we detect the presence of cRNA molecules, in agreement with the involvement of an RdRp. Thirdly, by analysing individual oocytes, we directly demonstrate a cell to cell heterogeneity of the RNA amplification which appears to be a stochastic process.
MATERIALS AND METHODS
Collection of oocytes and unfertilized eggs
Experiments were performed on axolotl (A.mexicamum shaw) stage VI oocytes (600–2200 μm in diameters) staged according to the classification of Beetschen and Gautier (19) and kept at 18–20°C in MBSH (Modified Barth Solution High-Salt) (20). Natural UFEs were obtained after stimulation of females by injection of 500 UI of human chorionic gonadotrophin (hCG; Chorulon, Intervet, France). Maturation was induced in vitro by incubating stage VI oocytes in 10–6 M progesterone (Sigma) in MBSH. UFEs or oocytes of a given series are from the same female.
Preparation of synthetic RNAs and microinjection into oocytes and UFEs
Capped RNAs were generated from linearized templates using an in vitro transcription kit (Ambion Inc., Austin, TX). After the transcription reaction, removal of the template (0.5 μg) was achieved by addition of RNase-free DNase (2 U) for 30 min at 37°C, and a PCR assay was performed to confirm the absence of residual template. All transcripts were resuspended in Gurdon’s injection buffer (88 mM NaCl, 15 mM HEPES, 1 mM KCl, 15 mM Tris–HCl pH 7.6). A 3'-untranslated region (UTR) XhoI–BamHI c-myc fragment from the pXLmyc X.laevis plasmid (21) was subcloned into pBluescript KS+ (Stratagene). Two transcripts of 0.7 kb were produced by T7 RNA polymerase after HindIII digestion (sense RNA) or by T3 RNA polymerase after XhoI digestion (antisense RNA). The Awnt-1 axolotl RNAs were synthesized in vitro from full-length cDNA clones ending with 8–10 adenosine residues (22). The Awnt-1 clone was linearized by PstI to generate a full-length 2.7 kb sense transcript by T3 RNA polymerase. For injection, the standard protocol was as follows: 60 axolotl stage VI oocytes or 60 UFEs were microinjected in MBSH with 2 ng per cell of one synthetic transcript (corresponding to 1/50–1/100 of the total mRNA amount present in one oocyte or UFE) in a volume of 80 nl. The time necessary to inject a transcript into these 60 cells varied between 10 and 20 min. Just after injection, all cells were pooled and a batch of 10 oocytes or UFEs was taken randomly, immediately frozen at –80°C and was considered as t = 0 (±15 min). In some experiments, 10 oocytes or UFEs were injected in <2 min (t = 2 min) and the corresponding batch was immediately frozen. The remaining oocytes or UFEs were maintained in MBSH. Ten oocytes or UFEs were taken randomly at each time point. 3'dUTP (TriLink, Inc.) or cordycepin 5'-triphosphate (Sigma) was co-injected with the transcripts at a final concentration of 2.5 mM. For the single cell analysis, UFEs from the same female were processed individually from injection to –80°C freezing at the indicated time. Injections of synthetic RNA were done in the cytoplasm of oocytes or natural UFEs where germinal vesicle breakdown occurs naturally after hormonal stimulation. Injection series were performed from oocytes or UFEs of at least three different females.
RNA extraction, northern analysis and RNase protection assays
Total RNA from each batch of oocytes or UFEs was extracted using the LiCl method (23). For northern analysis, total RNA was electrophoresed in 1.1% formaldehyde agarose gels and transferred to Hybond N+ membranes (Amersham) (24). Pre-hybridization and hybridization were done at 65°C in Church buffer (25); washes were performed at 65°C in 2x SSC, 0.1% SDS then in 0.2x SSC, 0.1% SDS. Filters were rehybridized with a 32P-labelled 220 bp axolotl 18S probe and radioactivity quantified with the PhosphorImager (Amersham, Molecular Dynamics). RNase protection assays were performed as described (26). For detection of the sense and antisense X.laevis c-myc 3'UTR molecules, the 210 bp ScaI–BamHI fragment of the X.laevis 3'UTR was subcloned in pSP73 plasmid (Promega). After digestion by NdeI, this recombinant plasmid was used for in vitro transcription of a 385 base antisense riboprobe using T7 RNA polymerase, allowing detection of the 210 base sense transcript. After HpaI digestion, a 317 base sense riboprobe was produced with the plasmid using SP6 RNA polymerase, allowing protection of the 210 base antisense RNA.
RESULTS
Fluctuations are abolished by 3'dUTP
We had previously observed that the co-injection of 3'dATP prevented the appearance of fluctuations (10). As 3'dATP (cordycepin) directly affects the metabolism of poly(A) tails and therefore could interfere with RNA stability, we repeated this experiment with another 3' deoxyribonucleotide. In Figure 1, axolotl oocytes were co-injected with c-myc 3'UTR transcripts and 3'dUTP. In the absence of any analogue, c-myc signal was 3-fold higher at 6 h than at 4 h. In contrast, the co-injection of 3'dUTP or 3'dATP led to a continuous decrease of c-myc signal. Additional experiments with other substrates confirmed that fluctuations were systematically abolished by co-injection of RNA synthesis inhibitors (six independent experiments were performed). In addition to the abolition of fluctuations, the presence of an RNA synthesis inhibitor resulted in a stabilization of the injected transcripts. Even if the fate of the detected RNA molecules could not be described by a single half-life because of the fluctuations, apparent half-lives could be obtained by restricting the analysis to continuously decreasing subsets of data. In the absence of inhibitor, the apparent half-life of the Xenopus c-myc 3'UTR RNA was 90 min between 1 and 4 h after injection and 5 h with 3'dUTP. The sensitivity of these two phenomena to 3'dUTP and 3'dATP strongly suggested that they were due to RNA synthesis and not to any interference with poly(A) tail metabolism. Thus, inhibition of RNA synthesis had two effects on the fate of injected RNA, eliminating apparent increases in abundance but also suppressing an accelerated degradation.
Figure 1. RNA chain terminators prevent fluctuations in RNA level. The X.laevis c-myc sense 3'UTR transcript (0.7 kb) was injected without (left panel) or with 3'dUTP or 3'dATP (cordycepin) into axolotl UFEs. Total RNA was extracted from batches of 10 oocytes at different times post-injection (10 min to 6 h), analysed (15 μg) by northern blot, then hybridized successively with the labelled c-myc 3'UTR cDNA probe and the 18S cDNA probes. NI = total RNA from non-injected UFEs. The histograms show the results of the quantification of the c-myc signals with a PhosphorImager, normalized with the 18S rRNA; the t = 10 min value was taken as 100% for each series.
cRNA molecules are detected
As indicated by their sensitivity to 3' deoxynucleotides, the observed fluctuations require an RNA synthesis activity within the oocytes. As this phenomenon has also been observed with transcripts that have no counterpart in the amphibian genome (10), one possible explanation for the amplification of exogenous transcripts in the absence of the corresponding gene is the involvement of an RdRp. Any mechanism of this type requires the presence of cRNA molecules during the amplification process. To further characterize the molecules involved in these fluctuations, we performed, in parallel, a northern blot and an RNase protection assay on samples obtained from the injection of the antisense transcript of the Xenopus c-myc 3'UTR (Fig. 2A and B). Both analyses yielded concordant results when the RNase protection assay was performed with a sense probe that could detect the major product of the in vitro transcription reaction. The observation of variable increased and decreased levels of RNA with a sense probe confirmed and extended our previous hypothesis (10) that the fluctuations observed in oocytes involved transcripts of the same polarity as the injected ones.
Figure 2. Detection of cRNA molecules following the injection of X.laevis c-myc antisense 3'UTR RNA. The 0.7 kb X.laevis c-myc RNA was injected into axolotl UFEs. Total RNA (15 μg) was extracted at different times post-injection (2 min to 13 h). (A) Northern blot analysis. The detection and quantification of the c-myc signal was as in Figure 1. Normalized results are indicated by the solid boxes in the histogram. (B) RNase protection assay. Total RNA corresponding to the experiment shown in (A) (3 μg: 0.5 UFE equivalent) was hybridized to the 317 base c-myc sense 3'UTR riboprobe protecting a 210 base antisense RNA fragment. Signals were expressed as the percentage of the t = 2 min and correspond to the hatched boxes (B) on the histogram. (C) Detection of RNA molecules complementary to the injected c-myc sense transcripts. Total RNA (10 μg from the samples presented in A) was hybridized to the antisense 385 base riboprobe which protected a 210 base fragment allowing detection of the sense c-myc RNA. M = pBR322 MspI-digested DNA marker; NI = total RNA (10 μg) from axolotl non-injected UFEs; tRNA = 10 μg; b = base.
We then analysed the samples of Figure 2B for the presence of cRNA molecules using an RNase protection assay with the appropriate antisense probe. cRNA molecules could be unambiguously detected by the RNase protection assay, albeit with weak signals (Fig. 2C). These cRNA molecules were present immediately after injection, in agreement with their detection among the in vitro transcription products (10). Because of the presence of these cRNA molecules, a reduction of the specific activity of the probe could be anticipated, interfering with the sensitivity and the quantification of the assay but not with its specificity. Importantly, an increase in the cRNA signal was observed between 2 and 15 min (+70%) which could not be ascribed to a bias of the analysis since these samples contained comparable amounts of injected molecules (Fig. 2A and B), yielding the same dilution of the probe in the RNase protection assay. Similar results were obtained with other transcripts, including a full-length axolotl wnt-1 RNA (data not shown). Thus, cRNA molecules increased rapidly in abundance after injection, before any significant change in the detection of RNA molecules of the same polarity as the injected ones.
cRNA detection is sensitive to RNA synthesis inhibition
Following the injection of Xenopus c-myc 3'UTR transcripts into axolotl oocytes, 3'dATP (cordycepin) was added or not and a time course was performed to analyse for the presence of molecules of both polarities (Fig. 3). As before, 3'dATP suppressed the fluctuations of the injected transcript and stabilized these RNA molecules with a half-life of the order of 8 h (Fig. 3A). Similarly, the detected cRNA molecules were stabilized in the presence of 3'dATP as no significant degradation was observed over 20 h, while in its absence they had an apparent half-life of 6 h (Fig. 3B). Thus, the stability of both the injected transcripts and the cognate cRNA molecules was increased when RNA synthesis was inhibited. Importantly, in spite of the stabilization of the injected transcripts, the presence of the inhibitor led to a decrease in the amount of each strand at the first time point (15 min), leading to a 25% reduction for the injected RNA molecules and a 50% reduction for the cRNA molecules. Together with the result of Figure 2, these data confirm that a wave of RNA synthesis affecting both strands takes place rapidly after injection.
Figure 3. Cordycepin stabilizes both the injected transcript and the cRNA molecules. The X.laevis c-myc antisense 3'UTR transcript (0.7 kb) was injected without (–; lanes 2–7) or with cordycepin (+; lanes 8–13) into axolotl stage VI oocytes. (A) Northern blot analysis. Total RNA (10 μg) was analysed as in Figure 1 and hybridized with the labelled c-myc 3'UTR cDNA probe. The histogram shows the c-myc signals normalized with the 18S rRNA levels and expressed as a percentage of the t = 15 min (lane 2). (B) RNase protection assay. Total RNA (lanes 2–7 and lanes 8–13: three and five oocyte-equivalents, respectively) was hybridized to the c-myc antisense 3'UTR 385 base riboprobe protecting a 210 base sense RNA fragment. Lanes 2–13: same times post-injection as in (A). Sense = 5 pg of Xenopus c-myc sense in vitro 3'UTR transcript; AS = 2 ng of c-myc antisense in vitro 3'UTR transcript. The histogram shows the c-myc sense RNA signals expressed as a percentage of the t = 15 min value (100%). M = pBR322 MspI-digested DNA marker; NI = total RNA from axolotl non-injected oocytes.
The RNA amplification process following RNA injection is heterogeneous from cell to cell
These results established the existence of a coupling between an RNA amplification process and an accelerated degradation of both the injected transcript and its cRNA. This negative feedback loop could generate oscillations in the RNA level and thus contribute to the fluctuation pattern. However, the stochastic nature of the fluctuations also suggested an absence of synchrony between cells. To investigate the fate of the injected RNA at the level of single cells, we performed an analysis of individual UFEs, using the Awnt-1 transcript as a substrate (Fig. 4). Individual UFEs were rapidly frozen either immediately or 1 h after injection, and RNA was subsequently extracted from each UFE. The Awnt-1 RNA content was then determined by northern blot, using the 18S rRNA to control for total RNA loading. Immediately after injection, the individual RNA levels exhibited a very narrow distribution (n = 11, mean = 100, SD = 15) excluding injection as a major source of variation in the experiments. This also indicated that UFEs contained similar amounts of rRNA, since these results were normalized with the 18S signal. In contrast, a wide dispersion of the Awnt-1 RNA content was observed in the group of UFEs analysed after 1 h incubation at 18°C (n = 31, mean = 70, SD= 68). For 18 UFEs, the signal was significantly lower than in the control group (P < 0.01). In nine UFEs, the signal was similar to that of the control group, while in four out of 31 UFEs it was significantly higher (P < 0.01), ranging from 170 to 265. Thus, at 1 h after injection, the fate of the detected RNA differed widely between eggs, even though they were derived from a single female. Remarkably, four out of 31 UFEs contained significantly more Awnt-1 RNA than they had received by injection. This net increase in Awnt-1 RNA content provides a direct demonstration of the amplification process. Thus, 1 h after injection, depending upon the egg, the amount of RNA could increase by up to 2.6-fold or decrease by up to 6-fold.
Figure 4. Analysis of individual UFEs. (A) The 2.7 kb Awnt-1 RNA was injected into 11 axolotl UFEs (A–K), each being individually and immediately frozen at – 80°C (t = 0). For each egg, the complete RNA content was analysed by northern blot. The blots were successively hybridized with the Awnt-1 and the 18S cDNA probes. (B) A total of 31 UFEs (1–31) were also injected with Awnt-1 RNA and individually frozen 1 h post-injection (t = 1 h). (C) The histogram presents the signal observed at t = 0 and 1 h, after normalization with the 18S rRNA levels, the mean value at t = 0 being 100.
To investigate the effect of these cell to cell differences when working on batches of cells, we injected the Awnt-1 transcript into UFEs and 1 h later we analysed the RNA content in four batches of six eggs and two batches of seven eggs. The results are presented in Figure 5A using the 18S signal to normalize for RNA loading. As expected, the observed values differed significantly between batches (a 1.6-fold difference between the extreme values). We then used the individual values of Figure 4C to simulate the result of grouping cells into batches (Fig. 5B). Importantly, when batches of 6–10 eggs were used, the heterogeneity of the individual cells could not be masked by the averaging process and an important dispersion of the results persisted. In addition, the experimental values of Figure 5A fell precisely within the range of the simulation in Figure 5B, establishing the coherence of our results whether analysed at the individual level or by batches. Thus, when working on batches of cells, the heterogeneity of the cellular behaviour following the injection of in vitro synthesized transcripts could significantly contribute, through the sampling procedure, to the fluctuations in the detection of RNA.
Figure 5. Analysis of the level of Awnt-1 RNA 1 h post -injection. (A) Experimental data for batches of either six or seven UFEs. (B) Simulation of the results of a batch analysis with batches of six, seven, eight or 10 UFEs, using the individual data of Figure 4. In each case, the results of five random selections of individual values are represented by diamonds. The theoretical upper and lower limits are indicated by bars.
DISCUSSION
In this study, we provide additional evidence for the role of an RNA amplification process in the fluctuations of the detection of exogenous RNA injected into axolotl oocytes or UFEs. First, these fluctuations are abolished by the co-injection of RNA chain terminators independently of any modification of the poly(A) tail metabolism. Secondly, complementary RNA molecules can be detected in the injected samples and their abundance increases during the early phase of the kinetics. The increase in abundance of cRNA as well as the cRNA degradation rate are sensitive to RNA chain terminators. Finally, analysis of individual UFEs provides a direct demonstration that an RNA amplification takes place in some cells and is extremely heterogeneous from cell to cell.
On the amplification mechanism
We have observed fluctuations for all the in vitro synthesized transcripts we have injected into axolotl oocytes or UFEs, whether or not the corresponding gene is present in the axolotl genome (9,10). Both the use of RNA chain terminators and the analysis of individual cells indicate that these fluctuations are associated with an in vivo RNA synthesis. We initially excluded the hypothesis of an RNA polymerase II involvement as -amanitin did not inhibit RNA fluctuations (9), and transcription is either limited or absent in these cellular models (5). In this study, we detected cRNA molecules, and their number increased between 2 and 15 min following injection, through a 3' deoxyribonucleotide-sensitive process. One pathway which could lead to an RNA amplification in the absence of the corresponding gene is an RNA replication by an RdRp via a double-stranded RNA intermediate. Our results strongly argue in favour of the implication of an RdRp in the amplification process. Moreover, the observation that no change in migration was observed even for the 0.7 kb fragment from the 3'UTR of the Xenopus c-myc gene indicates that the in vivo synthesized molecules are very similar in length to the injected ones. An RdRp activity has been characterized in tomato and Neurospora crassa (13,27), and homologous genes have been identified in the genomes of plants, fungi, yeast and nematode . It is of note that the cellular RdRp found in N.crassa possesses two distinct activities in vitro, one of which allows the synthesis of full-length copies from a single-strand input RNA (29). However, no RdRp homologue has been detected in the genomes of vertebrates or Drosophila (18). Attempts to detect an RdRp activity in murine and Drosophila cells by the extension of silencing to adjacent sequences (transitive RNAi) have been unsuccessful. It should be noted, however, that although the implication of RdRp genes in RNAi has been established in C.elegans, transitive RNAi is very inefficient in this organism, indicating that the activity of these enzymes is tightly controlled (15). In contrast, a cytoplasmic amplification of ?-globin mRNA has been reported in erythroleukaemia cells (30,31), suggesting the existence of an RdRp in vertebrates.
Coupling between RNA amplification and degradation
Co-injection of RNA synthesis inhibitors (3'dUTP or 3'dATP) abrogates RNA fluctuation presumably by inhibiting a cellular RNA amplification process. These inhibitors also stabilize the injected transcript, suggesting that the RNA amplification is coupled to an RNA degradation mechanism. cRNA molecules were detected among the injected transcripts (Fig. 2C), in agreement with previous observations on the products of in vitro transcription reactions (10,12). It is therefore likely that some double-stranded or partially double-stranded molecules are present among the products of the in vitro transcription reaction. These molecules could be the templates for the amplification reaction. The precise organization of these templates molecules, for instance at their extremities, could provide the basis for the asymmetric synthesis which we observed (Fig. 3). Here we show that the two complementary strands exhibit a differential behaviour towards the observed RNA amplification and degradation processes. A double-stranded RNA amplification has previously been observed to be asymetric in viral RNA replication (32) and for cytoplasmic ?-globin RNA in murine erythroleukaemia cells (31).
The implication of an RdRp and the necessary synthesis of double-stranded intermediates provides a simple framework for a coupling between amplification and degradation, as, in most eukaryotes, including Xenopus and axolotl, the presence of double-stranded RNA molecules induces RNAi, a potent and sequence-specific gene silencing . One prediction of this model is that both strands should be the target of the accelerated degradation since RNAi can target either strand. In this study, we observed that both strands of the Xenopus c-myc 3'UTR are stabilized by 3'dATP, in agreement with the involvement of double-stranded RNA in inducing the accelerated degradation of the transcripts.
On the cellular stochastic variability of RNA fluctuations
One limitation of most RNA studies including the ones reported here is their destructive nature. Consequently, independent samples are used to reconstruct the kinetics by assuming that they reflect an average behaviour. Because the values we observe when studying batches of injected cells are very close to those obtained by simulating virtual batches from individual cell observations (Fig. 5), this study indicates, on the contrary, that an important cell to cell variability exists in the biological process that underlies the observed fluctuations. Although very large batches of cells reduce the fluctuations, our validated simulation approach predicts that batches of up to 10 cells would still exhibit up to 2-fold differences between samples as individual cellular RNA levels really differ by >10-fold. Such a situation suggests that a limiting step exists in the initiation of the coupled RNA amplification/degradation processes: the stochastic pattern would then be due to a very tight control of the factor(s) leading to dramatic differences between cells. RdRp activities have already been implicated in similar stochastic cellular variability. In plants, some RNA molecules can be recognized as ‘aberrant’ and converted by an RdRp into double-stranded RNA, thereby inducing a co-suppression (11). To avoid the induction of interference against bona fide cellular mRNA, the initiation of such a process should be tightly controlled, and this has been proposed as the cause of the stochastic induction of co-suppression observed in plants. Other stochastic events occurring in animals, in particular during early development in axolotl, have been described, such as a remarkable variability of the individual cycle lengths from cell to cell (34) and, after introduction of a G1 phase for the first time at the MBT, a variable duration of this phase from one blastomere to another (35). This variability could account for lengthening of cell cycles and be required for zygotic transcriptions necessary for DNA replication (36).
In addition, the observation that fluctuations can persist for up to 24 h suggests that the heterogeneity among cells is not completely abrogated during this period. This suggests two possibilities. In the simplest model, once one amplification is initiated in a given cell, an accelerated degradation of transcripts should prevail but, from cell to cell, this initiation could proceed stochastically over to a 24 h period. Our results also open up the possibility that, within a cell, several cycles of amplification and degradation could take place over a 24 h time course. Such oscillations in RNA level could indeed occur provided that, during the accelerated degradation phase, not all the amplification templates are eliminated.
In summary, our results further establish the existence of an amplification mechanism in amphibian oocytes and eggs. The demonstration that after 1 h some eggs contain more than twice the amount of injected RNA also establishes that the amplification process can be efficient (of the order of one RNA molecule synthesized per minute assuming that the amplification templates represent 1.5% of the injected molecules). The detection of cRNA molecules indicates that an RdRp may be implicated. Any RdRp-mediated amplification could be expected to induce an RNAi provided that enough double-stranded RNA molecules are generated and are not sequestrated away from the interference machinery. Our observation that the inhibition of RNA synthesis leads to a greater stability of the detected molecules fully supports this idea. Since the initiation of RNA amplification is stochastic under our experimental conditions, our results also stress the existence of a limiting step in the initiation of the amplification phenomenon, presumably reflecting a tight control of the activity of an RdRp. Our work performed with in vitro synthesized transcripts suggests a possible role for RdRp in post-transcriptional regulations during late oogenesis and early development in amphibians, with the possibility that some endogenous maternal mRNA could be the target of this amplification process before the MBT.
ACKNOWLEDGEMENTS
The authors thank M.W. King and C. Séguin for providing the Xenopus c-myc cDNA clone and the Awnt-1 cDNA clone, respectively, and H. Pelczar and L. Dandolo for helpful discussions. This work was supported by grants from the ‘Ligue Régionale Ile de France contre le cancer’.
REFERENCES
Bashirullah,A., Cooperstock,R.L. and Lipshitz,H.D. (2001) Spatial and temporal control of RNA stability. Proc. Natl Acad. Sci. USA, 98, 7025–7028.
Day,D.A. and Tuite,M.F. (1998) Post-transcriptional gene regulatory mechanisms in eukaryotes: an overview. J. Endocrinol., 157, 361–371.
Yaniv,K. and Yisraeli,J.K. (2001) Defining cis-acting elements and trans-acting factors in RNA localization. Int. Rev. Cytol., 203, 521–539.
Signoret,J. and Lefresne,J. (1971) Contribution à l’étude de la segmentation de l’uf d’axolotl. I—Définition de la transition blastuléenne. Ann. Embryol. Morphol., 4, 113–123.
Andéol,Y. (1994) Early transcription in different animal species: implication for transition from maternal to zygotic control in development. Wilhelm Roux’s Arch. Dev. Biol., 204, 3–10.
Andéol,Y., Lefresne,J., Houillon,C. and Signoret,J. (1995a) Differential stability of Xenopus c-myc RNA during oogenesis in axolotl. Involvement of the 3' untranslated region in vivo. Wilhelm Roux’s Arch. Dev. Biol., 205, 182–191.
Andéol,Y., Lefresne,J. and Signoret,J. (1995b) Evidence for a nuclear factor involved in c-myc RNA degradation during axolotl oocyte maturation. Wilhelm Roux’s Arch. Dev. Biol., 205, 192–197.
Newport,J. and Kirschner,M. (1982) A major development transition in early Xenopus embryos: I—characterization and timing of cellular changes at the midblastula stage. Cell, 30, 675–686.
Andéol,Y., Lefresne,J., Simard,C., Séguin,C., Mouton,C. and Signoret,J. (1998) Post-transcriptional control of c-myc RNA during early development analyzed in vivo with a Xenopus–axolotl heterologous system. Differentiation, 63, 69–79.
Montreau,N., Vaur,S., Dautry,F. and Andéol,Y. (2003) Injection of exogenous RNA in amphibian oocytes leads to RNA level fluctuations which are sensitive to cordycepin, an RNA chain elongation terminator. C.R. Biol., 326, 1135–1147.
Vaucheret,H., Béclin,C. and Fagard,M. (2001) Post-transcriptional gene silencing in plants. J. Cell Sci., 114, 3083–3091.
Fire,A., Xu,S., Montgomery,M.K., Kostas,S.A., Driver,S.E. and Mello,C.C. (1998) Potent and specific genetic interference by double-stranded RNA in Caenorhabditis elegans. Nature, 391, 806–810.
Cogoni,C. and Macino,G. (1999) Posttranscriptional gene silencing in Neurospora by a RecQ DNA helicase. Science, 286, 2342–2344.
Smardon,A., Spoerke,J.M., Stacey,S.C., Klein,M.E., Mackin,N. and Maine,E.M. (2000) EGO-1 is related to RNA-directed RNA polymerase and functions in germ-line development and RNA interference in C. elegans. Curr. Biol., 10, 169–178.
Sijen,T., Fleenor,J., Simmer,F., Thijssen,K.L., Parrish,S., Timmons,L., Plasterk,R.H.A. and Fire,A. (2001) On the role of RNA amplification in dsRNA-triggered gene silencing. Cell, 107, 465–476.
Schiebel,W., Haas,B., Marinkovic,S., Klanner,A. and Sanger,H.L. (1993) RNA-directed RNA polymerase from tomato leaves: I Purification and physical properties. J. Biol. Chem., 263, 11851–11857.
Tijsterman,M., Ketting,R.F. and Plasterk,R.H.A. (2002) The genetics of RNA silencing. Annu. Rev. Genet., 36, 489–519.
Stein,P., Svoboda,P., Anger,M. and Schultz,R.M. (2003) RNAi: mammalian oocytes do it without RNA-dependent RNA polymerase. RNA, 9, 187–192.
Armstrong,J.B. and Malacinski,G.M. (1989) Developmental Biology of the Axolotl. Oxford University Press, New York.
Gurdon,J.B., Lane,C.D., Woodland,H.R. and Marbaix,H.R.W.G. (1971) Use of frog eggs and oocytes for the study of messenger RNA and its translation in living cells. Nature, 233, 177–182.
King,M.W., Roberts,J.M. and Eisenman,R.N. (1986) Expression of the c-myc proto-oncogene during development of Xenopus laevis. Mol. Cell. Biol., 6, 4499–4508.
Busse,U. and Séguin,C. (1993) Molecular analysis of the Wnt-1 proto-oncogene in Ambystoma mexicanum (axolotl) embryos. Differentiation, 53, 7–15.
Auffray,C. and Rougeon,F. (1980) Purification of mouse immunoglobulin heavy-chain messenger RNAs from total myeloma tumor RNA. Eur. J. Biochem., 107, 303–314.
Sambrook,J., Maniatis,T. and Fritsch,F. (1989) Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.
Church,G. and Gilbert,W. (1984) Genomic sequencing. Proc. Natl Acad. Sci. USA, 81, 1991–1995.
Neel,H., Weil,D., Giansante,C. and Dautry,F. (1993) In vivo cooperation between introns during pre-mRNA processing. Genes Dev., 5, 2194–2205.
Schiebel,W., Haas,B., Marinkovic,S., Klanner,A. and Sanger,H.L. (1993b) RNA-directed RNA polymerase from tomato leaves: II Catalytic in vitro properties. J. Biol. Chem., 268, 11858–11867.
Bosher,J.M. (2000) RNA interference: genetic wand and genetic watchdog. Nature Cell. Biol., 2, 31–36.
Makeyev,E.V. and Bamford,D.H. (2002) Cellular RNA-dependent RNA polymerase involved in posttranscriptional gene silencing has two distinct activity modes. Mol. Cell, 10, 1417–1427.
Volloch,V. (1986) Cytoplasmic synthesis of globin RNA in differentiated murine erythroleukemia cells: possible involvement of RNA-dependent RNA polymerase. Proc. Natl Acad. Sci. USA, 83, 1208–1212.
Volloch,V., Schweitzer,B. and Rits,S. (1987) Synthesis of globin RNA in enucleated differentiating murine erythroleukemia cells. J. Cell Biol., 105, 137–143.
Barton,D.J., Black,E.P. and Flanegan,J.B. (1995) Complete replication of poliovirus in vitro: preinitiation RNA replication complexes require soluble cellular factors for the synthesis of VPg-linked RNA. J. Virol., 69, 5516–5527.
Tuschl,T., Zamore,P.D., Lehman,R., Bartel,D.P. and Sharp,P.A. (1999) Targeted mRNA degradation by double strand RNA in vitro. Genes Dev., 13, 3191–3197.
Signoret,J. (1977) La cinétique cellulaire au cours de la segmentation du germe d’axolotl: proposition d’un modèle statistique. J. Embryol. Exp. Morphol., 42, 5–14.
Lefresne,J., Andéol,Y. and Signoret,J. (1998) Evidence for introduction of a variable G1 phase at the midblastula transition during early development in axolotl. Dev. Growth Differ., 40, 497–508.
Signoret,J. (1980) Evidence of the First Genetic Activity Required in Axolotl Development. Results and Problems on Cell Differentiation. Springer-Verlag, Berlin.(Nicole Montreau, Sabine Vaur, Francois D)
*To whom correspondence should be addressed. Tel: +33 1 44 27 27 82; Fax: +33 1 44 27 35 78; Email: andeol@ccr.jussieu.fr
ABSTRACT
The early development of amphibians takes place in the absence of significant transcription and is controlled at the post-transcriptional level. We have reported that in vitro synthesized transcripts injected into axolotl fertilized eggs or oocytes were not continuously degraded as their abundance apparently fluctuated over time, with detected amounts sometimes higher than initial injected amounts. To further characterize this phenomenon, we have co-injected RNA chain terminators to prevent RNA synthesis. This led to the suppression of fluctuations and to a regular decrease in the amount of transcripts that appeared to be more stable in the presence of inhibitors. These observations indicate a coupling between RNA synthesis and an accelerated degradation. Throughout the time course, cRNA molecules could be detected, and their abundance increased in the early phase of the kinetics, supporting the implication of an RNA-dependent RNA polymerase in an asymmetric amplification process. Finally, when the fate of the injected transcripts was investigated in individual oocytes, we observed an absolute increase in abundance in some but not all oocytes, supporting the existence of a limiting step in the initiation of the RNA amplification stochastic process.
INTRODUCTION
In eukaryotes, the spatial and temporal uncoupling of transcription, translation and RNA degradation provides multiple opportunities for post-transcriptional regulation of gene expression . These different regulations are critical for the early development of amphibians since this takes place in the absence of detectable transcription until the mid-blastula transition and is consequently controlled by post-transcriptional regulation acting on the pool of maternally inherited mRNA . We previously set up an original in vivo system allowing the investigation of the relationship between maternal mRNA stability/degradation before the appearance of zygotic transcription at the MBT (6,7). Our approach is based on the use of two amphibian species which differ in their developmental timing, Xenopus laevis and axolotl (Ambystoma mexicanum). In these two species, MBT occurs after 6 h (12 divisions) (8) and 21 h (10 divisions) (4), respectively. Following injection of Xenopus transcripts into axolotl eggs, early development was allowed to proceed and the fate of the injected RNA was monitored for up to 24 h. Unexpectedly, when using a probe specific for the injected synthetic RNA, the detected molecules did not continuously decrease in abundance following injection but exhibited an irregular pattern of decreases and increases (9). These fluctuations were observed in almost all experiments , and in most cases several increases were detected during a 24 h time course. Fluctuations in the level of the detected molecules were also observed in stage VI oocytes or unfertilized eggs (UFEs), indicating that this phenomenon was not restricted to early development but could also be observed in late oogenesis stages and throughout oocyte maturation. In these studies, fluctuations could be observed with a wide variety of in vitro synthesized transcripts, suggesting that they were not due to specific sequence elements (10). In addition, injection of axolotl transcripts into Xenopus oocytes also led to fluctuations. Nevertheless, whatever the animal model used, no intrinsic periodicity could be evidenced even when the injected cells were obtained from a single female. Because these kinetics were derived from independent groups of cells and did not represent the repeated observations of one set of cells, we proposed that a stochastic event might be involved in the generation of the fluctuations from the late stages of oogenesis and in the period preceding MBT, during early development.
Since RNA analysis is a destructive process, the above-mentioned observations could have simply reflected stochastic cell to cell differences in the RNA degradation rates. However, some of our observations led us to conclude that this hypothesis was too simple. First, labelling the in vitro synthesized RNA indicated that the injected transcripts accounted for only part of the RNA detected later. Secondly, cordycepin (3' deoxyadenosine), an RNA elongation inhibitor, prevented the appearance of fluctuations when co-injected with the RNA. Thirdly, in a few instances, the amount of detected RNA slightly exceeded what had been injected. Taken together, these data suggested that, in addition to RNA degradation, a hitherto unknown amplification process contributed to these fluctuations. Within this framework, the underlying stochastic event could be the initiation of an in vivo synthesis of RNA molecules contributing to the detected RNA molecules. In this respect, it is worth noting that a stochastic event has previously been invoked in the initiation step of co-suppression in plants and attributed to the genesis of double-stranded RNA by an RNA-dependent RNA polymerase (RdRp) (11).
Although the possibility of an RNA amplification independent of the presence of the gene has been considered for a long time, it is only through the study of RNA interference (RNAi) that its importance in cell biology has been established. RNAi is the process by which double-stranded RNA can induce a sequence-specific RNA degradation in almost all eukaryotes (12). In several organisms, including plants, fungi, Schizosaccharomyces pombe and Caenorhabditis elegans, an RdRp has been involved in the initiation of RNAi by generating double-stranded RNA (11). This enzymatic activity has also been implicated in the amplification of the silencing signal (13,14) and the extension of the silencing to adjacent sequences (transitive interference) (15). While the presence of a family of genes closely related to the tomato RdRp has been observed in the genome of plants (16), fungi and nematode, no homologue has been identified in the Drosophila or mammalian genomes. Thus it is unclear whether RdRp activities exist in vertebrates and what their contribution to post-transcriptional regulations could be (17,18).
In the present study, we further investigate the origin of the fluctuations of exogenous RNA in axolotl oocytes. First, as cordycepin (3'dATP) may also interfere with the poly(A) tail metabolism, we co-injected another inhibitor of RNA synthesis, 3'dUTP, with in vitro synthesized transcripts. We confirm the existence of an in vivo coupling between RNA amplification and accelerated degradation of the injected RNA molecules, and we establish that this phenomenon depends on an in vivo RNA synthesis activity. Secondly, we detect the presence of cRNA molecules, in agreement with the involvement of an RdRp. Thirdly, by analysing individual oocytes, we directly demonstrate a cell to cell heterogeneity of the RNA amplification which appears to be a stochastic process.
MATERIALS AND METHODS
Collection of oocytes and unfertilized eggs
Experiments were performed on axolotl (A.mexicamum shaw) stage VI oocytes (600–2200 μm in diameters) staged according to the classification of Beetschen and Gautier (19) and kept at 18–20°C in MBSH (Modified Barth Solution High-Salt) (20). Natural UFEs were obtained after stimulation of females by injection of 500 UI of human chorionic gonadotrophin (hCG; Chorulon, Intervet, France). Maturation was induced in vitro by incubating stage VI oocytes in 10–6 M progesterone (Sigma) in MBSH. UFEs or oocytes of a given series are from the same female.
Preparation of synthetic RNAs and microinjection into oocytes and UFEs
Capped RNAs were generated from linearized templates using an in vitro transcription kit (Ambion Inc., Austin, TX). After the transcription reaction, removal of the template (0.5 μg) was achieved by addition of RNase-free DNase (2 U) for 30 min at 37°C, and a PCR assay was performed to confirm the absence of residual template. All transcripts were resuspended in Gurdon’s injection buffer (88 mM NaCl, 15 mM HEPES, 1 mM KCl, 15 mM Tris–HCl pH 7.6). A 3'-untranslated region (UTR) XhoI–BamHI c-myc fragment from the pXLmyc X.laevis plasmid (21) was subcloned into pBluescript KS+ (Stratagene). Two transcripts of 0.7 kb were produced by T7 RNA polymerase after HindIII digestion (sense RNA) or by T3 RNA polymerase after XhoI digestion (antisense RNA). The Awnt-1 axolotl RNAs were synthesized in vitro from full-length cDNA clones ending with 8–10 adenosine residues (22). The Awnt-1 clone was linearized by PstI to generate a full-length 2.7 kb sense transcript by T3 RNA polymerase. For injection, the standard protocol was as follows: 60 axolotl stage VI oocytes or 60 UFEs were microinjected in MBSH with 2 ng per cell of one synthetic transcript (corresponding to 1/50–1/100 of the total mRNA amount present in one oocyte or UFE) in a volume of 80 nl. The time necessary to inject a transcript into these 60 cells varied between 10 and 20 min. Just after injection, all cells were pooled and a batch of 10 oocytes or UFEs was taken randomly, immediately frozen at –80°C and was considered as t = 0 (±15 min). In some experiments, 10 oocytes or UFEs were injected in <2 min (t = 2 min) and the corresponding batch was immediately frozen. The remaining oocytes or UFEs were maintained in MBSH. Ten oocytes or UFEs were taken randomly at each time point. 3'dUTP (TriLink, Inc.) or cordycepin 5'-triphosphate (Sigma) was co-injected with the transcripts at a final concentration of 2.5 mM. For the single cell analysis, UFEs from the same female were processed individually from injection to –80°C freezing at the indicated time. Injections of synthetic RNA were done in the cytoplasm of oocytes or natural UFEs where germinal vesicle breakdown occurs naturally after hormonal stimulation. Injection series were performed from oocytes or UFEs of at least three different females.
RNA extraction, northern analysis and RNase protection assays
Total RNA from each batch of oocytes or UFEs was extracted using the LiCl method (23). For northern analysis, total RNA was electrophoresed in 1.1% formaldehyde agarose gels and transferred to Hybond N+ membranes (Amersham) (24). Pre-hybridization and hybridization were done at 65°C in Church buffer (25); washes were performed at 65°C in 2x SSC, 0.1% SDS then in 0.2x SSC, 0.1% SDS. Filters were rehybridized with a 32P-labelled 220 bp axolotl 18S probe and radioactivity quantified with the PhosphorImager (Amersham, Molecular Dynamics). RNase protection assays were performed as described (26). For detection of the sense and antisense X.laevis c-myc 3'UTR molecules, the 210 bp ScaI–BamHI fragment of the X.laevis 3'UTR was subcloned in pSP73 plasmid (Promega). After digestion by NdeI, this recombinant plasmid was used for in vitro transcription of a 385 base antisense riboprobe using T7 RNA polymerase, allowing detection of the 210 base sense transcript. After HpaI digestion, a 317 base sense riboprobe was produced with the plasmid using SP6 RNA polymerase, allowing protection of the 210 base antisense RNA.
RESULTS
Fluctuations are abolished by 3'dUTP
We had previously observed that the co-injection of 3'dATP prevented the appearance of fluctuations (10). As 3'dATP (cordycepin) directly affects the metabolism of poly(A) tails and therefore could interfere with RNA stability, we repeated this experiment with another 3' deoxyribonucleotide. In Figure 1, axolotl oocytes were co-injected with c-myc 3'UTR transcripts and 3'dUTP. In the absence of any analogue, c-myc signal was 3-fold higher at 6 h than at 4 h. In contrast, the co-injection of 3'dUTP or 3'dATP led to a continuous decrease of c-myc signal. Additional experiments with other substrates confirmed that fluctuations were systematically abolished by co-injection of RNA synthesis inhibitors (six independent experiments were performed). In addition to the abolition of fluctuations, the presence of an RNA synthesis inhibitor resulted in a stabilization of the injected transcripts. Even if the fate of the detected RNA molecules could not be described by a single half-life because of the fluctuations, apparent half-lives could be obtained by restricting the analysis to continuously decreasing subsets of data. In the absence of inhibitor, the apparent half-life of the Xenopus c-myc 3'UTR RNA was 90 min between 1 and 4 h after injection and 5 h with 3'dUTP. The sensitivity of these two phenomena to 3'dUTP and 3'dATP strongly suggested that they were due to RNA synthesis and not to any interference with poly(A) tail metabolism. Thus, inhibition of RNA synthesis had two effects on the fate of injected RNA, eliminating apparent increases in abundance but also suppressing an accelerated degradation.
Figure 1. RNA chain terminators prevent fluctuations in RNA level. The X.laevis c-myc sense 3'UTR transcript (0.7 kb) was injected without (left panel) or with 3'dUTP or 3'dATP (cordycepin) into axolotl UFEs. Total RNA was extracted from batches of 10 oocytes at different times post-injection (10 min to 6 h), analysed (15 μg) by northern blot, then hybridized successively with the labelled c-myc 3'UTR cDNA probe and the 18S cDNA probes. NI = total RNA from non-injected UFEs. The histograms show the results of the quantification of the c-myc signals with a PhosphorImager, normalized with the 18S rRNA; the t = 10 min value was taken as 100% for each series.
cRNA molecules are detected
As indicated by their sensitivity to 3' deoxynucleotides, the observed fluctuations require an RNA synthesis activity within the oocytes. As this phenomenon has also been observed with transcripts that have no counterpart in the amphibian genome (10), one possible explanation for the amplification of exogenous transcripts in the absence of the corresponding gene is the involvement of an RdRp. Any mechanism of this type requires the presence of cRNA molecules during the amplification process. To further characterize the molecules involved in these fluctuations, we performed, in parallel, a northern blot and an RNase protection assay on samples obtained from the injection of the antisense transcript of the Xenopus c-myc 3'UTR (Fig. 2A and B). Both analyses yielded concordant results when the RNase protection assay was performed with a sense probe that could detect the major product of the in vitro transcription reaction. The observation of variable increased and decreased levels of RNA with a sense probe confirmed and extended our previous hypothesis (10) that the fluctuations observed in oocytes involved transcripts of the same polarity as the injected ones.
Figure 2. Detection of cRNA molecules following the injection of X.laevis c-myc antisense 3'UTR RNA. The 0.7 kb X.laevis c-myc RNA was injected into axolotl UFEs. Total RNA (15 μg) was extracted at different times post-injection (2 min to 13 h). (A) Northern blot analysis. The detection and quantification of the c-myc signal was as in Figure 1. Normalized results are indicated by the solid boxes in the histogram. (B) RNase protection assay. Total RNA corresponding to the experiment shown in (A) (3 μg: 0.5 UFE equivalent) was hybridized to the 317 base c-myc sense 3'UTR riboprobe protecting a 210 base antisense RNA fragment. Signals were expressed as the percentage of the t = 2 min and correspond to the hatched boxes (B) on the histogram. (C) Detection of RNA molecules complementary to the injected c-myc sense transcripts. Total RNA (10 μg from the samples presented in A) was hybridized to the antisense 385 base riboprobe which protected a 210 base fragment allowing detection of the sense c-myc RNA. M = pBR322 MspI-digested DNA marker; NI = total RNA (10 μg) from axolotl non-injected UFEs; tRNA = 10 μg; b = base.
We then analysed the samples of Figure 2B for the presence of cRNA molecules using an RNase protection assay with the appropriate antisense probe. cRNA molecules could be unambiguously detected by the RNase protection assay, albeit with weak signals (Fig. 2C). These cRNA molecules were present immediately after injection, in agreement with their detection among the in vitro transcription products (10). Because of the presence of these cRNA molecules, a reduction of the specific activity of the probe could be anticipated, interfering with the sensitivity and the quantification of the assay but not with its specificity. Importantly, an increase in the cRNA signal was observed between 2 and 15 min (+70%) which could not be ascribed to a bias of the analysis since these samples contained comparable amounts of injected molecules (Fig. 2A and B), yielding the same dilution of the probe in the RNase protection assay. Similar results were obtained with other transcripts, including a full-length axolotl wnt-1 RNA (data not shown). Thus, cRNA molecules increased rapidly in abundance after injection, before any significant change in the detection of RNA molecules of the same polarity as the injected ones.
cRNA detection is sensitive to RNA synthesis inhibition
Following the injection of Xenopus c-myc 3'UTR transcripts into axolotl oocytes, 3'dATP (cordycepin) was added or not and a time course was performed to analyse for the presence of molecules of both polarities (Fig. 3). As before, 3'dATP suppressed the fluctuations of the injected transcript and stabilized these RNA molecules with a half-life of the order of 8 h (Fig. 3A). Similarly, the detected cRNA molecules were stabilized in the presence of 3'dATP as no significant degradation was observed over 20 h, while in its absence they had an apparent half-life of 6 h (Fig. 3B). Thus, the stability of both the injected transcripts and the cognate cRNA molecules was increased when RNA synthesis was inhibited. Importantly, in spite of the stabilization of the injected transcripts, the presence of the inhibitor led to a decrease in the amount of each strand at the first time point (15 min), leading to a 25% reduction for the injected RNA molecules and a 50% reduction for the cRNA molecules. Together with the result of Figure 2, these data confirm that a wave of RNA synthesis affecting both strands takes place rapidly after injection.
Figure 3. Cordycepin stabilizes both the injected transcript and the cRNA molecules. The X.laevis c-myc antisense 3'UTR transcript (0.7 kb) was injected without (–; lanes 2–7) or with cordycepin (+; lanes 8–13) into axolotl stage VI oocytes. (A) Northern blot analysis. Total RNA (10 μg) was analysed as in Figure 1 and hybridized with the labelled c-myc 3'UTR cDNA probe. The histogram shows the c-myc signals normalized with the 18S rRNA levels and expressed as a percentage of the t = 15 min (lane 2). (B) RNase protection assay. Total RNA (lanes 2–7 and lanes 8–13: three and five oocyte-equivalents, respectively) was hybridized to the c-myc antisense 3'UTR 385 base riboprobe protecting a 210 base sense RNA fragment. Lanes 2–13: same times post-injection as in (A). Sense = 5 pg of Xenopus c-myc sense in vitro 3'UTR transcript; AS = 2 ng of c-myc antisense in vitro 3'UTR transcript. The histogram shows the c-myc sense RNA signals expressed as a percentage of the t = 15 min value (100%). M = pBR322 MspI-digested DNA marker; NI = total RNA from axolotl non-injected oocytes.
The RNA amplification process following RNA injection is heterogeneous from cell to cell
These results established the existence of a coupling between an RNA amplification process and an accelerated degradation of both the injected transcript and its cRNA. This negative feedback loop could generate oscillations in the RNA level and thus contribute to the fluctuation pattern. However, the stochastic nature of the fluctuations also suggested an absence of synchrony between cells. To investigate the fate of the injected RNA at the level of single cells, we performed an analysis of individual UFEs, using the Awnt-1 transcript as a substrate (Fig. 4). Individual UFEs were rapidly frozen either immediately or 1 h after injection, and RNA was subsequently extracted from each UFE. The Awnt-1 RNA content was then determined by northern blot, using the 18S rRNA to control for total RNA loading. Immediately after injection, the individual RNA levels exhibited a very narrow distribution (n = 11, mean = 100, SD = 15) excluding injection as a major source of variation in the experiments. This also indicated that UFEs contained similar amounts of rRNA, since these results were normalized with the 18S signal. In contrast, a wide dispersion of the Awnt-1 RNA content was observed in the group of UFEs analysed after 1 h incubation at 18°C (n = 31, mean = 70, SD= 68). For 18 UFEs, the signal was significantly lower than in the control group (P < 0.01). In nine UFEs, the signal was similar to that of the control group, while in four out of 31 UFEs it was significantly higher (P < 0.01), ranging from 170 to 265. Thus, at 1 h after injection, the fate of the detected RNA differed widely between eggs, even though they were derived from a single female. Remarkably, four out of 31 UFEs contained significantly more Awnt-1 RNA than they had received by injection. This net increase in Awnt-1 RNA content provides a direct demonstration of the amplification process. Thus, 1 h after injection, depending upon the egg, the amount of RNA could increase by up to 2.6-fold or decrease by up to 6-fold.
Figure 4. Analysis of individual UFEs. (A) The 2.7 kb Awnt-1 RNA was injected into 11 axolotl UFEs (A–K), each being individually and immediately frozen at – 80°C (t = 0). For each egg, the complete RNA content was analysed by northern blot. The blots were successively hybridized with the Awnt-1 and the 18S cDNA probes. (B) A total of 31 UFEs (1–31) were also injected with Awnt-1 RNA and individually frozen 1 h post-injection (t = 1 h). (C) The histogram presents the signal observed at t = 0 and 1 h, after normalization with the 18S rRNA levels, the mean value at t = 0 being 100.
To investigate the effect of these cell to cell differences when working on batches of cells, we injected the Awnt-1 transcript into UFEs and 1 h later we analysed the RNA content in four batches of six eggs and two batches of seven eggs. The results are presented in Figure 5A using the 18S signal to normalize for RNA loading. As expected, the observed values differed significantly between batches (a 1.6-fold difference between the extreme values). We then used the individual values of Figure 4C to simulate the result of grouping cells into batches (Fig. 5B). Importantly, when batches of 6–10 eggs were used, the heterogeneity of the individual cells could not be masked by the averaging process and an important dispersion of the results persisted. In addition, the experimental values of Figure 5A fell precisely within the range of the simulation in Figure 5B, establishing the coherence of our results whether analysed at the individual level or by batches. Thus, when working on batches of cells, the heterogeneity of the cellular behaviour following the injection of in vitro synthesized transcripts could significantly contribute, through the sampling procedure, to the fluctuations in the detection of RNA.
Figure 5. Analysis of the level of Awnt-1 RNA 1 h post -injection. (A) Experimental data for batches of either six or seven UFEs. (B) Simulation of the results of a batch analysis with batches of six, seven, eight or 10 UFEs, using the individual data of Figure 4. In each case, the results of five random selections of individual values are represented by diamonds. The theoretical upper and lower limits are indicated by bars.
DISCUSSION
In this study, we provide additional evidence for the role of an RNA amplification process in the fluctuations of the detection of exogenous RNA injected into axolotl oocytes or UFEs. First, these fluctuations are abolished by the co-injection of RNA chain terminators independently of any modification of the poly(A) tail metabolism. Secondly, complementary RNA molecules can be detected in the injected samples and their abundance increases during the early phase of the kinetics. The increase in abundance of cRNA as well as the cRNA degradation rate are sensitive to RNA chain terminators. Finally, analysis of individual UFEs provides a direct demonstration that an RNA amplification takes place in some cells and is extremely heterogeneous from cell to cell.
On the amplification mechanism
We have observed fluctuations for all the in vitro synthesized transcripts we have injected into axolotl oocytes or UFEs, whether or not the corresponding gene is present in the axolotl genome (9,10). Both the use of RNA chain terminators and the analysis of individual cells indicate that these fluctuations are associated with an in vivo RNA synthesis. We initially excluded the hypothesis of an RNA polymerase II involvement as -amanitin did not inhibit RNA fluctuations (9), and transcription is either limited or absent in these cellular models (5). In this study, we detected cRNA molecules, and their number increased between 2 and 15 min following injection, through a 3' deoxyribonucleotide-sensitive process. One pathway which could lead to an RNA amplification in the absence of the corresponding gene is an RNA replication by an RdRp via a double-stranded RNA intermediate. Our results strongly argue in favour of the implication of an RdRp in the amplification process. Moreover, the observation that no change in migration was observed even for the 0.7 kb fragment from the 3'UTR of the Xenopus c-myc gene indicates that the in vivo synthesized molecules are very similar in length to the injected ones. An RdRp activity has been characterized in tomato and Neurospora crassa (13,27), and homologous genes have been identified in the genomes of plants, fungi, yeast and nematode . It is of note that the cellular RdRp found in N.crassa possesses two distinct activities in vitro, one of which allows the synthesis of full-length copies from a single-strand input RNA (29). However, no RdRp homologue has been detected in the genomes of vertebrates or Drosophila (18). Attempts to detect an RdRp activity in murine and Drosophila cells by the extension of silencing to adjacent sequences (transitive RNAi) have been unsuccessful. It should be noted, however, that although the implication of RdRp genes in RNAi has been established in C.elegans, transitive RNAi is very inefficient in this organism, indicating that the activity of these enzymes is tightly controlled (15). In contrast, a cytoplasmic amplification of ?-globin mRNA has been reported in erythroleukaemia cells (30,31), suggesting the existence of an RdRp in vertebrates.
Coupling between RNA amplification and degradation
Co-injection of RNA synthesis inhibitors (3'dUTP or 3'dATP) abrogates RNA fluctuation presumably by inhibiting a cellular RNA amplification process. These inhibitors also stabilize the injected transcript, suggesting that the RNA amplification is coupled to an RNA degradation mechanism. cRNA molecules were detected among the injected transcripts (Fig. 2C), in agreement with previous observations on the products of in vitro transcription reactions (10,12). It is therefore likely that some double-stranded or partially double-stranded molecules are present among the products of the in vitro transcription reaction. These molecules could be the templates for the amplification reaction. The precise organization of these templates molecules, for instance at their extremities, could provide the basis for the asymmetric synthesis which we observed (Fig. 3). Here we show that the two complementary strands exhibit a differential behaviour towards the observed RNA amplification and degradation processes. A double-stranded RNA amplification has previously been observed to be asymetric in viral RNA replication (32) and for cytoplasmic ?-globin RNA in murine erythroleukaemia cells (31).
The implication of an RdRp and the necessary synthesis of double-stranded intermediates provides a simple framework for a coupling between amplification and degradation, as, in most eukaryotes, including Xenopus and axolotl, the presence of double-stranded RNA molecules induces RNAi, a potent and sequence-specific gene silencing . One prediction of this model is that both strands should be the target of the accelerated degradation since RNAi can target either strand. In this study, we observed that both strands of the Xenopus c-myc 3'UTR are stabilized by 3'dATP, in agreement with the involvement of double-stranded RNA in inducing the accelerated degradation of the transcripts.
On the cellular stochastic variability of RNA fluctuations
One limitation of most RNA studies including the ones reported here is their destructive nature. Consequently, independent samples are used to reconstruct the kinetics by assuming that they reflect an average behaviour. Because the values we observe when studying batches of injected cells are very close to those obtained by simulating virtual batches from individual cell observations (Fig. 5), this study indicates, on the contrary, that an important cell to cell variability exists in the biological process that underlies the observed fluctuations. Although very large batches of cells reduce the fluctuations, our validated simulation approach predicts that batches of up to 10 cells would still exhibit up to 2-fold differences between samples as individual cellular RNA levels really differ by >10-fold. Such a situation suggests that a limiting step exists in the initiation of the coupled RNA amplification/degradation processes: the stochastic pattern would then be due to a very tight control of the factor(s) leading to dramatic differences between cells. RdRp activities have already been implicated in similar stochastic cellular variability. In plants, some RNA molecules can be recognized as ‘aberrant’ and converted by an RdRp into double-stranded RNA, thereby inducing a co-suppression (11). To avoid the induction of interference against bona fide cellular mRNA, the initiation of such a process should be tightly controlled, and this has been proposed as the cause of the stochastic induction of co-suppression observed in plants. Other stochastic events occurring in animals, in particular during early development in axolotl, have been described, such as a remarkable variability of the individual cycle lengths from cell to cell (34) and, after introduction of a G1 phase for the first time at the MBT, a variable duration of this phase from one blastomere to another (35). This variability could account for lengthening of cell cycles and be required for zygotic transcriptions necessary for DNA replication (36).
In addition, the observation that fluctuations can persist for up to 24 h suggests that the heterogeneity among cells is not completely abrogated during this period. This suggests two possibilities. In the simplest model, once one amplification is initiated in a given cell, an accelerated degradation of transcripts should prevail but, from cell to cell, this initiation could proceed stochastically over to a 24 h period. Our results also open up the possibility that, within a cell, several cycles of amplification and degradation could take place over a 24 h time course. Such oscillations in RNA level could indeed occur provided that, during the accelerated degradation phase, not all the amplification templates are eliminated.
In summary, our results further establish the existence of an amplification mechanism in amphibian oocytes and eggs. The demonstration that after 1 h some eggs contain more than twice the amount of injected RNA also establishes that the amplification process can be efficient (of the order of one RNA molecule synthesized per minute assuming that the amplification templates represent 1.5% of the injected molecules). The detection of cRNA molecules indicates that an RdRp may be implicated. Any RdRp-mediated amplification could be expected to induce an RNAi provided that enough double-stranded RNA molecules are generated and are not sequestrated away from the interference machinery. Our observation that the inhibition of RNA synthesis leads to a greater stability of the detected molecules fully supports this idea. Since the initiation of RNA amplification is stochastic under our experimental conditions, our results also stress the existence of a limiting step in the initiation of the amplification phenomenon, presumably reflecting a tight control of the activity of an RdRp. Our work performed with in vitro synthesized transcripts suggests a possible role for RdRp in post-transcriptional regulations during late oogenesis and early development in amphibians, with the possibility that some endogenous maternal mRNA could be the target of this amplification process before the MBT.
ACKNOWLEDGEMENTS
The authors thank M.W. King and C. Séguin for providing the Xenopus c-myc cDNA clone and the Awnt-1 cDNA clone, respectively, and H. Pelczar and L. Dandolo for helpful discussions. This work was supported by grants from the ‘Ligue Régionale Ile de France contre le cancer’.
REFERENCES
Bashirullah,A., Cooperstock,R.L. and Lipshitz,H.D. (2001) Spatial and temporal control of RNA stability. Proc. Natl Acad. Sci. USA, 98, 7025–7028.
Day,D.A. and Tuite,M.F. (1998) Post-transcriptional gene regulatory mechanisms in eukaryotes: an overview. J. Endocrinol., 157, 361–371.
Yaniv,K. and Yisraeli,J.K. (2001) Defining cis-acting elements and trans-acting factors in RNA localization. Int. Rev. Cytol., 203, 521–539.
Signoret,J. and Lefresne,J. (1971) Contribution à l’étude de la segmentation de l’uf d’axolotl. I—Définition de la transition blastuléenne. Ann. Embryol. Morphol., 4, 113–123.
Andéol,Y. (1994) Early transcription in different animal species: implication for transition from maternal to zygotic control in development. Wilhelm Roux’s Arch. Dev. Biol., 204, 3–10.
Andéol,Y., Lefresne,J., Houillon,C. and Signoret,J. (1995a) Differential stability of Xenopus c-myc RNA during oogenesis in axolotl. Involvement of the 3' untranslated region in vivo. Wilhelm Roux’s Arch. Dev. Biol., 205, 182–191.
Andéol,Y., Lefresne,J. and Signoret,J. (1995b) Evidence for a nuclear factor involved in c-myc RNA degradation during axolotl oocyte maturation. Wilhelm Roux’s Arch. Dev. Biol., 205, 192–197.
Newport,J. and Kirschner,M. (1982) A major development transition in early Xenopus embryos: I—characterization and timing of cellular changes at the midblastula stage. Cell, 30, 675–686.
Andéol,Y., Lefresne,J., Simard,C., Séguin,C., Mouton,C. and Signoret,J. (1998) Post-transcriptional control of c-myc RNA during early development analyzed in vivo with a Xenopus–axolotl heterologous system. Differentiation, 63, 69–79.
Montreau,N., Vaur,S., Dautry,F. and Andéol,Y. (2003) Injection of exogenous RNA in amphibian oocytes leads to RNA level fluctuations which are sensitive to cordycepin, an RNA chain elongation terminator. C.R. Biol., 326, 1135–1147.
Vaucheret,H., Béclin,C. and Fagard,M. (2001) Post-transcriptional gene silencing in plants. J. Cell Sci., 114, 3083–3091.
Fire,A., Xu,S., Montgomery,M.K., Kostas,S.A., Driver,S.E. and Mello,C.C. (1998) Potent and specific genetic interference by double-stranded RNA in Caenorhabditis elegans. Nature, 391, 806–810.
Cogoni,C. and Macino,G. (1999) Posttranscriptional gene silencing in Neurospora by a RecQ DNA helicase. Science, 286, 2342–2344.
Smardon,A., Spoerke,J.M., Stacey,S.C., Klein,M.E., Mackin,N. and Maine,E.M. (2000) EGO-1 is related to RNA-directed RNA polymerase and functions in germ-line development and RNA interference in C. elegans. Curr. Biol., 10, 169–178.
Sijen,T., Fleenor,J., Simmer,F., Thijssen,K.L., Parrish,S., Timmons,L., Plasterk,R.H.A. and Fire,A. (2001) On the role of RNA amplification in dsRNA-triggered gene silencing. Cell, 107, 465–476.
Schiebel,W., Haas,B., Marinkovic,S., Klanner,A. and Sanger,H.L. (1993) RNA-directed RNA polymerase from tomato leaves: I Purification and physical properties. J. Biol. Chem., 263, 11851–11857.
Tijsterman,M., Ketting,R.F. and Plasterk,R.H.A. (2002) The genetics of RNA silencing. Annu. Rev. Genet., 36, 489–519.
Stein,P., Svoboda,P., Anger,M. and Schultz,R.M. (2003) RNAi: mammalian oocytes do it without RNA-dependent RNA polymerase. RNA, 9, 187–192.
Armstrong,J.B. and Malacinski,G.M. (1989) Developmental Biology of the Axolotl. Oxford University Press, New York.
Gurdon,J.B., Lane,C.D., Woodland,H.R. and Marbaix,H.R.W.G. (1971) Use of frog eggs and oocytes for the study of messenger RNA and its translation in living cells. Nature, 233, 177–182.
King,M.W., Roberts,J.M. and Eisenman,R.N. (1986) Expression of the c-myc proto-oncogene during development of Xenopus laevis. Mol. Cell. Biol., 6, 4499–4508.
Busse,U. and Séguin,C. (1993) Molecular analysis of the Wnt-1 proto-oncogene in Ambystoma mexicanum (axolotl) embryos. Differentiation, 53, 7–15.
Auffray,C. and Rougeon,F. (1980) Purification of mouse immunoglobulin heavy-chain messenger RNAs from total myeloma tumor RNA. Eur. J. Biochem., 107, 303–314.
Sambrook,J., Maniatis,T. and Fritsch,F. (1989) Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.
Church,G. and Gilbert,W. (1984) Genomic sequencing. Proc. Natl Acad. Sci. USA, 81, 1991–1995.
Neel,H., Weil,D., Giansante,C. and Dautry,F. (1993) In vivo cooperation between introns during pre-mRNA processing. Genes Dev., 5, 2194–2205.
Schiebel,W., Haas,B., Marinkovic,S., Klanner,A. and Sanger,H.L. (1993b) RNA-directed RNA polymerase from tomato leaves: II Catalytic in vitro properties. J. Biol. Chem., 268, 11858–11867.
Bosher,J.M. (2000) RNA interference: genetic wand and genetic watchdog. Nature Cell. Biol., 2, 31–36.
Makeyev,E.V. and Bamford,D.H. (2002) Cellular RNA-dependent RNA polymerase involved in posttranscriptional gene silencing has two distinct activity modes. Mol. Cell, 10, 1417–1427.
Volloch,V. (1986) Cytoplasmic synthesis of globin RNA in differentiated murine erythroleukemia cells: possible involvement of RNA-dependent RNA polymerase. Proc. Natl Acad. Sci. USA, 83, 1208–1212.
Volloch,V., Schweitzer,B. and Rits,S. (1987) Synthesis of globin RNA in enucleated differentiating murine erythroleukemia cells. J. Cell Biol., 105, 137–143.
Barton,D.J., Black,E.P. and Flanegan,J.B. (1995) Complete replication of poliovirus in vitro: preinitiation RNA replication complexes require soluble cellular factors for the synthesis of VPg-linked RNA. J. Virol., 69, 5516–5527.
Tuschl,T., Zamore,P.D., Lehman,R., Bartel,D.P. and Sharp,P.A. (1999) Targeted mRNA degradation by double strand RNA in vitro. Genes Dev., 13, 3191–3197.
Signoret,J. (1977) La cinétique cellulaire au cours de la segmentation du germe d’axolotl: proposition d’un modèle statistique. J. Embryol. Exp. Morphol., 42, 5–14.
Lefresne,J., Andéol,Y. and Signoret,J. (1998) Evidence for introduction of a variable G1 phase at the midblastula transition during early development in axolotl. Dev. Growth Differ., 40, 497–508.
Signoret,J. (1980) Evidence of the First Genetic Activity Required in Axolotl Development. Results and Problems on Cell Differentiation. Springer-Verlag, Berlin.(Nicole Montreau, Sabine Vaur, Francois D)