Feb 16, 1990 - degraded in the embryo after the midblastula transition (MBT), when gene ... The degradation ofthese mRNAs after the MBT was inhibited by ...
MOLECULAR AND CELLULAR BIOLOGY, Aug. 1990, p. 4123-4129
Vol. 10, No. 8
0270-7306/90/084123-07$02.00/0
Copyright C) 1990, American Society for Microbiology
Stability of Maternal mRNA in Xenopus Embryos: Role of Transcription and Translation C. DUVAL,t P. BOUVET, F. OMILLI, C. ROGHI, C. DOREL,T R. LEGUELLEC, J. PARIS,§ AND H. B. OSBORNE*
Laboratoire de Biologie et Gene'tique du Developpement, Centre National de la Recherche Scientifique UA 256, Universite de Rennes I, Campus de Beaulieu, 35042 Rennes Cedex, France Received 16 February 1990/Accepted 10 May 1990
The first 12 cell divisions of Xenopus laevis embryos do not require gene transcription. This means that the regulation of gene expression during this period is controlled at post transcriptional levels and makes Xenopus early development a potentially interesting biological system with which to study the mechanisms involved. We describe here the stability characteristics of several maternal Xenopus mRNAs which are deadenylated soon after fertilisation (J. Paris and M. Philippe, Dev. Biol., in press). We show that these mRNAs were only degraded in the embryo after the midblastula transition (MBT), when gene transcription was initiated. The kinetics with which the deadenylated maternal mRNAs decreased in the post-MBT embryos showed sequence specificity. The degradation of these mRNAs after the MBT was inhibited by cycloheximide but was not affected by dactinomycin. Therefore, the destabilization of these mRNAs does not appear to be initiated by new embryonic gene transcripts. Sequence comparisons of the 3' untranslated region of these mRNAs identified several motifs which may be involved in the posttranscriptional control of these gene products.
The control of mRNA decay is an important factor in the posttranscriptional regulation of several genes whose expression is only transitory, such as proto-oncogenes (8, 15, 17) and inflammatory response genes (12, 23, 27). Several cis-acting mRNA sequence motifs have been isolated and characterized which, when present in in vitro constructions with the sequence of an otherwise stable mRNA molecule, confer instability on these chimeric mRNAs when they are expressed in the cell (reviewed in reference 25). In addition, a shortening of the poly(A) tail has been shown to precede degradation of the body of at least certain polyadenylated mRNAs (4, 7, 14, 16, 18, 24, 29, 30). Indeed, the rate of shortening of the poly(A) tail has been proposed as a major factor in determining the decay rate of these mRNAs (3, 4, 30). However, little is known of the proteins or RNA molecules which promote the linkages between the cisacting instability-conferring sequences, the rate of poly(A) shortening, and the 3'-5' degradation of the mRNA body. Furthermore, it is not clear whether the degradation of the mRNA body occurs automatically once the poly(A) tail is reduced below a critical length or whether the degradation process is regulated by specific cis-acting sequence elements and/or trans-acting factors which can only act after deadenylation of the mRNA. In order to study these interrelationships, it is useful to be able to uncouple the various steps so that they can be studied separately. Early embryonic development in Xenopus laevis is a particularly suitable system for studying the posttranscriptional control of gene expression. During oogenesis, a large amount of mRNA is synthesized, so that the egg * Corresponding author. t Present address: Beckman France, Gagny, France. t Present address: Laboratoire de Gdnetique Moleculaire des Microorganismes, Institut des Sciences Appliqu6s, 69621 Villeurbanne C6dex, France. § Present address: Worcester Foundation for Experimental Biology, Shrewsbury, MA 01545.
4123
contains a stockpile of mRNAs and proteins. This allows the first 12 cell cycles which follow fertilization to occur in the absence of transcription (reviewed in reference 9). Gene transcription in the embryo is only clearly detected after the 12th division and is a marker of the midblastula transition (MBT). We have previously described the adenylation and deadenylation kinetics of several maternal Xenopus mRNAs (21). All of these mRNAs were perfectly stable in the embryo up to the MBT, after which there was some evidence for degradation of the deadenylated mRNAs. This suggested, at least for these mRNAs, that the processes of deadenylation and degradation may be uncoupled in the early Xenopus embryo. In the present study, we have characterized the stability of four maternal mRNAs, three of which are adenylated in the egg and pass into the nonpolyadenylated [poly(A)-] fraction after fertilization (Egl, 2.0 kilobases [kb]; Eg2, 2.2 kb; and Eg9, 6.0 kb) (J. Paris and M. Philippe, Dev. Biol., in press). In addition, the roles of transcription of the embryonic genome and translation of maternal mRNAs in the destabilization of the deadenylated mRNA after the MBT have been determined. Sequence analysis of the cDNAs corresponding to these mRNAs has shown that Egl is a Xenopus mRNA with high homology to the human and yeast cdc2 transcripts (J. Paris, R. LeGuellec, A. Couturier, K. LeGuellec, F. Omilli, and M. Philippe, unpublished data). The complete cDNA for Eg2 mRNA has an open reading frame of 1.2 kb which probably codes for a Ser-Thr protein kinase (C. Roghi, unpublished data); the cloned 1.5-kb Eg9 cDNA has a single polyadenylation signal and corresponds to the 3' extremity of this mRNA. The 5' end of the Eg9 cDNA starts with a 0.5-kb open reading frame. No homology to the amino acid sequences deduced from this open reading frame was found in either the National Biomedical Research Foundation or European Molecular Biology Laboratory (EMBL) data bank (C. Dorel, unpublished data). The other maternal mRNA corresponds to Xenopus ornithine decarboxylase
4124
DUVAL ET AL.
MOL. CELL. BIOL.
(XLODC) (T. Bassez, J. Paris, F. Omilli, C. Dorel, and H. B. Osborne, unpublished data), which initial experiments had shown was stable up to the gastrula stage of embryogenesis.
gels from all the other mRNAs studied, it was then used as an internal standard to normalize the signals from Egl, Eg2, and XLODC mRNAs on the densitometer tracings of the other autoradiograms.
MATERIALS AND METHODS Biological material. Eggs obtained from laboratory-reared X. laevis females were fertilized, and the embryos were allowed to develop as described previously (21). In order to correlate the developmental stage with time, the duration of the first and subsequent cell cycles up to the sixth (64-cell stage) was noted. The timing of the MBT was determined by detection on Northern (RNA) blots (as described below) of two early embryonic transcripts, GS17 (13) and DG42 (26). When dissociation of the blastomers was required, the embryos were transferred to Ca2'- and Mg2+-free medium containing 88 mM NaCl, 1 mM KCI, 2.4 mM NaHCO3, 7.5 mM Tris (pH 7.6) (10) 2 h after fertilization. Cycloheximide and dactinomycin were added to the incubation medium at the desired concentrations from neutralized stock solutions at 10 mg/ml. RNA preparation and Northern analysis. RNA was prepared by the LiCl-urea procedure (1), and polyadenylated [poly(A)+] RNA was isolated by oligo(dT)-cellulose chromatography (2). The RNA contained in the flowthrough volume was taken as the poly(A)- RNA fraction. RNA concentrations were evaluated from the A260 and controlled by gel analysis. RNA samples were fractionated by electrophoresis in 1.2% denaturing agarose gels containing 2.2 M formaldehyde and transferred onto nylon membranes (Hybond; Amersham). Hybridization to 32P-labeled Xenopus cDNA probe was carried out in 50% formamide-1% sodium dodecyl sulfate (SDS)-10x Denhardt solution (0.2% Ficoll, 0.2% polyvinylpyrrolidone, 0.2% bovine serum albumin)-10% dextran sulfate-0.1% PP,-1 M NaCI-0.05 M Tris hydrochloride (pH 7.5) at 42°C for 15 h. The filters were then washed twice in 2x SSC-0.5% SDS at 65°C for 30 min and twice in 0.1X SSC at room temperature for 30 min and autoradiographed at -70°C. The sizes and positions of the cDNAs used as probes, relative to the corresponding mRNAs, were as follows: Egl, 770 base pairs (bp) of the 3' untranslated region of this mRNA; Eg2, 930 bp which cover the whole of the 3' untranslated region and the last 250 bp of the coding region; Eg9, 1,500 bp corresponding to 400 bp of open reading frame and 1,100 bp of 3' untranslated region; XLODC, 1,900 bp which cover the whole of the coding and 3' untranslated regions and 240 bp of 5' untranslated region; p53, 300 bp corresponding to part of the coding region of this 2.2-kb mRNA. Quantification of the autoradiogram signals produced by the different mRNAs and normalization of the data were achieved as follows. First, the membranes were hybridized with a 32P-labeled cDNA probe to the 2.2-kb mRNA coding for the proto-oncogene p53. Since the amount of p53 mRNA in embryos does not change during early development (T. Soussi, These d'Etat, Universite de Paris VII, 1987), an observation that we have previously confirmed (21), the autoradiogram signals (surfaces on densitometer tracings) for this mRNA were used to normalize the amount of RNA loaded at each time point. Next, these membranes were hybridized with the 32P-labeled probe to the 6.0-kb Eg9 mRNA. The autoradiograms were scanned with a densitometer, and by using the normalization obtained with the probe to p53 mRNA, the changes during development in the amount of Eg9 mRNA were calculated. As Eg9 mRNA is clearly separated on agarose-formaldehyde
RESULTS Stability of deadenylated maternal mRNAs. In order to characterize the changes in the stability of the maternal Egl, Eg2, Eg9, and XLODC mRNAs, total RNA was prepared from unfertilized eggs and embryos of different ages. Since previous results had indicated that these deadenylated mRNAs may be degraded after the MBT (which occurs about 7.5 h after fertilization), samples of embryos were taken at hourly intervals from 6 to 10 h postfertilization. These RNAs were analyzed by electrophoresis in denaturating agarose gels, transferred to nylon membranes, and hybridized with 32P-labeled cDNA probes corresponding to the four maternal mRNAs Egl, Eg2, Eg9, and XLODC and to two mRNAs synthesized just after the MBT, DG42 (26) and GS17 (13). These latter two probes permitted the time at which embryonic gene transcription was initiated to be determined. The autoradiograms from four different agarose gels are shown in Fig. 1A. The intensity of the autoradiogram signals does not reflect the relative abundance of these mRNAs, as the specific activities of the labeled cDNA probes were not identical. On overexposed autoradiograms the embryonic DG42 (3.8 kb) and GS17 (0.75 kb) transcripts were first detected in RNA isolated from 8-h-old embryos, indicating that the MBT occurred between 7 and 8 h postfertilization in these experiments. This is in agreement with the time required for the first 12 divisions as calculated from the timing of the first 6 divisions (Fig. 1B). The Eg9 cDNA probe sometimes hybridized to an mRNA slightly larger than the 6.0-kb transcript (lanes E and 4 h). The irreproducibility of this signal indicates that it probably corresponds to another mRNA with only partial sequence homology to the Eg9 cDNA probe. The densitometer quantification for Eg2, Eg9, and XLODC mRNA, performed as described in Materials and Methods from several sets of autoradiograms, is given in Fig. 1B. The amount of the 2.4-kb XLODC transcript was constant in the embryo over the period shown. The mRNAs for Egl (2.0 kb), Eg2 (2.2 kb), and Eg9 (6.0 kb) were stable up to the MBT, confirming our previous results, and were subsequently degraded. This quantification of the autoradiograms showed that the degradation of Eg2 mRNA started concomitantly with the MBT and that this mRNA had decreased by about 50% at 2 h after the MBT. During this period, Eg9 mRNA had only decreased by about 25%. The apparently greater stability of Eg9 mRNA in the upper panel of Fig. 1A relative to that shown in the second and lower panels is due to an overexposure of the autoradiogram necessary to see the signals for the Egl and GS17 mRNAs. Such overexposed autoradiograms were not used for quantification. Egl mRNA, in general, gave signals which could not be accurately quantified (Fig. 1A, upper panel), although occasionally clean autoradiograms were obtained (see Fig. 2A). Qualitatively, Egl mRNA appeared to behave like Eg2 mRNA. However, as Egl mRNA could not be accurately and reproducibly quantified, the values are not shown here. We have no explanation for this variability. These results clearly confirm that the deadenylation and the degradation of at least certain maternal mRNAs are effectively uncoupled in the early Xenopus embryo. Only after the MBT does degradation of these previously deade-
- bXS-
STABILITY OF MATERNAL XENOPUS mRNAs
VOL. 10, 1990
4125
A
*WWWW
A
Eg 9
GS 17
1%
_*
-0
_ _
Egg
.*
w
t *@@@*W@ Co oc
Ow a*
we
A
-
E 4 6 7 8
2 4
16
MBT
lLJTh ll I
B 100
[
-T
rT
-
T~~r
--
T
S \*tT T
\
1\T
-
ODC
I
S
Eg 9
I \ T _1 \
60
U -I
Eg 2
z z
0
I
0
3
6
_ -~
9
TIME h! FIG. 1. (A) Northern analysis of maternal and embryonic transcripts during early development. Total RNA was isolated from unfertilized eggs (lane E) and embryos of different ages (4, 6, 7, 8, 9, and 10 h after fertilization); 20 ,ug of each of these samples was separated in agarose-formaldehyde gels, transferred to nylon membranes, and hybridized in stringent conditions with the 32P-labeled cDNA probes indicated on the right. The relevant parts of the autoradiograms from four different gels are shown. (B) Quantitative analysis of Eg2, Eg9, and ornithine decarboxylase (ODC) transcripts during early development. The relative amounts of these transcripts were calculated from densitometer scans of autoradiograms similar to those shown in panel A but exposed for different times. The timing of the first six divisions and the MBT, defined relative to the sample in which GS17 and DG42 transcripts were first detected in overexposed autoradiograms, are marked above the figure. Values are the averages + range from two separate experiments. F, Fertilization.
nylated mRNAs occur. These results imply that factors necessary for the degradation of these mRNAs are absent from or inactive in the pre-MBT embryo and that they are only synthesized or activated after the MBT. Role of transcription and translation. One of the major events that occurs at the MBT is the initiation of the transcription of many embryonic genes. Therefore, the deg-
Eg2 -
w
Egg9
_-
Eg 1
Egg ODC Eg 9
-
9 10
Eg I
_
7
8
Time
Eg g DG 42
9 10
64
n
Egl
T A A
@ _ _ _ _ E
Ih
Time
F
.
*W
*.
Eg 1
M.
B
_ _
(h)
FIG. 2. (A) Northern analysis of maternal transcripts in dissociated embryos. Total RNA was isolated from unfertilized eggs and embryos of different ages (7, 8, 9, and 10 h postfertilization) that had been transferred to Ca2+- and Mg2+-free medium 2 h after fertilization. Northern analysis was then performed as described in the legend to Fig. 1. The relevant parts of the autoradiograms from three different gels are shown. Similar results were obtained in two separate experiments. The time course of development was the same as that in Fig. 1. (B) Control of Eg9 deadenylation in dissociated embryos. Total RNA isolated from 6-h-old normal (upper) and dissociated (lower) embryos was separated into poly(A)+ and poly(A)- fractions by oligo(dT)-cellulose chromatography. Total RNA (T) (5 p.g) and poly(A)- RNA (A-) (5 F±g) and 100 ng of poly(A)+ RNA (A') were subjected to Northern analysis as described in the legend to Fig. 1 with 32P-labeled cDNA probes for Egl and Eg9 mRNAs.
radation-specific factors (or their activators) which are present in the embryo after the MBT could be coded for by these new transcripts. Alternatively, these factors (or their activators) could be coded for by maternal mRNAs whose translation is initiated at or just prior to the MBT. To test these various possibilities, experiments were performed in which either transcription or translation in the embryo was inhibited. In order to treat large numbers of embryos, "dissociated" embryos were incubated in the appropriate drug-containing medium. The blastomeres of embryos incubated in Ca2`- and Mg2'-free medium dissociate from each other but remain contained within the vitellin membrane (26). In these dissociated embryos, metabolic precursors and inhibitors enter the blastomeres (19), presumably through the exposed surfaces of the internal plasma membrane. To verify that the disruption of the cell-cell contacts in the dissociated embryos did not change the stability of the mRNAs studied, total RNA was purified from eggs and embryos of different ages, which had been transferred into Ca2`- and Mg2e-free medium 2 h after fertilization. The various RNA samples were subjected to Northern analysis with 32P-labeled probes to the maternal mRNAs for Egl, Eg2, Eg9, and XLODC. The timing of the MBT was also verified by using 32P-labeled probes to DG42 and GS17. The autoradiograms obtained (Fig. 2A) showed that dissociation of the blastomeres in the embryo affected neither the stability of Egl, Eg2, or Eg9 prior to the MBT nor the stablity of XLODC mRNA during the total 10-h postfertilization period studied. This observation was confirmed by densitometer quantification of the autoradiogram (data not shown). The apparent decrease in all the mRNAs between the samples from eggs and 7-h postfertilization samples is due to an overloading of the egg samples. As previously described by Sargent et al. (26), DG42 expression (between 7 and 8 h postfertilization) was unaffected by dissociation of the blastomeres. GS17 expression was initiated concomitantly with
4126
DUVAL ET AL.
MOL. CELL. BIOL.
T
A,:
a
>-\
s-.
(
D
6 10 10 10 1010 25 0-
-
Ti me 3f , .- ;:r?i *
Lft
-
C
