Abstract
The expression of the proto-oncogene c-myc during oogenesis and embryonic development was followed by in situ hybridization using a cytological protocol adapted to amphibian embryos. The c-myc RNA was highly expressed in the cytoplasm of young oocytes and was further diluted during oocyte growth without specific localization. From the neurula stage on, new myc transcripts were detected and the whole embryo appeared positive with antisense myc RNA probes relative to control sense RNA probes. In addition, a spatial localization of high levels of the transcript was also observed in specific areas of the developing embryo, including the epidermis, gill buds, optic vesicles and lens placodes. These observations might indicate a specific role of the c-myc gene during the differentiation of these tissues. Alternatively, this high level of myc expression might prevent such tissues from entering into terminal differentiation during the growth of the embryo.
Introduction
Expression of c-myc proto-oncogene has been correlated with cell proliferation in a number of experimental systems (for reviews, see Heldin & Wester-mark, 1984; Bishop, 1985). We previously reported the characterization of the Xenopus laevis c-myc cDNA sequence and showed that this proto-oncogene was closely homologous to its mammalian counterpart (Taylor et al. 1986). Northern blot analysis showed that this gene was expressed as a stable maternal mRNA during oogenesis. After fertilization, the myc mRNA level decreased until the gastrula stage, when it reached a minimum value per embryo. At the neurula stage, the myc mRNA content increased again and we assumed that it was a consequence of new transcription in the embryo (Taylor et al. 1986; King et al. 1986). This biochemical analysis of the steady state myc mRNA level was done on whole embryos and did not provide information about spatial localization of the transcript. Localization of specific transcripts during embryogen esis can provide information on early cell determination before any morphological change occurs. Thus specific localization of maternal mRNA during oogenesis has been recently demonstrated (Rebagliati et al. 1985; King & Barklis, 1985; Melton, 1987) and might, at least partly, account for mesoderm induction in the developing Xenopus embryo (Weeks & Melton, 1987; Kimelman & Kirschner, 1987). We used in situ hybridization to assay for localization of myc transcripts during oogenesis and embryonic development. For that purpose, we adapted to amphibian embryos a protocol designed for cytological studies in yolk-laden insect eggs and embryos (Pétavy, 1985). We confirm that the myc protooncogene is highly expressed in early oocytes and show that its transcript is uniformly distributed in the cytoplasm. From stage-III to stage-VI oocytes (Dumont, 1972), the myc mRNA content is diluted in the expanding cytoplasm and does not exhibit a detectable preferential localization. From the neurula stage on, when new myc transcripts are detected, we show that their distribution is not uniform throughout different tissues. The correlation between the localized c-myc expression in a developing embryo and the putative role of c-myc in cell proliferation and differentiation is discussed.
Materials and methods
Collection of oocytes and embryos
Xenopus laevis adults imported from South Africa (South Africa Farms, Fish Hoek) were accommodated and fed as described by Gurdon (1967). Oocytes, eggs and embryos were obtained as previously described (Taylor et al. 1986). Unless specified, in situ hybridizations were performed with albino Xenopus embryos to avoid confusion between pigment granules and silver grains.
Preparation of cut sections
Preliminary experiments using standard paraffin embedding for oocytes and early embryos were not successful in our hands. The technique, originally described by Pétavy (1985) for yolk-laden insect eggs and embryos, was adapted to Xenopus embryos. It gave satisfying results for both cytological studies and in situ hybridizations. The oocytes and embryos were fixed with freshly prepared 4 % paraformaldehyde in 0·5× PBS, pH7-4 for Ih at 4°C, washed three times in l× PBS for 30min each. Oocytes and embryos were then closely coated with 1·4 % (w/v) agarose at 60°C according to Pétavy (1985). Specimens were cast in molds filled with the agarose solution. After chilling, the agarose gel blocks were trimmed as little cubes containing specimens. They were then treated by sequential immersion in 50% ethanol (2×20min); 70% ethanol (2x20min); 50% ethanol, 50% ethoxy-ethanol (2×20 min); 66% ethanol, 33% 2-ethoxy-ethanol (2×20min), pure 2-ethoxy-ethanol (4×30min), and butanol-1 (several changes of 24h each). The specimens were placed 20min at 50°C in a 1:1 mixture of butanol-1 and ester wax (Gurr). This final embedding medium was then processed through five baths of pure ester wax at 50°C, under increasing vacuum in order to evaporate butanol-1 and facilitate infiltration with ester wax. Blocks were then cast in molds containing melted ester wax and stored at 4°C. 7·5 pm sections were mounted on poly-L-lysine-coated slides. After removal of ester wax by toluene and rehydration in decreasing ethanol concentrations, they were stained with 0·1 % toluidine blue.
RNA probes
A 680 bp Hindi-PvuII fragment from myc, including part of Exon II and Exon III (Taylor et al. 1986) was subcloned into the transcription vector Bluescribe (Stratagene) in both orientations, linearized with Bam HI and transcribed with T3 RNA polymerase (Stratagene). Thus, both sense and antisense probes were synthesized. The radioactive nucleotide was 35S-U I P, 400CimM−1 (Amersham). The probes had a specific activity of 4-6×108cts min−1μg−1 and were purified as described by Angerer et al. (1987a,b Template DNA was removed by digestion with RNase-free DNase (Promega). Reduction of probe fragment length was performed as described by Angerer et al. (1987a,b), to an average length of 100 nucleotides as determined by electrophoresis through a 6% denaturing polyacrylamide gel.
DNA probes
The 680 bp Hindl-Pvull myc insert separated by acrylamide gel electrophoresis was used for random-priming labelling according to Feinberg & Vogelstein (1983) using 35S-dATP, 400 Ci mmole−1 (Amersham). Reduction of probe length was carried out using digestion with HinQ and Pstl. Control probes were derived from pUC18 digested by Haell and Bgll and labelled by random-priming. Probe length was further reduced by digestion with Pvul. Fragments were purified by filtration on Biogel P100. The DNA probes had a specific activity of 3-5×108 cts min−1μg−1 and the length of fragments ranged from 70 to 200bp.
In situ hybridization
In situ hybridization was carried out on 7·5 pm sections following Cox’s et al. (1984) and Angerer’s et al. (1987) protocols, with minor modifications.
Prehybridization
Prehybridization treatment included removal of ester wax, treatment with 0·2N-HC1 for 20 min at room temperature, digestion with l μg ml−1 proteinase K (Merck) in 20mm-Tris-HCl, pH 7·4, 2mm-CaCl2 for 15 min at 37°C, treatment with 0·25% acetic anhydride in triethanolamine buffer, followed by two rinses in 2×SSC and dehydration in ethanol.
Hybridization
For DNA probes, hybridization was performed at 30°C in 50% formamide, 4×STE buffer pH 7·4 (0·6m-NaCl, 10 mm-Tris-HCl, pH7·4, 1mm-EDTA), 1×Denhardt’s, 8% dextran sulphate, 10mm-DTT, 100μgml−1E. coli sonicated DNA, 100 μgml−1 tRNA, 0-06μgml−1 35S-labelled DNA probe (5−7×105ctsmin−1 per slide). Slides were kept overnight at 30 °C in a moist chamber (Gall & Pardue, 1969). For RNA probes, hybridization was performed at 48°C in 50% deionized formamide, modified STE buffer (0·3m-NaCl, 10mm-Tris-HCl, pH7·5, 5mm-EDTA), l×Denhardt’s, 8% dextran sulphate, 10mm-DTT, 100·g ml−1 yeast tRNA, 0×06 μg ml−1 35S-labelled RNA probe (5− 7× 105ctsmin−1 per slide), for 16h.
Post-hybridization washes
For DNA probes, slides were washed twice in 50% formamide, 4× STE, 10mm-DTT at 30°C, twice in 50% formamide, 2× STE, 10mm-DTT at 30°C, once in 50% formamide l× STE, 10mm-DTT at room temperature. For RNA probes, slides were washed twice in 2× SSPE at room temperature for Ih, once in 2× SSPE, 50% formamide, 10mm-DTT at 60°C for 30min, treated with 20 μg ml−1 RNAse A and li.u.ml−1 RNAse T1 in 4 × SSPE at 37 °C for 30 min, washed in 4 × SSPE at 37 °C for 15min, 2× SSPE for Ih at room temperature, l× SSPE, 0-5× SSPE and 0-1 × SSPE for 30 min each at room temperature.
Autoradiography
Slides were dipped in Kodak NTB2 emulsion and exposed for 5 to 19 days at 4°C. After development, they were stained for 20− 30 s with 0· 04% Giemsa in lOmm-potassium phosphate buffer pH 7· 3.
Quantification of autoradiographic silver grains
Images of the sections formed by a Leitz Orthoplan microscope were scanned by a Plumbicon camera and analysed by a Quantimet 720 (Cambridge Instruments) connected to a PDP 11/34 computer (Digital equipment). The number of silver grains was counted over an area of the sections. The diameter of the oocytes was determined by encircling the sections with a lightpen on the screen.
Results
C-myc RNA is uniformly distributed in the cytoplasm of young oocytes and is diluted during oocyte growth
Precise localization of transcripts by in situ hybridization requires that good preservation of cell and tissue structure be compatible with mRNA accessibility and stability. The use of paraffin embedding has been found suitable for most tissue sections (Angerer et al. 1987; and our unpublished observations). Although used with success with oocytes and Xenopus embryos (Dworkin-Rastl et al. 1986; Melton, 1987; Kintner & Melton, 1987; Jamrich et al. 1987) this technique often gave poor oocyte sections in our hands. As yolk platelets appeared responsible for these drawbacks, we adapted for amphibians (Materials and methods) a technique originally described for insect yolk-laden eggs and embryos (Pétavy, 1985). This gave satisfactory results for all stages of oogenesis and embryogenesis (Figs 1, 4 and 5), preserving the integrity of both nuclear and cytoplasmic substructures such as nucleoli, yolk platelets and membranes (Fig. 1 and its legend).
Fig. 2A,B shows oocytes after hybridization with a 35S-DNA probe and autoradiography with a 7-day exposure. Stage-I oocytes show no signal from the nucleus but a strong signal over the cytoplasm, indicating a high concentration of myc mRNA. In all experiments, hybridizations with a control pUC18 DNA probe were performed in parallel and no signal above background was observed (Fig. 2G). Fig. 2C and D show stage-I oocytes after hybridization with a 35S-labelled antisense RNA probe. Although the specific activity and the exposure time were the same as for the experiments with the DNA probe, the signal is much higher, due to the higher sensitivity of RNA probes as compared to DNA probes. In all cases, hybridizations were performed in parallel with a control sense RNA probe; Fig. 2H shows the background signal with that control. With both DNA and RNA probes, the labelling was intense in pre-vitellogenic stage-I oocytes (Fig. 2A− D) and decreased in later stages (Fig. 2E,F). From stage II on, oocytes accumulate yolk while their diameter increases from 300 μm to 1200 μm. The concentration of myc mRNA decreased during this period of oocyte growth. However, as quantified and discussed below, the total number of myc transcripts per cell remained relatively unchanged. In all cases, myc RNA appeared uniformly distributed in the oocyte cytoplasm and no specific localization was detected whatever the stage of oocyte growth.
Quantification of c-myc-RNA autoradiographic signals during oogenesis
The signal observed on a series of sections hybridized with c-myc was quantified during oocyte growth, as described in Materials and methods. RNA probes reproducibly gave higher signals as compared to DNA probes (Fig. 3A) as observed by Cox et al. (1984). This difference might reflect the low sensitivity of DNA probes at low transcript density. Thus it is likely that when the silver grains density is close to the background level, the number of transcripts is underestimated. We also noted, in repeated experiments, that the background level is steady when the exposure time is increased. Values of 9× 10−3 grains per μ m2 and 8× 10−3 grains per μm2 were registered after 7- and 19-day exposures, respectively, with hybridizations using RNA probes. This observation might indicate that latent grains in the photographic emulsion are the source of most of total background (Cox et al. 1984). A stage-VI oocyte (109μ m3) contains 4· 8× 106 myc transcripts (Taylor et al. 1986), or 0· 48 transcripts per 100 μm3. In cut sections from such oocytes, a mean value of 1 grain per 100 μ m2 was registered with 7-day exposures with probes of 4 to 6× 108ctsmin− 1μ g−1. For a density of 1 transcript per 100 μm3 cell and with a probe of 2·5×109cts min− 1μg− 1, 1·6 grains per 100 μm2 per day would have been obtained. In another study using a 500 nucleotide probe (specific activity 2·5×106ctsmin− 1μg−1) versus 680 nucleotides in our case, the same calculation suggested a maximum of 1·8 grains per 100 μm2 per day (Angerer et al. 1987), which is similar to the sensitivity of the technique used here. Fig. 3A shows that the grain density reached a maximum level in small oocytes (100μm diameter). From that stage until stage-III oocytes (450·600μm), the grain density decreased about fourfold. During that period the volume of the oocyte increased from 0-523 to 113 nl, and so the actual number of silver grains per oocyte increased (Fig. 3B), and reached a near plateau level in oocytes above 600 μm diameter. In Fig. 3B, the number of grains was compared to the level of myc RNA as determined by slot blot and Northern blot quantification (Taylor et al. 1986). Accumulation of myc RNA per oocyte was first observed in 100 μgm diameter oocytes and reached a maximum level at stage III of oocyte growth (450−600 μm diameter). The in situ hybridization data were thus in agreement with the biochemical data previously reported,
C-myc RNA is highly expressed in specific tissues of the embryo
For all experiments described below, RNA probes were used and the results obtained have been repro duced in five independent experiments. The total amount of myc mRNA per embryo progressively decreases after fertilization (Taylor et al. 1986) and reaches a minimal value at the gastrula stage. During the early cleavage stages, the grain density of the whole embryo was low but a faint positive signal above background was reproducibly observed only at the morula stage in blastomeres of low yolk content in the animal hemisphere (not shown). We have not been able to detect any positive signal in the gastrula embryo. We previously showed that at neurula stage, the average myc RNA content per average embryonic cell started to increase (Taylor et al. 1986). Using in situ hybridization, a positive signal was also observed for the neurula stage (Fig. 4). The whole embryo appeared faintly positive but the grain density was higher in posterior epidermis, neural tube, notochord and endoblast (see legend to Fig. 4). At neurulation, a number of maternal RNAs that had been degraded during the cleavage stage are again synthesized (Shiokawa et al. 1981; Dawid et al. 1985). In addition, new specific genes are expressed (Dworkin et al. 1984). The average cell cycle is markedly lengthened as compared to the early blastula stage but the rate of cell division is still high in the embryo (Newport & Kirschner, 1982). The background of positive labelling observed was therefore expected, but the heterogeneity of labelling that was reproducibly detected was not foreseen. This heterogeneity was accentuated at later stages of development.
At the tailbud stage, a striking pattern of distribution of silver grains was observed (Fig. 5). Once again the whole embryo was positive as compared to control sense probe hybridization (not shown) but the truncal and caudal epidermis as well as the buds of the external gills displayed a higher concentration of myc RNA (Fig. 5A,D). Strong signals were also observed in the optic cups and lens placodes (Fig. 5A–C,E,F) which result from a thickening of the epidermis induced by the optic cup. No differentiated cell can be seen in these areas. The labelling also extended over the cephalic epidermis (Fig. 5E). The average grain density was about sixfold higher on the optic cups and lens placodes than on endoderm sections. Thus the concentration of myc RNA was sixfold higher in the optic cup cells. It is difficult to correlate these values with a given transcript density. As a tentative approach, however, oocyte sections were always mounted on the same slides as embryo sections so that hybridization conditions would be identical. Since the number of transcripts per oocyte is known quite acccurately (Taylor et al. 1986), comparison with the grain density on oocyte sections and tailbud sections enabled a rough estimate of the number of transcripts per cell volume in optic cup cells. This value corresponded to more than 100 transcripts per cell of the optic cup tissue.
Discussion
We previously reported that myc mRNA was uncoupled from cell proliferation during oogenesis and was accumulated as a stable maternal mRNA. The measure of the steady-state level of this transcript during development showed a degradation of 90 % of the stored myc mRNA during the early cleavage stage, followed by new transcription from the neurula stage onward (Taylor et al. 1986). These observations did not exclude the possibility of different patterns of variation within cells or between tissues. Therefore, in situ hybridization was applied to examine the in situ localization of myc transcripts during development.
It has been proposed that the localization of maternal determinants in unfertilized eggs may direct development of the early embryo (review by Davidson, 1987). We show here that myc RNA is uniformly distributed in the cytoplasm of oocytes, at all stages of their growth (Fig. 2), and thus cannot be considered as a localized determinant. The data obtained by quantification of autoradiographic silver grains are in accordance with biochemical data previously published (Taylor et al. 1986): the total number of grains per oocyte, which is related to the total number of myc transcripts, increases from young oocytes (100 μm diameter) through stage-III oocytes (450-600 μ m diameter) when it reaches a maximum level (Figs 2, 3). This myc RNA pool does not expand further during late oogenesis and is diluted in the cytoplasm of the still growing oocyte. Morula-stage embryos were weakly positive and we have not been able to detect a positive signal at the gastrula stage. This is in accordance with the decrease in myc RNA content induced by fertilization as previously reported (Taylor et al. 1986).
The observation of a high level of steady-state myc RNA in oocytes was interpreted as resulting from a specific regulation of the myc gene (Taylor et al. 1986). During the early cell cleavage stage of Xenopus development, the embryo divides every 30 min, in absence of detectable expression which first appears at the midblastula stage (Bachvarova & Davidson, 1966; Newport & Kirschner, 1982). The correlation between c-myc mRNA levels and cell proliferation is the most characteristic feature of the regulation of c-myc expression (Kelly et al. 1983; Cole, 1986 for a review), at least partly through post-transcriptional control (Dani et al. 1984; Dony et al. 1985; Knight et al. 1985). Thus, the accumulation of c-myc RNA during oogenesis could be interpreted as an uncoupling of c-myc expression and cell proliferation necessary to provide a myc RNA stockpile before the active cell proliferation period induced by fertilization. The absence of localization of this mRNA pool in the oocyte is in agreement with a role of c-myc in proliferation, as well as the observation of steadystate levels of c-myc expression in whole embryos from the neurula stage on since the embryo is growing at that stage.
The detection of high levels of myc RNA above this constitutive expression, spatially restricted to epidermis, optic vesicles and lens placodes, is more difficult to interpret. A spatial localization of c-myc expression was detected as early as the neurula stage and was more pronounced at the tailbud stage. It has been suggested that the transcripts appearing only after the gastrula stage are tissue-specific in tadpoles and code for differentiation-specific proteins (Dworkin et al. 1984). On the other hand, since most maternal transcripts are destroyed by the neurula stage, the high level of new transcription seen at that stage merely represents the replacement of homologous maternal transcripts (Shiokawa et al. 1980; Dawid et al. 1985). Indeed, the quantitative dominance among zygotic genes of maternally active genes is a general feature of most mid-stage embryos (Davidson, 1987 and references therein). If elevated myc RNA level is coordinated with rapid cell proliferation, then clusters of labelled and unlabelled cells might identify tissues with a high division rate at a given stage of embryogenesis. The different in situ signals between endodermic and epidermic cells of the tailbud might support such an interpretation (Woodland & Gurdon, 1968). The stronger signal observed in lens placodes is more difficult to reconcile with this supposition. Different studies have shown that constitutive c-myc expression has a major influence on cell differentiation. Thus c-myc expression in murine erythroleukemia cells blocks their terminal differentiation although their commitment is not altered (Coppola & Cole, 1986). Similarly, quail myogenic cells infected with a retrovirus fail to differentiate into myotubes (Falcone et al. 1985). A down regulation of c-myc expression has been correlated with cell differentiation both for hematopoietic cells (Westin et al. 1982) and for teratocarcinoma cells (Dony et al. 1985; Dean et al. 1986). Overall, these reports indicate that constitutive myc expression is incompatible with terminal cell differentiation. Microscopic examination of the tailbud shows that in the cephalic areas, optical vesicles grow rapidly to give rise to two large eyes in the young tadpole. However, at this stage, no cytological sign of differentiation is evident. Thus, the high level of myc-mRNA localized in these cells may prevent them from entering terminal differentiation in the still growing embryo.
Alternatively, c-myc expression may be a positive effector for the differentiation of this specific tissue. In a few instances, a transient increase in the c-myc RNA level was observed during cell differentiation (Lachman & Skoultchi, 1984; Greenberg & Ziff, 1984). More recently, a specific and transient elevation of c-myc mRNA was reported in differentiating chick lens cells and this has been correlated with a withdrawal from the cell cycle (Nath et al. 1987). Our observations on c-myc expression in the developing Xenopus embryo, together with these reports, suggest that, in addition to a specific role in cell proliferation, expression of the c-myc gene might have a distinct role in cell differentiation.
ACKNOWLEDGEMENTS
We thank C. Yanicostas for helpful discussions during the first stage of this work and Peter Brooks for critical reading of this manuscript. This work was supported by grants from the Association pour la Recherche Scientifique sur le Cancer and from INSERM. M.V.T. was supported by an EMBO postdoctoral fellowship.