A biochemical and ultrastructural study of RNA in yolk platelets of Xenopus gastrulae

The structure, chemical composition and role of yolk platelets during amphibian embryogenesis has been a subject of active investigation and lively controversy for several years. The ultrastructure of the yolk platelet was revealed by Ward (1962) and Karasaki (1963) to consist of a main body component with a distinct crystalline matrix, a superficial layer surrounding this main body, and a limiting membrane. Gross and Gilbert (1956) reviewed the chemical composition of the platelet in amphibians ; however, Lanzavecchia & Le Coultre (1958) were among the first to propose that these structures contained nucleic acids. Rounds & Flickinger (1958) suggested that nucleic acid found in the ‘yolk fraction’ of mesodermal cells during primary induction participated in early morphogenesis. However, in an effort to resolve and analyse yolk platelet components, Wallace (1963 a) concluded that amphibian yolk platelets in situ contained no RNA, but was uncertain of the content of DNA. Recently, Baltus, Hanocq-Quertier & Brachet (1968) have reported double-stranded, linear DNA in amphibian yolk platelets which they suggest plays a role in synthesis of enzymes for breaking down platelets during development. If this hypothesis is valid, then might yolk platelets contain RNA in addition to DNA? We chose techniques of electron-microscopic autoradiography, phenol extraction of RNA, and separation of RNA molecular species by means of sucrose-gradient centrifugation and acrylamide-gel electrophoresis in an attempt to answer this question.

Fertilized eggs were obtained from the South African clawed frog, Xenopus laevis, after the method of Gurdon (1967). Gastrulae (stage 11 ; Nieuwkoop & Faber, 1956) were carefully divested of surrounding jelly and fertilization coat in 2·0 % cysteine neutralized to pH 7·8 with NaOH and containing 0·2 % papain.

Entire embryos were preserved in a variety of fixatives (veronal acetate buffered osmium tetroxide, phosphate buffered osmium tetroxide, phosphate buffered formaldehyde, phosphate buffered glutaraldehyde, cacodylate buffered glutaraldehyde, and veronal acetate buffered glutaraldehyde). Best preservation was observed in specimens fixed with 1·75 % glutaraldehyde in 0.1 M phosphate buffer (Polysciences, Inc.) for 1 h at 4 °C, postfixed with 2·0 % osmium tetroxide in 0.1 M phosphate buffer at 4 °C for 1 h, rapidly dehydrated through an ethanol series to propylene oxide and embedded in Epon 812. Thin sections (mounted on uncoated grids) were stained for 1 h in saturated aqueous uranyl acetate at 35 °C, for 5 min in alkaline lead citrate at room temperature and examined in an Hitachi HU-11C electron microscope.

To determine the presence of RNA chemically, yolk platelets were isolated from gastrulae using modifications of the method of Wallace & Karasaki (1963). The following procedures were performed-in a cold room (5 °C). Embryos were washed with Niu-Twitty’s (1953) solution containing penicillin (1000 i.u.) and streptomycin (0·01 mg/ml) and were transferred to cold 0·25 M sucrose-5·0 % (w/v) polyvinyl pyrrolidone (PVP). Gradients were prepared by placing 20 ml of 1·0 M sucrose-5·0 % PVP in round-bottomed 35 ml Pyrex centrifuge tubes (Sorvall rotor SS-34). One thousand embryos were gently homogenized in 60 ml of the 0·25 M sucrose-5·0 % PVP medium, the grey homogenate then being divided into four equal aliquots and carefully layered on to the surface of each gradient.

Tubes were initially spun in a Sorvall RC2-B refrigerated centrifuge (0 °C) for 10 min (F_max= 600 g). This procedure produced a yellowish pellet of yolk platelets, a thin superficial layer of contaminating pigment on top of the pellet, and a grey supernatant. After decanting the supernatant (which was saved for analysis), walls of the tubes were rinsed with 0·25 M sucrose-5·0 % PVP, the pellets resuspended by gentle manual shaking, and centrifuged again at 590 rev/min for 10 min (F_max= 500 g). This procedure was repeated four times, the final pellet containing neither pigment nor contaminating particulate material (for some experiments, final resuspension was accomplished by shaking in a vortex mixer). Portions of each pellet after each spin were fixed for examination in the electron microscope.

To determine the localization of RNA in yolk platelets, gastrulae were incubated in [5-³H]uridine (New England Nuclear Corp.; specific activity 13·12 Ci/mmole, in sterile aqueous medium) for 2 h, and were transferred into unlabeled uridine (uracil riboside, Nutritional Biochemicals Corp. ; 10 mg/ml in sterile Niu-Twitty’s solution) for 1 h. Some embryos were prepared for electron-microscopic autoradiography after the method of Caro (1964), whereas others were homogenized, their yolk platelets isolated as described, and the pellets prepared for autoradiographic examination in the electron microscope. Control specimens (intact embryos and platelet pellets) were fixed in 3:1 ethanol:acetic acid and embedded in paraplast. These were sectioned at 6μm, mounted on gelatin-subbed slides, and subjected to one of the following procedures: (1) 3 h in DNase (Worthington, ×2 crystallized, 0·5 mg/ml in 0·1 M phosphate buffer with 0·1 % phenol as preservative); (2) 3 h in RNase (Worthington, 13 mg/ml in 0·1 M phosphate buffer with 0·1% phenol as preservative); (3) 3 h in buffer and in water, both without enzyme; and (4) 1 N-HC1 at 60 °C for 12 min. These slides were treated with cold 5·0 % trichloracetic acid for 15 min, rinsed in 80 and 95 % ethanol and allowed to dry.

Random samples of silver grains over control sections were counted using a grid (one square = 0·4 cm²) drawn on a transparent plastic sheet. All silver grains over yolk platelets within 100 adjacent squares of the grid were counted. Autoradiographs used for counting were at the same magnification.

Isolated yolk platelets were analysed by the following procedures :

After isolation of yolk platelets, pellets were resuspended in 5·0 ml of saline buffer (0·24 M NaCl, 0·01 M MgCl₂, 0·01 M Tris, pH 5·0) to which was added 0·5 ml of 10% sodium dodecyl sulfate (SDS), 0·5 ml of 2·5% bentonite in 0·01 M sodium acetate, and 5·0 ml of water-saturated phenol containing 0·1 % 8-hydroxyquinoline (Cline, 1966). After manual shaking for 20 min at 0 °C, phases were separated by centrifuging at 800 g for 10 min at 5 °C. The aqueous layer was saved whereas the phenol layer was re-extracted with 5·0 ml saline buffer by shaking for 5 min at 50 °C. The emulsion was chilled, centrifuged at 800 g for 10 min at 5 °C, the aqueous layer combined with the previous aqueous portion and re-extracted twice with 5-0 ml phenol at 50 °C for 5 min. After final, the aqueous fraction was centrifuged at 15000 g for 30 min to remove bentonite. One-tenth volume of 20 % sodium acetate and two volumes of 95 % ethanol were added to the supernatant and the RNA was allowed to precipitate overnight at −20 °C. The precipitate was treated with DNase (Worthington, 13 mg/ml in 0·1 M phosphate buffer with 0·1 % phenol as preservative) prior to sedimentation analysis and electrophoresis.

Ribonucleic acid in yolk platelet pellets was determined by standard orcinol colorimetric analysis as described by Shatkin (1969).

RNA was centrifuged at 12000 g for 30 min at 5 °C, the ethanol was decanted and the precipitate dried by inverting the tube at 5 °C for 30 min. RNA was dissolved in SDS buffer (0·1 M NaCl, 0·1 M EDTA, 0·01 M Tris-HCl, pH 7·4, 0·2 % SDS), layered on a 10–30 % linear sucrose gradient, and centrifuged in a Spinco L2-65 ultracentrifuge (SW 65 rotor) for 2·5 h at 65000 rev/min (300000 g) at 25 °C. The centrifuged gradients were displaced with a 40% sucrose solution and continuously monitored at 254 nm through a Model D Density Gradient Fractionator (Instrumentation Specialties Co., Inc., Lincoln, Nebraska).

RNA was extracted with phenol, purified as described above, and subjected to electrophoresis on an acrylamide gel according to Loening (1967). RNA from Escherichia coli (23 S and 16S) were utilized as markers in control gels.

Yolk platelets are present in random pattern throughout cells of Xenopus gastrulae (Fig. 1), the larger platelets (up to 50 μm in length) being present in prospective entodermal regions. Most platelets exhibit three basic components as described by Karasaki (1963) : a limiting membrane, a superficial layer and a main body component. However, platelets in presumptive neural ectoderm appear to have lost superficial layers and limiting membranes by the mid-gastrula stage.

Limiting membranes are generally lost during isolation procedures, whereas superficial layers and crystalline inner matrices remain (Fig. 2). The superficial layers consist of small electron-dense particles (50 Åin diameter), larger, more angular particles (150–250 Å in diameter) and an amorphous background substance. The main body contains a highly structured crystalline matrix (for review of fine structure in the main body component see Karasaki, 1963).

Pellets examined in the electron microscope after initial centrifugation revealed presence of membranous material, lipid droplets and disrupted mitochondria in addition to yolk platelets (Fig. 3). Superficial layers of the latter were fused with neighbouring platelets, creating a continuum of yolk substance. After resuspension and a second centrifugation, platelets regained their individuality, losing contaminating materials (Fig. 4). Following a third centrifugation, superficial layers began to separate from main body components (Fig. 5), revealing particulate material in the supernatant. Main body components were disrupted following suspension with a vortex mixer and a fourth centrifugation (Fig. 6).

Autoradiographs of gastrula cells cultured in [³H]uridine reveal silver grains over both superficial and central components of yolk platelets (Fig. 7) in addition to cytoplasmic particles, mitochondria and nuclei. In addition, label remains in isolated platelets.

Table 1 illustrates the effect of acid hydrolysis and RNase on grain counts over yolk platelets in situ. Thick sections of cells prepared for light-microscope autoradiography have fewer silver grains after incubation in RNase and treatment with 1 N-HCI than when treated with DNase, phosphate buffer, and water for similar periods of time at 38 °C.

Orcinol analysis (Shatkin, 1969) suggested that 50–60μg of RNA were present in yolk platelets isolated from 1000 Xenopus embryos (stage 11).

Optical-density measures of RNA from isolated pellets in sucrose-density gradients revealed three distinct peaks (Fig. 9). The major peak (fraction 1) at the top of the tubes represented contaminating sediments (nucleotides, degraded DNA, and some protein) and low-molecular-weight RNA. Fractions 2 and 3 were isolated and electrophoresed on acrylamide gels, revealing several distinct components (Fig. 10). Peaks corresponding to 23 S and 16S RNA from E. coli in control gels were noted and sedimentation values for peaks in experimental profiles were computed from these data.

RNA extracted from washings obtained during platelet isolation procedures yielded the following results. Fig. 11 represents RNA species separated by gel electrophoresis from collected supernatants after a single centrifugation of homogenate. Numerous peaks were present, representing high-and low-molecular-weight RNA in both nucleus and cytoplasm of cells from stage 11 embryos. However, after resuspension and a second centrifugation of platelet pellets, phenol extractions of supernatant did not provide optically active material demonstrable in acrylamide gels (Fig. 12, left). Extractions of washings from the third centrifugation, however (Fig. 12, right), yielded electrophoretic patterns similar to those obtained from isolated yolk platelets. Low-molecular-weight RNA as well as larger species were present. A 15S species was present in both third washings and isolated yolk platelet pellets. Analysis of washings from a fourth resuspension and centrifugation revealed electrophoretic profiles similar to those demonstrated after the third spin.

Questions concerning the presence and potential developmental role of nucleic acids within amphibian yolk platelets are recurrent topics in numerous papers (see Rounds & Flickinger, 1958; Yamada, 1961; Horn, 1962). Although the existence of RNA in yolk platelets has been challenged by Wallace (1963 a, b) and Ohno, Karasaki & Takata (1964), our results indicate that yolk platelets in Xenopus gastrulae (stage 11) do possess small quantities of RNA (0·05–0·06 μg/embryo).

Of principal concern is the problem of contamination or artifact. Bacteria would probably not contaminate an embryonic cell fraction to the extent of contributing significant optical density to the RNA population. Furthermore, bacteria can be easily removed from media. In addition, embryos were chemically dejellied, which has been shown by Brown (1967) to decrease significantly bacterial contamination. An additional potential contaminant is yolk phosphoprotein which has a molecular weight (30000) and phosphate content (8%) similar to low molecular weight (4 S) RNA (Wallace, 1963 a). Furthermore, phosphoprotein can be extracted into the aqueous phase and precipitated with ethanol. Fortunately removal of most of this contaminant is effected with 0·01 M MgCl₂ in the initial homogenate (Brown, 1967).

Are the 150–250 Å particles present in superficial layers of in vitro and in situ platelets ribonucleoprotein (RNP)? And if so, are they normal components of superficial layers or artifacts of the isolation procedure? Generally, particles that are 150–250 Å in diameter and either attached to membranes or free in the cytoplasm of adult tissues are designated ribosomes (Palade, 1955), whereas those larger in size (250–500 Å in diameter) and free in the cytoplasm are regarded as glycogen granules (Drochmans, 1962). It is thought then that larger particles present in superficial layers are RNP because of their appearance and affinity for aqueous uranyl acetate. In addition, the isolation procedure of Wallace & Karasaki (1963) used in this study provides the investigator with platelets which are reasonably free from contamination visible in the electron microscope (Figs. 4-6). Furthermore, the absence of RNA in the supernatant following a second resuspension and centrifugation (Fig. 12, left) and the associated intact appearance of isolated platelets (Fig. 4) suggest that 150–250 Å particles are normal components of superficial layers and not free cytoplasmic particles displaced by centrifugation.

Autoradiographs presented in this study of labeled (tritiated uridine) yolk platelets may provide morphological evidence for the presence of RNA in those structures (Figs. 7, 8). Although nucleosides are not thought to be precursors of nucleic acids in normal biosynthetic pathways, [³H]uridine has been used in a variety of embryos and appears to be an efficient precursor of nucleic acids. Bieliavsky & Tencer (1960) have demonstrated that during cleavage in amphibian embryogenesis, labeled uridine is incorporated into DNA rather than RNA, but the converse seems to be true during gastrulation. Since thick sections of cells, prepared for light-microscopic autoradiography as controls, have fewer silver grains over yolk platelets after incubation in RNase and 1 N-HCI than when treated with either DNase, buffer or water, it is concluded that most [³H]uridine is incorporated into RNA rather than DNA during this developmental period.

Failure to extract RNA from washings following a second resuspension and centrifugation (Fig. 12, left) followed by an increase in RNA content in supernatants of successive spins (Fig. 13, right) may correlate with progressive disruption of main body components (Fig. 6), superficial layers (Figs. 5 and 6), and the apparent release of RNP particles (Fig. 2) from the latter.

Hence, biochemical and autoradiographical evidence presented in this study supports the hypothesis that yolk platelets in amphibian gastrulae contain RNA. Acrylamide-gel profiles suggest that RNA species of high and low molecular weight are present. Furthermore, the RNA content of isolated platelets can be released to surrounding media during isolation procedures which supports observations of Rounds & Flickinger (1958). The developmental significance of yolk platelet RNA, however, remains uncertain.

Grateful acknowledgment is made to Professors A. J. Ladman and Leonard Napolitano for helpful discussions and critical reading of the manuscript, to Mrs Margo Goff for technical assistance, and to the United States Public Health Service for facilities through CA-10694.