Ultrastructural cytochemistry of the nucleus in zea mays embryos during germination

In Zea mays as in numerous flowering plants, the embryo enters the quiescent stage during ripening of seeds. The cells of the quiescent embryo are poorly hydrated and lose mitotic activity. Their metabolism is almost completely arrested. When seeds are germinated at 16 °C the radicle resumes growth after about 2 days, several hours before the other parts of the embryo. In root tissues the reactivation of nuclear functions is progressive. DNA replication is initiated 45 h after sowing and the first mitotic figures are observed a few hours later (Deltour & Jacqmard, 1974). Transcription of the different RNA species is re-established in sequence during early germination (Van de Walle, Bernier, Deltour & Bronchart, 1976).

In a previous ultrastructural study of maize embryo root cells it was shown that the most spectacular changes occur in the nucleus, at the beginning of germination (Deltour & Bronchart, 1971). The purpose of the present study is to describe more precisely the quiescent and the reactivating nucleus using cytochemical techniques.

Variations in the chromatin, particularly nucleolus-associated chromatin, have been studied. The surprising presence of a nucleolus organizer region (NOR) has been described previously in resting cells of Helianthus tuberosas at the level of the nucleolus-associated chromatin; the evolution of this NOR has also been described in the same cells during the reactivation of the tissues using methods of ultrastructural cytology (Jordan & Chapman, 1971). In the present study we have used a specific reaction for DNA to localize the possible presence of a NOR in quiescent cells of maize embryo and to follow it during germination.

As the genes coding for pre-rRNA are located in the NOR (Busch & Smetana, 1970), we investigated a possible relationship between ultrastructural changes in the NOR and the resumption of nucleolar pre-rRNA synthesis. In germinating maize root cells the timing of resumption of nucleolar activity is already known (Van de Walle et al. 1976).

Attention has also been paid to the ribonucleoprotein structures of the nucleus. By use of ultrastructural cytochemistry several classes of ribonucleoprotein-containing structures have been characterized in the nucleus of animal cells: perichromatin fibrils and granules, interchromatin granules and coiled bodies. Several cytological and biochemical studies have been performed to define the function of these structures (Georgiev & Samarina, 1971; Bouteille, Laval & Dupuy-Coin, 1974). At present, far less data are available concerning the nuclear ribonucleoprotein structures of the plant cell nucleus than those of the animal cell nucleus. Nevertheless an understanding of the fate of transcription products and of transport of genetic information is directly connected to an understanding of the function of the various nuclear ribonucleoprotein structures. Thus, it seems essential to undertake a systematic study of the ribonucleoprotein structures in different plant cells, in particular by using the preferential staining method for ribonucleoproteins (Bernhard, 1969). We describe here several ribonucleoprotein structures which we have observed in the radicle of the quiescent and germinating maize embryo.

Kernels of Zea mays L. (var. CiV₂) were germinated in the dark at 16 °C in Petri dishes over cotton wool and filter paper saturated with distilled water. We consider that the germination starts when the seeds are placed in contact with water. Under these conditions the resumption of growth of the embryo does not occur before 48 h of germination.

Embryos were excised from dry ungerminated kernels and from kernels germinated for 8, 24, 48 or 72 h. These embryos are respectively designated: quiescent, 8-h, 24-h, 48-h or 72-h embryos. Each batch contained 10 embryos.

After excision the embryos were rapidly immersed in a cold solution of 4 % glutaraldehyde buffered with 0·1 M Na-cacodylate (pH 7·0). Half an hour later the coleorhiza and the root cap were discarded from embryos maintained in fixative: the first millimetre of the root tip was then cut off and dipped in fresh cold fixative for 3 h. The total duration of fixation was about 4 h. The samples were washed in 0·1 M cacodylate buffer, dehydrated first in graded ethanol, and then in propylene oxide, and embedded in Epon. The same samples were used both for morphological and cytochemical observations. Cross-sections of roots were cut 1 mm from the tip; semithin sections (about 1 mm²) included all tissues of the roots, ultrathin sections the cortex only. Ultrathin sections were stained with saturated aqueous uranyl acetate and Reynold’s lead citrate.

Diameters of various ribonucleoprotein particles were measured on electron micrographs with a stereomicroscope with a Wild Censor micrometer. Mean diameters are calculated from 10–15 measurements taken on well defined particles.

Semithin sections were floated in a 5 M HC1 solution for 30 min at 20 °C, then in a 0·01 % solution of acriflavin, previously bubbled with SO₂, for 90 min at 20 °C. The sections were successively washed with distilled water, mounted on glass slides and observed under a fluorescence microscope with a Xenon lamp and filters BG38 and K530 (Gautier, 1976). As fluorescence micrographs are presented here in black and white reproductions, the colours observed are indicated in the figure captions.

Ultrathin sections were floated in a 5 M HC1 solution for 30 min at 20 °C, then in a o·o5–o·5 % solution of osmium ammines complex (OAC), previously bubbled with SO₂, for 60 to 120 min at 20 or 36 °C. Controls were non-hydrolysed, stained or hydrolysed, unstained preparations (Cogliati & Gautier, 1973; Gautier & Fakan, 1974; Gautier, 1976). Various batches of this reagent were used comparatively in the current study: they were chosen either to give a roughly-granular reaction end-product for the general detection of DNA-containing structures, or to give a very fine end-product to study ultrastructure.

Thin sections are floated in 5 % uranyl acetate solution for 10 min at 20 °C, washed, floated again in 0·2 M EDTA solution (pH 7·0) for 10–20 min at 20 °C, washed again, and finally treated with Reynold’s lead citrate. Controls are observed without EDTA treatment (Bernhard, 1969).

In all cells of the radicle of the quiescent embryo the whole nucleoplasm is filled with areas of dense chromatin. A narrow ring around the nucleolus makes an exception to this rule and very little dense chromatin may be seen in this nucleoplasmic region. Nevertheless, in some sections, one strand of the dense chromatinic network is seen in close connexion with the nucleolus. This strand consists of an external dense chromatinic knob, juxtaposed to an internal, less-dense chromatin zone (Fig. 1). This structure is analogous to the nucleolus organizer region (NOR) described in Zea mays microsporocytes by Gillies (1973).

At 8 h of germination, a decrease in dense chromatin is observed; this decrease is most apparent in the region close to the nucleolus (Fig. 2). At 24 h the dense chromatin has almost disappeared; only few isolated, contrasted aggregates, scattered in the nucleoplasm or in contact with the nuclear envelope are still visible (Fig. 3). At 48 and 72 h small areas of dense chromatin are again more numerous in the nucleoplasm (Fig. 4); they are, however, less abundant than in the nuclei of dry, ungerminated embryos.

In quiescent embryo cells the nucleolus is compact (Fig. 1): its structure is predominantly fibrillar up to 24 h. At high magnification, the granular components are randomly distributed among the fibrils; moreover, these granular components are sometimes packed as a very fine layer at the periphery of the nucleolus. In a few nucleoli, light areas, similar to nucleolar vacuoles are visible. The small size and scarcity of the light areas do not permit a good estimate of their content.

At 8 h of germination, the nucleolus is mostly vacuolated (Fig. 2): the nucleolar vacuoles contain fibrils and granules (which are more easily characterized after EDTA staining, see below). At 24 h, a drastic decrease in nucleolar vacuolation is observed and a significant decrease in the intravacuolar fibrils is also noticed. No further change in the structure of nucleolar vacuoles is detected later. At 72 h, a considerable increase in the granular component of the nucleolus is observed: the accumulation of this component at the periphery of the nucleolus forms a well differentiated and thick external layer (pars granulosa), while the central zone remains predominantly fibrillar (Fig. 4).

The network of dense chromatin shows a clear, positive response to H + ASO₂ and H + OACSO₂ reactions, both in cells of quiescent (Figs. 5, 8) and 8-h (Fig. 9) embryos. The perinucleolar zone does not react in these cells except in the NOR, i.e. the site where the network of dense chromatin joins the nucleolus (Fig. 1). Due to section thickness, these positively reacting regions are more often observed in semithin sections treated with H + ASO₂ (Fig. 5) than in ultrathin sections treated with H+OACSO₂ (Fig. 9). At 24 h the homogeneous, diffuse, and weak stain obtained with the H + ASO₂ reaction shows that almost all the nuclear DNA is spread within the nucleoplasm (Fig. 6); similarly, the H + OACSO₂ reaction shows only numerous very tiny aggregates, scattered throughout the nucleoplasm or bound to the nuclear membrane, indicating a nearly total dispersion of DNA (Fig. 10). At 48 and 72 h numerous small areas in the nucleoplasm show a positive reaction either with H + ASO₂ (Fig. 7) or with H + OACSO₂ (Fig. 11) indicating the restoration of the chromatin network, which nevertheless remains less pronounced than in the nucleus of the quiescent embryo. Between these restored areas of dense chromatin, the H+OACSO₂ reaction shows portions of the nucleoplasm where the dispersion of the DNA is still complete (Fig. 12).

The well delineated, rich DNA-containing region (NOR), which in the quiescent embryo and at 8 h bridges the nucleolus and the nucleoplasmic dense chromatin, is no longer seen at 24, 48 and 72 h (Figs. 6, 7, 10, 11).

DNA is not detectable with the H + OACSO_a reaction in the nucleolar matrix of quiescent and 8-h embryo cells or in the numerous nucleolar vacuoles at 8 h (Figs. 8, 9). From 24 up to 72 h, DNA-rich constituents can often be observed in nucleolar profiles (Figs. 10, 11). These constituents are localized in the matrix of the nucleolus and in some nucleolar vacuoles (Fig. 11).

In the quiescent, as well as in the germinated embryos up to 72 h, the H + OACSO₂ reaction regularly reveals cytoplasmic DNA (Fig. 8). Most likely this DNA is localized in mitochondria and/or proplastids, although these organelles are difficult to recognize after cytochemical staining as their own contrast is low.

After regressive staining with EDTA the dense chromatin appears bleached and the ribonucleoprotein material contrasted. Four distinct types of area are observed in the nucleoplasm of the quiescent embryo and in the germinating embryos up to 72 h. With time after germination, the relative proportions of these areas change. (1) Dense chromatin, filling large portions of the extranucleolar space, appears as bleached areas (Fig. 13). (2) Heavily contrasted material, probably of fibrillar texture, forms heavily contrasted areas located at the periphery of the bleached areas (Fig. 13). (3) Numerous spherical, heavily contrasted granules and irregular, rod-like structures form granular areas, or are dispersed free in the extranucleolar space (Fig. 14): these structures are intermingled with heavily contrasted material similar to the material located at the periphery of the bleached areas (Figs. 17, 18); in the granular areas, the relative proportion of granular- and fibrillar-like material is highly variable. (4) The remaining parts of the extranucleolar space are filled with thin, poorly contrasted fibrils (Figs. 14-23).

In the quiescent embryo, each type of area is well represented; the diameter of spherical granules, either isolated or in the granular area, reaches approximately 40 nm (Figs. 13, 19). At 8 h of germination the sizes of bleached areas and of heavily contrasted areas have markedly decreased; in the granular areas, the spherical granules become larger, approximately 50 nm in diameter; thin sections of these granular areas usually contain 3 to 10 granules, but may contain up to 50 (Fig. 14). At 24 b (Figs. 15, 16, 21), bleached chromatin areas, and heavily contrasted areas are dispersed in small patches in the whole nucleoplasm, and only some 50-nm granules are still found, freely dispersed in the extranucleolar space (Fig. 15): the greatest part of the nucleoplasm is therefore constituted of poorly contrasted fibrils, mixed with numerous very tiny bleached areas, corresponding probably to highly dispersed chromatin. At 48 and 72 h (Figs. 17, 18, 22), a considerable increase in bleached chromatin areas and associated heavily contrasted areas is noticed again; reappearance of granular areas is also noted. Thin sections of these areas usually contain 3 to 10 granules of 50 nm diameter, but isolated 50-nm granules are still dispersed in the whole nucleoplasm. Sometimes, 50-nm granules, surrounded by a clear halo, are found in heavily contrasted material (Fig. 18).

Due to their size and high contrast, 40-nm granules of the quiescent embryo and 50-nm granules of germinating embryos are easily recognized. Moreover, smaller granules of various sizes are scattered inside the nucleoplasm throughout germination, but mostly at 24 and 72 h (Figs. 15–18).

Fibrillar and granular constituents of the nucleolus, as well as other less usual ribonucleoprotein nucleolar constituents react positively to the EDTA regressive method.

In the quiescent embryo, the nucleolus is predominantly fibrillar (Fig. 19). At 8 h of germination, some 35-nm granules appear at the periphery of the predominantly fibrillar nucleolus; at the same time, in the newly formed nucleolar vacuoles, thin fibrils and a few granules are present; these granules, in the smallest vacuoles, measure 18 and 35 nm in diameter, whereas in the large vacuoles, only 35-nm granules are found (Fig. 20). At 24 h most nucleoli retain a predominantly fibrillar structure, although the matrix of a few contains more 18-nm granules than the quiescent nucleoli (Fig. 21); the sparse nucleolar vacuoles contain only small 18-nm granules and fibrils. From 48 to 72 h (Fig. 22) considerable enrichment of the nucleolar periphery {pars granulosa) is noticed.

In addition to the above-described ribonucleoprotein structures, from 8 h of germination, areas of less-contrasted material were often observed on the periphery of the nucleolus; some of these areas extend into the nucleoplasm. Their structure is fibrillar, although granules of various sizes are observed, principally in the protruding part of the area (Fig. 16).

The above-reported observations are summarized in Fig. 23.

During the first 24 h of germination, large blocks of dense chromatin, which characterize the quiescent nucleus, strongly disperse and convert, for several hours, into a fine network probably constituted of separate deoxyribonucleoprotein fibrils. This dispersion also has been observed in another variety of Zea mays (Deltour & Bronchart, 1971) and, in that species, appears thus as a normal pattern which may reflect early preparation for nuclear DNA replication (S-phase).

It has been observed earlier, in different varieties of maize, that the majority of nuclei of the quiescent embryo are arrested in the G₁ phase (Deltour & Jacqmard, 1974; Crèvecoeur, unpublished results). Moreover, in the variety under study here (CiV₂), autoradiography has shown that nuclear DNA replication starts at about 30 h (Crèvecoeur, unpublished results); thus our assumption that chromatin dispersion reflects preparation for nuclear DNA replication is supported by this observation. Moreover, a correlation between dispersion of dense chromatin and progression of cells to S-phase has also been observed in other plant (Lafontaine, 1974a) and animal cells (Bouteille et al. 1974).

Nevertheless, as DNA duplication requires several hours, not every portion of the genome is involved simultaneously in this DNA replication process (Lafontaine, 1974a). Therefore an additional phenomenon must explain the total dispersion of chromatin that we observed in root cells during early germination. A similar complete transformation of dense chromatin to dispersed chromatin is often observed in cells where gene derepression occurs, for instance, during activation of human blood lymphocytes incubated with phytohaemagglutinin (Tokuyasu, Madden & Zeldis, 1968). The total dispersion of chromatin has also been related to gene derepression which occurs at germination (Payne et al. 1976; Brooker, Tomaszewski & Marcus, 1978).

Until 8 h of germination only a small portion of the chromatin is associated with the nucleolus and forms, on its periphery, an eccentric knob which is similar to the nucleolus organizer region described in meiotic cells of Zea mays, by Gillies (1973), except that the elements of the synaptonemal complex are absent in the present case. In other plant species a similar nucleolar knob is also found: in quiescent cells of Allium cepa the nucleolus-associated chromatin forms a peripheral core, in continuity with the nucleoplasmic chromatin (Bal & Payne, 1972); in resting cells of Helianthus tuberosas tuber, ‘lightly staining material’ is usually found at the outside of the nucleolus and identified as the NOR (Jordan & Chapman, 1971). Jordan & Luck (1976) observed in meiotic cells of Endymion, an association of ‘lightly staining material’ with the synaptonemal complex. They postulated that ‘the pale-staining region of nucleoli is the nucleolus organizer and almost certainly the chromosome region containing the ribosomal cistrons’. The comparison with other species leads us to conclude that the nucleolus-associated chromatin in quiescent maize root cells is the NOR, or its equivalent.

An apparent contradiction to this conclusion lies in the observation that the NOR is usually only visible at metaphase near the secondary constriction of the chromosome responsible for nucleolus formation (Busch & Smetana, 1970). Nevertheless, other findings support our conclusion: in the quiescent cell nucleolus of maize embryo, cytochemical detection of DNA (here reported) and incorporation of [³H]uridine (revealed by high-resolution autoradiography, unpublished results) are observed only at the level of the nucleolus-associated chromatin.

The unusual presence of NOR in non-dividing cells of quiescent maize is probably linked to the high degree of chromatin condensation in quiescent tissues. Indeed, when maize kernels are dehydrated after 24 h of germination and the chromatin dispersion is complete, total recondensation of chromatin and reappearance of NOR are observed (Crèvecoeur, Deltour & Bronchart, 1976; Deltour, unpublished results). Moreover, 8 h after the start of germination, the NOR disappears just as the chromatin begins to disperse.

As shown by ultrastructural cytochemistry, no DNA is present within the nucleolus of the quiescent cells except in the NOR which lies in an external position. This situation lasts at least until 8 h of germination. Later the NOR is no longer detectable and the presence of DNA is clearly demonstrated inside the nucleolar matrix and vacuoles. As no amplification of ribosomal RNA-genes (rDNA) occurs during the germination of Zea mays (Ingle & Sinclair, 1972) the increasing content of intra-nucleolar DNA cannot result from gene multiplication. Therefore, it may be hypothesized that the NOR-associated DNA moves after 8 h from an external position to an internal one and becomes dispersed inside the nucleolus. A similar dispersion of NOR-associated DNA inside the fibrillar zone during reactivation of cells from Helianthus tuberosas tuber has been suggested by Jordan & Chapman (1971).

The dispersion of NOR-associated DNA inside the nucleolus of Zea mays could increase the accessibility of rDNA transcription sites and explain the acceleration of pre-rRNA synthesis which starts at about 8 h of germination. This resumption of high rate pre-rRNA synthesis in the nucleolus needs perhaps, in addition, the synthesis of new RNA polymerase I molecules, as in onion embryos (Bal & Payne, 1972).

Possibly the NOR which is clearly visible in maize cells during the first hours of germination is equivalent to the fibrillar centre observed in the nucleolus of several types of animal cells (Recher, Whitescarver & Briggs, 1969; Goessens, 1976). However, the fibrillar centre which is reported to contain the rDNA (Mirre & Stahl, 1978) persists unchanged during the whole mitotic cycle (Goessens & Lepoint, 1974). On the contrary, the NOR of quiescent maize cells disappears concomitantly with the reactivation of the nucleus and the nucleolus-associated DNA becomes dispersed in the nucleolar mass or associated with the nucleolar vacuoles (Fig. 11) as in some other somatic plant cells (LaCour, 1966; LaCour & Wells, 1967; Chouinard, 1970, 1975).

When the NOR is dispersed, no typical fibrillar centres, i.e. fibrillar light-staining spaces, are recognizable unequivocally within the nucleolus of maize root cell. All the light staining areas of the nucleolus contain ribonucleoprotein fibrils and granules and are very similar to the nucleolar vacuoles.

Among the structures which remain clearly visible inside the nucleoplasm after preferential staining procedure for ribonucleoproteins, we focused our attention on the heavily contrasted areas and on the 50-nm granules.

The heavily contrasted areas appear to have a fibrillar texture and can be compared either to micropuffs of plant cells (Lafontaine, 1965, 1974a, b; Risueño, Moreno Díaz de la Espina, Fernández-Gómez & Giménez-Martín, 1978), to centromeric heterochromatin of interphase nuclei (Gillies, 1973; Church & Moens, 1976), or to perichromatin fibrils of animal cells (Monneron & Bernhard, 1969).

Although the micropuffs also present a positive contrast after regressive EDTA staining (Risueño et al. 1978), their size (800 nm) and their round shape make them different from the heavily contrasted areas of maize embryo cells.

The comparison of the heavily contrasted areas with interphase centromeric heterochromatin is also doubtful as a large quantitative variation of the heavily contrasted areas is observed within the first 24 h of germination. During this period the root cells remain in G₁-phase (Deltour & Jacqmard, 1974) and the centromeric heterochromatin is expected to remain constant.

In animal cells perichromatin fibrils form thin ribonucleoprotein borders at the periphery of dense chromatin clumps and are probably the nucleoplasmic structures which contain the rapidly labelled heterogeneous nuclear RNA (HnRNA) (Bachellerie, Puvion & Zalta, 1975; Nash, Puvion & Bernhard, 1975; Fakan, Puvion & Spohr, 1976; Devilliers, Stevenin & Jacob, 1977). The heavily contrasted areas appear similar to perichromatin fibrils, because of their size and irregular strands, and location near dense chromatin. Nevertheless it must be pointed out that the compactness of these areas prohibits a clear demonstration that fibrils are the elementary component. Further studies are required to elucidate the nature of RNA present in these areas and thereby ascertain their relationship to perichromatin fibrils. This characterization is in progress, using high-resolution autoradiography combined with ribonucleoprotein cytochemistry (Fakan, 1970).

Dense granules of approximately the same size as those of Zea mays have been observed in the nucleoplasm of plant cells without identification of their role in the cell (Chaly & Setterfield, 1974; Lafontaine, 1974a; Rose, 1974). It may be suggested because of both the ribonucleoprotein nature and the size of these granules, that they correspond to perichromatin granules of mammalian cells (Swift, 1963; Mon-neron & Bernhard, 1969) or to Balbiani granules (Stevens & Swift, 1966). In maize cells these granules, when enclosed inside a heavily contrasted fibrillar area, appear surrounded by a clear halo characterizing perichromatin granules of animal cells. The function of perichromatin granules is unknown, but it has been suggested that they may be carriers of mRNA (Vazquez-Nin & Bernhard, 1971).

At 8 h of germination, ribonucleoprotein 35-nm granules are present inside the smallest nucleolar vacuoles and also on the nucleolar periphery. The significance of these granules is difficult to determine on the basis of their morphology.

Granules, larger than the usual 18-nm granules of the nucleolus, have been described in the nucleolus of Allium cepa during post-mitotic nucleologenesis (Chouinard, 1975) as well as in the fibrillar zone of the nucleolus of plant (Hyde, 1967) or animal cells (Shankaranarayan & Busch, 1965) but their function remains unknown. Recently, an interruption of pre-rRNA processing and a correlated appearance of large ribonucleoprotein granules associated with the nucleolus have been observed in animal cells treated with cordycepin (Puvion, Moyne & Bernhard, 1976) or by camptothecin (Gajkowska, Puvion & Bernhard, 1977). These unusual granules are considered by these authors to be the result of abnormal storage of pre-rRNA. In maize cells the transitory presence of ribonucleoprotein 35-nm granules associated with the nucleolus may be explained in the same manner. In the embryo, processing of pre-rRNA is slow during the first hours of germination (Van de Walle et al. 1976). Consequently, it is possible that pre-rRNA does accumulate. The pre-rRNA molecules may originate from both old pre-rRNA synthesized during the maturation of embryo, and from new pre-rRNA synthesized at a low rate during the first hours of germination.

In addition to usual granular and fibrillar components of the nucleolus, 18-nm granules are seen inside and also at the periphery of nucleolar vacuoles. Arrays of fibrils are seen inside the same vacuoles. With regard to size and staining characteristic, the small granules are apparently identical to the granular component of the pars granulosa. As the nucleolar vacuoles are frequently located inside the fibrillar zone of the nucleolus, we suggest that they may be sites where processing of fibrillar components to granular components of the nucleolus occurs. This hypothesis is supported by the observation of ribonucleoprotein 35-nm granules in the smallest vacuoles when processing of pre-rRNA is slow.

We wish to thank Dr S. Fakan for critical review of the manuscript and Mrs F. Nyffenegger for its preparation. This work was supported by grants of the Fonds de la Recherche Fondamentale Collective (grant No. 2.4505.78) and of the Swiss National Science Foundation (grants No. 3.1880.73 and No. 3.055.76). Part of this work was communicated at the annual meeting of the French Society for Electron Microscopy, Nizza 1977 (Deltour, Gautier & Fakan, 1977).