Different patterns of rDNA organization at interphase in nuclei of wheat and rye

The organization and structure of the nucleolus is of considerable interest, since it is one structure in the interphase nucleus where gene expression can be studied at both the molecular and cytogenetic levels. The ultrastructure of nucleoli (Goessens, 1984; Risueño and Medina, 1986; Schwarzacher and Wachtler, 1986; Deltour and Motte, 1990), their proteins (e.g. see Hernandez-Verdun, 1991) and the molecular characteristics of the rDNA loci (e.g. see Flavell and O’Dell, 1990) have been extensively reviewed. However, the location of the rDNA loci within the nucleus and nucleolus is controversial (Jordan, 1991), and the patterns of gene expression are still unclear.

The activity of the rRNA genes in cereals has been examined using nucleolar volumes and numbers (Jor dan et al. 1982; Martini and Flavell, 1985), silver staining (Moreno et al. 1990; Vieira et al. 1990) and in situ hybridization (Appels et al. 1986a; Gustafson et al. 1988). DNA:DNA in situ hybridization of rDNA sequences enables gene sequences to be localized at interphase when cells are transcriptionally active (Mukai et al. 1991). Chinese Spring wheat has ten rDNA loci (Mukai et al. 1991), but only chromosomes 6B and IB are normally actively involved in nucleolus formation (Jordan et al. 1982; Martini and Flavell, 1985). In cereals regulation of gene activity occurs by the suppression of whole loci and genes within individual loci (Flavell and O’Dell, 1990).

In this study, we aimed to examine where genes are expressed along the tandem array of rDNA sequences in two cereal species, wheat and rye. We have applied high-resolution in situ hybridization methods using probes for rDNA on spreads examined with the light microscope (LM) and on thin sections observed in the electron microscope (EM).

Root tips were obtained from seedlings of rye (Secale cereale L.) cv. Petkus Spring and two wheat (Triticum aestivum L. emend. Thell.) cultivars, Chinese Spring (Fig. 1A, B; Fig. 2A, C, D) and Beaver (Fig. 1C, D; 2B). Seeds were germinated on moist filter paper for 48 h at 25°C, transferred to 4°C for 24 h and then returned to 25°C for 26 h. The seedlings were then kept on ice for 24 h. This treatment does not injure cereals but delays anaphase in the root tip meristem and increases the mitotic index.

Cell spreads were made from enzyme-softened root tip meristems that were squashed under a coverslip onto a slide using methods described by Schwarzacher et al. (1989).

Root tips were fixed in 1% (v/v) glutaraldehyde and 0.25% (v/v) saturated aqueous picric acid solution in phosphate buffer (0.05 M Na₂HPO₄, 0.05 M KH2PO4) for 1 h at room temperature, dehydrated through a graded ethanol series, embedded in LR White (medium) resin and sectioned to 0.1 μm thickness using methods described by Leitch et al. (1990). Sections were picked up on gold 400 mesh grids with a carboncoated support film prepared from 4% (w/v) pyroxylene in amyl acetate. Grids with sections were incubated overnight at 37°C prior to in situ hybridization.

pTa71 from wheat is a complete 9 kb rDNA gene unit containing the 5.8 S, 18 S, 25 S genes and the intergenic spacer (Gerlach and Bedbrook, 1979; kindly provided by R. B. Flavell and M. O’Dell).

pScR4.Tl is a TaqI fragment of rye rDNA containing the intergenic spacer sequences (Appels et al. 1986b).

Total genomic DNA from Triticum durum or salmon sperm was sometimes used as blocking DNA to increase probe specificity and reduce non-specific background labelling (Anamthawat-Jônsson et al. 1990).

The DNA was labelled with digoxigenin-11-dUTP (Boehringer Mannheim) by nick translation. The probe solution was prepared by mixing labelled DNA (100 ng ml^-1) in a solution of unlabelled blocking DNA (5 μg ml^-1), 50% (v/v) formamide, 10% (w/v) dextran sulphate, 0.1% (w/v) SDS (sodium dodecyl sulphate), and 2×SSC (0.3 M NaCl, 0.03 M sodium citrate) before denaturation.

Slides with cell spreads were incubated in 100 μg ml^-1 DNase-free RNase A in 2xSSC at 37°C for 1 h, washed twice in 2×SSC for 10 min at room temperature, refixed in freshly depolymerized 4% (w/v) paraformaldehyde in water for 5 min at room temperature and washed in 2xSSC for 5 min. Grids with ultrathin sections were treated with RNase, proteinase K and paraformaldehyde using methods described by Leitch et al. (1990).

The probe solution was denatured at 70°C for 10 min, loaded onto slides with cell spreads and covered with a coverslip. Grids were placed in a drop of the denatured probe solution on a glass slide and covered with a coverslip. Cell spreads and grids were transferred to a humid chamber and denatured in the probe mixture at 90°C for 10 min.

Hybridization and post-hybridization washes followed Schwarzacher et al. (1989). The most stringent wash allowed DNA sequences with more than 85-90% sequence identity to remain hybridized if parameters used by Meinkoth and Wahl (1984) apply to hybridization in situ.

Slides were transferred into detection buffer (4×SSC, 0.2% (v/v) Tween 20) for 5 min, treated with BSA block (5% (w/v) bovine serum albumin in detection buffer) for 5 min and then incubated in 20 μg ml⁻¹ sheep anti-digoxigenin-FTTC IgG (Boehringer Mannheim, in BSA block) antibodies for 1 h at 37°C. After incubation the slides were washed in detection buffer at 37°C. The signal was amplified, after 5 min of goat serum block (5% (v/v) normal goat serum in detection buffer) at room temperature, using 10 μg ml⁻¹ FITC-conjugated rabbit anti-sheep IgG (Dakopatts, in goat serum block) antibodies for 1 h at 37°C. After washing in detection buffer the slides were counterstained with 2 μg ml⁻¹ DAPI (4’,6-diamidino-2-phenylindole) in Mcllvaine’s citrate buffer (0.01 M citric acid, 0.08 M Na₂HPO₄, pH 7). Slides were sometimes counterstained with 1 μg/ml propidium iodide in PBS (phosphate-buffered saline: 120 mM NaCl, 2.7 mM KC1, in phosphate buffer, pH 7.4). All slides were mounted in antifade solution (AF1, Citifluor). Slides were examined with a Leitz Aristoplan epifluorescence microscope and photographs were taken on Kodak Ektar 1000 colour print film.

Grids were transferred into detection buffer, treated with BSA block for 5 min at room temperature and then incubated in 7.5 units ml^−-1 sheep anti-digoxigenin/horseradish peroxidase (HRPO, Boehringer Mannheim, in BSA block) antibodies for 1 h at 37°C. After incubation the grids were washed in detection buffer at 37°C. After treating the grids for 5 min in goat serum block two different approaches were used. (A) Incubation in HRPO-conjugated rabbit anti-sheep (Dakopatts in goat serum block) antibodies for 1 h at 37°C (Figs 2B; 3A, B). (B) Incubation in 150 μg ml⁻¹ unlabelled rabbit antigoat (Dakopatts, in goat serum block) antibodies for 1 h at 37°C followed by washing in detection buffer at 37°C, reincubation in BSA block for 5 min at room temperature and incubation with 1.2μg ml⁻¹ peroxidase/anti-peroxidase (PAP-raised in goat, Dakopatts, in BSA block) antibodies for 30 min at 37°C (Fig. 2A, C, D).

After three washes in detection buffer the signal was visualized by diaminobenzidine (DAB) reduction using methods of Wachtler et al. (1990). The grids were incubated for 20 min at 4°C in 50 mM Tris-HCl, pH 7.4, 0.05% (w/v) DAB and 0.015% (v/v) fresh hydrogen peroxide. The reaction was stopped in several rinses of water and the grids were air dried. Grids were examined after 10 min of staining in saturated aqueous uranyl acetate at 80 kV using a Jeol JEM 1200EX transmission electron microscope.

The location of rDNA sequences in cell spreads of wheat was shown by in situ hybridization of digoxigenin-labelled pTa71 detected with FITC fluorescence (Fig. 1A, C). A gradient in DAPI fluorescence occurred across the nucleus, and nucleoli were present in areas with weak or no DAPI fluorescence (Fig. 1B, D). Within the nucleoli there were many fluorescent, punctate sites of probe hybridization that were occasionally joined by a fine fluorescent line. Larger probe hybridization sites tended to occur outside the nucleolus and particularly over the domain of the nucleus with brightest DAPI fluorescence (compare Fig. 1A and B, C and D).

Using electron microscopy the sites of probe hybridization were visualized on sections by electron-opaque deposits of reduced DAB. Sectioned metaphase chromosomes had a cross-sectional diameter of about 1 μm. Chromosomes with a nucleolar organizing region could be identified by extensive DAB labelling at the sites of probe hybridization (Fig. 2A). At interphase, condensed chromatin with in situ hybridization signal was seen outside the nucleolus (Fig. 2B), against the nuclear envelope (data not shown), adjacent to the nucleolus (perinucleolar, Fig. 2C) and within the nucleolus (Fig. 2B-D). Some sections showed labelled condensed chromatin extending into the nucleolus (Fig. 2C). The intranucleolar condensed chromatin was always labelled. It had a cross-sectional diameter up to 350 nm, and was surrounded by an electron-translucent zone (Fig. 2B-D). The number of labelled, condensed chromatin sites observed in sectioned nucleoli could be predicted from the number of probe hybridization sites seen in spreads of wheat nucleoli (cf. Fig. 1A and Fig. 2B-D; an arbitrary transect through nuclei in Fig. 1A usually includes 1 to 3 punctate fluorescent sites, the number of condensed chromatin sites seen in typical EM sections). In the EM, additional diffuse in situ hybridization signal occurred within the nucleolus (Fig. 2B-D). In control experiments where no probe was added chromatin was visible but no DAB precipitate was specifically localized within the cell.

The location of rDNA in interphase cell spreads of rye was shown by in situ hybridization of digoxigenin-labelled pScR4.Tl detected with FITC fluorescence. Similar in situ labelling patterns were found with the probe pTa71 (cf. Fig. 1E with Fig. 2B of Leitch et al. 1991). The probe hybridized predominantly to two discrete regions that fluoresced brightly (Fig. 1E). Some preparations showed fine fluorescent lines of in situ hybridization signal that emanated from the main sites of probe hybridization into the nucleolus. DAPI fluorescence revealed a polarity in the nuclei with a domain showing bright fluorescence and a domain with less overall fluorescence. In all nuclei, the latter domain contained the subtelomeric heterochromatin (DAPI dots) and nucleoli (Fig. IF) but no strong hybridization sites. The two predominant sites of probe hybridization in each nucleus tended to lie adjacent to nucleoli towards the poles with the highest DAPI fluorescence.

Using electron microscopy, many sections of rye nuclei probed with pTa71 showed one or two labelled 800 nm diameter perinucleolar chromatin segments (Fig. 3A). When two such segments were in a nuclear section they were associated with the same single nucleolus or with two separate nucleoli. Most unlabelled chromatin in the nucleus had a diameter of about 130 nm except for some segments of about 800 nm diameter against the edge of the nucleus in the domain containing a low proportion of chromatin (identified as subtelomeric heterochromatin; Anamthawat-Jónsson and Heslop-Harrison, 1990). The labelled 800 nm chromatin segments (Fig. 3A) occurred in the same relative position within the nucleus as the major in situ hybridization sites seen on spread preparations (Fig. 1E). Unlabelled 130 nm diameter perinucleolar chromatin was found on the opposite side of the nucleolus to the labelled perinucleolar chromatin (Fig. 3A, arrow). Within the nucleolus, diffuse in situ hybridization signal was seen (Fig. 3B). Intensity and position of the diffuse signal within the nucleolus corresponds well with the fine lines of in situ hybridization signal shown in LM preparations (Fig. 1E). In contrast to wheat, the rye nucleolus did not contain detectable condensed chromatin.

Sites of hybridization away from the nucleolus were detected in wheat (Figs 1A-D; 2B) but not rye. The sites in wheat are exclusively associated with condensed chromatin that is resolved in the EM (Fig. 2B) and represent loci that are inactive. In rye, where both rDNA loci are normally expressed, all rDNA sites were associated with a nucleolus (Figs 1E, F; 3A, B).

In rye and wheat, condensed rDNA-containing chromatin was associated with the periphery of the nucleolus. The hybridization sites seen by light microscopy (Fig. 1A, C, E) were shown by electron microscopy to be chromatin with a cross-sectional diameter of about 800 nm in rye (Fig. 3) and up to 350 nm in wheat (Fig. 2C). In rye, the diameter is close to that of metaphase chromosomes (1 pm, Fig. 2A), suggesting that little decondensation of the tandem array of rRNA genes has occurred during interphase. Lafontaine et al. (1991) speculate on the nature of nucleolus-associated bodies in Leucaena glauca; they appear similar to the condensed perinucleolar rDNA in rye. In wheat, the diameter of the perinucleolar rDNA was that of the average condensed interphase chromatin (350 nm). Chromatin of such diameters is probably transcriptionally inactive (Heslop-Harrison et al. 1988). In pea, Rawlins and Shaw (1990) found perinucleolar rDNA about 1 pm diameter in LM preparations. The chromatin fibre widths of transcriptionally inactive, condensed perinucleolar rDNA can thus vary.

The LM observations of the rye variety Snoopy (Appels et al. 1986a; Gustafson et al. 1988) indicated that expressed loci were entirely dispersed, which is in contrast to our results. Although the use of different varieties may account for some of the difference, our observations indicate that it is difficult to visualize different condensation states without the combination of in situ hybridization data from the LM and EM. Analysis of optical (Rawlins and Shaw, 1990) or physical EM (Figs 2, 3) sections enables unequivocal spatial localization of labelled perinucleolar rDNA.

In wheat, condensed rDNA within the nucleolus was visualized by LM as punctate fluorescence (Fig. 1A, C) and by EM as labelled chromatin axes up to 350 nm in diameter. In rye neither punctate fluorescence (Fig. 1E) nor intranucleolar condensed rDNA (Fig. 3) was seen.

Regions of in situ hybridization signal within the nucleolus of wheat and rye, which are not associated with resolvable chromatin in the EM (Figs 2B-D; 3B), detect highly decondensed rDNA. It is accepted that decondensation is correlated with expression (e.g. see Appels et al. 1986a). The fact that only highly decondensed rDNA occurs within nucleoli of rye (as well as occurring in wheat) confirms that the intranucleolar highly decondensed rDNA is transcribed.

Our data in cereals indicate that the rDNA from each active locus extends into the nucleolus from the perinucleolar condensed rDNA chromatin (Figs 1C, E; 3B). The same rDNA-carrying chromosome arm can be traced through the nucleolus and shown to emerge away from the entry point, usually on the opposite side (Fig. 3B). We propose calling the model where the NOR chromosome runs through the nucleolus, and does not loop back within it, the “extension model”. The extension model differs from all nucleolar models illustrated in detail by Jordan (1991), where rDNA from one locus is shown to enter and exit the nucleolus at adjacent sites (i.e. a loop). Rawlins and Shaw (1990) suggest from LM data that loops of chromatin extend into the nucleolus from the perinucleolar sites in pea, but their illustrations can be interpreted to support the extension model. The reconstructions of Jordan and Rawlins (1990) can also be interpreted to support the extension model. The universality of the extension model has still to be determined.

The data presented here directly relate the location of rDNA at interphase in the LM with the location and ultrastructure of the rDNA locus in the EM. Jordan and Rawlins (1990) examined the DNA fluorescence within Spirogyra nucleoli using optical sectioning. They stated that nearly all the fluorescence arises from bright foci and considered that the foci are the sites of rDNA (“nucleolar”) transcription. Noting that chloroplast DNA gives rise to punctate fluorescence despite being decondensed and transcribed, they concluded that the foci are “therefore not necessarily condensed”. The Spirogyra foci are similar to the wheat foci (Fig. 1A) that we show at the EM level to be condensed rDNA. The lack of such foci of rDNA in rye (Fig. 1E) implies that they are not needed for rDNA gene transcription and that the decondensed rDNA associated with transcription lies between the foci of condensed rDNA. This contrasts with the views of Jordan and Rawlins (1990), who consider that DNA causing low levels of DNA fluorescence “may simply correspond to DNA that links transcription sites together”.

Onion differs from the cereal species studied here because all the rDNA is intranucleolar (Martin et al. 1989). Martin et al. (1989) consider the intranucleolar rDNA location to be “strongly suggesting that all the genes are involved in nucleolar activity”. Our findings in wheat indicate that unexpressed genes can occur within the nucleolus.

We conclude that the intranucleolar foci (punctate sites) seen in the LM (Fig. 1A, C) and EM (Fig. 2B-D) micrographs of wheat nuclei are condensed and non-transcriptionally active sites of rDNA. The numerous condensed sites are separated by decondensed sites, a model that we term fragmented decondensation. In rye no condensed chromatin was observed in the nucleolus, implying that the distal end of the locus was entirely decondensed, a situation that we term terminal decondensation. In both rye and wheat, the thin lines of rDNA seen in the LM (Fig. 1A, C, E), and diffuse hybridization sites seen in the EM (Figs 2B-D, 3), are the sites of transcriptional activity of the locus within the nucleolus.

The nucleolar organizer region (NOR) is a tandem array of rDNA genes and intergenic spacers (Flavell, 1986). The molecular control and regulation of rDNA gene expression have been studied by Flavell et al. (1988, 1990). Their studies of aneuploids in which the number of major rDNA loci varied between two and six showed that the number of genes unmethylated at CCGG sites in the intergenic region is highly regulated and correlates with the relative activity of the locus. The evidence we present indicates that, in rye, the gene units at the distal (telomeric) end of the locus are those most often expressed, with unexpressed genes occurring outside the nucleolus at the proximal end (at least in the cultivar studied here). In wheat, activity probably occurs at genes interspersed along each locus, with the expressed sites separated by condensed chromatin within the nucleolus. It will be interesting to test whether increased ribosomal biosynthesis in rye is correlated with decondensation (and demethylation) of rDNA sequences outside the nucleolus, and their incorporation into the nucleolus. In wheat, genes from outside the nucleolus may be activated or the number or size of the sites of intranucleolar condensed chromatin might decrease when transcriptional activity increases.

We thank Prof. H.G. Schwarzacher, Drs T. Schwarzacher, F. Wachtler and I.J. Leitch for assistance, and Mr B. Allen for photography. We thank BP and Venture Research International for enabling this research to be done and are grateful to “Österreichische Fonds zur Fôrderung der wissenschaftlichen Forschung” (grant P7820 MED)for additional support.