Journal of Cell Science 104, 1199-1205 (1993)
Printed in Great Britain © The Company of Biologists Limited 1993
1199
A study on nucleolar DNA: isolation of DNA from fibrillar components and
ultrastructural localization of different DNA probes*
Pavel Hozák1,†, Christian Schöfer2, James Sylvester3 and Franz Wachtler2
1Institute of Experimental Medicine, Academy of Sciences, Legerova 61, 120 00 Prague
2Histologisch-Embryologisches Institut der Universität Wien, Schwarzspanierstrasse 17,
3Department of Pathology, Hahnemann University, Philadelphia, PA 19102, USA
2, Czech Republic
A-1090 Wien, Austria
*This article is dedicated to the 65th birthday of Prof. H. G. Schwarzacher
†Author for correspondence
SUMMARY
The nature and localization of DNA contained in the fibrillar centres and the dense fibrillar component (the fibrillar complex) in the nucleoli, was studied in human
LEP cells, Sertoli cells, spermatogonia A and in mitotic
chromosomes of stimulated lymphocytes. A novel procedure for isolating the intact fibrillar complex from
LEP cells was used; the complex contains DNA that
hybridizes to secondary constrictions of mitotic chromosomes and to 28 S rDNA sequences, on Southern
blots. Electron microscopic DNA-DNA in situ hybridization was performed, with (a) a probe prepared from
DNA extracted from the fibrillar complex of LEP cells,
(b) a probe for human total genomic DNA, and (c) a
probe for the transcribed part of human rDNA. On the
basis of the results obtained we conclude that the ribosomal RNA genes in human Sertoli cells and spermatogonia A are predominantly associated with the dense
fibrillar component, including the border region
between fibrillar centres and the dense fibrillar component. The ribosomal RNA genes are the main, if not
exclusive, DNA type present in the fibrillar complex in
the studied cell types.
INTRODUCTION
rillar component to contain most of the rDNA (Wachtler et
al., 1989; Ghosh and Paweletz, 1990; Stahl et al., 1991;
Wachtler et al., 1992).
There may be several explanations for these discrepancies. The nucleolar morphology is highly variable in miscellaneous cell types and obviously depends on rRNA transcription intensity: the size and number of fibrillar centres
and the spatial distribution of individual nucleolar components are different among different types of cells. In DNADNA in situ hybridization experiments, the fine morphology is not always easy to recognize in ultrathin sections,
therefore it is important to use cell types with a well established and stable nucleolar morphology. Also, it can not be
excluded that the fibrillar centres and/or the dense fibrillar
component, although appearing morphologically homogenous in electron micrographs, contain functionally different zones, and that different nucleoli may have different
structure-function organizations.
In most approaches to the study of structure-function
relationships in nucleoli, the localization of distinct molecules in nucleolar components was performed. In this study,
we take the opposite route, by characterizing the molecular contents of a defined nucleolar structure, to more specifically test the nature of the DNA contained in the fibrillar
centres and in the dense fibrillar component, in human LEP
cells, and to characterize more closely this DNA. We used
In most active nucleoli, four morphologically distinct components can be recognized: fibrillar centres, a dense fibrillar component, a granular component and nucleolar chromatin. The first two components, both somewhat fibrillar
in structure, have been established to play a crucial role in
the transcription of rRNA. However, the functional organization of nucleoli, and the exact localization of rRNA transcription in particular, still remain under question. Autoradiographic detection of radiolabelled RNA precursors have
not given results with sufficient spatial resolution. Ultrastructural immunocytochemistry was able to show the presence of enzymes necessary for rRNA transcription (polymerase I and topoisomerase I) in fibrillar centres and in the
dense fibrillar component, but could not discriminate
between active and pool molecules (for reviews see Derenzini et al., 1990; Raška et al., 1990; Reeder, 1990; Scheer
and Benavente, 1990; Hernandez-Verdun, 1991; Jordan,
1991 and Thiry et al., 1991). Likewise, the rDNA distribution was studied by DNA-DNA in situ hybridization without any universally accepted conclusion. In the few cell
types studied so far, one group of papers claims the fibrillar centres to be the exclusive site of rDNA (Thiry and
Thiry-Blaise, 1989; Puvion-Dutilleul et al., 1991; Thiry et
al., 1991), whereas the second group found the dense fib-
Key words: nucleolus, rDNA, transcription, in situ hybridization
1200 P. Hozák and others
nucleolar disintegration, induced by a brief hypotonic treatment, followed by the isolation of the two fibrillar parts
(the fibrillar complex) of the nucleolus. Using this procedure, nucleolar fibrillar complexes can be isolated with a
high level of integrity (Hozák et al., 1990; Hozák et al.,
1992). The DNA found to be present in the fibrillar complexes was extracted, tested by Southern blotting with a 28
S rDNA probe, and with a probe for alphoid satellite DNA,
which was known to be a non-ribosomal sequence present
in nucleoli (Kaplan et al., 1992), and itself used as a probe
for DNA-DNA in situ hybridization. Morphometric analysis and statistical evaluation were performed in order to
obtain precise information on the location of hybridization
signals with respect to nucleolar components.
MATERIALS AND METHODS
Cells
For the isolation of DNA from nucleoli, the human diploid cell
line LEP, originating from embryonic lung epithelium (ÚSOL,
Czech Republic), was used. The cells were grown in MEM supplemented with 10% fetal calf serum, glutamine and antibiotics.
For in situ hybridization studies, three cell types were used:
human lymphocytes from healthy 25 year old male donors, after
72 h of phytohaemagglutinin stimulation; human Sertoli cells and
human spermatogonia A. The testicular samples originated from
patients who underwent orchidectomy without any clinical treatment prior to surgery; the testicular histology was normal.
Conventional electron microscopy
Samples of LEP-cells or cell fractions (see below) were fixed in
2.5% glutaraldehyde in 0.1 M Sörensen buffer (SB), pH 7.2 for
1 h at 4°C; testicular tissue was fixed in 2.5% glutaraldehyde and
0.5% paraformaldehyde, in 0.2 M phosphate buffer, pH 7.2. Samples were postfixed in 1% OsO4, dehydrated in a graduated series
of ethanol rinses and embedded in Epon.
Preparation of cells for in situ hybridization
For light microscopy, spreads of stimulated lymphocytes were prepared using a standard cytogenetic protocol (Schwarzacher and
Wolf, 1974). For electron microscopy, the testicular tissue was
fixed in 2.5% glutaraldehyde and 0.5% paraformaldehyde, in 0.2
M phosphate buffer pH 7.2, at 4°C, dehydrated in a graded series
of ethanol rinses and embedded in LR White medium according
to the manufacturer’s instructions. Ultrathin sections were
mounted on gold grids.
Isolation of DNA from nucleolar fibrillar
complexes
The LEP cells were rinsed briefly in PBS, harvested by trypsinization, washed twice in 0.1 M SB, pH 7.2, 37°C and incubated for
5 min in 0.035 M SB at 37°C. All of the following procedures
were carried out at 0-4°C in the presence of 0.05 mM MgCl2, 0.1
mM PMSF, 10 µg/ml leupeptin, 10 µg/ml pepstantin and 1 µg/ml
aprotinin. Cells were spun down at 400 g for 5 min and homogenized in 40 volumes of 0.05 M SB with 0.4% Nonidet P-40 by
20 strokes in a Dounce homogenizer. The fraction containing the
nuclei was collected by 100 g centrifugation and purified by centrifugation on a saccharose gradient (0.25 M/0.88 M). Nuclei were
then sonicated in 0.34 M sucrose, on ice, three times for 3 s (UZD
21, A=40, 20% power output). Released fibrillar complexes were
purified by spinning through a 0.88 M saccharose solution at 2500
g for 10 min. All individual steps of the isolation procedure were
monitored by electron microscopy. The DNA was extracted by
the standard phenol/chloroform method and precipitated by
ethanol and sodium acetate.
Characterization of fibrillar complex DNA
DNA extracted from nucleolar fibrillar complexes was restricted
with EcoRI and, following electrophoresis, transferred onto a
membrane, then probed for the presence of 28 S rDNA using a
pA probe (Erickson et al., 1981) and alphoid satellite DNA, originating from human chromosome 21 and known to be at the centromere of chromosomes.
DNA-DNA in situ hybridization on cells
Three different probes were used: (1) DNA extracted from nucleolar fibrillar complexes and digested with EcoRI, (2) human total
genomic DNA, prepared by EcoRI and HaeII digestion of placental DNA type XIII (Sigma), and (3) a probe for the transcribed
part of human rDNA namely an EcoRI-BglII fragment of the transcription unit (includes parts of 18 S and 28 S rDNA, 5.8 S rDNA
and both internal transcribed spacers; for details see Wachtler et
al., 1989). All probes were labelled by biotin or digoxigenin, using
a nick-translation kit (Boehringer) according to the manufacturer’s
instructions.
The hybridization protocol was as described in Wachtler et al.
(1989). Briefly, the slides, or ultrathin sections were pre-treated
with RNase A (100 µg/ml in 2×SSC, 1 h at 37°C) and with proteinase K (0.5 µg/ml in 20 mM Tris-HCl, 2 mM CaCl 2, pH 7.4,
7 min at 37°C), then washed and the hybridization mixture (consisting of 2 ng/µl of the probe, 100 ng/µl salmon sperm DNA,
100 ng/µl yeast RNA and 10% dextran sulphate in 50% deionized formamide in 2×SSC) was placed onto them. DNA denaturation of the probes and the specimens was carried out simultaneously by incubation at 90°C for ten min followed by an
overnight hybridization at 37°C, in a moist chamber. In the case
of chromosome preparations, hybridization was performed simultaneously with both the DNA probe originating from DNA isolated from fibrillar complexes and labelled with digoxigenin; and
with the DNA probe for the transcribed part of the rDNA, labelled
with biotin. In order to remove DNA/RNA hybrids, preparations
were, after washing, treated with RNase H, 40 µg/ml, according
to the manufacturer’s instructions (Boehringer).
Detection of the hybridization signal
For simultaneous detection of different probes and amplification
of the signal, after repeated washes, the slides were incubated in
4×SSC with 0.05% Tween 20, 5% non-fat dry milk, 5 µg/ml
avidin-FITC (Vector) and 50 µg/ml rhodamine-labelled antidigoxigenin Fab fragments (Boehringer), for 30 minutes at room
temperature. Following a few more washes, slides were incubated
with 5 µg/ml biotinylated goat anti-avidin (Vector) and rhodamine-labelled rabbit anti-sheep IgG (50 µg/ml, Biomedica), then
washed again and finally incubated with avidin-FITC and rhodamine-labelled goat anti-rabbit IgG (Souther Biotechnology
Associates). The slides were mounted in a hydrophilic (10 µg/ml)
medium containing 0.1 µg/ml DAPI (Boehringer) and observed
in a Leitz epifluorescence microscope, using three optical channels selective for rhodamine (digoxigenin-labelled probes), FITC
(biotin-labelled probes) and DAPI (total DNA). For ultrastructural
studies, the sections were incubated with a goat anti-biotin antibody (Vector Anti-Biotin, affinity purified, 1:300) followed by an
incubation with a rabbit anti-goat antibody (1:20) coupled with 5
nm gold particles (Amersham).
Controls
In all hybridization experiments, two control hybridizations were
carried out in parallel: one with a biotinylated plasmid DNA of
DNA distribution in nucleoli 1201
similar GC-content to the human rDNA probe (pGEM, Promega)
and another with no DNA probe. The light microscope analysis
system was tested by applying both the digoxigenin and the biotin
detection systems to slides hybridized with one probe only; a possible leak between the two optical channel was tested. No inappropriate signal could be detected in all these checks. Treatment
with RNase H after in situ hybridization did not affect the intensity or distribution of the signal.
Statistical evaluation of morphometric results
The densities of gold particles over the nucleolar components, the
non-nucleolar area of the nuclei, and the cytoplasm were determined using the software Sigmascan (Jandel-Scientific), in 15-35
cells for each set of experiments. The statistical significance of
the results was assessed using Student’s two-sided t-test.
RESULTS
Isolated fibrillar complexes
LEP cell nucleoli possess well defined nucleolar components with a typical spatial arrangement: the fibrillar centres are surrounded by dense fibrillar component layers and
together, they are embedded in the granular component
(Fig. 1). After isolation, the resulting fraction is highly
enriched in fibrillar complexes, although small residues of
other nuclear material can occasionally also be found (Fig.
2).
Characteristics of DNA extracted from fibrillar
complexes
In order to assess the chromosomal localization of the DNA
that we had extracted from isolated fibrillar complexes, a
double-labelling in situ DNA hybridization was performed
on human metaphase chromosomes. Fig. 3 gives a com-
parison of the hybridization signal of the DNA probes for
the transcribed part of the human rDNA (Fig. 3A), and for
the DNA extracted from the fibrillar complexes (Fig. 3B).
The signal consists mostly of double dots; when compared
to total DNA staining (Fig. 3C), these doublets can be localized to the secondary constrictions of acrocentric chromosomes. Within the limits of light microscopy, the distribution of the signal for the ribosomal genes and for the DNA
from the fibrillar complexes is identical.
Further proof that the fibrillar complex DNA contains
ribosomal gene sequences was obtained by the fact that a
probe for 28S rDNA hybridized to it, but not one for alphoid
satellite DNA (data not shown).
Nucleolar distribution of different DNA probes in
Sertoli cells and in spermatogonia A
Sertoli cell nucleoli exhibit a very stable and unique spatial organization. They contain, usually, one large fibrillar
centre; the dense fibrillar component surrounds the fibrillar
centre, forming a coat with radial protrusions. Next to the
fibrillar centre, a large block of granular component can be
found. Some dense fibrillar component branches are also
frequently found at the periphery of the granular block,
especially on the opposite side to the fibrillar centre (Fig.
4A).
The label from all three DNA probes used for the in situ
hybridizations were shown by electron microscopy to be
distributed in a similar way over the nucleoli. The probes
for total genomic DNA (Fig. 4B), DNA from fibrillar complexes (Fig. 4C), and for the transcribed part of ribosomal
genes (Fig. 4D) hybridized predominantly to the dense fibrillar component. Gold particles are visible in different
zones of the dense fibrillar component; they are seen in the
layer around the fibrillar centre. An intense label is also
Fig. 1. The LEP cells possess nucleoli with well distinguishable fibrillar centres (C), dense fibrillar component (D), and granular
component (G). Magnification, ×20,800.
Fig. 2. Fibrillar complexes, isolated after short hypotonic treatment of LEP cells. The electron micrograph shows a high level of purity of
isolated fraction. Magnification, ×35,000.
1202 P. Hozák and others
rillar component (Fig. 5A). Again, if hybridized with the
DNA probe from fibrillar complexes, most of the label is
found in the dense fibrillar component attached to the fibrillar centres. However, it is more difficult than in Sertoli
cells to see the border line between the fibrillar centre and
the dense fibrillar component; nevertheless, there is virtually no label in the interior of the fibrillar centre. Significant labelling is found in those threads of the dense fibrillar component that are remote from the fibrillar centres
(Fig. 5B).
DISCUSSION
Fig. 3. Detection of two different DNA probes after in situ
hybridization to human metaphase chromosomes. Arrows indicate
signal pattern. (A) Pattern obtained with the probe for the
transcribed part of rDNA (FITC labelling) and (B) pattern
obtained with the probe isolated from the nucleolar fibrillar
complexes (rhodamine labelling). The two signal patterns can be
seen to be identical. Typical doublets (nucleolar organizers) can
be seen in secondary constrictions of chromosomes, when
comparing A and B to total DNA stained with DAPI (C).
found associated with the trabeculae of the dense fibrillar
component located far away from the fibrillar centres. On
the contrary, the fibrillar centres were almost devoid of
gold particles; few were seen on their periphery (Fig. 4D).
The label over the fibrillar centres was not significantly
different from background values (Fig. 6). In addition,
when the probe for total genomic DNA was used (Fig. 4B),
clusters of gold particles were visible throughout the chromatin outside of nucleolus.
Human spermatogonia A nucleoli also have a very stable
and characteristic structure (Hartung et al., 1990). The
prominent fibrillar centres are always attached to the
nuclear envelope and are surrounded by meandering
threads of the dense fibrillar component that extend into
the nuclear interior. Typically, some granular component
is found as a continuation of the strands of the dense fib-
DNA distribution in nucleoli
Condensed nucleolar DNA can be seen morphologically as
peri- or intranucleolar chromatin clumps. Using cytochemical or immunoelectron microscopy approaches, decondensed DNA has been localized in both the fibrillar centres
and the dense fibrillar component of different cell types.
All papers presented so far are in agreement on the presence of DNA in the fibrillar centres, although the intensity
and localization of label varies and, in some cell types, only
the peripheral zone of the fibrillar centres was found to contain a significant amount of DNA. However, contradictory
results were obtained regarding the dense fibrillar component: some authors did not detect DNA in this component
in Ehrlich tumour cells or in mouse hepatocytes, but others
detected DNA predominantly in the dense fibrillar component of Sertoli cells or in stimulated lymphocytes (for
reviews see Wachtler et al., 1989; Ghosh and Paweletz,
1990; Stahl et al., 1991 and Wachtler et al., 1992). This
discrepancy has not yet been sufficiently explained. It was
shown recently that the DNA found in nucleoli is not exclusively ribosomal DNA. For example, long interspersed
(LINE) KpnI elements which are absent from the ribosomal gene repeat, and a tandemly-repeated simple sequence
cluster derived from the short arm of chromosome 15, were
found, using light microscopy, in human lymphocyte nucleoli (Kaplan et al., 1992). Centromeres were also identified
as a frequent part of nucleoli (Ochs and Press, 1992). Until
now, the exact localization and function of non-ribosomal
DNA in nucleoli was unknown. In our experiment, a difference in the distribution of hybridization signal between
total genomic DNA and transcription unit DNA probes
would indicate that, besides rDNA, other DNA is involved.
However, we did not detect such a difference in Sertoli
cells; both probes hybridized predominantly and significantly to the dense fibrillar component and not to the fibrillar centres. Our results confirm previous conclusions
about predominant rDNA distribution in the dense fibrillar
component in human Sertoli cells (Wachtler et al., 1989;
Wachtler et al., 1992) and extend this conclusion, at the
ultrastructural level, to human spermatogonia A. However,
these results do not necessarily indicate an absolute absence
of DNA from fibrillar centres. It was occasionally possible
to find small clusters of gold particles in the border region
between the two components in question. Also, the concentration of DNA in the relatively large volume of fibrillar centres could be very low and not all of the DNA accessible to the probes. As a consequence, small amounts of
DNA distribution in nucleoli 1203
Fig. 4. Nucleoli of human Sertoli cells. (A) General electron microscopic view; (B,C,D) after in situ hybridization to sectioned material
with three different DNA probes: (B) for total genomic DNA, (C) for DNA isolated from the fibrillar complex of nucleoli, (D) for the
transcribed part of rDNA. The gold particles are found almost exclusively over the dense fibrillar component of the nucleolus
(arrowheads). Magnification, (A) ×21,000; (B) ×32,000; (C) ×24,000; (D) ×24,900.
DNA could escape detection by in situ hybridization, given
the low signal/background ratio of this technique at the
electron microscopic level.
We conclude therefore that ribosomal RNA genes in
human Sertoli cells and spermatogonia A are predominantly
associated with the dense fibrillar component, including the
1204 P. Hozák and others
Fig. 5. Nucleoli of human spermatogonia A. (A) General electron microscopic view; (B) localization of the probe for DNA isolated from
the fibrillar complex of nucleoli. Most labelling is concentrated over the dense fibrillar component of the nucleolus (arrowheads).
Magnification, (A) ×33,600; (B) ×36,800.
Total genomic DNA
Fibrillar complex DNA
Transcription unit
Fig. 6. Statistical evaluation of gold label density in different cell portions, as estimated in Sertoli cells after in situ hybridization with
probes for (A) total genomic DNA, (B) DNA isolated from fibrillar nucleolar components, and (C) the transcribed part of rDNA. Data
show the mean number of gold particles per µm 2 and the standard deviation, *P < 0.01, **P < 0.001.
border region between the fibrillar centres and the dense
fibrillar component.
The nature of DNA in the fibrillar complex
The hypotonic isolation procedure allowed us to study the
DNA contained in the two nucleolar fibrillar components.
Southern blot hybridization with the 28 S rDNA probe indicated that the isolated DNA contained ribosomal RNA
genes, but it did not hybridize with a probe for alphoid
satellite DNA, known to be a non-ribosomal sequence present in nucleoli (Kaplan et al., 1992). The results show that
we were predominantly isolating the fibrillar complexes and
that the method we used is a promising tool for their further study.
If hybridized in situ, fibrillar complex DNA hybridized
mainly to chromosomal regions recognizable as nucleolar-
DNA distribution in nucleoli 1205
organizing regions at the light microscope level, and to the
dense fibrillar component at the electron microscope level.
This can be interpreted (regarding the discussed problems
above) in a similar way to that in the previous paragraph
and points again to the ribosomal RNA genes as being the
main, if not exclusive, DNA type present in the fibrillar
complex and, more specifically, in the dense fibrillar component.
In conclusion, we would like to stress that it is probably
impossible to explain easily all of the current discrepancies
in the localization of rRNA transcription in various cell
types. It is questionable as to how much we can rely on the
apparent morphological homogeneity of individual nucleolar components. For example, if studying silver staining or
a variation in DNA hybridization signal, usually, we do not
see a diffuse labelling pattern. More typically, areas with a
diverse labelling intensity in the fibrillar centres and the
dense fibrillar component can be found. Also, the borderline between these two components is variable; it can be
sharp but also more or less gradual. We speculate therefore
that the two fibrillar nucleolar components could have a
zonal arrangement, possibly with a different activity or
function. Moreover, there are several types of nucleoli;
whether it is possible to reduce all varieties to one universal model still remains to be answered.
This work was supported by the ‘Österreichischer Fonds zur
Förderung der wissenschaftlichen Forschung’ (P 7820-MED).
REFERENCES
Derenzini, M., Thiry, M. and Goessens, G. (1990). Ultrastructural
cytochemistry of the mammalian nucleolus. J. Histochem. Cytochem. 38,
1237-1256.
Erickson, J. M., Rushford, C. L., Dorney, D. J., Wilson, G. H. and
Schmickel, R. D. (1981). Structure and variation of human ribosomal
DNA: molecular analysis of cloned fragments. Gene 16, 1-9.
Ghosh, S. and Paweletz, N. (1990). Localization of ribosomal cistrons at
the ultrastructural level by in situ hybridization technique. Cell Biol. Int.
Rep. 14, 521-525.
Hartung, M., Wachtler, F., de Lanversin, A., Fouet, C., Schwarzacher,
H. G. and Stahl, A. (1990). Sequential changes in the nucleoli of human
spermatogonia with special reference to rDNA location and transcription.
Tiss. and Cell 22, 25-37.
Hernandez-Verdun D. (1991). The nucleolus today. J. Cell Sci. 99, 465471.
Hozák, P., Zatsepina, O. V. and Likovský, Z. (1990). In situ separation of
nucleolar components by hypotonic treatment of cells. Biol. Cell 68, 167170.
Hozák, P. Zatsepina, O.V. and Likovský, Z. (1991). Nucleolar changes in
hypotonic treatment and the possibility of isolation of nucleolar fibrillar
complexes. In Nuclear Structure and Function (ed. J. Harris and I. B.
Zbarsky), pp. 197-202, Plenum: London, New York.
Hozák, P., Géraud, G. and Hernandez-Verdun, D. (1992). Revealing
nucleolar architecture by low ionic strength treatment. Exp. Cell Res. 203,
128-133.
Jordan, E. G. (1991). Interpreting nucleolar structure: where are the
transcribing genes? J. Cell Sci. 98, 437-442.
Kaplan, F. S., Murray, J., Sylvester, J. E., Gonzales, I. L., O’Connor, P.,
Doering, L., Muenke, M., Emanuel, B. S. and Zasloff, M. A. (1992).
Repetitive element domains in human lymphocyte nuclei. Genomics, in
press.
Ochs, R. L. and Press, R. I. (1992). Centromere autoantigens are associated
with the nucleolus. Exp. Cell Res. 200, 339-350.
Puvion-Dutilleul, F. , Laithier, M. and Puvion, E. (1991). Nucleolar
organization of HeLa cells as studied by in situ hybridization.
Chromosoma 100, 395-409.
Ras̆ka, I., Ochs, R. L. and Salamin-Michel, L. (1990).
Immunocytochemistry of the cell nucleus. Elect. Microsc. Rev. 3, 301353.
Reeder, R. H. (1990). rRNA synthesis in the nucleolus. TIG 6, 390-395.
Scheer, U. and Benavente, R. (1990). Functional and dynamic aspects of
the mammalian nucleolus. BioEssays 12, 14-21.
Schwarzacher, H. G. and Wolf, U. (1974). Methods in human
cytogenetics. Springer-Verlag: Berlin, Heidelberg, New York.
Sollner-Webb, B. and Mougey, E. B. (1991). News from the nucleolus:
rRNA gene expression. TIBS 16, 58-62.
Stahl, A., Wachtler, F., Hartung, M., Devictor, M., Schoefer, C.,
Mosgoeller, W., Delanversin, A. and Fouet, C. (1991). Nucleoli,
nucleolar chromosomes and ribosomal genes in the human spermatocyte.
Chromosoma 101, 231-244.
Thiry, M. and Thiry-Blaise, L. (1989). In situ hybridization at the electron
microscopic level: an improved method for precise localization of
ribosomal DNA and RNA. Eur. J. Cell Biol. 50, 235-243.
Thiry, M., Scheer, U. and Goessens, G. (1991). Localization of nucleolar
chromatin by immunocytochemistry and in situ hybridization at the
electron microscopic level. Electron Microsc. Rev. 4, 85-110.
Wachtler, F., Hartung, M., Devictor, M., Wiegant, J. and Stahl, A.
(1989). Ribosomal DNA is located and transcribed in the dense fibrillar
component of human Sertoli cell nucleoli. Exp. Cell Res. 184, 61-71.
Wachtler, F., Schoefer, C., Mosgoeller, W., Weipoltshammer, K.,
Schwarzacher, H. G. Guichaoua, M., Hartung, M. and Stahl, A.
(1992). Human ribosomal RNA gene repeats are localized in the dense
fibrillar component of nucleoli - light and electron microscopic in situ
hybridization in human Sertoli cells. Exp. Cell. Res. 198, 135-143.
(Received 29 October 1992 - Accepted 7 January 1993)