Nucleic Acids Research, Vol. 19, No. 1 11
Locating transcribed and non-transcribed rDNA spacer
sequences within the nucleolus by in situ hybridization
and immunoelectron microscopy
Marc Thiry and Lydia Thiry-Blaise1
Laboratoire de Biologie cellulaire et tissulaire, Institut A. Swaen, University de Liege, 20 rue de
Pitteurs, B-4020-Liege, Belgium and 1 Laboratoire de Genetique appliquee, Universite Libre de
Bruxelles, Centre de Recherches industrielles, 24 rue de I'lndustrie, B-1700-Nivelles, Belgium
Received October 16, 1990; Revised and Accepted December 4, 1990
ABSTRACT
Immunoelectron microscopy and in situ hybridization
have been used to investigate the precise location of
transcribed and non-transcribed rDNA spacer
sequences. Whereas a 5'-extemal transcribed spacer
sequence is predominantly visualized in the flbrillar
centers of nucleoli, a non-transcribed spacer sequence
is preferentially detected in the interstices, in close
contact with the flbrillar centers and which interrupt the
surrounding dense fibrillar component. Occasionally
these two spacers are also observed In clumps of
dense nucleolus-associated chromatln. These
observations provide insights into the organization of
rlbosomal repeats within the nucleolus.
INTRODUCTION
The rRNA genes of higher eukaryotes are present in multiple
copies and are organized as tandem repeats (1,2). Each repeating
unit comprises a transcribed region flanked by a stretch of DNA
known as the non-transcribed spacer. The transcription unit
contains, in addition to the genes coding for 18, 5.8 and 28S
rRNAs, an external transcribed spacer at the 5' end and two
internal transcribed spacers separating the coding sequences.
In interphase cells, rRNA genes are expressed in the nucleolus.
Extensive electron microscopic studies have provided in-depth
knowledge of this nuclear compartment's ultrastructure. These
studies have led to recognize the following elements as general
components of the mammalian interphase nucleolus (3, 4): the
flbrillar centers, the dense flbrillar component, the granular
component, the nucleolar interstices, and nucleolus-associated
chromatin. On the other hand, the precise relationships between
these structural compartments and their molecular functions have
been but partially elucidated.
We have recently investigated the location of rRNA genes
within Ehrlich tumor cells by in situ hybridization at the electron
microscope level (5). Using a 1.95 kb probe containing 1.45 kb
of 18S rDNA, we detected hybrids principally at the periphery
of the flbrillar centers.
In this report, we have extended our analysis to other regions
of rDNA repeats. We have investigated the precise location of
two rDNA spacer sequences: a transcribed external spacer and
a non-transcribed spacer. The latter represents about two-thirds
of the (approximately) 44 kb mouse rDNA repeat (6).
MATERIALS AND METHODS
Strains
Ehrlich ascites tumor cells from the peritoneal cavity of C 57
Bl mice were cultured as monolayers in Petri dishes according
to Lepoint and Bassleer (7). The Escherichia coli strain C 600
PyrFA(Pro-ArgF-Lac) Argl PyrF thi r~ m~ was used for
transformation (8).
Preparation of cells for electron microscopy
The monolayer cultures were scraped off the dishes and
centrifuged at low speed to form a pellet. Small fragments of
the pellet were fixed in 0.2% glutaraldehyde or in 4%
formaldehyde in 0.1 M Sorensen's buffer (pH 7.4) at 4°C for
15 min or 60 min, respectively. After fixation, the cells were
washed in Sorensen's buffer, dehydrated through graded ethanol
solutions and then processed for embedding in Lowicryl K4M
using the technique of Roth et al. (9). Ultrathin sections were
collected in platinum rings and stored on distilled water until use.
Enzymes
Restriction endonucleases were purchased from Boehringer
Mannheim/FRG and Bethesda Research Laboratories Inc. (BRL;
Gibco Europe SA, Gent, Belgium) and were used as suggested
by the manufacturer. T4 ligase was from Boehringer and E. coli
DNA polymerase I was from BRL.
DNA
The mouse ribosomal fragment Mr974(10) was a kindly gift, as
a 11.35 kb EcoRI insert in pUC9 from I. Grummt (University
of Wurzburg, Wurzburg, FRG). DNA manipulations were as
described by Maniatis et al. (11).
In vitro DNA labeling
Biotinylated plasmid DNA was prepared by nick-translation (12)
with Bio-11-dUTP. Purification of the different labeled probes
was achieved by gel filtration on Sephadex G-50, tracking the
12 Nucleic Acids Research, Vol. 19, No. 1
DNA by 0.5% blue dextran, and a subsequent ethanol
precipitation.
In situ hybridization
Sections of cells were either heated in water at 100 C for 5 min
or incubated for 30 min at 37 °C in 0.1 mg/ml protease of Bacillus
polymyxa (Boehringer Mannheim) in 0.1 M Sorensen's buffer
(pH 7.4) followed for 120 min in 0.1 N NaOH at room
temperature and immediately cooled to 4°C before being
hybridized as previously described (5).
These denaturation conditions have previously been shown to
provide an intense labeling after the application of an
immunocytochemical approach involving a monoclonal antisingle-stranded DNA (5).
After hybridization, biotinylated hybrids were detected
according to Thiry and Thiry-Blaise (5).
Finally, the sections were stained with uranyl acetate and lead
citrate before examination in JEOL CX 100 II electron
microscope at 60 KV.
of indirect immunoglobulin-gold methods. However, the analysis
of about a hundred nucleoli reveals that the location of label within
the nucleolus differs for the tworibosomalrepeat spacer sequence
probes. With pBRMrE (Fig. 2), gold is found distributed in small
clusters located preferentially in the fibrillar centers, frequently
at the periphery near the densefibrillarcomponent and interstices.
With pBRMrA (Figs. 3 and 4), on the other hand, label is
preferentially distributed over the small clumps of intranucleolar
chromatin which occupy the interstices interrupting the layer of
dense fibrils. It is rarely detected in the fibrillar centers. Neither
probe ever gives rise to labeling over the dense fibrillar
component or over the granular component. A few gold particles
can occasionally be seen over the perinucleolar condensed
chromatin.
The reaction's specificity was tested in several ways. First of
all, no labeling occurs when the spacer-sequence probe is replaced
by pBR322. Secondly, labeling is completely abolished if the
denaturation step or the probe is omitted. Finally, no labeling
occurs when only the colloidal-gold-coupled secondary antibody
is used.
RESULTS
A few years ago, Grummt et al. (10) cloned a 11.35 kb EcoRI
fragment Mr974 containing parts of the 18S region and adjacent
spacer sequences of mouse rDNA. Figure 1 illustrates the
organization of transcribed and non-transcribed regions on this
cloned ribosomal fragment and shows the Sail cleavage site
positions determined by these authors (13). Fragment C contains
1.45 kb of the 18S sequence and has previously been used as
an hybridization probe (5).
Here, we have biotinylated two fragments: large fragment A,
representing 3.8 kb of non-transcribed spacer, and small fragment
E, which contains 0.55 kb of external transcribed spacer. For
this purpose, the EcoRI-Sail fragment A was subcloned in
pBR322 (=pBRMrA) and the Sail fragment E was inserted into
pBR325 (=pBRMrE).
Both labeled probes were hybridized with denatured Lowicryl
ultrathin cell sections and the biotinylated hybrids were then
visualized by means of an indirect immunogold labeling
procedure using an anti-biotin antibody in consort with secondary
antibody coupled to colloidal gold particles 5 nm in diameter.
Under these conditions, observed labeling of many Ehrlich tumor
cells is restricted to the nucleoli, whatever the ribosomal repeat
probe or the fixation used. The labeling is represented by small
clusters of gold particles. This clustering label results from the
fact that each hybrid contains several biotinylated nucleotides
which may be independently detected by our immunogold
procedure. In addition, each biotin molecule itself may be
revealed by several gold particules due to amplification effect
Direction of transcription
NTS
3.8
EcoRI
I
Sail
1.75
1.75
II
Sill
3.3
B
EcoRI
E
C
Figure. 1. Schematicrepresentationof the mouse rDNA fragment Mr974.
DISCUSSION
As already suggested by previous biochemical data (14) and
recent in situ hybridization experiments (5), these results show
that Ehrlich-tumor-cell ribosomal repeating units are
predominantly located inside the nucleolus. They specifically
reveal that an external transcribed spacer sequence is located
preferentially in the fibrillar centers whereas a non-transcribed
spacer sequence is to be found primarily in the interstices at the
boundaries of the fibrillar centers.
Although autoradiographic studies seemed to point to the dense
fibrillar component as the site of pre-rRNA synthesis (4, 15),
recent data strongly suggest that rDNA transcription occurs,
rather, in the fibrillar centers. Among these data is the fact that
DNA has never been clearly visualized in the dense fibrilllar
component of mammalian-cell nucleoli (16, 17, 18, 19). In
contrast, many studies based on a variety of methods have
demonstrated its presence in the fibrillar centers (4, 16), especially
at their periphery (17, 18). Furthermore, electron microscopy
has shown that the enzymes RNA polymerase I and DNA
topoisomerase I, involved in transcribing rRNA genes, are
selectively concentrated in the peripheral regions of the fibrillar
centers (20, 21). Our results, reported here and previously (5),
indicate that two different transcribed rDNA sequences are
located preferentially in the fibrillar centers, especially near their
periphery. The fibrillar centers are thus the sole nucleolar
structures where transcribed regions of ribosomal repeats and
the enzymes involved in rDNA transcription are located together.
This report further shows that a non-transcribed ribosomal
repeat spacer sequence is predominantly detected outside the
fibrillar center, in the interstices which interrupt the layer of dense
fibrils. The presence of DNA in these small nucleolar cavities
bordering on the fibrillar centers has been evidenced by various
ultrastructural techniques (17, 18, 22, 23). It is also interesting
to note that, contrary to the DNA contained in thefibrillarcenters,
this DNA is generally quite conspicuous and exhibits an electron
density similar to that of perinucleolar chromatin. Moreover, an
in situ Feulgen-like osmium-ammine reaction reveals this
intranucleolar DNA as chromatin clumps or fibers with a
nucleosomal configuration, whereas the fibrillar-center DNA
looks more like extended non-nucleolsomal filaments (24).
Nucleic Acids Research, Vol. 19, No. 1 13
However, one has to ask whether the absence of gold particules
in the fibrillar centers observed with a non-transcribed probe
could not result from detection problems rather than from a true
absence of non-transcribed rDNA in these nucleolar regions. As
a first point, a lower density of DNA in the fibrillar center could
limit its detection, but this restriction would also apply to the
two transcribed regions and it is not the case. Secondly, the probe
size could affect the labeling efficiency. Nevertheless, whatever
iff^'M
Figure. 2. Electron microscopic immunolocalization ofribosomalexternal transcribed spacer sequence on sections of Ehriich tumor cell nucleoli after in situ hybridization.
Each of the two fibrillar centers (FC) exhibits a cluster of gold particles (arrowheads). F: dense fibrillar component; G: granular component; I: nucleolar interstices.
0.2% glutaraldehyde/Lowicryl K4M/ procease-NaOH. Bar - 0.1 jim.
14 Nucleic Acids Research, Vol. 19, No. 1
•; V .
«
Si
> -
•
•
•
.
.
"
.
Figures 3 and 4. Electron microscopic immunolocalizatioa of ribosomal non-transcribed spacer sequence. General view (Fig. 3) or details (Fig. 4) showing a cluster
of gold particles (arrowheads) over interstice (I) contiguous to the fibrillar components. C: condensed chromatin; FC: fibrillar center; F: dense fibrillar component;
G: granular component. 0.2% glutaraldehyde/Lowicryl K4M/ protease-NaOH. Bar = 0.1 /im.
Nucleic Acids Research, Vol. 19, No. 1
the original fragment size, the nick-translation provides small
single-stranded probe of the same average size (12). Moreover,
the non-trancribed fragment is longer than the two transcribed
ones and offers so more hybridization possibilities. Finally, a
non-transcribed spacer of ribosomal repeats engaged in
transcriptional activity might be thought inaccessible to the probe
due to the tight association with the nucleolar structural proteins
(25). However, 5'-external transcribed spacer sequences are also
present in the nucleolar matrix of mouse L-cells (25).
In electron microscopic spread preparations of transcribing ribo
somal repeats, transcription-unit chromatin invariably appears as
a thin, smooth axis and is clearly distinguishable from spacer
chromatin which is often characterized by variable numbers of
particles, irregular in shape and distribution, and generally
somewhat smaller than those of nucleosomes (26, 27). In some
species (28, 29, 30), particularly in mammalian cells (31), spacer
chromatin reportedly looks identical to the bulk of
transcriptionally inactive chromatin, whose regular, beads-ona-string appearance indicates a nucleosomal organization (26, 27).
Furthermore, DNase I-hypersensitive sites have recently been
visualized in situ in the fibrillar centers as well as in the interstices
surrounding them (32). In previous biochemical experiments,
such sites were identified in the actively transcribed ribosomal
genes of various species, especially in their promoter sequences
(33), which have been mapped in the non-transcribed spacer (34,
35).
Taken together, these observations strongly suggest that the
non-transcribed spacer sequences visualized in the interstices are
indeed of two kinds: they belong to both transcribing and inactive
ribosomal repeats.
In conclusion, a new level of compartmentalization of
ribosomal repeats appears to emerge. First, there is the
morphological distinction between inactive repeats contained in
the nucleolus-associated dense chromatin and transcriptionally
active repeats to be found inside the nucleolus. The active repeats
are further distributed into distinct nucleolar compartments: the
transcribed part in the fibrillar centers and the non-transcribed
part in the interstices in contact with the fibrillar centers.
ACKNOWLEDGEMENTS
The authors are grateful to Prof. G. Goessens (University of
Liege) for encouraging discussions and for critical reading of the
manuscript, and to Prof. I. Grummt (University of Wurzburg)
for providing the Mr974 fragment. They also achowledge the
skillful technical assistance provided by Miss F. Skiv£e. M.T.
is 'Charge" de Recherches aupres du Fonds National de la
Recherche Scientifique de Belgique'.
REFERENCES
1. Long, E.O., and Dawid, I.B. (1980) Ann. Rev. Biochem., 49, 727-764.
2. Hadjiolov, A. (1985) Cell Biol. Monographs, Springer-Verlag, Wien, New
York, 12, 1-263.
3. Smetana, K., and Busch, H. (1974) In Busch, H. (ed.), The Cell Nucleus.
Academic Press, New York.l, 73-147.
4. Goessens, G. (1984) Int. Rev. Cyt., 87, 107-158.
5. Thiry, M., and Thiry-Blaise, L. (1989) Eur. J. Cell Biol., 50, 235-243.
6. Cory, S., and Adams, J. (1977) Cell, 11, 795-805.
7. Lepoint, A., et Bassleer, R. (1973) C.R. Acad. Sci, 276, 441-443.
8. Messing, J. (1983) In Grossman, L., and Moldave, K. (eds.), 10, 20-78.
9. Roth, J., Bendayan, M., Carlemalm, E., Villiger, W., and Garavito, M.
(1981) J. Histochem. Cytochem., 29, 663-671.
10. Grummt, I., Soellner, C , and Scholz, I. (1979) Nucleic Acids Res., 6,
1351-1369.
IS
11. Maniatis, T., Frisch, E., and J. Sambrook (1982) Molecular cloning : a
laboratory manual. Cold Spring Harbor Laboratory Press, Cold Spring
Harbor, New York, 368-369.
12. Rigby, P., Dieckmann, M., Rhodes, C., and Berg, P. (1977) J. Mol. Biol.,
113, 237-251.
13. Grummt, I., and Gross, H. (1980) Molec. gen. Genet., 177, 223-229.
14. Bachellerie, J.-P., Nicoloso, M., and Zalta, J.-P. (1977) Eur. J. Biochem.,
79, 23-32.
15. Fakan, S. (1986) Meth. Achiev. Exp. Path., 12, 105-140.
16. Scheer, U., Messner, K., Hazan, R., Raska, I., Hansmann, P., Falk, H.,
Spiess, E., and Franke, W. (1987) Eur. J. Cell Biol., 43, 358-371.
17. Thiry, M., Scheer, U., and Goessens, G. (1988) Biol. Cell, 63, 2 7 - 3 4 .
18. Thiry, M. (1988) Exp. Cell Res., 179, 2O4-213.
19. Goessens, G. (1976). Exp. Cell Res., 100, 88-94.
20. Scheer, U., and Rose, K. (1984) Proc. Natl., Acad. Sci., 81, 1431-1435.
21. Scheer, U., and Benavente, R. (1990) Bioessays, 12, 14-21.
22. Goessens, G. (1979) Cell Biol. Int. Rep., 3, 337-343.
23. Thiry, M., and Goessens, G. (1986) Exp. Cell Res., 164, 232-242.
24. Derenzini, M., Hemandez-Verdun, D., Farabegoli, F., Pession, A., and
Novello, F. (1987) Chromosoma, 95, 63-70.
25. Bolla, R., Braaten, D., Shiomi, Y., Brenda Hebert, M., and Schlessinger,
D. (1985) Mol. Cell Biol., 5, 1287-1294.
26. Franke, W., Scheer, U., Spring, H., Trendclenburg, M., and Zentgraf, H.
(1979) In Busch, H. (ed.), The Cell Nucleus. Academic Press, New York,
7, 4 9 - 9 5 .
27. Scheer, U., and Zentgraf, H. (1982) In Busch, H., and Rothblum, L. (eds.),
The Cell Nucleus. Academic Press, New York, 11, 143-176.
28. Foe, V. (1978) Cold Spring Harbor Symp. Quant. Biol., 42, 732-739.
29. Reeder, R., Me Knight, S., and Miller, O. (1978) Cold Spring Harbor Symp.
Quant. Biol., 42, 1174-1177.
30. Greimers, R., and Deltour, R. (1981) Eur. J. Cell Biol., 23, 303-311.
31. Puvion-Dutilleul, F., and Bachellerie, J.-P. (1979) J. Ultras. Res., 66,
190-199.
32. Thiry, M. (1990) Arch. Intern. Physiol. Biochem., 98, B153.
33. Gross, D., and Garrard, W. (1988) Ann. Rev. Biochem., 57, 159-197.
34. Grummt, I. (1981) Nucleic Acids Res., 9, 6093-6102.
35. Tower, J., Henderson, S., Dougherty, K., Wejksnora, P., and SoUner-Webb,
B. (1989) Mol. Cell Bid., 9, 1513-1525.