Anat Embryol (1995) 191:359-367 �9 Springer-Verlag 1995

M. Lafarga �9 M. A. Andrds �9 C. Fernfindez-Viadero J. Villegas �9 M. T. Berciano

Number of nucleoli and coiled bodies and distribution of fibrillar centres in differentiating Purkinje neurons of chick and rat cerebellum

Accepted: 3 October 1994

Abstract We used differentiating chick and rat Purkinje cells to investigate in homologous neurons the influence of the number of nucleolar organizer regions (two in the chick and six in the rat) on the behaviour of the nucleo- lus and coiled bodies. We employed specific silver-stain- ing methods on smear preparations and on semithin and ultrathin sections. In chick Purkinje cells the number of nucleolar silver-staining granules increased from 15.7+3 (mean_+SD) at embryonic day 13 to 23.8_+3 at post-hatch- ing day 7. These nucleolar granules were unevenly dis- tributed between the two nucleoli of binucleolated cells. Electron-microscopic cytochemistry showed that nucleo- lar granules are equivalent to the fibrillar centres with their associated shell of dense fibrillar component. A re- duction in the number of nucleoli was found during the differentiation of both chick and rat Purkinje cells, al- though in mature cells the average number of nucleoli per cell was higher in the chick (1.60) than in the rat (1.07). The number of coiled bodies decreased from 1.33 in newborn rats to 0.47 at postnatal day 90 in the rat. Coiled bodies were not observed in homologous chick Purkinje cells. The dynamic behaviour ,of nucleoli and coiled bodies during neuronal differentiation and the re- lationship of these two nuclear organelles with the num- ber of nucleolar organizer regions is discussed.

Key words Nucleoli �9 Coiled bodies - Nucleolar organizer regions �9 Fibrillar centres

Introduction

The nucleolus plays a basic role in the synthesis of ribo- somal RNA (rRNA) and is therefore essential for normal cell differentiation, which involves notable changes in

M. Lafarga (4) �9 M.A. Andrds. C. Fernfindez-Viadero J. Villegas - M.T. Berciano Department of Anatomy and Cell Biology, University of Cantabria, Avda. Cardenal Herrera Oria s/n, E-39011 Santander, Spain

the cytoplasmic demand for nucleolar pre-rRNAs for ri- bosome biogenesis (Hadjiolov 1985). In higher eukary- otes, the genes that encode rRNAs are repeated tandemly in the nucleolar organizer regions (NORs) of metaphase chromosomes. While most mammals contain several NORs per haploid genome, the chicken haploid genome has a simpler organization, with only one NOR located on a microchromosome (Muscarella et al. 1985). Diploid chicken cells have two NORs, one or two nucleoli and about 290 rRNA gene copies (Muscarella et al. 1985, 1987). Since each NOR can direct the formation of a sin- gle nucleolus, the binucleolated diploid cells of the chicken provide a good model for investigating whether a differential behaviour exists between the two NORs, resulting in an asymmetric nucleolar distribution of fi- brillar centres.

The aim of this work is to analyse the quantitative changes in the number of nucleoli per neuron and in the number of silver-stained granules (SSGs) per nucleolus in mono- and binucleolated cells during critical phases of differentiation of chick Purkinje cells (PCs). It is well established that SSGs are equivalent to the fibrillar cen- tres plus the associated shell of dense fibrillar component of the nucleolus (Daskal et al. 1980; Hernandez-Verdun 1983; Goessens 1984), which contain silver-staining pro- teins of the NORs (Ag-NOR-proteins) and ribosomal genes (for review see Ploton et al. 1984; Robert-Fortel et al. 1993; Schwarzacher and Wachtler 1993). A second goal is to compare the nucleolar behaviour of differenti- ating PCs in the chick with that of PCs in the maturing rat, i.e. homologous mammalian neurons. Rat PCs have the same morphological and functional features as chick- en PCs and undergo similar stages of differentiation, but they have a different genomic organization of ribosomal genes, which are clustered in six NORs (Kano et al. 1976). Finally, in both kinds of homologous maturing PCs we will analyse the pattern of expression of a type of nuclear body commonly found in neuronal nuclei, the "accessory body" of Cajal (1903), or its ultrastructural counterpart, the "coiled body" (Monneron and Bernhard 1969). Coiled bodies (CBs) appear frequently associated

360

with the nuc leo lus in neuronal nucle i (Hardin et al. 1969; Sei te et al. 1982; Lafa rga et al. 1983). The metabo l i c role of the CB and its poss ib l e funct ional re la t ionship to the nucleolus r ema in to be es tabl ished. However , recent im- m u n o c y t o c h e m i c a l and in situ hybr id iza t ion studies have shown that CBs conta in the nuc leo la r pro te in f ibr i l lar in, a CB-spec i f i c protein, p80-coi l in , and smal l nuc lear ri- bonuc leopro te ins ( snRNPs) involved in m R N A spl ic ing (for review see Andrade et al. 1991; R a s k a et al. 1991; L a m o n d and C a r m o - F o n s e c a 1993), but that they lack r R N A or the U3 smal l nuc leo la r R N A ( C a r m o - F o n s e c a et al. 1993). On the bas is of these and other exper imenta l data, L a m o n d and C a r m o - F o n s e c a (1993) have p roposed that CBs serve as nuc lear domains involved in the assem- b ly or d i s a s sembly o f mu l t i - snRNP complexes . Our re- suits indica te that CBs are dynamic structures in differ- ent ia t ing neurons and also sugges t that their fo rmat ion m a y be par t i a l ly in f luenced by the number and nuclear pos i t ion o f NORs.

Materials and methods

Animals

The cerebella from 35 chicks (Rhode Island Red) and 47 rats (Sprague-Dawley) were used. Chick embryos from incubation days El3, El5 and El7 and young chicks from hatching (H) and 7 days post-hatching (7PH) were employed. Rat cerebella were sam- pled at postnatal days P0, P7, P14, P21, P30, P60 and P90. After ether anaesthesia, the animals were killed and processed for the different procedures. The animals were housed, supervised and handled according to the approved national guidelines for animal c a r e .

Quantitative analysis of nucleoli, silver-stained granules and coiled bodies

The number of nucleoli, SSGs and CBs per cell nucleus was deter- mined on smear preparations from block-impregnated cerebella according to the procedure of Lafarga et al. (1991). The animals were fixed by perfusion with 4% paraformaldehyde and 0.075% glutaraldehyde in 0.12 M phosphate buffer. After decapitation, the cerebella were immediately removed and the vermes were dissect- ed and processed with the silver-staining method. Briefly, cerebel- lar fragments were treated with 3:1 methanol/acetic acid for 30 min, impregnated for 3-5 h at 37 ~ C in 2.5% silver nitrate solution containing 1% gelatin, and reduced in an aqueous solution of 3.7% formaldehyde containing 1% pyrogallic acid for 12 h. After this, small fragments of cerebellar folia were dissected from 300- gm-thick tissue-chopper slices. These tissue fragments were treat- ed with 1% of trypsin in PBS at 37 ~ C for 1 h, and then dissociat- ed by trituration with a Pasteur pipet and centrifuged (10 min, 12000 rpm). The supernatant was removed, and the pellet of disso- ciated neurons was resuspended in a drop of PBS. Smears of these dissociated neurons were made on cover glasses, mounted on mi- croscope slides and observed with a microscope with a xl00 oil objective. For the quantitative analysis, five animals from each age group were used, and at least 50 neurons per animal were sam- pled. Smear preparations were examined without previous knowl- edge of the experimental group. As this silver method provides good preservation of neuronal perikarya, only large cells with a pear-shaped soma, identified as PCs, were selected. The numbers of nucleoli and coiled bodies per cell were counted by direct ex- amination of the different focal planes throughout PC nuclei. In the chick samples, the number of SSGs per nucleolus and nucleus

was also determined. For the comparative analysis of the number of SSGs between the two nucleoli of chick binucleolated PCs, nu- cleolus 1 was distinguished from nucleolus 2 by having the greater number of SSGs. The statistical tests used were analysis of vari- ance and Student's t-test. Significance was established at P<0.05.

Ultrastructural analysis

The animals were perfused with 3% glutaraldehyde in 0.12 M phosphate buffer, then the cerebella were removed and left over- night in the same fixative. Small fiagments were dissected from sagittal, 200-pro-thick slices from the cerebellar vermis, The blocks were then postfixed in 2% buffered osmium tetroxide, de- hydrated, and embedded in Araldite. Ultrathin sections stained with uranyl acetate and lead citrate were employed for the ultra- structural analysis.

Silver staining of nucleolar organizer regions

A one-step silver-staining method that is specific for the Ag-NOR proteins was used on semithin and ultrathin sections following the procedure of Moreno et al. (1985). The animals were fixed by per- fusion with 1% paraformaldehyde and 1% glutaraldehyde in 0.12 M phosphate buffer, at 4 ~ C. Blocks of cerebellar cortex were postfixed in 3:1 ethanol/acetic acid for 10 min at 4 ~ C. Then the blocks were embedded in Lowicryl K4M and polymerized with ul- traviolet irradiation at 4 ~ C. Ultrathin sections were mounted on gold grids and then treated for 5 rain at room temperature with a mixture containing one volume of 2% gelatin in 1% formic acid and two volumes of 50% silver nitrate aqueous solution. Grids were rinsed in distilled water. Ultrathin sections were slightly con- trasted with uranyl acetate. Semithin sections, 1 gm thick, were silver-impregnated with the same procedure.

Results

N u m b e r of nuc leo lar s i lver-s ta ined granules in matur ing Purk in je cel ls of the chick

Dis soc ia t ed Purkin je cells, f rom cerebe l la r f ragments p rev ious ly impregna ted by means of the p rocedure de- sc r ibed above, exh ib i ted g o o d preserva t ion of neuronal bod ies and exce l len t s taining of nuc leo lar SSGs (Fig. 1). Since the whole nucleus could be examined in these prepara t ions , an accurate es t imat ion o f the number of SSGs per nucleolus and per nucleus was poss ib le . The par t ia l ly unrave l led nuc leo la r organiza t ion character is t ic o f ch ick PCs, with large d i spe rsed nuc leo lar SSGs (Fig. 1), fac i l i ta ted the count of nuc leo lar granules. By con- trast, the SSGs in the rat PC nucleolus are so smal l that it is very diff icult to make an accurate count by direct l ight m ic ro sc opy examinat ion . The d i f ference in dis t r ibut ion pat terns o f nuc leo lar SSGs be tween the chick and rat was conf i rmed on 1-pm semithin sect ions of Lowic ry l - e m b e d d e d samples s ta ined with the one-s tep A g - N O R p rocedure (Figs. 3, 4).

In d i f ferent ia t ing ch ick PCs, the quant i ta t ive analys is revea led that the average number of SSGs per nucleus increased dur ing neuronal d i f ferent ia t ion f rom 15.78+3.25 (mean+_SD) at E l 3 to (23.85+3.78) at PH7, when the mature state is achieved. For each deve lopmen- tal stage, the mean number of SSGs per nucleus was sig- n i f icant ly h igher in b inuc leo la t ed neurons than in mono-

361

Fig. IA-C Smear preparations of silver-stained chick PCs at em- bryonic day El7. x1100. Nucleolar SSGs appear specificially ira- pregnant as black dots. A Mononucleolated neuron. B, C The same binucleolated neuron with two different focal planes, one for each of the two nucleoli

Fig. 2 Smear preparations of silver-stained rat PCs at postnatal days P7 (A) and PI4 (B, C). A Note two nucleoli of different size and two smaller intensely impregnated CBs free in the nucleo- plasm, x1600 B A mononucleolated neuron with a nucleolus-at- tached CB. x1250. C A mononucleolated neuron with a free CB. x1250

nucleolated ones (Fig. 5). In addition, SSGs were un- evenly distributed between the two nucleoli of binucleo- lated neurons, with one nucleolus containing a signifi- cantly greater number (Table 1).

Number of nucleoli and coiled bodies in differentiating Purkinje cells from chick and rat cerebella

The quantitative analysis of dissociated neurons showed a progressive reduction in the mean number of nucleoli per cell during the first stages of PC differentiation in

both chick and rat cerebella. The mature pattern in the number of nucleoli was attained at embryonic day E 17 in the chick and at postnatal day P14 in the rat (Table 2). Furthermore, in the chick cerebellum the average num- ber of nucleoli per mature PC was 1.60, and more than half of the neuronal population (about 56%) was binucle- olated, while the average in the rat was lower (1.07) with about 90% of the PCs being mononucleolated (Fig. 6).

Coiled bodies were prominent nuclear inclusions in differentiating and mature PCs of the rat in silver-stained preparations (Fig. 2), but they were not observed in ho- mologous chick neurons. These nuclear bodies appeared as intensely impregnated black spots, located free in the nucleoplasm (Fig. 2A, C) or attached to the nucleolar surface (Fig. 2B). They were clearly distinguishable from nucleoli by their small size and homogeneous tex- ture. Counts of Purkinje cell CBs revealed a progressive decrease in the average number per nucleus during cere- bellar neurogenesis from 1.3 at P0 to 0.5 at P30, when it became constant (Table 2). The highest frequency of neurons with 2 (36%) or 3 (3%) CBs was detected at P0, whereas CBs were absent in about 50% of mature PCs at P90.

362

Figs. 3, 4 Semithin sections from chick (3) and rat (4) PCs of Lowicryl-embedded sam- ples, stained with the Ag-NOR procedure, x1400. Fig. 3 Note the beaded appearance of the nucleolar SSGs; chick 7 days post-hatching. Fig. 4 The nu- cleolar SSGs appear closely clustered; rat 30 days postna- tal

25 ~lg

100 1

=~ 22 7 5

O

18 25

15

12 i I i , t ~ , i

E13 E15 E17 H PH7

100 - Fig. 5 Distribution of the silver stained granules (SSGs) per nu- cleus in mono- (0) and binucleolated (R) Purkinje cells during the embryonic (E13-EIT), hatching (H) and 7 days posthatching (PHT) developmental stages of chick cerebellar cortex 75 -

Table 1 Comparative analysis of the number of nucleolar silver- 50 - stained granules (SSGs) in the two nucleoli of binucleolated Pur- kinje cells during the chick cerebellar developtment (Mean• In all cases, nucleolus 1 represents the nucleolus containing the greater number of SSGs. The nucleoli of 250 cells were estimated 25 - for each animal group. * P<0.001when compared to El3, El5 and El7; ** P<0.001 when compared to other groups

SSGs/Nucleolus 1 SSGs/Nucleolus 2 Student's t-test

El3 8.74• 7.53• P<0.001 E15 9.81• 8.84• P<0.001 El7 11.38• 8.02• P<0.001 H 12.45• 10.47• P<0.001 PH7 12.80• 11.04• P<0.001

PO P7 P14 P21 P30 P60

Fig. 6 Distribution of Purkinje rat cells with one (I) , two (~) and three ([]) nucleoli during chick and rat neurogenesis

363

Fig. 7 Typical nucleolus of a chick PC at hatching. Note promi- nent fibrillar centres (asterisks) which appear as rounded fibrillar areas of low electron density, x24000

Fig. 8 Detail of a chick PC nucleolus at hatching after cytochemi- cal staining with the Ag-NOR procedure. The silver grains are specifically concentrated on the fibrillar centres and the surround- ing shell of dense fibrillar component, x24000

Fig. 9 Nucleolus of a rat PC at postnatal day P21, Note numerous small fibrillar centres (asterisks), a nucleolus-attached CB (arrow) and a clump of perinncleolar chromatin (arrowhead). x23000

Fig. 10 Detail of a typical CB composed of coiled dense fibrillar strands and an intervening low-density matrix. Rat PC at P7. x40000

Fig. 11 Electron micrograph of a CB stained with the Ag-NOR procedure. Note the silver grains selectively located on the coiled strands of dense fibrillar material. Rat PC at P14. x47000

364

Table 2 Number of nucleoli and coiled bodies (CBs) in differenti- ating Purkinje cells from chick and rat cerebellum (Mean_+SE). The nuclei of 300 Purkinje neurons were analysed for each animal group

Chick No. of nucleoli No. of CBs per nucleus per nucleus

El3 1.88_+0.04 0 n.s.

El5 1.81+0.03 0 P<0.00I

El7 1.61+_0.03 0 n.s.

H 1.53_+0.03 0 n.s.

PH7 1.60+.0.04 0

Rat No. of nucleoli No. of CBs per nucleus per nucleus

PO 1.91+0.01 1.33+_0.05 P<0.01 P<O.01

P7 1.48_+0.05 1.19_+0.07 P<0.01 P<O.01

P14 1.16+0.04 1.00_+0.05 n.s. P<0.0I

P21 1.18+0.04 0.89+0.07 n.s. P<0.01

P30 1.13+.0.03 0.51+.0.05 n.s. n.s.

P60 1.10_+0.02 0.50+0.05 n.s. n.s.

P90 1.07_+0.01 0.47_+0.02

Ultrastructural analysis

Electron microscopy of differentiating chick PCs re- vealed irregular nucleolar profiles with a compact or re- ticulated configuration. The organization of the nucleolar components was characterized by the presence of promi- nent large fibrillar centres - round fibrillar areas of low electron density - surrounded by a thin shell of dense fi- brillar component (Fig. 7). The granular component and some nucleolar interstices were distributed in the areas between fibrillar centres. Nuclear inclusions with the morphology of CBs were not observed. By using the Ag- NOR procedure on ultrathin sections of Lowicryl-em- bedded samples, the silver reaction could be selectively detected on the fibrillar centres and the associated layer of dense fibrillar component. As can be seen in Fig. 8, the silver staining was more intense on the dense fibrillar component.

Maturing rat PCs commonly exhibited reticulated nu- cleoli with numerous small fibrillar centres (Fig. 9). Typ- ical CBs were the only category of nuclear inclusion found in this neuronal population. They appeared free in the nucleoplasm or associated with the nucleolus (Figs. 9, 10) and displayed the usual structural organization of coiled strands of dense fibrillar material embedded in a low-density fibrillar matrix (Fig. 10). The application of the silver procedure for staining argyrophilic proteins of the NOR revealed silver grains exclusively located on the dense coiled strands of CBs (Fig. 11).

Discussion

Our results show that PC differentiation in the chick in- volves a significant increase in the total number of SSGs per nucleus and a reduction in the mean number of nu- cleoli per cell, with no CBs being detected. In homolo- gous rat neurons, a reduction in the number of both nu- cleoli and CBs is associated with PC maturation.

The number of SSGs increases in differentiating Purkinje neurons

Chick PCs from embryonic day El3 to post-hatching day 7 provide a good model of postmitotic (GO) maturing neurons for analysing the response of the nucleolus to the dramatic changes in the structure and function that accompany terminal neuronal differentiation. During this developmental period, PCs undergo prominent perikaryal and dendritic growth and active synaptogenesis (Mug- naini 1969; Foelix and Oppenheim 1974), achieving a fully mature dendritic tree in the course of the first post- hatching week (Palacios-Pru et al. 1976; Bouvet et al. 1987).

Our results show a progressive increase in the number of nucleolar SSGs as differentiation proceeds. The SSGs of PCs are similar to those previously reported in cancer cells and activated lymphocytes, where the number of SSGs has been positively related to the transcription rate of rRNA genes (Busch et al. 1979; Wachtler et al. 1980; Likovsky and Smetana 1981). This relationship is con- sistent with the well-known association of nucleolar ar- gyrophilic proteins of SSGs with transcribing rRNA genes (Angelier et al. 1982; Hubbell 1985), and also with the immunocytochemical demonstration that nucle- olar SSGs contain RNA polymerase I (Scheer etal . 1984; Ochs and Smetana 1989), the enzyme that directs nucleolar transcription. By electron microscopic analy- sis, our results with the Ag-NOR procedure, and those of other authors (for review see Daskal et al. 1980; Hernan- dez-Verdun 1983; Goessens 1984), indicate that fibrillar centres with their associated shell of dense fibrillar com- ponent are the ultrastructural counterpart of SSGs. More- over, a quantitative relationship between the fibrillar cen- tres and changes in cell activity has been reported by Jor- dan and McGovern (1981) in fibroblasts using serial ultrathin sections. In this context, we propose that the progressive increase in the number of SSGs in maturing PCs reflects an elevation in the rate of pre-rRNA synthe- sis associated with the increasing requirements for ribo- some biogenesis and protein synthesis in growing PCs (Patel et al. 1984).

In some cell types that exhibit unravelled nucleoli, as occurs naturally in activated lymphocytes (Ochs and Smetana 1989), or in cells treated with the adenosine an- alogue 5,6-dichloro-l-b-D-ribofuranosylbenzimidazole (Scheer et al. 1984; Haaf et al. 1991), it has been ob- served that the number of SSGs corresponds approxi- mately to the number of rRNA gene copies, supporting

the view that a SSG, or its ultrastructural equivalent, rep- resents a single transcription unit. In the case of chick diploid cells, the rDNA copy number is about 290 (Mus- carella et al. 1985, 1987), whereas the largest average number of SSGs counted in fully mature PCs is about 23. If each SSG corresponds to a single rDNA copy, this would imply that in chick PCs about 92% of rRNA gene copies remain inactive. This interpretation seems to be impossible in PCs, which are characterized by high pro- tein synthesis activity (Patel et al. 1984); therefore, we suggest that in this neuronal type an SSG may result from the transcriptional activity of several copies of rRNA genes. Another point for discussion is the signifi- cant difference in the number of SSGs between the two nucleoli of binucleolated PCs which was detected in all developmental stages studied. Since diploid chick PCs have two NORs, each nucleolus of these binucleolated cells presumably results from the activity of a single NOR. If this is true, our results suggest a differential be- haviour of the two NORs, resulting in one nucleolus be- ing transcriptionally more active than the other.

The number of nucleoli declines differently in maturing chick and rat Purkinje cells

Since the structural and functional organization and the stages of differentiation of PCs are highly conserved in vertebrate evolution (Mugnaini 1969; Palay and Chan- Palay 1974), the number of nucleoli can be compared in homologous neurons from avian and mammalian cere- bella, in order to analyse the possible relationship be- tween this parameter and the number of NORs. Our re- sults reveal a significant reduction in the mean number of nucleoli per nucleus during differentiation of both chick and rat PCs. Although other authors have also re- ported a high proportion of mononucleolated PCs in adult rodents (Shea and Leblond 1966; Brodsky et al. 1988), to our knowledge no previous work has reported a quantitative analysis of the number of nucleoli during the cytodifferentiation of PCs.

It is well established that the theoretical maximum number of nucleoli per cells is equivalent to the number of NORs: two and six in the diploid cells of the chick and rat, respectively. There are two general cellular pro- cesses that explain the lower mean number found here in PCs of both species: (1) transcriptionally active NORs tend to associate in only one or two nucleolar domains during cell differentiation and (2) there are inactive NORs that do not organize nucleoli (for review see Wachtler et al. 1986; Schwarzacher and Wachtler 1993). Both mechanisms might account for the progressive de- crease in the number of nucleoli reported here in matur- ing PCs, and previously in other neuronal (Buschmann and LaVelle 1981; Crespo et al. 1988) and non-neuronal (Shea and Leblond 1966; Wachtler and Schwarzacher 1984) cell types. Concerning the NOR association pat- tern, it is worth noting the difference between the aver- age number of nucleoli observed in mature PCs from the

365

chick (1.60) and the rat (1.07), which involves a propor- tion of mononucleolated cells of 45% versus 90%, re- spectively, indicating a different pattern of NOR conver- gence in these two homologous neurons. In this context, several authors have shown a non-random intranuclear lo- cation of active NORs in various cell types (Jordan et al. 1982; Manuelidis 1984; Brodsky et al. 1988), supporting the view that the spatial position of NORs is an important factor in determining the extent of nucleolar fusion. In addition to this factor, our results in homologous PCs, with similar morphological and functional phenotypes, suggest that the pattern of NOR association is also strongly influenced by the number of NORs present in the cell. Thus, most of the active NORs of chick PCs seem to remain disassociated during cell differentiation, resulting in a predominant proportion of binucleolated neurons in the mature state. By contrast, there must be a strong tendency for NOR aggregation to occur in PCs of the rat and other mammals that contain an elevated num- ber of NORs (six in the rat and mouse, and ten in hu- mans) and are predominantly mononucleolated.

On the other hand, regarding the possible existence of inactive NORs that do not organize nucleoli (for review see Schwarzacher and Wachtler 1993), it has been dem- onstrated by autoradiography following rDNA in situ hy- bridization that the characteristic pattern of hybridization signal in mature PCs of the rat consists of one large nu- cleolar cluster of silver grains (Brodsky et al. 1988). On the basis of this finding, these authors proposed that in rat PCs, which represent a neuronal model of high pro- tein synthesis activity (Patel et al. 1984), the bulk of rRNA genes are concentrated in the nucleolus.

Finally, in chick PCs we have observed that the ma- ture pattern in the number of nucleoli was attained at El7, whereas the mean number of SSGs per nucleolus continued to increase in later stages of PC differentia- tion. Previous studies in mature supraoptic neurons have shown great stability in the number of nucleoi per cell following changes in neuronal activity (Lafarga et al. 1991) or even after a cycloheximide-induced alteration of nucleolar rRNA synthesis and processing (Lafarga et al. 1994). This indicates that the number of nucleoli is not affected by the metabolic status of the neurons. By contrast, the number of SSGs seems to be a sensible in- dicator of the level of cellular activity, particularly of the transcriptional rate of the nucleolus for ribosome biogen- esis. In fact the dynamic behaviour of SSGs, or their ul- trastructural counterpart, the fibrillar centres, has been demonstrated in neurons and non-neuronal cells in re- sponse to changes in cell activity (for review see Jordan and McGovern 1981; Lafarga et al. 1991; Schwarzacher and Wachtler 1993).

Coiled bodies are dynamic structures in maturing rat Purkinje cells

In maturing rat PCs we have observed that the number of CBs decreases from a mean of 1.33 at P0 to 0.47 in the

366

mature state (P90). A similar reduction in the average number of CBs has recently been reported during termi- nal erythroid differentiation of murine erythroleukemia cells in vitro (Antoniou et al. 1993). This dynamic be- haviour of CBs supports the view that their formation is related to the dramatic changes in gene expression that accompany neuronal differentiation. Moreover, the stabi- lization of the genome in fully differentiated neurons is associated with an average number of CBs per cell that is significantly lower and constant. This interpretation is consistent with previous evidence revealing that the for- mation of CBs is "up regulated" when changes in gene expression are induced in stimulated or transformed cells (Lafarga et al. 1991; Brasch and Ochs 1991; Spector et al. 1992; Carmo-Fonseca et al. 1993; Andrade etal. 1993).

Interestingly, CBs were not observed in differentiat- ing and mature chick PCs with structural and functional characteristics similar to those of homologous neurons in rat cerebellum. This suggests that the formation of CBs not only depends upon the level of gene expression or the functional activity of the cell, but may also be influ- enced by other factors, such as the pattern of chromo- somal arrangement within the nucleus, particularly of NORs and adjacent chromatin domains. Several clues ar- gue in favour of this possibility: (1) CBs are frequently associated with the nucleolus in mammalian neurons having high metabolic activity such as rat PCs, which concentrate multiple NORs in one nucleolar domain (Brodsky et al. 1988); (2) conversely, CBs are absent in the granule cells (microneurons) of mammalian cerebella which exhibit a dissociated ribosomal genome (Brodsky et al. 1988) and ring-shaped micronucleoli (Lafarga et al. 1989); (3) CBs are not observed in chick PCs, in which only one or two microchromosomes containing NORs converge in a nucleolar domain, and (4) changes in the frequency of CBs and an abnormal intranucleolar loca- tion have been reported in certain breast cancer cell lines (Ochs et al. 1994) where mutational events and chromo- somal rearrangements must occur. Further studies with specific molecular markers for CB components are re- quired to establish the distribution pattern of these com- ponents within nuclear subcompartments in cells lacking this nuclear organelle.

Acknowledgements This work received financial support from the "Comisidn Mixta Caja Cantabria-Universidad de Cantabria" and from the DGICYT (PB93-0369).

References

Andrade LEC, Chan EKL, Raska I, Peebles CL, Roos G, Tan EM (1991) Human autoantibody to a novel protein of the nuclear coiled body: immunological characterization and cDNA clon- ing of p-80-coilin. J Exp Med 173:1407-1419

Andrade LEC, Tan EM, Chan EKL (1993) Immunocytochemical analysis of the coiled body in the cell cycle and during cell proliferation. Proc Natl Acad Sci USA 90:1947-1951

Angelier N, Hernandez-Verdun D, Bouteille M (1982) Visualiza- tion of Ag-NOR proteins on nucleolar transcriptional units in molecular spreads. Chromosoma 86:661-672

Antoniou M, Carmo-Fonseca M, Ferreira J, Lamond AI (1993) Nuclear organization of splicing snRNPs during differentiation of murine erythroleukemia cells in vitro. J Cell Biol 123:1055-1068

Bouvet J, Usson Y, Legrand J (1987) Morphometric analysis of the cerebellar Purkinje cell in the developing normal and hy- pothyroid chick, lnt J Dev Neurosci 5:345-355

Brasch K, Ochs RL (1992) Nuclear bodies (NBs): a newly "redis- covered" organelle. Exp Cell Res 202:211-223

Brodsky VY, Marshak TL, Karavanov AA, Zatsepina OV, Nosikov VV, Korochkin LI, Braga EA (1988) Cell differentiation as as- sayed by topography and number of ribosomal genes. Cell Differ 24:201-208

Busch H, Daskal Y, Gyorkey F, Smetana K (1979) Silver staining of nucleolar granules in tumor cells. Cancer Res 39:857-863

Buschmann MB, La Velle A (1981) Morphological changes of the pyramidal cell nucleolus and nucleus in hamster frontal cortex during development and aging. Mech Ageing Dev 15: 385-397

Carmo-Fonseca M, Ferreira J, Lamond AI (1993) Assembly of snRNP-containing coiled bodies is regulated in interphase and mitosis - evidence that the coiled body is a kinetic nuclear structure. J Cell Biol 102:841-852

Crespo D, Viadero CF, Villegas J, Lafarga M (1988) Nucleoli numbers and neuronal growth in supraoptic nucleus neurons during postnatal development in the rat. Dev Brain Res 44:151-155

Daskal Y, Smetana K, Busch H (1980) Evidence from studies on segregated nucleoli that nucleolar silver-staining proteins C23 and B23 are in the fibrillar component. Exp Cell Res 127:285-291

Foelix RF, Oppenheim R (1974) The development of synapses in the cerebellar cortex of the chick embryo. J Neurocytol 3:277-294

Goessens G (1984) Nucleolar structure. Int Rev Cytol 87:107-157 Haaf T, Hayman DL, Schmid M (1991) Quantitative determination

of rDNA transcription units in vertebrate cells. Exp Cell Res 193:78-86

Hadjiolov AA (1985) The nucleolus and ribosome biogenesis. Springer, Berlin Heidelberg New York

Hardin JW, Spicer SS, Green WB (1969) The paranucleolar struc- ture, accessory body of Cajal, sex chromatin, and related struc- tures in nuclei of rat trigeminal neurons: a cytochemical and ultrastructural study. Anat Rec 164:403-432

Hernandez-Verdun D (1983) The nucleolar organizer regions. Biol Cell 49:191-202

Hubbell HR (1985) Silver staining as an indicator of active ribo- somal genes. Stain Technol 60:285-294

Jordan EG, McGovern JH (1981) The quantitative relationship of the fibrillar centres and other nucleolar components to changes in growth conditions, serum deprivation and low doses of acti- nomycin D in cultured diploid human fibroblasts (strain MRC- 5). J Cell Sci 52:373-389

Jordan EG, Martini G, Bennett MD, Flavell RB (1982) Nucleolar fusion in wheat. J Cell Sci 56:485-495

Kano Y, Maeda S, Sugiyama T (1976) The location of ribosomal cistrons (r-DNA) in chromosomes of the rat. Chromosoma 55:37-42

Lafarga M, Hervas JR Santa-Cruz MC, Villegas J, Crespo D (1983) The "accessory body" of Cajal in the neuronal nucleus. A light and electron microscopic approach. Anat Embryol 166:19-30

Lafarga M, Berciano MT, Hervas JP, Villegas J (1989) Nucleolar organization in granule cell neurons of the rat cerebellum. J Neurocytol 18:19-26

Lafarga M, Andres MA, Berciano MT, Maquiera E (1991) Organi- zation of nucleoli and nuclear bodies in osmotically stimulated supraoptic neurons of the rat. J Comp Neurol 308:329-339

Lafarga M, Berciano MT, Andres MA, Testillano PS (1994) Ef- fects of cycloheximide on the structural organization of the nu- cleolus and the coiled body in normal and stimulated supraop- tic neurons of the rat. J Neurocytol 23:500-513

Lamond AI, Carmo-Fonseca M (1993) The coiled body. Trends Cell Biol 3:198-204

Likovsky Z, Smetana K (1981) Further studies on the cytochemis- try of the standardized silver staining of interphase nucleoli in smear preparations of yoshida ascitic sarcoma cells in rats. Histochemistry 72:30 l-313

Manuelidis L (1984) Active nucleolns organizers are precisely po- sitioned in adult central nervous system cells but not in neuro- ectodermal tumor cells. J Neuropathol Exp Neurol 43:225-241

Monneron A, Bernhard W (1969) Fine structural organization of the interphase nucleus in some mammalian cells. J Ultrastruct Res 27:266-288

Moreno FJ, Hernandez-Verdun D, Masson C, Bouteille M (1985) Silver staining of the nucleolar organizer regions (NORs) on Lowicryl and cryo-ultrathin sections. J Histochem Cytochem 33:389-399

Mugnaini E (1969) Ultrastructural studies on the cerebellar histo- genesis. II. Maturation of nerve cell populations and establish- ment of synaptic connections in the cerebellar cortex of the chick. In: Llimis R (ed) The neurobiology of cerebellar evolu- tion and development. American Medical Association, Chica- go, pp 749-782

Muscarella DE, Vogt VM, Bloom SE (1985) The ribosomal RNA gene cluster in aneuploid chickens: evidence for increased gene dosage and regulation of gene expression. J Cell Biol 101:1749-1756

Muscarella DE, Vogt VM, Bloom SE (1987) Charakterization of ribosomal RNA synthesis in a gene dosage mutant: The rela- tionship of topoisomerase I and chromatin structure to tran- scriptional activity. J Cell Biol 105:1501-1513

Ochs RL, Smetana K (1989) Fibrillar center distribution in nucleo- li of PHA-stimulated human lymphocytes. Exp Cell Res 184:552-557

Ochs RL, Stein TW Jr, Tan EM (1994) Coiled bodies in the nucle- olus of breast cancer cells. J Cell Sci 107:385-399

Palacios-Prti E, Palacios L, Mendoza RV (1976) In vitro vs in situ development of Purkinje cells. J Neurosci Res 2:357-361

Palay S, Chan-Palay V (1975) Cerebellar cortex: cytology and or- ganization. Springei, New York Berlin Heidelberg

Patel N, Cohen J, Balazs R (1984) Protein synthesis in cells isolat- ed from the developing rat cerebellum. Int J Dev Neurosci 2:287-299

367

Ploton D, Menager M, Adnet JJ (1984) Simultaneous high resolu- tion localization of Ag-NOR proteins and nucleoproteins in in- terphasic and mitotic nuclei. Histochem J 16:897-906

Ram6n y Cajal S (1903) Un sencillo m6todo de coloraci6n select- iva del ret/culo protopl~ismico y sus efectos en los diversos 6rganos nerviosos de vertebrados e invertebrados. Trab Lab Invest Biol 2:129-221

Raska I, Andrade LEC, Ochs RL, Chan EKL, Chang CM, Roos G, Tan EM (1991) Immunological and ultrastructural studies of the nuclear coiled body with autoimmune antibodies. Exp Cell Res 195:27-37

Robert-Fortel 1, Junera HR, Gerand G, Hernandez-Verdun D (1993) Three-dimensional organization of the ribosomal genes and Ag-NOR proteins during interphase and mitosis in PtK1 cells studied by confocal microscopy. Chromosoma 102:146-157

Scheer U, Htigle B, Hazan R, Rose KM (1984) Drug-induced dis- persal of transcribed rRNA genes and transcriptional products: immunolocalization and silver staining of different nucleolar components in rat cells treated with 5,6-dichloro-fi-o-ribofura- nosylbenzimidazole. J Cell Biol 99:672-679

Schwarzacher HG, Wachtler F (1993) The nucleolus. Anat Em- bryol 188:515-536

Seite R, Pebusque MJ, Vio-Cigna M (1982) Argyrophilic proteins on coiled bodies in sympathetic neurons identified by Ag- NOR procedure. Biol Cell 46:97-100

Shea JR Jr, Leblond CP (1966) Number of nucleoli in various cell types of the mouse. J Morphol 119:425-434

Spector DL, Lark G, Huang S (1992) Differences in snRNP local- ization between transformed and nontransformed cells. Mol Biol Cell 3:555-569

Wachtler F, Schwarzacher HG (1984) On the fusion nucleoli in in- terphase. Eur J Cell Biol 34:190-192

Wachtler IF, Ellinger A, Schwarzacher HG (1980) Nucleolar changes in human phytohaemagglutinin-stimulated lympho- cytes. Cell Tissue Res 213:351-360

Wachtler F, Hopman AHN, Wiegant J, Schwarzacher HG (1986) On the position of nucleolar organizer regions (NORs) in in- terphase nuclei. Exp Cell Res 167:227-240