Localization of centromeric satellite and telomeric DNA sequences in

dorsal root ganglion neurons, in vitro

FILIO BILLIA* and UMBERTO DE BONIf

Department of Physiology, Faculty of Medicine, University of Toronto, Toronto, Ontario M5S 1A8, Canada

* Present address: Department of Medical Biophysics, University of Toronto, Ontario Cancer Institute, Toronto, Ontario M4X 1K9,Canadat Author for correspondence

Summary

Chromatin domains of interphase nuclei are organ-ized in a tissue-specific, non-random manner. In thepresent work, the spatial arrangement of satellite(sDNA) and telomeric (tDNA) DNA was examined innuclei of murine Dorsal Root Ganglion (DRG) cells,maintained in vitro. In situ hybridization in conjunc-tion with three-dimensional reconstruction wasemployed. A mean number of 8.02 ±0.40 sDNAsignals/nucleus was detected, of which 41.65±0.59%were associated with the nucleolus. The remainingfraction of signals was localized between the nu-cleolus and the nuclear membrane. sDNA signalswere reproducibly localized at a mean distance of3.15±0.06/mi from the nuclear center and measured1-2 fan in diameter. Given a centromere complementof 40 per murine nucleus, the relatively low number

of signals detected and their large signal volumeswere interpreted to reflect clustering of centromeres,a phenomenon common in mammalian cells. Anaverage of 37.00±1.52 tDNA signals was detected pernucleus. Of these, and in contrast to sDNA signals,only 18.45±0.41%> of these signals were associatedwith the nucleolus while the remainder was distrib-uted between the nucleolus and the nuclear mem-brane. Both centromeric and telomeric signals oftenoccurred in pairs and were distributed throughoutthe nucleoplasm. No evidence for a classical Rablconfiguration was found.

Key words: cell nucleus, neurons, centromere, telomeres,nuclear organization.

Introduction

The spatial organization of chromatin in interphase nucleiis currently of much interest. Early work by Rabl (1885)indicated that interphase chromatin may retain a telo-phase-like arrangement of chromosomes with centromeresand telomeres associated with opposite poles of thenucleus. Agard and Sedat (1983), Hochstrasser et al. (1986)and Hochstrasser and Sedat (1987a,b) found evidence forthis model in the specialized, polytene nuclei of Drosophilasalivary gland cells. In contrast, a non-Ttabl' chromosometopography is frequently reported for mammalian celltypes (Moroi et al. 1981; Brinkley et al. 1986; Manuelidis,19846, 1985a; Haaf and Schmid, 1989; Haaf et al. 1990).

Individual interphase chromosomes have been shown tooccupy distinct territories within mammalian interphasenuclei (Murray and Davies, 1979; Cremer et al. 1982;Schardin et al. 1985) as well as within diploid and polyteneDrosophila melanogaster cells (Hochstrasser and Sedat,1987a). Individual human chromosomes occupy relativelycompact interphase domains with homologous chromo-some domains not necessarily adjacent (Manuelidis andBorden, 1988; Hilliker and Appels, 1989). In contrast,cultured PtKl and Indian muntjac cells exhibit homolo-gous chromosomes in close proximity (Hadlaczky et al.1986). Moreover, chromatin domains of interphase neur-ons of the mouse and human central nervous systems havebeen shown to be organized in a tissue-specific, non-

Journal of Cell Science 100, 219-226 (1991)Printed in Great Britain © The Company of Biologists Limited 1991

random manner (Manuelidis, 1984a,b, 1985a,6; Manueli-dis and Borden, 1988; Arnoldus et al. 1989). A unifyingtheory for chromosomal arrangement within interphasenuclei has not yet been proposed. This is related todifficulties in the unequivocal identification of individualchromosomes, which has been limited to those in polyteneand diploid Drosophila nuclei (Agard and Sedat, 1983;Hochstrasser et al. 1986; Hochstrasser and Sedat, 1987a,6;Hiraoka et al. 1990). It is possible, however, to identifyspecific, well-defined chromosome regions such as centro-meres and telomeres. The three-dimensional topology ofsatellite DNA (sDNA) sequences, localized to centromeres(Jones, 1970; Pardue and Gall, 1970; Joseph et al. 1989;Matsumoto etal. 1989), has been extensively studied. Thiswas achieved either by indirect immunofluorescence,which detects kinetochore proteins associated with sDNAsequences (Moroi et al. 1981; Earnshaw et al. 1984;Hadlaczky et al. 1986; Chaly and Brown, 1988), or,alternatively, by in situ hybridization (Jones, 1970;Pardue and Gall, 1970; Rae and Franke, 1972; Manuelidis,1982, 19846, 1985a,6; Joseph et al. 1989). While aconsiderable body of evidence shows that telomeres areoften attached to the nuclear membrane in polytene nucleiand in some plant nuclei (Mathog et al. 1984; Hochstrasseret al. 1986; Katsumata and Lo, 1988; Rawlins and Shaw,1990a,6), little is known regarding the spatial arrange-ment of telomeric DNA (tDNA) in mammalian interphasenuclei.

219

DNA localized to centromeric and telomeric regions isenriched in highly repetitive DNA sequences (Horz andAltenburger, 1981; Manuelidis, 1981a,6; Moyzis et al.1988). Therefore, these sequences can be effectively probedto determine their spatial, intranuclear positions (Hsu etal. 1971; Rae and Franke, 1982; Manuelidis, 1982, 19846,1985a,6; Manuelidis and Borden, 1988; Moyzis et al. 1988;Rawlins and Shaw, 1990a,6).

In a test of the hypothesis that centromeric andtelomeric sequences are reproducibly and nonrandomlycompartmentalized in Dorsal Root Ganglion (DRG) neur-ons in vitro, the arrangement of these domains wasexamined by in situ hybridization and computer-assistedthree-dimensional reconstruction. DRG neurons arehighly differentiated and, as they are permanentlyarrested in interphase, the spatial topology of chromo-somes is not subject to events related to the cell cycle.

Materials and methods

Tissue preparationDRG neurons were cultured as previously described (Fung and DeBoni, 1988). Briefly, DRGs were aseptically removed fromneonatal mice, collected in a drop of cold Hank's Balanced SaltSolution (HBSS) and then transferred to 0.25% trypsin (GIBCO,lOmin, 37 °C). Upon inactivation of the latter with fetal bovineserum, DRG neurons were dissociated by trituration, centrifuged(500 g, 5 mm) and then resuspended in a few drops of culturemedium containing 89 % (v/v) Minimum Essential Medium withHank's Salts (GIBCO), 10% (v/v) fetal calf serum (GIBCO),lOOngml"1 7S nerve growth factor (Collaborative Research) and1 % (w/v) of a D-glucose concentrate to raise the final glucoseconcentration to 600 mg/100 ml. Cells were seeded onto22 mm x 22 mm Corning no. 1 glass coverslips contained withinthe center of plastic Petri dishes with tight-fitting lids (Falcon no.1006). Covershps had been previously coated with approximately0.3ml of calf skm collagen (Calbiochem; O.Smgml"1 m 0.1%(v/v) glacial acetic acid), which was permitted to dry beforeseeding of cells. Neurons attached to the collagen substratumwithin 2—3 h after which 0.5 ml of culture medium was added.Cultures were maintained for 5—6 days with the culture mediumreplaced every 3 days.

For use in experiments, cultures were washed with HBSS(prewarmed to room temperature CRT)) and fixed (4% (w/v)paraformaldehyde in PBS, 30min). After washing in PBS(3x6min), the fixed cells were exposed to a series of pretreat-ments. Cultures were incubated in 0.1 M HC1 (5-20 min, RT; timedepended on cell type, see Results). After washing in PBS(3X5 min), the cells were permeabilized with 1 % Triton X in PBS(lh, RT) and incubated in O^mgml"1 RNAase A (BoehringerMannheim, 37°C, lh), previously boiled (10min) to eliminateendogenous DNAses. Following washing in PBS (3x5 min),nuclear DNA was further digested for 10-15 min, with0.6—O.Smgml"1 Pronase (Boehringer Mannheim; in 0.05MTris-HCl, 5mM EDTA, pH7.6; concentration and time dependedon cell type, see Results). Pronase was autodigested before use at37°C for lh . After washing with 2mgml"1 glycine in PBS(3x5 min) to inhibit further Pronase activity, coverslips andattached cells were dehydrated in a graded ethanol series and airdried.

Probe preparationpMSAT5, a gift from J. Rossant (University of Toronto, Toronto)contains a 1.9 kb (kilobase) repeat of mouse centromeric sDNAsubcloned in the vector pTZ194 between EcoBl and HindBIrestriction sites. The methods of isolation of murine sDNA(Manuelidis, 19816) and characterization of its consensus se-quence have been described by Manuelidis (1981a) and Horz andAltenburger (1981).

For tDNA, a 246 bp (base pair) repeat was subcloned in thevector pBR322 at the Pstl restriction site and designated pHUR93

(Moyzis et al. 1988). To use 'vector-tails' as an aid in amplificationof the hybridization signal (Singer et al. 1986), the 1.8 kb Taqldigest fragment containing the insert was used for in situhybridization.

A 200 ng sample of probe was random-primed with a biotin-16-deoxy-UTP nucleotide (Boehringer Mannheim) in a final volumeof 50/<l. Samples (5/ig) of herring sperm DNA (BoehringerMannheim) and E. coli DNA were added to labelled inserts ofpMSAT5 and pHUR93, respectively. After ethanol precipitation,the labelled probes were redissolved in 50/il of deionizedformamide.

In situ hybridizationCellular DNA was denatured in 70% formamide in 2xSSC (80 °C,3 min; SSC is 0 .15 M NaCl, 0 .015 M sodium citrate, pH7.0)immediately after which the coverslips were immersed in a coldethanol series (80%, 95%, 100% ethanol; 5 min each) and airdried. The probes (5 /<l/coverslip) were similarly denatured (80 °C,10 min) and then chilled on ice (5 min). An equal volume ofhybridization buffer (4xSSC, 0.1% Triton X, 20% dextransulfate) was added and the mixture was vortexed briefly andplaced back on ice. A 10 /A sample of probe-hybridization mixturewas then added to each coverslip, which was placed, cell-sidedown, onto parafilm in a plastic Petri dish. The humidity in thedish was maintained by including a paper tissue saturated withspent denaturation solution. Dishes were sealed with parafilmand incubated at 37 °C, for 2=16 h.

Detection of hybridized sequencesFollowing washing (5x5 min, 2xSSC, RT), bound probe wasreacted with 1:100 rabbit anti-biotin antibody (Enzo) in 2xSSC,0.1% Triton X (2h, RT), washed again and incubated with 1:50fluorescein-conjugated goat anti-rabbit antibody (Jackson) in2xSSC, 0.1% Triton X (2h, RT), followed by washing as before.Coverslips were mounted on glass microscope slides usingantifading mountant (Johnson and Nogueira Aguaro, 1981).Neurons exhibiting circular nuclei were photographed withphase-contrast optics (xlOO, oil-immersion) and fluorescentsignals were visualized in a standard fluorescent microscope (Aexcit. 495 nm, bandpass 540 nm) and photographed on Kodak Tmzp3200 film. Nuclei were optically sectioned at 1 /an intervals, 6 ftmabove and below the nuclear midplane, as previously described(De Boni and Mintz, 1986). The film was push-developed to 6400ASA using D-76 developer (Kodak).

Three-dimensional analysis of the spatial position ofintranuclear lociNegatives were projected onto paper and the outlines of nucleiand nucleoli derived from phase-contrast micrographs, as well asfluorescent sDNA and tDNA signals, were traced for each frame.Nuclei of these cells, previously shown to be spherical (De Boniand Mintz, 1986), had their center assigned as the origin of athree-dimensional XYZ coordinate system. The projections weredigitized, using an in-house computer program, and the distancefrom the nuclear center for each detected locus was calculated.Distances are expressed as mean ± standard error of the mean(S.E.M.). Signals were compartmentalized according to criteriapreviously reported (Borden and Manuelidis, 1988), so thatsignals found to be >l/on from the nucleolus or the nuclearmembrane were defined to be nucleoplasmic signals. In contrast,signals clearly apposed to the nucleolus were counted asnucleolus-associated signals. Computer-assisted three-dimen-sional reconstructions of some nuclei illustrated the arrangementof signals within the nucleus.

Results

Morphology

DRG neurons routinely exhibited a large vesicularnucleus with a clearly defined nuclear envelope andnucleolus. DRG neurons were readily distinguishable

220 F. Billia and U. De Boni

from other cell types on the basis of size and cellmorphology. As neuronal cells dedifferentiate in culture,they typically exhibit three to four nucleoli. However, asthe culture matures and cells redifferentiate, nucleoli areobserved to fuse into one central, large nucleolus. OnlyDRG neurons with a single, large nucleolus were analyzedin the present study.

Despite the extensive pretreatments and denaturationconditions employed, nuclear morphology was sufficientlypreserved to permit clear identification of cellular andnuclear outlines and of nucleoli. Only nuclei that did notoverlap with other nuclei of DRG neurons or of back-ground, non-neuronal cells were chosen. Ambiguous, faintsignals were not included in the analysis.

Probe penetration was improved by prolonged pretreat-ments with 0.1 M HC1 and Pronase and by use of increaseddenaturation times. This resulted in poor nuclear mor-phology. Therefore, conditions were optimized to permitadequate probe penetration while maintaining nuclearmorphology. Optimal conditions for in situ hybridizationfor non-neuronal, 'fibroblast'-like nuclei resulted fromlOmin incubations with O.Smgml"1 Pronase, in theabsence of HC1 treatment. For detection of sDNA in nucleiof DRG neurons, cultures were incubated with 0.1 M HC1for 5min and exposed to O.Smgml"1 Pronase for lOmin.In contrast, more prolonged pretreatment was required forDRG neurons probed for tDNA sequences. These includedincubation with 0 . 1 M HC1 for 20min and O.Smgml"1

Pronase for 15 min. It should be noted that treatment withPronase had previously been shown not to result in areorganization of chromatin (Manuelidis, 19846).

The average nuclear diameter in all DRG neuronsexamined (n=73) was determined to be 10.08±0.17fim(range 7.00-14.54 jan). In vitro studies have previouslyestablished the average diameter of an unfixed DRGneuronal nucleus as 10.72±0.32 jim (Fung and De Boni,unpublished). The fixation protocol, dehydration andpretreatments thus altered the dimensions of the nucleusby a mean factor of 0.94. Owing to the range of nucleardiameters encountered in the DRG nuclei examined, allnuclear dimensions were normalized to a nuclear diameterof 10.72 jtm.

Control experiments for in situ hybridizationSeveral control experiments were carried out to eliminatethe possibility that the observed signals were due to non-specific hybridization of probe to endogenous denaturedDNA and to exclude the possibility that no unincorpor-ated, biotinylated nucleotides were detected. In a typicalcontrol, 5-10 //I of hybridization buffer was used instead ofdenatured probe, while in others, the primary antibodywas omitted. In both of these controls only low levels ofcytoplasmic autofluorescence were detected, with thenucleus clearly outlined as a dark, circular area, withoutany fluorescent signals (not shown).

Arrangement of sDNA sequencesThe spatial arrangement of sDNA sequences in neuronswas examined in 40 cells (319 signals; Fig. 1A,B). A meannumber of 8.02±0.40 signals/nucleus (range 5-13signals/nucleus) was observed. The hybridization signalsfrom sDNA sequences measured 1-2 im in width andwere clearly in focus in more than one optical section(lftm). Of all sDNA signals assayed, 41.65±0.59% werefound closely associated with the nucleolus, ranging from2 to 5 signals/nucleus (Figs 2, Fig. 3A,C). The remainingfraction of 57.23±0.56% or 3-8 signals/nucleus was

Fig. 1. Phase-contrast (A,C) and corresponding fluorescence(B,D) micrographs of representative neuron (A,B) and non-neuronal cells (C,D). In the neuron (A,B), note association ofsome sDNA signals with the nucleolus while others arelocalized to the nucleoplasm. In the non-neuronal cell, note thelarge number of sDNA signals resolved in the given focalplane and absence of apparent clustering. Bar, 5 /jm.

detected between the nucleolus and the nuclear periphery(Figs 3C, 4). A small fraction of these signals, approxi-mately 10 %, was found in close proximity to the nuclearmembrane. Overall, sDNA signals were found at a meandistance of 3.15±0.06/<m from the center of the nucleus,representing 63.81±1.07 % of the nuclear radius (Fig. 3A).

In non-neuronal interphase nuclei of background,'fibroblast'-like cells, 37-40 sDNA signals/nucleus weredetected (n=10; Fig. 1C,D). These sDNA signals werefound throughout the nucleus, were often observedoccurring in pairs, were smaller in size, approximately0.5 ,um in width, and in focus only within one opticalsection of 1 f.tm.

Arrangement of tDNA sequencesThe spatial pattern of telomeric DNA sequences wasexamined in 33 cells (1219 signals; Fig. 5). In contrast tosDNA sequences, an average of only 18.45±0.41% of thetotal number of signals was associated with the nucleolus(Figs 3B,C, 4B). The remaining fraction of 81.79±0.41 % ofsignals was either associated with the nuclear periphery(12.81±0.50%) or observed at positions between thenucleolus and the nuclear membrane (Figs 3B, 6). Signalswere found at a mean distance of 3.57 ±0.03 ,um from thenuclear center, corresponding to 66.59±0.61% of thenuclear radius (Fig. 3B). Although this mean is notsignificantly different from the mean distance of sDNAsignals from the nuclear center, the frequency distri-butions differ significantly from that found for sDNA(Fig. 3) as determined by the non-parametric Kolmogrov-Smirnov two-group test (P=0.0126). This statistical test isideal for determining the level of significance of afrequency profile (Young, 1977; van Dekken etal. 1990). Incontrast to sDNA signals, tDNA signals were much

Satellite and telomeric DNA in neurons 221

40 n

30-

% 20-

10

II

401

30-

% 20-

10-

0

II. B

0-1 1-2 2-3 3-4 4-5Distance from nuclear centre

5-6

80

60

40-

20

CUsDNA

Nucleotar Nuclsoplatmlc

0.0

Distance from nuclear center (yum)

Fig. 2. Diagrammatic representation of spatial positions ofnucleoli and corresponding sDNA signals in neurons. Y axisrepresents 40 individual cells. X axis shows spatial position, asdistance from nuclear center, of nucleoli (bold) and of sDNAsignals (fine). Extent of bold bar defines widths of nucleoli withfilled circle defining the nucleolar center. Similarly, size andcenter of sDNA signals are denned by fine bar and filledtriangle, respectively.

Fig. 3. Histograms showing distributions (mean±s.E.M.) ofspatial positions of sDNA (A) and tDNA (B) with respect to thenuclear center and their compartmentalization (C). Notesignificant shift (P=0.01, see text) of tDNA signals to a moreperipheral distribution (B,C), compared to sDNA (A,C). Forcomparison, insets (A,B; reduced in size) show correspondingspatial distributions of nucleoli, derived from same data set(n=73).

smaller in size, <0.5jrm in width, were frequentlyobserved in pairs and were in focus only within one opticalsection. An average of 37.00±1.52 signals/nucleus (range26-54) was detected.

Discussion

Chromosomes occupy distinct territories within the inter-phase nucleus of many cell types. Several investigatorshave shown that chromatin arrangement is cell type-

222 F. Billia and U. De Boni

specific, an arrangement that may be influenced bydifferentiation, pathological conditions or gene expression(Barr and Bertram, 1949; Manuelidis, 1985a; Blobel, 1985;Borden and Manuelidis, 1988; Arnoldus et al. 1989). Thefinding reported here that sDNA signals in DRG neuronalnuclei are found to be associated primarily with thenucleolus and at positions intermediate between thenucleolus and the nuclear membrane, is similar to the

Fig. 4. Representative, three-dimensional reconstructions ofneurons showing spatial positions of sDNA (A) and tDNA (B),relative to nucleoli (filled) and nuclear periphery.Reconstructions are of neurons shown in Fig. 1 and in Fig. 5.Halves are shown rotated —35 degrees and -150 degrees,respectively. Note large size and lower number of sDNAsignals (A) compared to tDNA (B).

J

Fig. 5. Micrographs showing labelled tDNA signals (B) andcorresponding phase-contrast image (A) of representativeneuron. Note 5 clearly defined signals in this focal plane,others out of focus. Bar, 5 fan.

arrangement of sDNA sequences reported for interphasenuclei in human cancer cell lines (Haaf and Schmid, 1989;van Dekken et al. 1990), in Chinese Hamster Ovarian andB-lymphoid cell lines (Moroi et al. 1981), in mouse livercells (Katsumata and Lo, 1988), in mouse Sertoli cells (Raeand Franke, 1972) and in mouse and human centralnervous system cells (Manuelidis, 19846, 1985a; Manueli-dis and Borden, 1988).

Non-neuronal 'fibroblast'-like cells examined in thepresent work exhibited a number of sDNA signals close tothe diploid complement of the mouse genome, 40 signals/nucleus. This confirmed that all centromeric sDNA targetswere detected under the conditions employed, and standsin contrast to the reduced number of signals/nucleus

A A A. A_^ A AAA* ? i*A*AA A l l A* A* A A AA I

: • •

. A *A

._. A

* A A

A

A A*

< '

A A _ \ A""

A A

_ A A A

\ A

UA

A A

A A

A A

A AA

A A

A M

A

t

AA A

i "

A

/ A

A A A

A t A

A A

A*A

AA m

AA>A

' A »,

A A \

A

tAAA

/ A

AA A

A / „

A A*

A *

\ AA A AA A

^ A V * AA*

AA *A I S * A A AA At A

A A A A AA A

'AA *M \ A.A

A A AA A A A * A

A \ *A A \

A

A A

A t

A

A

A A A

A A

\ A

. - / ;

* * * A AA AA

A A A A * / P u

A A _ A * S *AA** A

. 1 / A - A . A V ' . -

A. A O A ^

4

AA 1 A

. : . . » .

. A A

A

• A * .

A

A A

A A A / . . 1

A

AAA*AA A i t

1fc-A 1 1

M V A I «AA I

* *LA A AAA A f

1*1 AAAA A

. 1 - A A l

»* 1 A * T * A I

AA A A A

A A A

4 .A

1 A A l i

AA *k AAA

A AA AA A A A

A *

i*kA t " A A * fA

AA A r A A

AAA * A A . ' A A

^ • U it

• t AA. A*

AA. S i A «A

A A A > / A t

A l u r f t 1

A A A A * A A

0.0 1.0 2.0 3.0 4.0 5.0 6.0

Distance from nuclear center (/u.m)

Fig. 6. Diagrammatic representation of the spatial positions ofnucleoli and corresponding tDNA signals (triangles) in 26neurons. See legend to Fig. 2 for details, with the exceptionthat tDNA signals are small and their actual size is containedwithin the filled triangle. Note spread of signal positionsbetween nucleolus and the nuclear periphery.

Satellite and telomeric DNA in neurons 223

observed in DRG neuronal nuclei. This is likely to be dueto clustering of sDNA signals, resulting in the large signalvolumes observed in neurons. Given an ultrastructurallydetermined average width of centromeric regions ofapproximately 0.2-0.3 Jim (Sedat and Manuelidis, 1978;Manuelidis, 1982), it may be considered that centromericregions of several chromosomes cluster to form the signalsobserved, with the signal size potentially proportional tothe number of centromeres contained in the cluster. Incontrast, the larger number and smaller size of sDNAsignals observed in nuclei of the 'fibroblast'-like cells,suggest that little or no clustering occurred in this celltype. Clustering of centromeres, as reported here forneuronal nuclei in DRGs, is a common, well-documentedphenomenon in neurons (Manuelids, 1984a,6; Arnoldus etal. 1989). It has also been described in plant cell nuclei(Fussel, 1975; Church and Moens, 1976; Evans and Filion,1982). The mechanism that underlies the condensation ofcentromeres into clusters remains unclear. It is thought,however, that interactions between kinetochore proteinscould mediate such clustering (Haaf and Schmid, 1989).This is supported by results of ongoing work in ourlaboratory that have clearly shown that kinetochoreproteins are present in post-mitotic DRG neurons. More-over, these proteins present as clusters, are quantitativelysimilar in distribution to sDNA signals (Holowacz and DeBoni, 1991).

In contrast to the extensive body of work carried out onsDNA patterns, relatively little work has addressed thetopology of tDNA in interphase nuclei. In contrast toresults of work by Rawlins and Shaw (1990a,6), whorecently reported an association of telomeres with nucleoliand the nuclear membrane in Pisum sativum, the workshown here indicates that approximately 80 % of tDNAsignals are not localized to these loci but, rather, are foundin the nucleoplasm, intermediate between the nucleolusand the nuclear membrane.

The observation that the average number of telomeresignals detected was half that expected for a diploidcomplement may result either from an inconsistency inlabelling or, alternatively, may be due to a distinctlabelling pattern of mouse acrocentric chromosomes(Narayanswami et al. 1989). In that work, in situhybridization to telomeres also resulted in labelling ofonly half of the potential labelling sites. This wasinterpreted to indicate that labelling occurred to onlythose telomeres associated with the short segment ofmouse acrocentric chromosomes and that either hybridiz-ation at the other chromosome end was less intense ortDNA was possibly inaccessible to the probe. On this basis,it could be predicted that the distances for sDNA andtDNA signals would be similar. This is supported byresults reported here. In addition, the observation thatsignals of both centromeres and telomeres often occur inpairs may be interpreted as an indication of pairing ofhomologous chromosomes, as reported by Brinkley et al.(1986), Hadlaczky et al. (1986) and Arnoldus et al. (1989).

The finding that satellite and telomeric DNA sequencesare distributed throughout the nucleoplasm and notassociated with any particular pole of the nucleus standsin contrast to the topology of chromosomes exhibiting aclassical Rabl configuration, as seen in polytene nuclei(Agard and Sedat, 1983; Hochstrasser et al. 1986; Hoch-strasser and Sedat, 1987a,6) and in nuclei of several plantspecies (Fussell, 1975; Church and Moens, 1976; Avivi andFeldman, 1980). This is in keeping with evidence fromother mammalian systems where Rabl configurations are

not generally found (Pardue and Gall, 1970; Rae andFranke, 1972; Manuelidis, 1982, 19846, 1985a,6; Schardinetal. 1985; Manuelidis and Borden, 1988; Arnoldus, 1989).

Order within the interphase nucleus is thought to resultfrom attachment of specific chromosome segments, i.e.centromeres, telomeres and heterochromatic regions, tothe nuclear envelope, to the nuclear matrix and to thenucleolus. The observation that 47 % of the sDNA signalsin neuronal nuclei are associated with the nucleolusindicated that nucleolar-organizing region (NOR)-bearingchromosomes may be contained within these sDNAclusters. The nucleolus is a major structure in the nucleuswherein its spatial position is tissue-specific (Bourgeois etal. 1979; Hubert and Bourgeois, 1986; Wachtler et al. 1986;Hochstrasser et al. 1987a; Bourgeois and Hubert, 1988). Ininterphase, it represents the position of active NORs (Hsuet al. 1971; Bourgeois et al. 1984; Wachtler et al. 1986;Rawlins and Shaw, 1990a) and is thought to be the moststriking example of nonrandom organization of chromatindomains (Comings, 1968, 1980; Hsu et al. 1971; Manueli-dis, 1984a; Hochstrasser et al. 19876). In fact, it has beenpostulated that the attachment of the NOR-bearingchromosomes to the nuclear matrix is the focal point forthe formation of the fibrillar center of the nucleolus(Hilliker and Appels, 1989). Moreover, Haaf and Schmid(1989) have found that the majority of centromeres intumor cell nuclei were closely associated with thenucleolus, presumably representing the acrocentric NOR-bearing chromosomes.

If centromere, telomeric and other repetitive sequencesact to anchor selected regions of chromosomes to thenuclear matrix (Comings, 1980) and/or nuclear envelope,a general model of chromatin organization could beenvisioned. Compartmentalization of chromatin couldprovide a structural framework for efficient processing ofnuclear events. In fact, Manuelidis and Borden (1988) andSchardin et al. (1985) proposed such models in whichrepetitive sequences could act as a structural center for theextension and condensation of chromatin to and from atranscriptionally competent nuclear compartment. It ispossible that this organization reflects the physiologicalstate of a given cell (Manuelidis and Borden, 1988; Bordenand Manuelidis, 1988) and is thought to remain dynamicin highly differentiated cells. Rearrangement of specificchromatin domains has been repeatedly observed in vitro,represented by a phenomenon termed Nuclear Rotation(Pomerat et al. 1967; De Boni and Mintz, 1986; De Boni,1988; Fung and De Boni, 1988; Hay and De Boni, 1991;Park et al. 1991; Park and De Boni, unpublished data). Inaddition, rearrangement of chromatin domains has beenimplied to occur in vivo (Manuelidis, 1985a; Borden andManuelidis, 1988; Haaf and Schmid, 1991). It remains tobe shown in which manner the tissue-specific, non-randompattern of chromatin organization is established. NuclearRotation, the motion of chromatin domains, has beenproposed to play such a role during differentiation. In fact,recent observations indicate that Nuclear Rotation is mostpronounced during differentiation of cells in vitro and isless pronounced as cells assume their fully differentiatedstate (Park and De Boni, unpublished data; Park et al.1991).

In summary, the results presented here confirm thatspecific chromatin domains are reproducibly compartmen-talized. While this has been postulated to represent onelevel of control of gene expression (Comings, 1980), itremains to be shown whether the nuclear location of agene effects its expression.

224 F. Billia and U. De Boni

References

AOARD, D. A. AND SEDAT, J. W. (1983). Three-dimensional architectureof a polytene nucleus. Nature 302, 676-681.

ARNOLDUS, E. P. J., PETERS, A. C. B., BOTS, G. T. A M., RAAP, A. KAND VAN DER PLOEO, M. (1989). Somatic pairing of chromosome 1centromeres in interphase nuclei of human cerebellum Hum. Genet83, 231-234.

AVIVI, L. AND FELDMAN, M. (1980) Arrangement of chromosomes in theinterphase nucleus of plants. Hum. Genet. 55, 281-295.

BARR, M. L. AND BERTRAM, E. G. (1949). A morphological distinctionbetween neurones of the male and female, and the behavior of thenucleolar satellite during accelerated nucleoprotein synthesis. Nature163, 676-677

BLOBEL, G. (1985). Gene gating: A hypothesis Proc. natn. Acad. Sci.U.S.A. 82, 8627-8529.

BORDEN, J. AND MANUELIDIS, L. (1988). Movement of the X chromosomein epilepsy. Science 242, 1687-1691.

BOURGEOIS, A. AND HUBERT, J. (1988). Spatial relationship between thenucleolus and the nuclear envelope In International Review ofCytology, vol. I l l (ed. H. Busch), pp.1-52 New York: Academic Press.

BOURGEOIS, C. A., COSTACLIOLA, D., LAQUERRIEREE, F., BARD, F., HEMON,D. AND BOUTEILLE, M. (1984). In situ arrangement of NOR-bearingchromosomes in the interphase nucleus of Aotus trwirgatus. J. CellSet. 69, 107-115.

BOURGEOIS, C. A., HEMON, D. AND BOUTEILLE, M. (1979). Structuralrelationship between the nucleolus and the nuclear envelope JUltrastruct. Res. 68, 328-340

BRINKLEY, B. R., BRENNER, S. L., HALL, J. M , TOUSSON, A, BALCZON, R.D. AND VALDIVIA, M. M. (1986). Arrangements of kinetochores inmouse cells during meiosis and spermiogenesis. Chromosoma 94,309-317.

CHALY, N. AND BROWN, D. L. (1988) The prometaphase configurationand chromosome order in early mitosis. J. Cell Sci. 91, 325-335.

CHURCH, K. AND MOENS, P. B. (1976) Centromere behavior duringinterphase and meiotic prophase in Alhum fistulum from 3-D E.Mreconstruction. Chromosoma 56, 249-262.

COMINGS, D. E. (1968). The rationale for an ordered arrangement ofchromatin in the interphase nucleus. Am. J. hum. Genet. 20, 440-460.

COMINGS, D. E. (1980) Arrangement of chromatin in the nucleus. Hum.Genet. 53, 131-143.

CREMER, T., CREMER, C, BAUMANN, H., LEUDTKE, E.-K., SPERLING, K.,TEUBER, V. AND ZORN, C. (1982). Rabl's model of the interphasechromosome arrangement tested in Chinese hamster cells bypremature chromosome condensation and laser-UV-microbeamexperiments. Hum. Genet. 60, 46-56.

DE BONI, U. (1988). Chromatin motions in interphase nuclei, itsmodulation and its potential role in gene expression. Anticancer Res.8, 885-898.

DE BONI, U. AND MINTZ, A. H. (1986). Curvilinear, three-dimensionalmotion of chromatin domains and nucleoli in neuronal interphasenuclei. Science 234, 863-866.

EARNSHAW, W C, HALLIGAN, N., COOKE, C. AND ROTHFIELD, N. (1984).The kinetochore is part of the metaphase chromosome scaffold. J. CellBiol. 98, 362-357.

EVANS, K. J. AND FILION, W. G. (1982). The distribution of chromatin inthe interphase nucleus of Zebrina pendula. Can. J. Genet. Cytol. 24,583-591.

FUNG, L. C. AND DE BONI, U. (1988). Modulation of nuclear rotation inneuronal interphase nuclei by nerve growth factor, by gamma-aminobutync acid, and by changes in intracellular calcium. CellModi. Cytoskel. 10, 363-373.

FUSSELL, C. P. (1975). The position of interphase chromosomes and latereplication DNA in centromere and telomere regions of Allium cepaL. Chromosoma 50, 201-210.

HAAF, T., GRUNENBERG, H. AND SCHMID, M. (1990). Paired arrangementof nonhomologous centromeres during vertebrate spermiogenesis. ExplCell Res. 187, 157-161.

HAAF, T AND SCHMID, M. (1989). Centromeric association and non-random distribution of centromeres in human tumour cells. Hum.Genet. 81, 137-143

HAAF, T. AND SCHMID, M. (1991). Chromosome topology in mammalianinterphase nuclei. Expl Cell Res. 192, 325-332.

HADLACZKY, G. Y., WENT, M. AND RINQERTZ, N. R. (1986). Directevidence for the non-random localization of mammalian chromosomesin the interphase nucleus. Expl Cell Res. 167, 1-15.

HAY, M. AND DE BONI, U. (1991). Chromatin motion in neuronal nuclei1

Changes induced by disruption of intermediate filaments. Cell Motit.Cytoskel. 18, 63-75.

HILLIKER, A. J. AND APPELS, R. (1989). The arrangement of interphasechromosomes: Structural and functional aspects. Expl Cell Res 185,297-318.

HIRAOKA, Y., AGARD, D. A. AND SEDAT, J. W. (1990) Temporal and

spatial coordination of chromosome movement, spindle formation, andnuclear envelope breakdown during prometaphase in Drosophilamelanogaster embryos J Cell Biol 111,2816-2828.

HOCHSTRASSER, M., MATHOO, D., GRUENBAUM, Y., SAUMWEBER, J. ANDSEDAT, J. W. (1986). Spatial organization of chromosomes in thesalivary gland nuclei of Drosophila meUinogaster. J. Cell Biol. 102,112-123.

HOCHSTRASSER, M AND SEDAT, J. W (1987a). Three-dimensionalorganization of Drosophila melanogaster interphase nuclei I Tissue-specific aspects of polytene nuclear architecture. J Cell Biol 104,1455-1470.

HOCHSTRASSER, M. AND SEDAT, J. W. (19876) Three-dimensionalorganization of Drosophila melanogaster interphase nuclei.II. Chromosome spatial organization and gene regulation. J. Cell Biol.104, 1471-1483.

HOLOWACZ, T. AND DE BONI, U. (1991). Arrangement of kinetochoreproteins and satellite DNA in neuronal interphase nuclei' Changesinduced by gamma-aminobutyric acid (GABA). Expl Cell Res. (inpress).

HORZ, W. AND ALTENBURGER, W. (1981). Nucleotide sequence of mousesatellite DNA. Nucl. Acid Res. 9, 683-696.

Hsu, T. C, COOPER, J. E. K., MACE, M. L. JR AND BRINKLEY, B R.(1971). Arrangement of centromeres in mouse cellB. Chromosoma 34,73-87.

HUBERT, J. AND BOURGEOIS, A. (1986) The nuclear skeleton and thespatial arrangement of chromosomes in the interphase nucleus ofvertebrate somatic cells. Hum. Genet. 74, 1-15.

JOHNSON, G. D. AND NOGUEIRO AGUARO, G. M C. (1981) A simplemethod of reducing the fading of immunofluorescence. J. immun.Meth. 43, 349-350.

JONES, K. W. (1970). Chromosomal and nuclear location of mousesatellite DNA in individual cells. Nature 225, 912-915.

JOSEPH, A , MITCHELL, A R. AND MILLER, O. J. (1989). The organizationof the mouse satellite DNA at centromeres Expl Cell Res 183,494-500.

KATSUMATA, M AND LO, C. W. (1988). Organization of chromosomes inthe mouse nucleus' Analysis by in situ hybridization. J. Cell Sci. 90,193-199.

MANUELIDIS, L (1981a). A simplified method of preparation of mousesatellite DNA Analyt Biochem 78, 661-568

MANUELIDIS, L. (19816). Consensus sequence of mouse satellite DNAindicates it is derived from tandem 116 basepair repeats. FEBS Lett.129, 26-28.

MANUELJDIS, L (1982) Repeated DNA sequences and nuclear structures.In Genome Evolution (ed G. A. Dover and R. B Flavell), pp. 263-285.London: Academic Press.

MANUBLIDIS, L. (1984a). Active nucleolus organizers are preciselypositioned in adult central nervous system cells but not inneuroectodermal tumour cells. J. Neuropath, exp. Neurol. 43, 225-241.

MANUELIDIS, L. (19846) Different central nervouB system cell typesdisplay distinct and nonrandom arrangements of satellite DNA. Proc.natn. Acad. Sci. U.S.A. 81, 2123-3127.

MANUELIDIS, L. (1985a). Indications of centromere movement dunnginterphase and differentiation. Ann. N. Y. Acad. Sci. 450, 205-220.

MANUELIDIS, L. (19856). Individual interphase chromosome domainsrevealed by in situ hybridization. Hum Genet 71, 288—293.

MANUELIDS, L. AND BORDEN, J (1988). Reproduciblecompartmentalization of individual chromosome domains in humanCNS cells revealed by in situ hybridization and three-dimensionalreconstruction. Chromosoma 96, 397-410.

MATHOG, D , HOCHSTRASSER, M., GRUENBAUM, Y., SAUMWEBER, H. ANDSEDAT, J. W. (1984). Characteristic folding pattern of the polytenechromosomes in Drosophila salivary gland nuclei. Nature 308,414-421.

MATSUMOTO, H., SUQIMOTO, K. AND OKAZAKI, T. (1989). Alphoid satelliteDNA is tightly associated with centromere antigens in humanchromosomes throughout the cell cycle. Expl Cell Res. 181, 181-196.

MOROI, Y., HARTMAN, A. L., NAKAND, P. K. AND TAN, E M (1981).Distribution of kinetochore (centromere) antigen in mammalian cellnuclei. J. Cell Biol 90, 254-259

MOYZIS, R. K., BUCKINGHAM, J. M., CRAM, L S., DANI, M., DEAVEN, L.L., JONES, M. D., MEYNE, J., RATUPT, R. L. AND WU, J.-R. (1988). Ahighly conserved repetitive DNA sequence, (TTAGGG)n, present atthe telomeres of human chromosomes Proc. natn. Acad. Sci. U.S.A.85, 6622-6626.

MURRAY, A. B. AND DAVIES, H. G. (1979). Three-dimensionalreconstruction of the chromatin bodies in the nuclei of matureerythrocytes from the newt Triturus cristatus The number of nuclearenvelope-attachment sites. J. Cell Sci. 35, 59-66.

NARAYANSWAMI, S., MOYZIS, R. K. AND HAMKALO, B. A. (1989). Highresolution mapping of higher eukaryotic telomeres. J. Cell Biol. 107,739a.

Satellite and telomeric DNA in neurons 225

PAHDUE, M. L. AND GALL, J G (1970) Chromosomal localization ofmouse satellite DNA. Science 168, 1356-1358

PARK, P. C, DORJUNGTON, C. AND DE BONI, U (1991). Saltatory andperiodic motion of nucleoh in interphase nuclei during neuronaldifferentiation in vitro. Proc. Can. Fed. Biol. Soc. Symp 34th AnnualMeeting, p. 126, 464.

POMERAT, C M., HENDBLMAN, W. J., RAIBORN, C. W. AND MASSEY, J. F(1967). Dynamic activities of nervous tissue in vitro In The Neuron(ed. H. Hyden), pp. 119-178. New York: Elsevier Publishing.

RABL, C. (1885). In Morphologisches Jahrbuch, vol. 10 (ed. C.Gegangauur), pp. 214-330.

RAK, P. M. M. AND PRANKE, W W (1972) The interphase distribution ofsatellite DNA-containing heterochromatin in mouse nuclei.Chromosome 39, 443-456.

RAWLINS, D. J. AND SHAW, P. J. (1990a). Localization of ribosomal andtelomeric DNA sequences in intact plant nuclei by in-situhybridization and three dimensional reconstruction. J. Microsc. 157,83-89.

RAWLINS, D. J. AND SHAW, P. J. (19906) 3-D arrangement of genes ininterphase by in situ hybridization and confecal microscopy. J CellBiol. I l l , 504a.

SCHARDIN, M., CRKMER, T., HAGER, H. D AND LANG, M. (1985). Chinese

hamsterxman hybrid cell lines demonstrates interphase chromosometerritories. Hum. Genet. 71, 281-287.

SEDAT, J. AND MANUELJDIS, L. (1978). A direct approach to the structureof eukaryotic chromosomes. Cold Spring Harbor Symp. quant. Biol.42, 331-350

SINGER, R H., LAWRENCE, J. B. AND VILLNAVE, C. (1986). Optimizationof in situ hybridization using isotopic and non-isotopic detectionmethods Biotechniques 4, 230-250

VAN DEKKEN, H., VAN ROTTERDAM, A , JONKER, R., VAN DER VOORT, H. TM. AND BRAKENHOFF, G. J. (1990) Confocal microscopy as a tool forthe study of the intranuclear topography of chromosomes. J. Microsc.158, 207-214.

WACHTLER, F., HOPMAN, A. H. N., WIEGANT, J. AND SCHWARZACHER, H.G. (1986) On the position of nucleolus organizer regions (NORs) ininterphase nuclei. Expl Cell Res. 167, 227-240.

YOUNG, I. T (1977). Proof without prejudice: Use of the Kolmogrov-Smirnov test for the analysis of histograms from flow systems andother sources. J. Histochem. Cytochem. 25, 935-941.

(Received 29 April 1991 - Accepted 7 June 1991)

226 F. Billia and U. De Boni