J. Cell Sci. 71, 75-93 (1984) 75Printed in Great Britain © The Company of Biologists Limited 1984

STRUCTURAL AND BIOCHEMICAL ASPECTS OF

LIVER CELL NUCLEAR ORGANIZATION IN NORMAL

AND SCALDED RATS

LJILJANA SEVALJEVIC, MIODRAG PETROVIC, GORANPOZNANOVlC, ANDJELIJA gKARO AND DRAGOLJUBPANTELICInstitute for Biological Research, Beograd, 29 Novembra 142, Yugoslavia

SUMMARY

The response of rat liver nuclei to thermal injury was studied at the ultrastructural and bio-chemical levels using nuclei isolated in the presence of either Mg2"1" or polyamines and metalchelators.

The extent of chromatin condensation, as revealed by electron microscopy, increased in theorder: nuclei in situ, Mg-stabilized nuclei, polyamine-stabilized nuclei. In addition, we observedthat thermal injury caused an increase in the number of nuclear pores, an enlargement of nucleoliand an accumulation of ribonucleoprotein material. Along with this, greater amounts of protein,DNA and RNA were retained in nuclei from scalded rats.

The salt-resistant residue from Mg-stabilized nuclei was a spherical proteinaceous structure of thenuclear matrix, whereas that of the polyamine-stabilized nuclei was amorphous and deprived ofthree major constituents of the spherical matrix, the 60-70 X 10? Mr lamina proteins. However,exposure of the polyamine-stabilized nuclei to Ca2+ and Mg24 rendered the 60-70 X 103 M, proteinssalt-insoluble and thus enabled the extraction of a spherical residual nuclear structure. This struc-ture was highly enriched in DNA, RNA and non-histone proteins and, hence, more like dehis-tonized nuclei than the minimal residual nuclear structure. It retained 70 % of DNA in the controlsbut virtually all the DNA in scalded rats. This difference was interpreted to reflect activation-relatedchanges in chromatin organization.

INTRODUCTION

There is a great deal of evidence that the cell nucleus is composed of an internalproteinaceous nuclear framework contiguous with the nuclear membrane (Berezney& Coffey, 1977) and that the chromatin is anchored to this structure at both thenuclear envelope and internal matrix sites (Lebkowski & Laemmli, 1982a,6; Small,Nelkin & Vogelstein, 1982). The interactions of chromatin with the nuclear matrixmight thus determine the higher order of chromatin structure and have a role in themechanisms by which a 'message' promotes a change in the chromatin conformation,which introduces changes in the pattern of gene expression. The problem hinderinga more direct study of chromatin-matrix interactions concerns the nature of thenuclear matrix. This structure has not yet been visualized in situ despite the convinc-ing evidence for a masking effect of chromatin and other nucleoplasmic constituents(Brasch, 1982). Also the reported variations in the structural and biochemical charac-teristics of the isolated nuclear framework structures (Berezney, 1979; Riley, Keller& Beyer, 1975; Kaufmann, Coffey & Shaper, 1981; Herlan, Quevedo & Wunderlich,

76 L. Sevaljevic and others

1978; Sevaljevic", Petrovi<5, Poznanovid & Konstantinovic", 1981) make the occurrenceof artifactual distortions of the original structure during isolation likely.

In this work we investigated the ultrastructural and biochemical characteristics ofnuclei and nuclear matrices from normal and activated liver cells, giving particularemphasis to the effects of different nuclear isolation methods on the properties ofisolated structures. Rat liver hepatocytes 24 h after a sublethal scalding offer anappropriate experimental model of cells with enhanced nuclear activity. At that perioda remarkable increase in the polyploidy of hepatic nuclei (bevaljevid, Petrovic" &Pantelic", 1983) and in the rates of liver-tissue protein synthesis (Sevaljevic", Petrovic",Savid & Pantelic", 1982) were found to coincide with the appearance of a large amountof newly synthesized acute-phase reactant proteins in the blood stream (Sevaljevid etal. 1983a). The results presented here demonstrate that the internal organization andgross composition of nuclei and nuclear matrices depended strongly both on thefunctional state of the cells and on the method of nuclear isolation.

MATERIALS AND METHODS

Infliction of scaldingThe experiments were performed on male albino rats of our own breed, weighing 220-250g. A

standardized amount of skin surface, representing 20 % of the total skin surface, was clipped on bothlateral sides of the body one day before the experiment. Scalding was inflicted as described byArturson (1961), under ether anaesthesia. Clipped skin was exposed to hot water (83°C) for 30s.At 24h after scalding a volume equal to 10% of the body weight of 153-8mM-NaCl was injectedperitoneally. Thereafter, commercial chow and tap water were given regularly. The injured animalswere accommodated in a room with constant temperature variations (20°C-22°C). Except for thescalding, the animals from the control groups were subjected to the same experimental procedure.

Isolation of nucleiTwo different procedures for the isolation of rat liver nuclei were followed.Isolation in the presence of metal cations. Liver nuclei were prepared as described by Berezney

& Coffey (1977). Livers were quickly excised, cut with scissors and homogenized with a Janke andKunkel homogenizer at 30 % of the maximum speed for 60 s with a tissue ratio of 1 g of liver to2ml of 0-25M-sucrose, 005M-TrisHCl (pH7-4), 5mM-MgCl2, 1 mM-phenylmethansulphonyl-fluoride (PMSF). After filtration through four layers of cheesecloth, the homogenate was centri-fuged at 1000 £ for lOmin in a Sorvall HB-4 rotor. After a wash with the same buffer, the crudenuclear pellet was resuspended in 2-2M-sucrose, 0-05M-Tris-HCl (pH7-4), 5mM-MgCl2, 1 min-PMSF and centrifuged in a Beckman SW27 rotor at 60000|f for 60min. The purified nuclei weresubsequently washed twice with the homogenization buffer.

Isolation in the presence of metal chelators and poly amines. Nuclei were isolated according to theprocedure of Burgoyne, Hewish & Mobbs (1974). Minced livers were homogenized under the sameconditions as described above, in the following medium: 0-34M-sucrose, 2mM-ethylendiaminetetraacetic acid (EDTA), 0-5mM-ethylenglycolbis(/3-aminoethyl ether)-Nffjy'JV'-tetraaceticacid (EGTA) in buffer A (60mM-KCl, 15mM-NaCl, 0-5 mM-spermidine, 0-15 mM-spermine,lmM-PMSF, 15mM-TrisHCl, pH 7-4). The homogenate was centrifuged at 1000 gfor lOmin ina Sorvall HB-4 rotor. The crude nuclear pellet was then resuspended in buffer A with 1-37M-sucrose, 1 mM-EDTA and 0-25 M - E G T A , layered over an equal volume of the same medium andcentrifuged at 16 000 # for 15 min. The nuclear pellet was resuspended in 2-l M-sucrose, 0-1 mM-EDTA and 0-l rriM-EGTA in buffer A, layered over an equal volume of the same medium andcentrifuged in a Beckman SW27 rotor at 60 000£ for 60 min. The purified nuclear pellet was thenwashed twice with 0'34M-sucrose in buffer A.

Liver cell nuclear organization in rats 77

Incubation of nuclei with metal cations and DNaseThe nuclei were used for isolating the residual nuclear structure 16 h after preparation. Nuclei

isolated in the presence of Mg2"1" were kept on ice in the Mg-homogenization buffer. The nucleiisolated in the presence of metal chelators and polyamines were resuspended in 0-34M-sucrose inbuffer A and divided into five samples. One was kept on ice overnight in the 0-34 M-sucrose/bufferA medium without any additions, the remaining four nuclear samples were incubated overnight inthe following concentrations of cations and DNase II: (1) lmM-CaCb; (2) 10mM-MgCl2; (3)

and 10mM-MgCl2; 50 units of DNase II (Worthington)/unit A260 •

Isolation of residual nuclear structuresThe nuclear matrices were isolated from different nuclear samples as described by Berezney &

Coffey (1977). After the overnight incubation step described above, the nuclei were centrifuged at1000|f for 15min and resuspended in a solution of 1 % Triton X-100 either in the Mg-homogenization buffer or in 0-34 M-sucrose in buffer A depending on whether the nuclei in questionwere initially isolated in the presence of Mg2"1" or polyamines and metal chelators. After incubationon ice for 15 min the nuclei were centrifuged for 15 min. This extraction was repeated once more.The pellet obtained was resuspended in the above buffer and an equal volume of 4M-NaCl in thesame medium was slowly added to the nuclear suspension with constant stirring. The nuclearstructures were kept in the 2M-NaCl solution (HS) on ice for 15 min and centrifuged at 1500|*for15 min. This extraction was repeated twice more. In some experiments, the nuclear matrices wereincubated with 200 units of DNase I/unit A26O for 60 min at 22°C.

Buoyant density centrifugationThe nuclei and nuclear matrices were analysed in the sucrose/glucose gradient described by

Raynaud &Ohlenbush (1972). Ten units olAwa of nuclei and nuclear matrices per 1 ml were layeredon the gradients prepared in the tubes of the Beckman SW25.1 rotor. In the course of our work wefound that centrifugation at high g forces, as described by those authors, caused the nuclei andnuclear matrices to collapse. Thus, we performed the centrifugation at 2500£ and monitored theintactness of the nuclear spheres by light microscopy. Under these conditions the buoyant densitiesof the nuclear spheres were reached after 90 min of sedimentation. Longer periods of centrifugationdid not change the buoyancy positions of the nuclear structures.

Electron microscopyThe samples (tissue, nuclei and nuclear matrices) were fixed in 1-6% Sbrensen solution in

glutaraldehyde (pH 7-3-7-4) at 4°C for 1 h and stained either by the standard procedure, or forribonucleoprotein (RNP) with EDTA. For standard staining the samples were postfixed in 2%OsC>4, dehydrated with increasing concentrations of ethanol and embedded in Epon according tothe usual procedure. Ultrathin sections of embedded material were stained with uranyl acetatefollowed by lead citrate. For RNP staining the postfixation with OsO4 was omitted, whereas theultrathin sections were subsequently exposed to 5 % uranyl acetate, 0 2 M - E D T A and lead citrate.The stained samples were examined in a Philips 201C electron microscope.

Chemical determination and electrophoresisProteins were determined by the method of Lowry, Rosebrough, Farr & Randall (1951) and

analysed in a polyacrylamide gradient gel as described by O'Fan-ell (1975). DNA was determinedby the method of Burton (1956) and RNA by the method of Munro & Fleck (1969). The nuclei andthe nuclear matrix spheres were counted in a Biirker-Turk counting chamber.

RESULTS

The in situ aspect of ultras true tural changes in nuclei from scalded ratsAs compared to the control (Fig. 1A), the nuclei from scalded rats appeared ir-

regular, and enriched in pores and enlarged nucleoli (Fig. 1B). Condensed material

L. Sevaljevic and others

1A

\^^m^.

Fig. 1. Hepatoeyte nucleus from control rat (A) and animal 24 h after scalding (B). Notethe invaginations (large arrowheads), enlarged nucleoli (nu) and numerous pores (smallarrowheads), A, X9632; B, X9600.

Fig. 2. Nuclei from scalded rats. Perichromatin granules (pg) were observed at the bor-ders of the condensed chromatin in the peripheral as well as the internuclear regions.X27 300.

Liver cell nuclear organization in rats 79

was preferentially clustered in the peripheral and perinucleolar regions (Fig. 3A).After staining for RNP, the bleached areas of the DNA-containing structures in theperipheral and peri- and intranucleolar regions were clearly delineated from the sur-rounding densely staining border of RNP material (Fig. 3B). The latter was alsodispersed throughout the nucleus as a meshwork composed of more or less denselypacked fibrils and clusters of interchromatin granules (Fig. 3B). The peripheral andinternuclear regions from scalded rats were also enriched in perichromatin granules(Fig. 2). Since the boundary of the condensed chromatin is presumed to be the regionof active chromatin, while the perichromatin granules and fibrils are the primaryproducts of RNA synthesis (Fakan, 1980; Puvion & Viron, 1981), the enrichment ofnuclei from scalded rats with RNP material and with enlarged nucleoli, was inter-preted as a thermal injury-induced increase in the RNA synthetic activity of liver cells.This is consistent with the reported increase in the synthetic activity of liver cells andwith the increase in amount of nuclear RNA in liver cells from scalded rats (Sevaljevic"et al. 1983). The data are also in general agreement with the finding of an increase inperichromatin fibrils and granules in regenerating rat liver (Derenzini, Novello &Pession-Brizzi, 1978) and in heat-shocked H.Ep-2 cells (Cervera & Montero, 1980),probably reflecting the accumulation of newly synthesized RNPs.

The scalding-related changes in ultras tructure and gross composition of isolated nuclei

Nuclei stabilized by Mg2+. The ultrastructure of these nuclei (Fig. 4A) resembledthat of nuclei in situ (Fig. 1A); however, the heterochromatin clusters appeareddenser and the RNP fibrils more clearly delineated in the condensed material. Ascompared to the control, the nuclei from scalded rats (Fig. 4B) contained two or threeenlarged nucleoli more loosely associated with the perinucleolar heterochromatin.The RNP fibrils appeared more compact and enriched with interchromatin granules.The observed enrichment of the nucleoplasm with RNP material agreed with thebiochemical data, which showed a twofold increase in the RNA content per nucleusfrom scalded rats (Table 1) and an 80% increase in the rate of incorporation oflabelled orotic acid into nuclear RNA (unpublished data). As reported previously(Sevaljevic" et al. 1983) and confirmed again (Table 1), the amount of DNA pernucleus from scalded rats was almost double, whereas the proteins increased byapproximately 20 %. The mean value and standard deviation for the buoyant positionof the control nuclei on the sucrose/glucose gradient was 1-27 ± 0*015 g/cm3 (Fig. 5).The nuclei from the injured rats resolved into a smaller and a major peak at densitiesof 1-26 and l-30g/cm3, the reproducibility of the buoyant positions being similar tothat of the control nuclei. The heterogeneous appearance of isolated nuclei was alsoobserved on the sucrose gradient, where nuclei involved in the heavier fraction werefound enriched in DNA and [/nef/ry/-3H]thymidine label (Sevaljevic" et al. 1983).

Structural and biochemical characteristics of the salt-resistant residue from theMg2+-stabilized nuclei (Table 1, Fig. 6) were similar to those originally described byBerezney & Coffey (1977). The matrix from the nuclei of scalded rats was similar tothat of the control with respect to the structure and relative proportions of the nuclearmacromolecules retained in the residue (Table 1). It did, however, differ from the

SO L. Sevaljevic and others

control matrix in containing slightly more protein and D N A , and twice as much RNA(Table 1). In accordance with this, the banding densities of the matrices from thescalded rats shifted from 1-26 to l - 28g /cm 3 (Fig. 7). Th i s network was also enriched

S899 •«J

Liver cell nuclear organization in rats 81

Table 1. The effect of the isolation medium on the macromolecular composition ofnuclei and nuclear matrices from control and scalded rats

tIsolationmedium

Nuclei

MgCl2

PolyaminesEDTA-EGTA

Matrices

MgCl2

PolyaminesEDTA-EGTA

NucleiK

Treatment

—

DNase I

—

DNase IICa2 +

Mg 2 +

Ca2+ + Mg2+

Ca2+ + Mg24-+ DNase 11

ControlScalded

ControlScalded

Control

Scalded

Control

ControlScaldedControlControlControlControl

Scalded

Control

f

Protein

34-4 + 0-9944-1 ±1-59

44-0 ±2-0854-3

15-1 ± 0 1 3(44%)

17-810-56(40%)

11-1 ± 0-41(32%)

29-2(66%)44-0

(81%)

18-0(40%)

Spheres (fig X 10"6)A

DNA

13-4 ±0-9425-1 ±0-84

13-2±l-0820-3

0-9210-07(7%)

117 ± 0 1 2(5%)

0-50 + 008(4%)

Gelatinous residueGelatinous residueGelatinous residueGelatinous residueGelatinous residue

9-5(67%)21-6

(100%)

1-2(9%)

36

68

0-

1-

0-

RNA

•0 ± 0-11•0±0-50

•5 ±0-24•4

60±0-01(20%)

30 ±0-26(27%)

60±0-16(20%)

3-6(55%)

4-5(46%)

0-5(8%)

The values are means ± S.E. calculated from data obtained in several separate experiments. Thevalues lacking standard errors refer to two determinations. The figures in parenthesis are themacromolecules retained relative to the nuclei (100%).

Fig. 3. A. Condensed material (arrowhead) was preferentially clustered close to the innermembrane of the nucleus and at the periphery of the nucleoli (arrows). Fibrillar if) andgranular (gr) components of the nucleolus can be clearly distinguished, B. Portion of ahepatocyte nucleus from scalded rats after EDTA staining. Bleached areas of DNA-containing material in peripheral and perinucleolar (arrowhead) and internucleolar(arrow) regions were clearly delineated from other heavily staining structures. Most of thechromatin clumps were surrounded by perichromatin fibrils (pf). In internuclear regionsclusters of interchromatin granules (ig) were visible, A, X18225; B, X18225.

Fig. 4. Mg^-stabilized nucleus from control (A) and scalded (B) rats, nu, nucleolus; he,heterochromatin; ig, interchromatin granules;/, interchromatin fibrils, A, X9600; B,x 13 140.

L. Sevaljevic and others

c

8CM

0-9

0-8

0-7

0-6

0-5

c.2 0-4o

0-3

0-2

0-1

1-20 1-30p(g/cm3)

Fig. 5

Fig. 6

Liver cell nuclear organization in rats 83

0-8

0-7-

0-6

Ec8 o&CM

c 0-4CD

•s

1 0-3

0-2

0-1

1-20 1-30p(g/cm3)

Fig. 7. Density profiles of matrices from Mg2+-stabilized nuclei in a sucrose/glucosegradient. Control ( • • ) and scalded ( • • ) rats.

in 30-40 X \$M, polypeptides (Fig. 8A), which correspond to the proteins ofheterogeneous nuclear RNP (hnRNP) particles (Peters & Comings, 1980). As shownin Table 2, the scalding-induced activation of nuclei was accompanied by a moreintense labelling of nuclear proteins with [3SS]methionine, the increase being muchhigher for the matrix than for other nuclear proteins. Fig. 8B shows that several, butin particular the 50, 66 and 180 X KfiMr, polypeptides accounted for the enhancedradioactivity of the isolated matrix. The revealed enrichment of the nuclear matrixwith labelled proteins was correlated with a decrease in radioactivity of the crudenucleosomal fraction (Table 2). This finding suggests that activation of nuclei afterscalding was accompanied by a transfer of some nuclear proteins from the DNA to thenuclear matrix.

Nuclei stabilized by polyamines

In these experiments nuclei were isolated in the presence of spermine, spermidine,EDTA and EGTA as described by Burgoyne et al. (1974). Under these conditions

Fig. S. Density profiles of the Mg2+-stabilized nuclei from control ( • • ) and scalded( • • ) rats in a sucrose/glucose gradient.

Fig. 6. Nuclear matrix isolated from the Mg2+-stabilized nuclei of a control rat. Nosignificant differences were observed between these matrices and those isolated fromnuclei of scalded rats. X24000.

m L. Sevaljevic and others

X10- 3

-150 Mr

66

50

-" if

-33-29

*»-21--16

41-14-

5

3

8A1 S- 1

Fig. 8. Coomassie Blue-staining patterns (A) and fluorographic profiles (B) of nuclearmatrix proteins from control (lanes 1) and scalded (lanes 2) rats. Protein samples (250 fig)were analysed in a polyacrylamide gel as described by O'Farrell (1975). The pattern ofproteins that incorporated [35S]methionine was revealed according to the procedure ofBonner & Laskey (1974). S, marker proteins (MrX\0~3): bovine serum albumin (67),ovalbumin (43), soybean trypsin inhibitor, Worthington (21-5) and lysozyme egg white,Worthington (14-3).

the activity of endogenous nucleases is inhibited, whereas in the Mg2+-stabilizednuclei they are stimulated.

The control nuclei stabilized by polyamines retained 20 % more protein and twiceas much RNA than the control Mg2"1"-stabilized nuclei, whereas the DNA contentswere similar for both nuclear preparations (Table 1). In experiments where [14C]sper-midine was added to buffer A, the recovery of the label was twice as high in the controlnuclei as in the nuclei from scalded rats (not shown).

At the ultrastructural level, the polyamine-stabilized nuclei (Fig. 9) differed fromthose stabilized by Mg2"1" (Fig. 4) by having a greater extent of chromatin condensa-tion. They banded at a higher density (1-31 g/cm3) in a sucrose/glucose gradient

Liver cell nuclear organization in rats SS

Table 2. The effecct of scalding on the incorporation of PsS]methionine into proteinsof rat liver nuclei

Protein (c.p.m.//^g)

Control Scalded

NucleiMatrix

Micrococcal nuclease-treated nuclei:

1 mM-EDTA extract(crude nucleosomes)

Protein/DNA inEDTA extract

Matrix

10091

T3Z

(49)

45

145214

101

(29)

67

At 90min before killing, the animal received an intraperitoneal injection of 0-3 mCi of [35S]-methionine (sp. act. 1195 and 1223 Ci/mmol, Amersham). The isolated nuclei were digested for15 s at 37 °C with 300 units of micrococcal nuclease (Worthington)/108 nuclei per ml. The crudenucleosomal fraction was prepared by extraction of nuclei with 1 miu-EDTA (pH7), containing1 mM-PMSF, and recentrifugation of the extracted material at 10 000^. The nucleosome-depletednuclei were further processed for the isolation of the nuclear matrix according to the method usedfor intact nuclei (see Materials and Methods). The data are averages from three to five experiments.

Fig. 9. Polyamine-stabilized nucleus from control (A) and scalded (B) rats. These nucleiappeared on the whole denser than the Mg2+-stabilized nuclei (Fig. 4). As compared tothe controls, the nuclei from scalded rats seemed more disperse and the nucleolus enlargedand impoverished, with perinucleolar heterochromatin (arrows), A, X9600; B, X 13 240.

86 L. Sevaljevic and others

(Fig. 10) than did the Mg2"1"-stabilized nuclei (Fig. 5), which was probably aconsequence of both the enrichment of nuclei in protein and RNA and the increasedextent of condensation of the nuclear material. The effect of scalding on the nuclearultrastructure was similar but more pronounced than that observed in the Mg2"1"-stabilized nuclei (Figs 9B, 4B). This was revealed as an increased dispersion of thechromatin and an enrichment of the nucleoplasm with granulated material (Fig. 9B).The nucleoli were enlarged, more reticulated and discontinuously surrounded by alayer of loosely attached perinucleolar heterochromatin. The banding density of theactivated nuclei was also shifted from 1-31 to 1-27g/cm3 (Fig. 10).

When subjected to the same sequence of salt extractions, the Mg2"1"- and polyamine-stabilized nuclei released a very different set of proteins. Fig. 11 shows that the lowsalt extract (LM) of the Mg2+-stabilized nuclei (lane 1) contained primarily histones,whereas the one from the polyamine-stabilized nuclei (lane 4) was completelydepleted of histones but enriched in 30-40 X l&Mt, 60-70 X 103Mr and 100 X103Mr polypeptides. Although the high salt extracts (HS) of both nuclear prepara-tions contained large amounts of histones (lanes 2 and 5) the morphology (Fig. 12)and the composition (Fig. 11, lane 6) of the salt-resistant residue from the

1-20 1-30p(g/cm3)

Fig. 10. Density profiles of polyamine-stabilized nuclei on a sucrose/glucose gradient.Control nuclei before ( ^ ^ ) and after incubation with 1 mM-CaCl2, 10mM-MgCl2( • • ) . The density of the nuclei from scalded rats ( • # ) did not change aftersuch treatment.

Liver cell nuclear organization in rats W

polyamine-stabilized nuclei was essentially different from that of the well-knownmatrix structure isolated from Mg2+-stabilized nuclei (Fig. 6; Fig. 11, lane 3). Insteadof matrix spheres, an amorphous, lamina-free network was obtained (Fig. 12). Thisresidue was greatly enriched in histone proteins and depleted of the major constituentsof the spherical matrix, the 60-70x10^MT polypeptides (Fig. 11, lanes 6 and 3).

In order to obtain the spherical shape of the salt-resistant residue, polyamine-stabilized nuclei were preincubated either with DNase II or calcium and/or mag-nesium ions and then subjected to a sequence of salt extractions. As shown in Table

1LM

2HS

3 4 5M LM HS

Mg2+

Ca2+, Mg 2 +

PolyaminesEDTA-EGTA

Fig. 11. Electrophoretic patterns of proteins extracted by low (LM) and high (HS) saltsolutions, and of salt-resistant proteins (M) from Mg2+-stabilized nuclei (lanes 1-3) andpolyamine-8tabilized nuclei (lanes 4—10) before (lanes 4—6) and after (lanes 7—10)pretreatment with 1 mM-CaCk and 10 mM-MgCk. Lane 10, matrix proteins from DNaseI-treated nuclei.

L. Sevaljevic and others

12 ' A - *

Fig. 12. The ultrastructure of salt-resistant residue from polyamine-stabilized nuclei,x 14 000.

1, the pretreatment with DNase II was without any effect, whereas neither of the twometal ions exerted an effect until they were added together in the incubation buffer.The salt-resistant residue isolated under these conditions was spherical (Fig. 13A) butretained twice as much protein and RNA and a greater amount of DNA than thematrices from the Mg2"1"-stabilized nuclei (Table 1). The electrophoretic patterns inFig. 11 (lanes 7-10) show that the pretreatment of nuclei with Ca2+ and Mg2*rendered most of the nuclear non-histone proteins salt-insoluble, which enabled analmost quantitative recovery of histones in the salt extracts (lanes 7 and 8) and non-histone proteins in the salt-resistant residue (lanes 9 and 10). These findings stronglysuggest that the spherical residue can be isolated only if several nuclear non-histoneproteins, in particular the 60— 70 X103 Mr polypeptides, occur in a salt-insoluble state.Under such conditions, the residual structure was surrounded by a layer that resem-bles the residual lamina (Fig. 13A). Under the assumption that the 60—70xl03Mr

proteins, which correspond in size to the lamins (Kaufmann et al. 1981), are indeedlamina constituents, it might be speculated that the removal of these proteins duringsalt extractions of polyamine-stabilized nuclei prevented the formation of a salt-resistant lamina residue and, hence, the formation of a spherical residual nuclearstructure.

The pretreatment of nuclei with metal ions did not introduce significant changesin their ultrastnictural features, except that the chromatin appeared more condensed(not shown). However, the banding density of control nuclei on the sucrose/glucose

Liver cell nuclear organization in rats

Fig. 13. The salt-resistant residue of polyamine-stabilized nuclei after pretreatment ofnuclei with 1 mM-CaC^ and 10mM-MgCl2^ A. The residue from the controls; B, theresidue from nuclei of scalded rats. Note the more disperse texture of the matrix from theactivated nuclei but also a more densely staining texture of the controls as compared tothose isolated from the Mgz+-stabilized nuclei (Fig. 6). A, X13 140; B, X9600.

gradient was shifted from 1-31 to 1-26 g/cm3, whereas that of activated nuclei remainedunchanged (Fig. 10). Matrices from the metal ion-pretreated nuclei of scalded ratsretained virtually all the nuclear DNA, this amount being 30% higher than thatrecovered from matrices of control nuclei (Table 1). This difference strongly suggeststhat the organization of at least one third of the nuclear DNA was different in the controland activated nuclei. Under the electron microscope the matrix from activated nuclei(Fig. 13B) appeared more disperse than the matrix from the control nuclei (Fig. 13A).This was consistent with the observed lower density of the matrices from nuclei of scal-ded rats as compared to that of the matrices from the controls (Fig. 14).

DISCUSSION

In response to a thermal injury rat liver cell nuclei underwent changes, whichresulted, in situ as well as in isolated nuclei, in the appearance of invaginations of thenuclear membrane, an enlargement of the nucleoli and an accumulation of RNPmaterial in the nucleoplasm. Nuclei were isolated in two different media containingeither Mg ions or polyamines as stabilizing agents. While both ionic environmentsenabled the observation of the scalding-induced changes, they profoundly affected theultrastructural and biochemical characteristics of the isolated nuclei.

90 L. Sevaljevic and others

1 20 1-30p(g/cm3)

Fig. 14. Density profiles of the matrix spheres isolated from polyamine-stabilized nucleipretreated with 1 mM-CaCl2 and 10mM-MgCl2. Control nuclei ( • • ) ; nuclei fromscalded rats ( # • ) .

The procedure employing polyamines as stabilizing agents and metal chelators wasoriginally developed by Burgoyne et al. (1974) and used for the characterization of aCa-Mg endonuclease(s) in liver cells (Burgoyne et al. 1974; Hewish & Burgoyne,1973a,b). Since the presence of Mg ions in the nuclear isolation buffer was found topromote the rapid accumulation of nicks in the DNA (Cook & Brazell, 1975), thepolyamine-stabilized nuclei had an advantage in those studies that required a completeinactivation of endogenous deoxyribonuclease activity (Hewish & Burgoyne, 19736;Horz&Zachau, 1980; Burgoyne & Skinner, 1981). Our results showed that the extentof chromatin condensation was lowest in situ and greater in polyamine- than in Mg-stabilized nuclei. Along with this, a larger amount of protein and RNA was retainedin the polyamine-stabilized nuclei. The condensation of chromatin in the presence ofcations could be explained by the fact that chromatin is maintained in a dispersed stateby a repulsion between those phosphodiester groups of DNA that are not alreadybound to basic proteins, and that neutralization of these groups by bound cationscaused condensation of chromatin (Leake, Trench & Barry, 1972). The greaterefficiency of polyamines in condensation of chromatin is in agreement with the findingthat the concentration at which contraction of the chromatin was half-complete wassignificantly lower for spermine than for magnesium cations (Leake et al. 1972). Itmight also be related to the hypothesis that the polyamine-DNA complex is stabilized

Liver cell nuclear organization in rats 91

not only by electrostatic interactions between the phosphate and cationic groups butalso by hydrophobic bonds between the non-polar regions of DNA and polyamines(Liquory et al. 1967).

The differences between the two nuclear preparations became particularly evidentwhen nuclei were subjected to sequential salt extractions. Nuclear matrix spheressimilar to those isolated and characterized by Berezney & Coffey (1977) were isolatedfrom the Mg-stabilized nuclei only, whereas the same treatment of the polyamine-stabilized nuclei resulted in the appearance of an amorphous residue, which wasenriched in histones and depleted of the major constituents of the spherical matrix,the 60—70xl03Mr proteins. Since the incubation of nuclei with Ca and Mg ionsrendered the non-histone proteins salt-insoluble and the residue spherical, it is likelythat the removal of nuclear bivalent metals with metal chelators (rather than thebinding of polyamines with nucleic acids) caused the histones to resist and the non-histone proteins to dissolve in low salt solutions. This implies that bivalent metalswere necessary for the organization of the 60-70 X103 Mr proteins into a salt-insolublelamina, which was an apparent prerequisite for the isolation of the matrix spheres.Different from the protein, the DNA in the polyamine-stabilized nuclei was sig-nificantly more resistant to salt extractions than the DNA in nuclei stabilized by Mgions. Thus, the bulk of the DNA from the polyamine-stabilized nuclei was recoveredin the amorphous salt-resistant pellet as well as in the matrix spheres obtained whenthe salt extractions were preceded by incubation of the nuclei with Ca and Mg ions.These matrices represented dehistonized nuclei rather than a 'minimal residualnuclear structure' since the residue from the scalded rats retained virtually all of thenuclear non-histone proteins and DNA, whereas that of the controls retained slightlyless protein and 70 % of the nuclear DNA. This difference between the DNA contentof the salt-resistant residue from the control and that from activated nuclei was amongthe most intriguing results of this work. It indicated that the activation of nuclei anda concomitant increase in ploidy (Sevaljevic" et al. 1983) was accompanied by the'strengthening' of the association between the salt-insoluble non-histone proteins andDNA; at least that part of the DNA which, in the control nuclei, occurred in a salt-insoluble state. Activation-related redistribution of proteins also resulted in the Mg-stabilized nuclei in an enrichment of the matrix and a depletion of the nucleosomalfractions with 35S-labelled nuclear proteins. An understanding of the interactions thatconfer distinct properties on certain protein and DNA fractions might contributegreatly to the elucidation of the mechanisms regulating gene expression.

REFERENCES

ARTURSON, G. (1961). Pathophysiological aspects of the burn syndrome. Ada chir. scand. 274(suppl.) 12.

BEREZNEY, R. (1979). Effect of protease inhibitors on matrix proteins and the association of replicat-ing DNA. Expl Cell Res. 123, 411-414.

BEREZNEY, R. & COFFEY, D. (1977). Nuclear matrix. Isolation and characterization of a frameworkstructure from rat liver nuclei. J. Cell Biol. 73, 616—637.

BONNER, W. & LASKEY, R. (1974). A film detection method for tritium-labelled proteins andnucleic acids in polyacrylamide gels. Eur.J. Biochem. 46, 83-88.

92 L. Sevaljevic and others

BRASCH, K. (1982). Fine structure and Iocali2ation of the nuclear matrixinsitu. ExplCellRes. 140,161-171.

BURGOYNE, L., HEWISH, D. & MOBBS, J. (1974). Mammalian chromatin substructure studies withthe calcium-magnesium endonuclease and two-dimensional polyacrylamide-gel electrophoresis.Biochem.J. 143, 67-72.

BURGOYNE, L. & SKINNER, J. (1981). Chromatin superstructure: the next level of structure abovethe nucleosome has an alternating character. Biochem. biophys. Res. Commun. 99, 893-899.

BURTON, K. (1956). A study of the condition and mechanism of the diphenylamine reaction for thecolorimetric estimation of deoxyribonucleic acid. Biochem. J'. 62, 315-322.

CERVERA, J. & MONTERO, M. (1980). Effect of thermic shock on H.Ep-2 cells. III . Accumulationof perichromatin granules. J. Ultrastruct. Res. 71, 1-13.

COOK, P. & BRAZELL, I. (1975). Supercoils in human DNA.J . Cell Sri. 19, 261-279.DERENZINI, M., NOVELLO, F. & PESSION-BRIZZI, A. (1978). Perichromatin fibrils and chromatin

ultrastructural pattern. Expl Cell Res. 112, 443-454.FAKAN, S. (1980). Ultrastructural visualization of transcription at the cellular and molecular level.

Biol. Cell 39, 113-116.HERLAN, G., QUEVEDO, R. & WUNDERLICH, F. (1978). Structural transformation of the nuclear

matrix in situ. Expl CellRes. 115, 103-110.HEWISH, D. & BURGOYNE, L. (1973a). The calcium dependent endonuclease activity of isolated

nuclear preparations. Relationships between its occurence and the occurence of other classes ofenzymes found in nuclear preparations. Biochem. biophys. Res. Commun. 52, 475-481.

HEWISH, D. & BURGOYNE, A. (19736). Chromatin sub-structure. The digestion of chromatin DNAat regularly spaced sites by a nuclease. Biochem. biophys. Res. Commun. 52, 504—510.

HORZ, W. & ZACHAU, H. (1980). Deoxyribonuclease II as a probe for chromatin structure. I.Location of cleavage sites. J. molec. Biol. 144, 305-327.

KAUFMANN, S., COFFEY, D. & SHAPER, J. (1981). Considerations in the isolation of rat livernuclear matrix, nuclear envelope and pore complex lamina. Expl Cell Res. 132, 105-121.

LEAKE, R., TRENCH, M. & BARRY, J. (1972). Effect of cations on the condensation of hen erythro-cyte nuclei and its relation to gene activation. Expl CellRes. 71, 17-26.

LEBKOWSKI, J. & LAEMMLI, U. (1982a). Evidence for two levels of DNA folding in histone-depleted HeLa interphase nuclei. J . molec. Biol. 156, 309-324.

LEBKOWSKI, J. & LAEMMLI, U. (19826). Non-histone proteins and long-range organization ofHeLa interphase DNA. J . molec. Biol. 156, 325-344.

LIQUORY, A., CONSTANTINO, L., CRESCENZI, V., ELIA, V., GIGLIO, E., PULITI, R., D E SANTIS

SAVINO, M. & VITAGLIANO, V. (1967). Complexes between DNA and polyamines: A molecularmodel. J. molec. Biol. 24, 113-122.

LOWRY, O. H., ROSEBROUGH, W. J., FARR, A. L. & RANDALL, R. J. (1951). Protein measurementswith the Folin reagent. J. biol. Chem. 193, 265-275.

MUNRO, N. H. & FLECK, A. (1969). The determination of nucleic acids. In Methods in BiochemicalAnalysis (ed. Glick), vol. 14, p. 159. New York: Wiley.

O'FARRELL, H. P. (1975). High resolution two-dimensional electrophoresis of proteins. J. biol.Chem. 250, 4007-4021.

PETERS, K. & COMINGS, D. E. (1980). Two-dimensional gel electrophoresis of rat liver nuclearwashes, nuclear matrix and hnRNA proteins. J. Cell Biol. 86, 135-155.

PUVION, E. & VIRON, A. (1981). In situ structural and functional relationships between chromatinpattern and RNP structures involved in non-nucleolar chromatin transcription. J. Ultrastruct.Res. 74, 351-360.

RAYNAUD, A. & OHLENBUSH, H. (1972). Buoyant density of native chromatin. J. molec. Biol. 63,523-537.

RILEY, D., KELLER, J. & BEYER, B. (1975). The isolation and characterization of nuclear ghostsfrom cultures HeLa cells. Biochemistry 14, 3005-3013.

SMALL, D., NELKIN, B. & VOGELSTEIN, B. (1982). Nonrandom distribution of repeated DNAsequences with respect to supercoiled loops and the nuclear matrix. Proc. natn. Acad. Sri. U.SA.79, 5911-5915.

SEVALJEVIC, L J . , PANTELIC, D., STOJANOVIC, R., PETROVIC, M. & RADOJCIC, C. (1983). Ther-mal trauma-induced changes in the synthesis of rat serum proteins. Comp. Biochem. Physiol. 76B,227-233.

Liver cell nuclear organization in rats 93

SEVALJEVIC, LJ . , PETROVIC, M. & PANTELIC, D. (19836). Thermal injury response of rat livernuclei. Int.J. Biochem. 15, 225-231.

SEVALJEVIC, LJ . , PETROVIC, M., POZNANOVI6, G. & KONSTANTINOVIC, M. (1981). On the

similarity between the nuclear network and chromatin nonhistone proteins of sea urchin embryos.Cell, molec. Biol. 27, 147-157.

SEVALJEVIC, Lj., PETROVKJ, M., SAVIC, J. & PANTELIC, D. (1982). The effects of repeated thermalinjury on rat liver and serum proteins synthesis rates. Comp. Biochem. Physiol. 73B, 379-384.

(Received 5 March 1984 -Accepted 25 April 1984)