Nuclear matrices, associated with over 80% of the chromosomal DNA, could be isolated from BHK nuclei by extraction with 2M-NaCl. The matrices were found to impose at least two levels of structural order upon nuclear DNA.

From sedimentation studies it was inferred that metal depletion of the salt-extracted nuclei generated matrix structures, which sedimented at significantly lower rates than control matrices. Fluorescence microscopy revealed that the reduced sedimentation rate is a consequence of the increase in the radius of the DNA halo, i.e. the DNA loops emanating from the residual nucleus. Addition of Cu ions to nuclei prior to salt extraction was found to induce contraction of this DNA halo. These results indicate that Cu ions may play an important role in stabilizing one level of DNA folding.

When metal depletion had been brought about by thiol agents, a second effect was observed to occur. Within 15 min, salt-extracted nuclei disintegrated, generating irregularly shaped, slowly sedimenting structures. Disintegration only occurred when the full complement of DNA was still attached to the nuclear matrices.

Analysis by sodium dodecyl sulphate-polyacrylamide gel electrophoresis revealed that treatment with thiols did not detectably alter the polypeptide composition of DNA-depleted residual nuclei.

Results of these experiments suggest that both metal-protein interactions and disulphide bonds are important in maintaining higher-order structure in the nucleus. A model to account for these observations is discussed.

In both interphase nuclei and metaphase chromosomes higher-order structure of chromatin is mediated by non-histone chromosomal proteins. For the packing of the chromatin fibre in chromosomes a ‘radial loop’ model has been proposed (Laemmli et al. 1978; Marsden & Laemmli, 1979), i.e. chromatin fibres are attached to an axial proteinaceous scaffolding in a loop-wise fashion. Comparable structural order is thought to exist in interphase, the framework being provided by the nuclear matrix (Berezney & Coffey, 1974; Cook & Brazell, 1976; Wanka et al. 1977).

Since the nuclear matrix has been shown to be involved in a variety of nuclear functions, such as DNA replication (Berezney & Coffey, 1975; Dukwel et al. 1979; Pardoll et al. 1980), RNA transcription and processing (Jackson et al. 1981;Mariman et al. 1982), hormone-receptor binding (Barrack & Coffey, 1980) and binding of carcinogens (Hemminki & Vainio, 1979; Tew et al. 1980), considerable effort has been invested in order to elucidate nuclear matrix structure and establish its protein composition. In this approach, the nuclear matrix has been operationally defined as the residual structure obtained after treatment of nuclei with non-ionic detergent and extraction with high concentrations of NaCl. Generally, the residual structure consists of approximately 10% of total nuclear protein (Wankaet al. 1977; Lebkowski & Laemmli, 1982a), of which the lamins are the most prominent class (Mullenders et al. 1982; Kaufmann et al. 1981; Lebkowski & Laemmli, 1982b), and varying amounts of chromosomal DNA, depending on whether or not nuclease treatment is included in the isolation procedure.

Though some insight into the three-dimensional architecture of the nuclear matrix has been provided (Brasch, 1982; Capco & Penman, 1983), the precise way in which matrix proteins are ordered to form a nuclear scaffolding remains to be established. This greatly complicates the attribution of functions to specific nuclear constituents. A novel attempt to establish the positioning of subnuclear structures, to which DNA is attached, was recently reported (Lebkowski & Laemmli, 1982a,b). It was reported that bivalent cations, notably Cu2+, are required for one level of DNA organization in the nucleus. Removal of this ion resulted in expansion of the DNA halo and loss of several proteins (Lebkowski & Laemmli, 1982b). The effect could be observed both when metal chelators such as orthophenantroline, and thiol agents such as mercaptoethanol and dithiothreitol, were added. However, a study reported previously (Cook et al. 1976), in which HeLa nucleoids were found to be unstable in dithiothreitol-containing media, seems to question the results of the study mentioned earlier as far as the thiol effect is concerned. The aim of the present study was to evaluate the relevance of the data presented to date concerning the effects of thiol agents on the nuclear matrix.

Cell culture and preparation of nuclear matrices

BHK-A3 cells (Cupido & Simons, 1984), obtained from the Department of Radiation Genetics and Chemical Mutagenesis at Leiden University, were maintained in monolayer culture in Minimal Essential Medium (MEM; Flow Labs) supplemented with 8% (v/v) foetal calf serum (Gibco).

Cells were labelled for three to five generations with 1μCiml−1 L-[4,5-3H]leucine (sp. act. 46 Cimmol−1; Amersham) and 0·02μCiml−1 [2-14C]thymidine (sp. act. 52·8 Ci mol−1; Amersham).

Nuclei were isolated essentially as described (Wanka et al. 1977). Monolayers were rinsed sequentially with 0·9% NaCl in 5 mw-Tris HC1 (pH 8·0) and with 0’ 1% Triton X-100 in 5 mM-Tris’HCl (pH 8’0) (TT buffer). Cells were then washed from the glass surface with TT buffer. Nuclei were isolated by forcing the suspension twice through a hypodermic needle (0·8 mm diameter) and centrifuging the suspension subsequently for 3 min at 800 g. The nuclear pellet was resuspended in 50mM-Tris-HCl (pH8·0) and checked for contamination by phase-contract microscopy. Finally, an equal volume of 4M-NaCl in 50mM-Tris’HC1 (pH 8·0) was added to the nuclear suspension.

All steps were performed on ice.

Centrifugation and determination of radioactivity

After appropriate treatment with thiol agents or metal chelators, 5-ml samples of nuclear lysate in 2M-NaCl were carefully layered on 5% to 25% sucrose gradients containing 2M-NaCl and 50mM-Tris-HCl (pH8·0). The gradients were prepared on 65% sucrose shelves containing 0·4gml−1 CsCl. Centrifugation was performed at 4°C in Beckman SW 2711 rotors at 5000 rev. min−1 for different periods of time.

After centrifugation, fractions were collected starting from the bottom of the tube. Trichloroacetic acid was added and the acid-insoluble material was pelleted by centrifugation, washed once with 70% ethanol and dried. Subsequently, the pellet was allowed to dissolve in 0·2M-NaOH at 80°C. After 30 min an equal volume of 1 M-perchloric acid was added and hydrolysis was allowed to proceed for another 30 min at 80°C. After addition of scintillation mix (Packard) the radioactivity in the samples was determined in a Philips LSA.

Fluorescence microscopy

From the nuclear lysates prepared for sedimentation analysis samples were taken to which 4μgml−1 ethidium bromide was added. The sample was then viewed with a Zeiss Photomikroskop III, using a 50 W HBO-mercury lamp. Photographs were taken with high-speed films (Kodak VR 1000) to minimize illumination time.

Polyacrylamide gel electrophoresis

For analysis of the polypeptide composition of nuclear matrices, nuclei were DNA-depleted either by treatment with DNase I in the presence of 1 mM-MgC12 or by digestion with staphylococcal nuclease in the presence of 0·1mM-CaC12-The nuclei were then collected, suspended in 50mM-Tris-HC1 (pH 8·0) and extracted with 2M-NaCl. Protein samples were prepared as described (Pieck et al. 1985). Protein samples were dissolved in sample buffer (Laemmli, 1970) supplemented with 6M-urea and analysed on 4% to 18% polyacrylamide slab gels according to Laemmli (1970).

When the effect of thiols and the polypeptide composition of nuclear matrices was analysed, all solutions used were supplemented with either β-mercaptoethanol or dithiothreitol in the appropriate concentrations. Gels were stained with Coomassie Blue.

Effects of thiol agents

As from a variety of eukaryotic cells, matrices could be prepared from nuclei of BHK cells by extraction with 2M-NaCl. Fig. 1A shows the sedimentation profile on neutral sucrose gradients of salt-extracted nuclei, of which DNA and protein were labelled with [14C]thymidine and [3H]leucine, respectively. Approximately 80% of the DNA and 20% of the nuclear proteins are found to sediment to the lower half of the gradient heterogeneously, as indicated by the broad peak. As expected for different labels residing in the same structure, the distribution of the protein label closely follows that of the DNA label. Residual nuclei were observed to sediment with average values between 6000 and 9000 S. From the top of the gradients, most of the protein and a small amount of DNA were invariably recovered. Both fractions were unable to enter the gradient under the conditions chosen.

Fig. 1.

Disintegration of residual nuclei in high-salt buffer containing dithiothreitol. Cells were continuously labelled with [3H]leucine and [14C]thymidine. Nuclei were isolated and extracted with 2M-NaCl. A. Sedimentation of salt-extracted nuclei on a linear 5% to 35% sucrose gradient; centrifugation for 60 min at 5000 rev. min−1 at 4°C. (○ —○) [3H]leucine, 21000disintsmin−1; (• — •) [14C]thymidine, 57700disints min−1. B. Sedimentation of salt-extracted nuclei prepared in the presence of 10 mM-DTT (as for A). (○ — ○) [3H]leucine, 21000disintsmin−1; (•-— •) [I4C]thymidine, 60300 disintsmin−1. Fluorescence visualization of residual nuclei, stained with 4μg ml−1 ethidium bromide in: C. 2M-NaCl; D, 2M-NaCl, immediately after addition of lOmM-DTT; E, 2M-NaCl, 20 min after addition of 10 mM-DTT.

Fig. 1.

Disintegration of residual nuclei in high-salt buffer containing dithiothreitol. Cells were continuously labelled with [3H]leucine and [14C]thymidine. Nuclei were isolated and extracted with 2M-NaCl. A. Sedimentation of salt-extracted nuclei on a linear 5% to 35% sucrose gradient; centrifugation for 60 min at 5000 rev. min−1 at 4°C. (○ —○) [3H]leucine, 21000disintsmin−1; (• — •) [14C]thymidine, 57700disints min−1. B. Sedimentation of salt-extracted nuclei prepared in the presence of 10 mM-DTT (as for A). (○ — ○) [3H]leucine, 21000disintsmin−1; (•-— •) [I4C]thymidine, 60300 disintsmin−1. Fluorescence visualization of residual nuclei, stained with 4μg ml−1 ethidium bromide in: C. 2M-NaCl; D, 2M-NaCl, immediately after addition of lOmM-DTT; E, 2M-NaCl, 20 min after addition of 10 mM-DTT.

Previous studies have shown that inclusion of β-mercaptoethanol and dithiothreitol (DTT) in the isolation procedure resulted in a decrease in the sedimentation rate of residual nuclei in sucrose gradients. Fluorescence microscopy revealed an increase of the average radius of the DNA halo by a factor of 2. Most probably, the DNA-unfolding occurred as a consequence of the collapse of internal matrix structure as a result of chelation of either copper or calcium ions (Lebkowski & Laemmli, 1982a).

We also observed that addition of thiol agents, in concentrations ranging from 0·5M to 50mM, to nuclei suspended in 2M-NaCl, induced significant changes. As shown in Fig. IB, addition of DTT destroys the rapidly sedimenting component. Almost all label applied to the gradient is recovered at the uppermost fractions. Only a small fraction, approximately 15% of the DNA cosedimenting with 5% of the protein, was found to have entered the gradient with an s value of 1500. Similar results were obtained when thiol agents had been present in the isolation procedure from the nuclear isolation step onwards. Neither addition of MgC12 in millimolar amounts, nor the inclusion of 0·5 mM-phenylmethylsulphonyl fluoride (PMSF) was found to affect these observations.

To establish more directly to what extent residual nuclei are morphologically altered in thiol-containing high-salt buffers, the samples were also subjected to fluorescence microscopy. Prior to addition of thiols, salt-extracted nuclei were characterized by a uniformly fluorescing residual nucleus surrounded by a less intensely fluorescing DNA halo (Fig. 1C). Addition of 10mM-DTT had two immediate effects. Concomitant with the loss of the uniform appearance of the residual nuclear core, an expansion of the DNA halo was observed (Fig. ID). These structures were found to be unstable. Consequently, it is difficult to estimate the radius of the expanded halo. Nevertheless, our data indicate a 1·5-fold increase in the radius (control nuclei, 12±2μm, n — 11; DTT nuclei, 19±3μm, n — 9). Finally, approximately 15 min after addition of thiol agents to the salt-extracted nuclei, only the structures shown in Fig. IE could be observed. In these structures the nuclear matrix appears to have disintegrated. Consequently, these structures have no characteristic shape. They seem to consist of aggregates of material originating from the former residual nucleus spread out over a background of diffusely fluorescing DNA. Furthermore, as a consequence of their dramatically increased diameters, these structures are expected to sediment at rates considerably lower than those of control salt-extracted nuclei.

Effects of metal chelators and reducing agents

As thiols are bifunctional, i.e. metal chelators as well as reducing agents, we tried to unravel these effects by using EDTA and 1,10-orthophenantroline (OP) on the one hand and NaBH4 on the other.

Addition of 1 mM or 10mM-EDTA to salt-extracted nuclei had no effect on sedimentation characteristics or on morphology. However, OP, tested in concentrations between OT mM and 50 mM, did affect residual nuclei. Compared to control matrices, sedimenting at 9000 S (Fig. 2A), matrices prepared in 1 mM-OP sedimented at a reduced rate of 5000 S (Fig. 2B). The amount of DNA and protein in these structures was not detectably altered by OP.

Fig. 2.

Reversible unfolding of DNA induced by 1,10-orthophenantroline and Cu2+. Residua] nuclei were prepared as described. A. Sedimentation of salt-extracted nuclei on a linear 5% to 25% sucrose gradient; centrifugation for 60min at 5000rev. min−1 and at 4°C. (○ —○) [3H]leucine, 53200 disintsmin−1; (• —•) [14C] thymidine, 16300 disintsmin−1. B. Sedimentation of salt-extracted nuclei prepared in the presence of 1 mM-OP (as for A). (○ — ○) [3H]leucine, 67000disintsmin’ (• — •) [14C]thymidine,22500 disints min−1. C. Sedimentation of salt-extracted nuclei prepared in the presence of 1 mM-CaClz (as for A). (○— ○) [3H]leucine, 43 100disintsmin−1; (• — •)[14C]thymidine, 17 800 disints min−1. For fluorescence visualization cells, grown on coverslips, were treated with 0·5% NP-40, 0·0, 0·4, 0·8, 1·2, 1·6 and 2·0M-NaCl in 5 mM-Tris. HC1 (pH 8· 0) containing 5 mM-MgCl2 for 30 s sequentially. Then the residual cells were stained with 4 μg ml−1 ethidium bromide, either in the absence (C) or presence (D) of 1 mM-OP. For analysis of the effect of Cuz+ by fluorescence microscopy, residual nuclei were prepared in the usual way: F, a residual nucleus in 2M-NaCl; G, a residual nucleus, to which 1 mM-CuCl2 had been added prior to extraction.

Fig. 2.

Reversible unfolding of DNA induced by 1,10-orthophenantroline and Cu2+. Residua] nuclei were prepared as described. A. Sedimentation of salt-extracted nuclei on a linear 5% to 25% sucrose gradient; centrifugation for 60min at 5000rev. min−1 and at 4°C. (○ —○) [3H]leucine, 53200 disintsmin−1; (• —•) [14C] thymidine, 16300 disintsmin−1. B. Sedimentation of salt-extracted nuclei prepared in the presence of 1 mM-OP (as for A). (○ — ○) [3H]leucine, 67000disintsmin’ (• — •) [14C]thymidine,22500 disints min−1. C. Sedimentation of salt-extracted nuclei prepared in the presence of 1 mM-CaClz (as for A). (○— ○) [3H]leucine, 43 100disintsmin−1; (• — •)[14C]thymidine, 17 800 disints min−1. For fluorescence visualization cells, grown on coverslips, were treated with 0·5% NP-40, 0·0, 0·4, 0·8, 1·2, 1·6 and 2·0M-NaCl in 5 mM-Tris. HC1 (pH 8· 0) containing 5 mM-MgCl2 for 30 s sequentially. Then the residual cells were stained with 4 μg ml−1 ethidium bromide, either in the absence (C) or presence (D) of 1 mM-OP. For analysis of the effect of Cuz+ by fluorescence microscopy, residual nuclei were prepared in the usual way: F, a residual nucleus in 2M-NaCl; G, a residual nucleus, to which 1 mM-CuCl2 had been added prior to extraction.

To visualize the effect of OP by fluorescence microscopy an alternative method was adopted. Salt-extracted nuclei were prepared, attached to coverslips (Vogelstein et al. 1980). This makes gentle removal of the OP-containing high-salt buffer possible, which was imperative because of the high background fluorescence of OP under these circumstances. Fig. 2E shows that OP-treated residual nuclei are surrounded by a more-extensive halo than control matrices (Fig. 2D). Moreover, and as in the initial stage of thiol-mediated expansion, features within the residual core can be observed. However, structures prepared in OP were stable. From this we conclude that the decrease of the sedimentation rate is a consequence of the expansion of the DNA halo.

Therefore, a role for metal ions in nuclear structure is likely. This conclusion is supported further by results of experiments in which, immediately prior to saltextraction, 1 mM-CuC12 was added to nuclei. Nuclear matrices obtained from these nuclei were found to sediment at higher rates (Fig. 2C) than the corresponding controls (Fig. 2A). It should be noted, however, that the residual structures prepared in the presence of copper ions, contained significantly more protein than the control matrices.

Results of fluorescence microscopy were consistent with the sedimentation data. Fig. 2F shows a residual nucleus prepared as a control. Salt-extraction of nuclei prepared in the presence of 1 mM-CuC12, yielded structures with a significantly contracted halo (Fig. 2G). Moreover, copper ions were found to counteract the effect of OP on residual nuclei (data not shown).

From the data presented to date it can be concluded that in addition to metal chelation, thiol agents have another effect on nuclear matrices. This effect could arise as a consequence of reductive disruption of disulphide bonds, which might stabilize the structure. This possibility was assessed by treating salt-extracted nuclei with NaBH4. However, 1 mM-NaBH4 did not convert salt-extracted nuclei to a slowly sedimenting form. On the contrary, a slight increase in sedimentation rate was observed (data not shown), consistent with data presented by Lebkowski & Laemmli (1982a). As expected, morphological changes were not found to be induced by NaBH4, even when the concentration was raised to 10mM (data not shown).

Since neither metal chelators nor reducing agents, when used separately, can induce the thiol-induced disintegration of the nuclear matrix, it must be concluded that the bifunctional character of thiols is a prerequisite for their action.

The role of DNA in thiol-mediated disintegration

In contrast to nuclear matrices, nuclei appeared to be stable in thiol-containing buffers (data not shown). The most obvious effect of addition of 2M-NaCl to nuclei is removal of histones from the DNA. Consequently, the now naked DNA will adopt a configuration minimizing repulsion. As DNA is anchored to the nuclear matrix at more or less regular intervals, the well-known halo surrounding the proteinaceous matrix will be the endpoint of this process. In this configuration, though minimized, repulsion is not reduced to zero. Consequently, destabilization of the residual nucleus, for instance by the action of the bifunctional thiols, might result in further expansion and, eventually, disintegration.

To assess whether this is the case, salt-extracted nuclei were prepared; some of these were subsequently treated with DNase I to remove most of the DNA attached. Sedimentation analysis revealed that the DNA-rich matrices disintegrate in the presence of DTT (namely, Fig. 1). However, DTT was found not to affect the sedimentation behaviour of DNA-depleted matrices (Fig. 3A,B). Fluorescence microscopy showed that, in contrast to residual nuclei not treated with DNase I, DNA-depleted matrices (Fig. 3C) were stable in DTT-containing high-salt buffer (Fig. 3D).

Fig. 3.

Effect of dithiothreitol on DNA-rich and DNA-depleted residual nuclei. Residual nuclei were prepared as described for Fig. 1. A. Sedimentation of salt-extracted nuclei incubated for 10 min with 50 units ml−1 DNase I prior to salt extraction, on a linear 5% to 25% sucrose gradient. Centrifugation was for 45 min at 5500 rev. min−1, at 4°C. (O O) [3H]leucine, 23 150disintsmin−1; (• — •) [14C]thymidine,11800disintsmin−1. B. As for A; sedimentation in the presence of 10mM-DTT. (○ — ○) [3H]leucine, 25400disintsmin−1; (• — •) [14C]thymidine, 13000disintsmin−1. Fluorescence visualization of a residual nucleus in: C, 2M-NaCl, after removal of DNA with DNase I; D, 2M-NaCl, containing IOmM-DTT, after removal of DNA with DNase I.

Fig. 3.

Effect of dithiothreitol on DNA-rich and DNA-depleted residual nuclei. Residual nuclei were prepared as described for Fig. 1. A. Sedimentation of salt-extracted nuclei incubated for 10 min with 50 units ml−1 DNase I prior to salt extraction, on a linear 5% to 25% sucrose gradient. Centrifugation was for 45 min at 5500 rev. min−1, at 4°C. (O O) [3H]leucine, 23 150disintsmin−1; (• — •) [14C]thymidine,11800disintsmin−1. B. As for A; sedimentation in the presence of 10mM-DTT. (○ — ○) [3H]leucine, 25400disintsmin−1; (• — •) [14C]thymidine, 13000disintsmin−1. Fluorescence visualization of a residual nucleus in: C, 2M-NaCl, after removal of DNA with DNase I; D, 2M-NaCl, containing IOmM-DTT, after removal of DNA with DNase I.

These results indicate that, though thiols have an effect on salt-extracted nuclei, this by itself is not sufficient to destabilize the structures to such an extent that disintegration occurs.

The effect of thiols on the polypeptide composition of residual nuclei

Destabilization of salt-extracted nuclei by thiol agents might be related to dissociation of certain proteins from these structures. Consequently, nuclei were isolated both in the absence and in the presence of DTT. Chromatin was then removed by nuclease digestion after which the nuclei were salt-extracted.

Lane b of Fig. 4 shows the polypeptide composition of nuclear matrices isolated in the absence of thiols. The prominent bands in the 60–70 (× 103)Mr range represent the lamin proteins. Comparison with lane c shows that the presence of DTT during the isolation procedure does not change the protein composition, except for a protein of high molecular weight. The relative amounts of this protein were found to vary considerably from one matrix preparation to the other, indicating it might consist of a complex of matrix proteins. Disulphide bonding most probably generates this complex, as indicated by its greatly reduced abundance in matrices prepared in the presence of DTT.

Fig. 4.

Effect of thiols on the polypeptide composition of DNA-depleted residual nuclei. Nuclei were isolated, digested with nuclease and extracted with 2M-NaCl. Nuclear matrix proteins were subsequently prepared for electrophoresis on 4% to 18% polyacrylamide slab gels. The Coomassie Blue staining patterns are depicted: lanes A,D, marker proteins; lane B, nuclear matrix proteins obtained from matrices isolated in the absence of thiols; lane C, nuclear matrix proteins obtained from matrices isolated in the presence of 10mM-DTT.

Fig. 4.

Effect of thiols on the polypeptide composition of DNA-depleted residual nuclei. Nuclei were isolated, digested with nuclease and extracted with 2M-NaCl. Nuclear matrix proteins were subsequently prepared for electrophoresis on 4% to 18% polyacrylamide slab gels. The Coomassie Blue staining patterns are depicted: lanes A,D, marker proteins; lane B, nuclear matrix proteins obtained from matrices isolated in the absence of thiols; lane C, nuclear matrix proteins obtained from matrices isolated in the presence of 10mM-DTT.

Furthermore, from the banding pattern depicted in lane c it can be inferred that salt-extracted nuclei are essentially free of proteolytic activity which is activated by sulphydryl groups. Consequently, the disintegration observed to occur in thiol-containing media cannot be ascribed to protein degradation by enzymes of this type.

In this paper the nature of the protein-protein interactions stabilizing nuclear matrix structure was probed with thiol agents, metal chelators and reducing agents. Effects of these substances on salt-extracted nuclei were analysed by sedimentation analysis and fluorescence microscopy.

Nuclear matrices were found to sediment rather heterogeneously on sucrose gradients. The close resemblance of the sedimentation profiles of DNA label and protein label indicate that this is a consequence of the heterogeneity of the matrix population. Fluorescence microscopy revealed residual nuclei to be surrounded by the characteristic DNA halo.

Addition of the metal chelator OP to salt-extracted nuclei led to a decrease in the average sedimentation rate of the structures. As neither the relative amount of DNA nor the proportion of protein residing in the residual nuclei was affected by OP, the decrease in the sedimentation rate was considered to be a consequence of DNA unfoldirtg. This assumption was proved to be correct by fluorescence microscopy. Concomitant with DNA decompaction, internal features of the salt-extracted nuclei became visible.

Our results confirm data presented by Lebkowski & Laemmli (1982a), indicating involvement of divalent cations, presumably copper, in long-range order of DNA in matrix structures. Our study also indicates that copper is important for the maintenance of matrix structure. Salt extraction of nuclei, suspended in CuCl2-containing buffers, gave rise to matrices sedimenting at increased rates. These matrices were found to be surrounded by contracted halos. Furthermore, our observations suggest that the heterogeneous sedimentation behaviour of control matrices might be accounted for by the fact that nuclei lose their divalent ions to different extents during isolation.

Though copper ions appear to be involved in maintaining long-range DNA order in nuclei, some care has to be taken in interpretating the results along these lines. Copper-treated residual nuclei were found to contain significantly more protein than control matrices, while the amount of DNA was found unchanged. Increase in mass could therefore contribute to the increase in the sedimentation rate. Moreover, merely as a consequence of twice the regular amount of protein in the matrix, unfolding of DNA might be hampered physically. Consequently, the radius of the DNA halo would be reduced.

Addition of thiol agents to nuclei suspended in 2M-NaCl initially had effects comparable to those of OP. The radius of the DNA halo was found to increase by a factor 1·5. However, in contrast to experiments in which the metal chelator had been used, this situation did not represent an endpoint. In the course of 15 min all residual nuclei expanded greatly losing the shape characteristic of the original nucleus. Since the nuclear lamina, a prominent feature of the nuclear matrix, appeared disrupted in the expanded structures, loss of the original morphological appearance might be a consequence of this disruption.

The lamina as a target for thiols has been made plausible by several studies (Shelton & Cochran, 1978; Cobbs & Shelton, 1978; Lamm & Kasper, 1979), which indicate that nuclear envelope proteins can be reversibly crosslinked. Whether or not internal matrix structure is in some way affected by disruption of disulphide bonds could not be established (Kaufmann et al. 1981).

Since Lebkowski & Laemmli (1982a) did not observe the disintegrating effect of thiol agents on HeLa nuclear matrices, we investigated whether our observations were specific for BHK cells. It was found, however, that salt-extracted nuclei obtained from CHO, bovine liver and HeLa cells all responded to thiols similarly. The disintegration observed to occur in thiol-containing high-salt media might therefore be a universal phenomenon. Our conclusion is supported by an observation of Cook et al. (1976). They reported destruction of HeLa nucleoid integrity in media containing 50mM-DTT.

DNA-depleted residual nuclei were stable in thiol-containing 2M-NaCl. It is therefore concluded that, after metal depletion and disruption of disulphide bonds, nuclear matrix proteins are still able to interact to such an extent that a structure reminiscent of that of the original nucleus is preserved. This is consistent with previous results, which indicate that after removal of DNA, residual nuclei could be obtained from /J-mercaptoethanol-containing high-salt buffers (Kaufmann et al. 1981). However, when forces, such as are generated by the dehistonized DNA complement, are exerted on these structures the matrices fall apart.

Assuming that chromosomal DNA is attached to the nuclear matrix both internally and peripherally, our results suggest that the nuclear lamina might be stabilized by disulphide bonding while divalent cations stabilize the internal matrix. Metal depletion leads to a residual structure, in which the lamina remains intact, but as suggested by Lebkowski & Laemmli (1982a) the internal matrix has collapsed. Consequently, the radius of the DNA halo increases and the nuclear interior acquires an empty appearance. When both metal depletion and disruption of disulphide bonds have been induced, the lamina is also destabilized. Disruption follows subsequently if a considerable amount of DNA is still attached to the residual nucleus. The model further suggests that mere disruption of disulphide bonds by reducing agents will have no observable effect. The lamina might be disrupted, but as the internal matrix is thought to be unaffected, indicated by the absence of a change in the radius of the DNA halo, the residual nucleus will be prevented from disintegrating. Alternatively, our observations might also fit the dimercaptide model that was proposed recently (Jeppesen & Morten, 1985). In agreement with this study it was also found that for disintegration to occur, thiol-mediated elimination of specific proteins from the residual nuclei does not seem to be required.

We are grateful to J. Eijgensteijn and J. Poddighe for performing the protein analyses and to Dr F. Wanka for his advice and discussions during the preparation of this manuscript. This study was supported by the Dutch Cancer Foundation (KWF), grant SNUKC 81–10 to Dr F. Wanka.

Barrack
,
E. R.
&
Coffey
,
D. S.
(
1980
).
The specific binding of estrogens and androgens to the nuclear matrix of sex hormone responsive tissues
.
J. biol. Chem
.
255
,
7265
7275
.
Berezney
,
R.
&
Coffey
,
D. S.
(
1974
).
Identification of a nuclear protein matrix. Biochem. biophys
.
Res. Commun
.
60
,
1410
1417
.
Berezney
,
R.
&
Coffey
,
D. S.
(
1975
).
Nuclear protein matrix: association with newly synthesized DNA
.
Science
189
,
291
293
.
Brasch
,
K.
(
1982
).
Fine structure and localization of the nuclear matrix in situ
.
ExplCell Res
.
140
,
161
171
.
Capco
,
D. G.
&
Penman
,
S.
(
1983
).
Mitotic architecture of the cell: the filament networks of the nucleus and cytoplasm
.
J. Cell Biol
.
96
,
896
905
.
Cobbs
,
C. S.
&
Shelton
,
K. R.
(
1978
).
Major oligomeric structural proteins of the HeLa nucleus
.
Archs Biochem. Biophys
.
189
,
323
335
.
Cook
,
P. R.
&
Brazell
,
I. A.
(
1976
).
Conformational constraints in nuclear DNA
.
J. CellSci
.
22
,
287
302
.
Cook
,
P. R.
,
Brazell
,
I. A.
&
Jost
,
E.
(
1976
).
Characterization of nuclear structures containing superhelical DNA
.
J. Cell Sci
.
22
,
303
324
.
Cupido
,
M.
&
Simons
,
J. W. I. M.
(
1984
).
A re-appraisal of the BHK 21 clone-13 cell transformation system with respect to the assay conditions and the mutational origin of the transformed phenotype
.
Carcinogenesis
5
,
857
862
.
Dukwel
,
P. A.
,
Mullenders
,
L. H. F.
&
Wanka
,
F.
(
1979
).
Analysis of the attachment of replicating DNA to a nuclear matrix in mammalian interphase nuclei
.
Nucl. Acids Res
.
6
,
219
230
.
Hemminki
,
K.
&
Vainio
,
H.
(
1979
).
Preferential binding of benzopyrene into nuclear matrix fraction
.
Cancer Lett
.
6
,
167
173
.
Jackson
,
D. A.
,
Mccready
,
S. J.
&
Cook
,
P. R.
(
1980
).
RNA is synthesized at the nuclear cage
.
Nature, Lond
.
292
,
552
555
.
Jeppesen
,
P.
&
Morten
,
H.
(
1985
).
Effects of sulphydryl reagents on the structure of dehistonized metaphase chromosomes
.
J. Cell Sci
.
73
,
245
260
.
Kaufmann
,
S. H.
,
Coffey
,
D. S.
&
Shaper
,
J. H.
(
1981
).
Considerations in the isolation of rat liver nuclear matrix, nuclear envelope and pore-complex-lamina
.
Expl Cell Res
.
132
,
105
132
.
Laemmli
,
U. K.
(
1970
).
Cleavage of structural proteins during the assembly of the head of bacteriophage T4
.
Nature, Lond
.
221
,
680
685
.
Laemmli
,
U. K.
,
Cheng
,
S. M.
,
Adolph
,
K. W.
,
Paulson
,
J. R.
,
Brown
,
J. A.
&
Baumbach
,
W. R.
(
1978
).
Metaphase chromosome structure: the role of non-histone proteins
.
Cold Spring Harbor Symp. quant. Biol
.
42
,
109
118
.
Lamm
,
K. S.
&
Kasper
,
C. B.
(
1979
).
Electrophoretic analysis of the three major nuclear envelope polypeptides
.
J. biol. Chem
.
254
,
11713
11720
.
Lebkowski
,
J. S.
&
Laemmli
,
U. K.
(
1982a
).
Evidence for two levels of DNA-folding in histone- depleted HeLa interphase nuclei
.
J. molec. Biol
.
156
,
309
324
.
Lebkowski
,
J. S.
&
Laemmli
,
U. K.
(
1982b
).
Non-histone proteins and long-range organization of HeLa interphase DNA
.
J. molec. Biol
.
156
,
325
344
.
Mariman
,
E. C. M.
,
Van Eekelen
,
C. A. G.
,
Reinders
,
R. J.
,
Berns
,
A. J. M.
&
Van Venroou
,
W. J.
(
1982
).
Adenovirus heterogeneous nuclear RNA is associated with the host nuclear matrix during splicing../
,
molec. Biol
.
154
,
103
119
.
Marsden
,
M. P. F.
&
Laemmli
,
U. K.
(
1979
).
Metaphase chromosome structure: evidence for a radial loop model
.
Cell
17
,
849
858
.
Mullenders
,
L. H. F.
,
Eugensteun
,
J.
&
Wanka
,
F.
(
1982
).
Composition and DNA-binding properties of the nuclear matrix proteins from mammalian cell nuclei
.
Biochem. biophys. Acta
698
,
70
77
.
Pardoll
,
D. M.
,
Vogelstein
,
B.
&
Coffey
,
D. S.
(
1980
).
A fixed site of DNA-replication in eukaryotic cells
.
Cell
19
,
527
536
.
Pieck
,
A. C. M.
,
Van Der Velden
,
H. M. W.
,
Rdken
,
A. A. M.
,
Neis
,
J.
&
Wanka
,
F.
(
1985
).
Protein composition of the chromosomal scaffold and the interphase nuclear matrix
.
Chromosoma
91
,
137
144
.
Shelton
,
K. R.
&
Cochran
,
D. L.
(
1978
).
In vitro oxidation of intrinsic sulfhydryl groups yields polymers of two predominant polypeptides in the nuclear envelope fraction
.
Biochemistry
17
,
1212
1216
.
Tew
,
T. D.
,
Wang
,
A. L.
&
Schein
,
P. S.
(
1980
).
The nuclear matrix as a target of chloroethylnitrosoureas
.
Proc. Am. Assoc. Cancer Res
.
21
,
297
306
.
Vogelstein
,
B.
,
Pardoll
,
D. M.
&
Coffey
,
D. S.
(
1980
).
Supercoiled loops and eucaryotic DNA-replication
.
Cell
22
,
79
85
.
Wanka
,
F.
,
Mullenders
,
L. H. F.
,
Bekers
,
A. G. M.
,
Pennings
,
L. J.
,
Aelen
,
J. M. A.
&
Eugensteun
,
J.
(
1977
).
Association of nuclear DNA with a rapidly sedimenting structure
.
Biochem. biophys. Res. Commun
.
74
,
739
747
.