Subnuclear localization of S/MAR-binding proteins is differently affected by in vitro stabilization with heat or Cu2+

The nuclear matrix, a proteinaceous entity thought to be a scaffolding structure that determines the higher order organization of eukaryotic chromatin, is usually prepared from intact nuclei by a series of extraction steps. In most cell types investigated, the nuclear matrix does not spontaneously resist these extractions, but must rather be stabilized before the application of extracting agents such as high salt solutions or lithium diiodosalicylate. We have examined the effect of two widely used stabilization procedures on the localization of nuclear matrix proteins. Four individual polypeptides were studied, all of which are scaffold or matrix-associated region (S/MAR)-binding proteins: SATB1, SAF-A/hnRNP-U, NuMA , and topoisomerase II α. Nuclei were isolated from K562 human erythroleukemia cells in a buffer containing spermine, spermidine, KCl and EDTA, and the nuclear matrix or scaffold was obtained by extraction with lithium diiodosalicylate after stabilization by heat treatment (37° or 42°C) or incubation with Cu²⁺ ions. When the localization of individual proteins was determined by immunofluorescent staining and confocal scanning laser microscopy, markedly different consequences of the two stabilization strategies became evident, ranging from a total maintenance of the localization (NuMA and topoisomerase II α) to a marked redistribution (SATB1 and SAF-A/hnRNP-U). Our results seem to indicate that a reevaluation of stabilization protocols employed for the preparation of the nuclear matrix is desirable, especially by performing morphological controls. Received: 22 January 1997; in revised form: 17 February 1997 / Accepted: 21 February 1997     

Changes in the subnuclear distribution of two RNA metabolism-related proteins can be detected in nuclear scaffold or matrix prepared by different techniques

 The nuclear scaffold or matrix is a mainly proteinaceous structure thought to act as a nucleoskeleton determining the higher order organization of eukaryotic chromatin. These structures are prepared from isolated nuclei by a series of extraction steps involving the use of ionic detergents or high salt, and restriction enzymes or non-specific nucleases to remove chromatin and other loosely bound components. Since these treatments are harsh and unphysiological, the question remains open as to whether or not these structures, isolated in vitro, correspond to a nucleoskeleton existing in vivo. Recently, it has been demonstrated that the majority of nuclear matrix proteins are involved in RNA metabolism. In this study we have employed a morphological approach involving the use of confocal laser scanning microscopy and indirect immunofluorescence techniques to analyze whether two widely employed methods to prepare the nuclear scaffold or matrix can maintain the spatial distribution of two polypeptides involved in RNA metabolism, i.e., a 105-kDa component of spliceosomes and a ribonucleoprotein antigen. We demonstrate that the localization of these polypeptides changes, in some cases dramatically, in the final nucleoskeletal structures when compared with intact cells. Only when isolated nuclei were stabilized in vitro with the cross-linking agent sodium tetrathionate (NaTT) prior to extraction with 2 M NaCl and DNase I digestion, were the immunofluorescent patterns displayed by the nuclear matrix indistinguishable from those detected in intact cells. These results emphasize the usefulness of NaTT in studying putative nucleoskeletal structures, but also show that the methods currently employed to prepare the nuclear scaffold or matrix may create in vitro artifacts. Accepted: 12 May 1997     

Lamin A is part of the internal nucleoskeleton of human erythroleukemia cells

Nuclear lamins are the most abundant components of the nuclear lamina, a 10–50-nm-thick fibrous layer underlying the inner nuclear envelope membrane. Nevertheless, a number of recent investigations performed on epithelial and fibroblast cells have suggested that nuclear lamins are also present within the nucleoplasm and could be important constituents of the nucleoskeleton. We have studied the subnuclear distribution of lamins A and B1 in human erythroleukemia cells by using immunoblotting analysis and immunofluorescent staining of fractionated nuclei. In intact cells and isolated nuclei, antibodies to lamins A and B1 mainly stained the nuclear periphery, although some immunoreactivity was detected in the nuclear interior. However, when chromatin was removed by nuclease digestion and extraction with nonionic detergent or solutions of high ionic strength, a previously masked immunoreactivity for lamin A, but not for lamin B1, became evident in the internal part of the residual structures representing the nuclear matrix or scaffold. Preferential localization of lamin A to the inner part of the nucleus was also demonstrated by the presence of the majority of lamin A in the solubilized inner nuclear network subfraction. In contrast, lamin B1 was mainly recovered in the fraction corresponding to the nuclear periphery. Double labeling experiments showed that lamin A, but not lamin B1, colocalized with coiled and GATA-1 bodies. Thus, our results support the hypothesis that lamin A, but not lamin B1, may be a component of an internal nucleoskeleton in human erythroleukemia cells. J. Cell. Physiol. 178:284–295, 1999. © 1999 Wiley-Liss, Inc.     

Biochemical and morphological characterization of the nuclear matrix from apoptotic HL-60 cells

We have characterized the nuclear matrix-intermediate filament fraction from control and apoptotic HL-60 cells. Apoptosis was induced by exposure to the topoisomerase I inhibitor, camptothecin. By means of two-dimensional polyacrylamide gel electrophoresis, striking qualitative and quantitative differences were seen in the protein composition of the nuclear matrix-intermediate filament fraction obtained from apoptotic cells in comparison with controls. Western blotting analysis of apoptotic nuclear matrix proteins revealed degradation of some (topoisomerase IIalpha, SAF-A) but not other (SATB1 and nucleolin) components. Moreover, immunofluorescent staining for typical matrix antigens (NuMA protein, lamin B, SC-35) showed that in 35-40% of the structures prepared from apoptotic samples, marked changes in the subnuclear distribution of these proteins were present. Striking morphological differences between control and apoptotic samples were also detected at the ultrastructural level. These results demonstrate that both biochemical and morphological changes can be detected in the nuclear matrix prepared from apoptotic HL-60 cells.     

Spatial distribution of lamin A and B1 in the K562 cell nuclear matrix stabilized with metal ions.

When the nucleus is stripped of most DNA, RNA, and soluble proteins, a structure remains that has been referred to as the nuclear matrix, which acts as a framework to determine the higher order of chromatin organization. However, there is always uncertainty as to whether or not the nuclear matrix, isolated in vitro, could really represent a skeleton of the nucleus in vivo. In fact, the only nuclear framework of which the existence is universally accepted is the nuclear lamina, a continuous thin layer that underlies the inner nuclear membrane and is mainly composed of three related proteins: lamins A, B, and C. Nevertheless, a number of recent investigations performed on different cell types have suggested that nuclear lamins are also present within the nucleoplasm and could be important constituents of the nuclear matrix. In most cell types investigated, the nuclear matrix does not spontaneously resist the extraction steps, but must rather be stabilized before the application of extracting agents. In this investigation, by immunochemical and morphological analysis, we studied the effect of stabilization with different divalent cations (Zn(2+), Cu(2+), Cd(2+)) on the distribution of lamin A and B1 in the nuclear matrix obtained from K562 human erythroleukemia cells. In intact cells, antibodies to both lamin A and B1 mainly stained the nuclear periphery, although some immunoreactivity was detected in the nuclear interior. The fluorescent lamin A pattern detected in Cu(2+)- and Cd(2+)-stabilized nuclei was markedly modified, whereas Zn(2+)-incubated nuclei showed an unaltered pattern of lamin A distribution. By contrast, the distribution of lamin B1 in isolated nuclei was not modified by the stabilizing cations. When chromatin was removed by nuclease digestion and extraction with solutions of high ionic strength, a previously masked immunoreactivity for lamin A, but not for lamin B1, became evident in the internal part of the residual structures representing the nuclear matrix. Our results indicate that when metal ions are used as stabilizing agents for the recovery of the nuclear matrix, the distribution of both lamin A and lamin B1 in the final structures, corresponds to the pattern we have very recently reported using different extraction procedures. This observation strengthen the concept that intranuclear lamins may act as structural components of the nuclear matrix.     

Localisation of DNA topoisomerase IIalpha in mouse erythroleukemia cells.

The presence of DNA topoisomerase IIalpha was investigated in interphase and metaphase mouse erythroleukemia (MEL) Friend-S cells, and in extracted with 25 mM lithium diiodosalicylate buffer (Lis) nuclei using indirect immunofluorescence. The results showed that DNA topoisomerase IIalpha is localised in the nuclei. In the metaphase cells, we found high concentrations of this enzyme in the mitotic chromosomes. Our results support the idea of the accumulation of DNA topoisomerase IIalpha at the end of the cell cycle. The extractions of nuclei with 25 mM Lis led to the complete depletion of DNA topoisomerase IIalpha from the residual nuclear matrix. Using a high dilution of the first antibody, we established that the high level of heterochromatin compactisation in the interphase nuclei is caused by the high concentration of DNA topoisomerase IIalpha.     

Preparation and characterization of nuclear matrix/chromosome scaffold in situ

A method of nuclear matrix and chromosomal scaffold preparation from cultured animal cells was developed. After the high-salt extraction, interphase and mitotic cells were not detached from the coverslips that enabled us to analyse the nuclear matrix and chromosomal scaffold in cells at all mitotic phases. Morphological methods (phase contrast microscopy and electron microscopy of ultrathin sections) did not reveal any structures that could be identified as a chromosomal scaffold. However, after staining with antibodies to XCAP-E and topoisomerase IIalpha some structures were revealed in metaphase cells having both localization and morphology of a chromosomal scaffold. The cell residuals were not stained with antibodies to XCAP-E and topoisomerase IIalpha, if the nuclear matrix and chromosomal scaffold were destabilized by addition of beta-mercaptoethanol.     

Macromolecular domains containing nuclear protein p107 and U-snRNP protein p28: Further evidence for an in situ nuclear matrix

Summary Polyclonal antibodies have been produced which react with a nuclear protein having a molecular weight of 107kD and a pl of 8.7–8.8 (designated p107). This protein is shown to be a component of the residual ribonucleoprotein (RNP) network of the nuclear matrix. P107 localized exclusively to the nuclear interior but not within nucleolar or chromatin domains. We have taken advantage of this unique probe to examine whether the RNP network of the isolated nuclear matrix has a physical counterpart in situ. We show that RNA, p107, divalent cations and the 28 kD Sm antigen of U-snRNPs are components of in situ macromolecular assemblies. While the morphology and intranuclear distribution of these assemblies are insensitive to the removal of chromatin, they are markedly altered by degradation of RNA. Digestion in situ of RNA in the presence of EDTA followed by extraction with high ionic strength buffers solubilized the components of these assemblies. Electron microscopic and immunobiochemical data are presented which support the concept that the residual RNP network of the nuclear matrix is an isolate of a pre-existing structure, and that perturbations in this internal network can be created by RNA degradation, depletion of essential metal ions and proteolysis.Abbreviations CRLM polyclonal chicken antibody raised against rat liver nuclear matrix - Sm monoclonal antibody specific for the 28 kd protein antigen of U1, U2, U4, U5 and U6 snRNPs - hnRNP ribonucleoprotein particles containing hnRNA - snRNP ribonucleoprotein particles containing snRNA - PBS phosphate buffered saline - PMSF phenylmethylsulfonyl fluoride - PAGE polyacrylamide gel electrophoresis - EDTA ethylenediaminetetraacetic acid - VRC vanadium ribonucleoside complex - BSA bovine serum albumin - DMSO dimethylsulfoxide - HS high salt buffer - LS low salt buffer     

The Plant Nucleoskeleton: Ultrastructural Organization and Identification of NuMA Homologues in the Nuclear Matrix and Mitotic Spindle of Plant Cells

In the present work we investigate the structural organization of the nucleoskeleton ofAllium cepameristematic root cells. Resinless sections reveal for the first time a residual filamentous network in plant nuclei. This network is composed of branched knobbed filaments with associated globular structures, connected to the lamina and to the dense aggregates of different sizes. Results of immunoblotting show that many components of this network are homologues of intermediate filament-type proteins. NuMA, a coiled-coil protein related to intermediate filaments, found in animal cells, can also be detected in this plant nuclear matrix system. Immunofluorescence reveals a diffuse distribution of the animal NuMA homologues in plant nuclear core filaments in interphase. Resinless immunoelectron microscopy further reveals a distribution along the extended filaments and the dense aggregates. During mitosis, in contrast to the accumulation at the poles in animal cells, NuMA homologues in plant onion cells show a diffuse pattern, which may correspond to the spindle matrix. Our data are the first report of the conservation in plants of NuMA proteins, which may be involved in both nuclear and mitotic spindle organizations.     

植物细胞中NuMA类似蛋白的分布及其在细胞周期中的变化

免疫荧光染色结果说明植物细胞核内含有与抗动物ＮｕＭＡ多抗呈阳性交叉反应的多肽。选择性抽提并结合免疫荧光染色结果说明这种多肽位于核基质纤维蛋白网络上。免疫印迹反应显示胡萝卜（ＤａｕｃｕｓｃａｒｏｔａＬ．）悬浮培养细胞核基质蛋白与抗动物ＮｕＭＡ蛋白多抗的阳性反应条带为７４ｋＤ和７６ｋＤ。有丝分裂各期免疫荧光染色的结果表明植物细胞中的ＮｕＭＡ类似蛋白在有丝分裂过程中呈现有规律的变化。结合选择性抽提的有丝分裂各期的免疫荧光染色的结果表明核基质在此过程中也发生明显变化。应用选择性抽提并结合ＤＧＤ包埋去包埋电镜技术对植物细胞间期及有丝分裂期核基质的形态结构进行了观察。结果显示胡萝卜悬浮培养细胞间期核内存在一个非染色质性的纤维蛋白网络体系,而在正处于分裂的细胞中则未观察到。以上结果说明ＮｕＭＡ类似蛋白是核基质的组分之一并与有丝分裂密切相关。     

Effects of wild-type p53 expression on the quantity and activity of topoisomerase IIalpha and beta in various human cancer cell lines.

The p53 null HL-60 cell line was transfected with plasmids coding for either the wild-type p53 or mutant p53 gene. The stable expression of wild-type p53 resulted in a significant increase in sensitivity to the topoisomerase II poisons etoposide and doxorubicin, but not to the topoisomerase II inhibitors razoxane and ADR-529. HL-60 cells expressing wild-type p53 demonstrated 8- to 10-fold more VP-16 induced DNA breaks by the alkaline elution assay. The effect of inducible expression of wild-type p53 was also studied in the p53 null erythroblastoid cell line K562 and in the human squamous carcinoma cell line SqCC. The inducible expression of wild-type p53 in the K562 cell line resulted in a 3-fold increase in sensitivity to VP-16. The quantity of topoisomerase IIalpha was not altered by the transfection as determined by immunoblotting, while the amount of the beta isoform was increased 2.5-fold in HL-60 cells. The topo II catalytic activity present in nuclear extracts was measured as the decatenation of kinetoplast DNA, and found to be unaltered by p53 expression. Immunostaining for topoisomerase IIalpha was substantially diminished in both stable and inducible wild-type p53 expressing cells when three different antibodies were used (two polyclonal and one monoclonal). However, the addition of VP-16 resulted in a rapid appearance of nuclear fluorescence for topoisomerase IIalpha. No changes in topoisomerase IIbeta immunostaining were observed. These results suggest that an epitope for topoisomerase IIalpha is concealed in cells expressing wild-type p53 and that a complex between topoisomerase IIalpha and p53 may be disrupted by the addition of antitumor drugs.     

Active DNA topoisomerase IIalpha is a component of the salt-stable centrosome core

Recently, we reported that the monoclonal antibody specific for human DNA topoisomerase IIalpha, Ki-S1, stains not only the nuclei of human A431 cells but also extranuclear structures suggestive of centrosomes (Meyer, K. N., Kjeldsen, E., Straub, T., Knudsen, B. K., Kikuchi, A., Hickson, I. D., Kreipe, H., and Boege, F. (1997) J. Cell Biol. 136, 775-788). Here, we confirm colocalization of Ki-S1 with the centrosomal marker gamma-tubulin. In addition, we show labeling of centrosomes by peptide antibodies against the N and C termini of human topoisomerase IIalpha. Probing Western blots of isolated centrosomes with topoisomerase IIalpha antibodies, we demonstrate a protein band of 170 kDa. Moreover, isolated centrosomes exhibited DNA decatenation and relaxation activity correlated to the amount of topoisomerase IIalpha protein in the same way as seen in the pure recombinant enzyme. Topoisomerase IIalpha epitopes could not be removed from centrosomes by salt extraction, DNase treatment, or RNase treatment, procedures that completely removed the enzyme from nuclei. Taken together, these observations suggest that active topoisomerase IIalpha is bound tightly to the centrosome in a DNA-independent manner. Because such centrosomal topoisomerase IIalpha was also present in quiescent lymphocytes devoid of topoisomerase IIalpha in the nuclei, we assume that it might be a long-lived storage form.     

DNA topoisomerase IIalpha interacts with CAD nuclease and is involved in chromatin condensation during apoptotic execution

Apoptotic execution is characterized by dramatic changes in nuclear structure accompanied by cleavage of nuclear proteins by caspases (reviewed in [1]). Cell-free extracts have proved useful for the identification and functional characterization of activities involved in apoptotic execution [2-4] and for the identification of proteins cleaved by caspases [5]. More recent studies have suggested that nuclear disassembly is driven largely by factors activated downstream of caspases [6]. One such factor, the caspase-activated DNase, CAD/CPAN/DFF40 [4,7,8] (CAD) can induce apoptotic chromatin condensation in isolated HeLa cell nuclei in the absence of other cytosolic factors [6,8]. As chromatin condensation occurs even when CAD activity is inhibited, however, CAD cannot be the sole morphogenetic factor triggered by caspases [6]. Here we show that DNA topoisomerase IIalpha (Topo IIalpha), which is essential for both condensation and segregation of daughter chromosomes in mitosis [9], also functions during apoptotic execution. Simultaneous inhibition of Topo IIalpha and caspases completely abolishes apoptotic chromatin condensation. In addition, we show that CAD binds to Topo IIalpha, and that their association enhances the decatenation activity of Topo IIalpha in vitro.     

Modifications of the chromatin arrangement induced by ethidium bromide in isolated nuclei, analyzed by electron microscopy and flow cytometry

Ethidium bromide (EB) is widely used for investigating the DNA conformation in chromatin both with conventional and cytofluorimetric techniques. Since the interaction of the dye with DNA should result in structural deformations which can be different in isolated or in situ chromatin, a study has been performed on the effects caused by different amounts of EB and the analogous propidium iodide on isolated nuclei, in which chromatin maintains its native relationships with the other nuclear structures (envelope, nucleolus, interchromatin RNP, nuclear matrix). The results obtained by comparing ultrastructural observations in thin sections and in freeze-fracturing with conformational analysis in multiparameter flow cytometry indicate that the phenanthridinic fluorochromes, especially at the high concentrations used for cytofluorimetric analyses, cause deep rearrangements of the chromatin in situ. These effects consist both in aggregation and condensation of the fibers into the dense chromatin domains, and in an increase of the supernucleosomal configuration associated with an enlargement of interchromatin spaces in which the RNP particles appear particularly evident. These results, discussed with those available on isolated chromatin, suggest that any unwinding effect of the intercalating dyes on the DNA cause a general condensation of chromatin as a consequence of the constraints which characterize the organization of the chromatin inside the nucleus.     

Characterization of proteins associated with nuclear ribonucleoprotein particles by two-dimensional polyacrylamide gel electrophoresis.

Rat liver nuclear ribonucleoprotein particles were prepared by two different methods and defined as 40S ribonucleoprotein (40S RNP) and heterogeneous nuclear ribonucleoprotein (HnRNP) particles. The RNP particles were either solubilized in 8 M urea--6 mM 2-mercaptoethanol--20 mM glycine--20 mM Tris--HCl (pH 8.4) or subjected to removal of RNA by phenol extraction prior to solubilizing the proteins in the urea buffer. The proteins associated with 40S RNP and HnRNP were heterogeneous and very similar in their electrophoretic patterns when analyzed by two-dimensional PAGE, except a protein with molecular weight of 62 000 and an isoelectric point (pI) of 6.2 was present only in HnRNP particles. At least 12 major and 22 minor components could be identified in both preparations. The major proteins were found at pI values varying from 6.0 to 8.5 and with molecular weights from 32 000 to 42 000, and a group of proteins with molecular weight approximately 65 000 were more prominent in HnRNP than in 40S RNP. The other components were found mainly at pI ranges from 5.0 to 6.5 with molecular weights from 43 000 to 65 000. The phenol method extracted essentially all proteins associated with either 40S RNP and HnRNP, but was less effective in extracting a group of proteins with pI values from 5.0 to 5.5 and more efficient for proteins with pI values from 7.5 to 8.5. When chromatin proteins isolated by phenol extraction were compared with HnRNP particle proteins isolated by the same method, the electrophoretic mobilities of the HnRNP particle proteins were found to be identical with a fraction nonhistone chromatin proteins. The 40S RNP particles were further purified by metrizamide isopycnic density gradient centrifugation. The electrophoretic patterns of these proteins were very similar to those prepared by sucrose density gradient centrifugation. Therefore, we concluded that the proteins of RNP particles constituted part of the chromatin proteins.     

Visualization of the chromosome scaffold and intermediates of loop domain compaction in extracted mitotic cells

A novel extraction protocol for cells cultured on coverslips is described. Observations of the extraction process in a perfusion chamber reveal that cells of all mitotic stages are not detached from coverslips during extraction, and all stages can be recognized using phase contrast images. We studied the extracted cell morphology and distribution of a major scaffold component - topoisomerase IIalpha, in extracted metaphase and anaphase cells. An extraction using 2M NaCl leads to destruction of chromosomes at the light microscope level. Immunogold studies demonstrate that the only residual structure observed is an axial chromosome scaffold that contains topoisomerase IIalpha. In contrast, mitotic chromosomes are swelled only partially after an extraction using dextran sulphate and heparin, and it appears that this treatment does not lead to total destruction of loop domains. In this case, the chromosome scaffold and numerous structures resembling small rosettes are revealed inside extracted cells. The rosettes observed condense after addition of Mg2+-ions and do not contain topoisomerase IIalpha suggesting that these structures correspond to intermediates of loop domain compaction. We propose a model of chromosome structure in which the loop domains are condensed into highly regular structures with rosette organization.     

Extracellular signal-regulated kinase activates topoisomerase IIalpha through a mechanism independent of phosphorylation

The mitogen-activated protein (MAP) kinases, extracellular signal-related kinase 1 (ERK1) and ERK2, regulate cellular responses by mediating extracellular growth signals toward cytoplasmic and nuclear targets. A potential target for ERK is topoisomerase IIalpha, which becomes highly phosphorylated during mitosis and is required for several aspects of nucleic acid metabolism, including chromosome condensation and daughter chromosome separation. In this study, we demonstrated interactions between ERK2 and topoisomerase IIalpha proteins by coimmunoprecipitation from mixtures of purified enzymes and from nuclear extracts. In vitro, diphosphorylated active ERK2 phosphorylated topoisomerase IIalpha and enhanced its specific activity by sevenfold, as measured by DNA relaxation assays, whereas unphosphorylated ERK2 had no effect. However, activation of topoisomerase II was also observed with diphosphorylated inactive mutant ERK2, suggesting a mechanism of activation that depends on the phosphorylation state of ERK2 but not on its kinase activity. Nevertheless, activation of ERK by transient transfection of constitutively active mutant MAP kinase kinase 1 (MKK1) enhanced endogenous topoisomerase II activity by fourfold. Our findings indicate that ERK regulates topoisomerase IIalpha in vitro and in vivo, suggesting a potential target for the MKK/ERK pathway in the modulation of chromatin reorganization events during mitosis and in other phases of the cell cycle.     

Organization of higher-level chromatin structures (chromomere,chromonema and chromatin block) examined using visible light-induced chromatin photo-stabilization

The method of chromatin photo-stabilization by the action of visible light in the presence of ethidium bromide was used for investigation of higher-level chromatin structures in isolated nuclei. As a model we used rat hepatocyte nuclei isolated in buffers which stabilized or destabilized nuclear matrix. Several higher-level chromatin structures were visualized: 100nm globules-chromomeres, chains of chromomeres-chromonemata, aggregates of chromomeres-blocks of condensed chromatin. All these structures were completely destroyed by 2M NaCl extraction independent of the matrix state, and DNA was extruded from the residual nuclei (nuclear matrices) into a halo. These results show that nuclear matrix proteins do not play the main role in the maintenance of higher-level chromatin structures. Preliminary irradiation led to the reduction of the halo width in the dose-dependent manner. In regions of condensed chromatin of irradiated nucleoids there were discrete complexes consisting of DNA fibers radiating from an electron-dense core and resembling the decondensed chromomeres or the rosette-like structures. As shown by the analysis of proteins bound to irradiated nuclei upon high-salt extraction, irradiation presumably stabilized the non-histone proteins. These results suggest that in interphase nuclei loop domains are folded into discrete higher-level chromatin complexes (chromomeres). These complexes are possibly maintained by putative non-histone proteins, which are extracted with high-salt buffers from non-irradiated nuclei.