REVIEW The nuclear matrix: structure and composition

RON VERHEIJEN 1 '*, WALTHER VAN VENROOIJ 2 and FRANS RAMAEKERS 1

'Department of Pathology and department of Biochemistry, University of Sijmegen, The Xetherlands

Author for correspondence at: Department of Pathology, University Hospital of Nijmcgcn, Geert Grootcplein Zuid 24, 6525 GA Nijmcgcn, The Netherlands

Summary

Introduction The pore complex-lamina The nuclear pore complexes The nuclear lamina The nucleolar residue The internal matrix Morphological and biochemical aspects Heterogeneous nuclear RNP particles Small nuclear RNP particles Nuclear actin

Enzymes involved in DNA and RNA metabolism Virus-specific proteins Associated transcripts The nuclear matrix and RNA transport Behaviour of nuclear matrix components during mitosis
Key words: nuclear matrix, pore complex-lamina, nucleolar matrix, internal matrix, chromosomal scaffold.

Introduction The term nuclear matrix was first introduced by Berezney & Coffey (1974) to denote a highly structured residual framework obtained from rat liver nuclei by sequential salt extractions, detergent and nuclease treatments. The isolated three-dimensional structure consisted almost entirely of protein. Subsequent studies showed that when protease inhibitors were included in all isolation steps and ribonuclease (RNase) was omitted, the isolated nuclear matrix contained RNA as the second most abundant component (Herman et al. 1978; Miller et. al. 1978a; Shaper <?</. 1979; Berezney, 1980; van Eekelen & van Venrooij, 1981; Mariman et al. 1982; Fey et al. 1986a,b). In this review the term nuclear matrix is defined as the biochemical entity that can be isolated after sequential extraction of cells with non-ionic detergents, nucleases and high-salt buffers (Shapere/ al. 1979). With respect to the nucleases, it should be stated here that several authors use only deoxyribonucleases (DNases), while others use DNases in combination with RNases. Nuclear matrices have been isolated from a wide variety of mammalian and non-mammalian cell types (reviewed by Shaper et al. 1979; Barrack & Coffey, 1982).
Journal of Cell Science 90, 11-36 (1988) Printed in Great Britain The Company of Biologists Limited 1988

However, it has been shown not to be an obligatory nuclear component (Lafond & Woodcock, 1983). In general, nuclear matrix preparations from different cells and tissues possess some common structural entities: (1) the residual elements of the nuclear envelope, also designated the pore complex-lamina; (2) the residual nucleoli; (3) a granular and fibrous internal matrix structure that extends throughout the interior of the nucleus. In recent years various studies have implicated the nuclear matrix as being involved in nuclear activities such as DNA metabolism (reviewed by Berezney, 1984; Jackson et al. 1984; Razin et al. 1985; Vogelstein et al. 1985; Zehnbauer & Vogelstein, 1985; Nelson et al. 1986), DNA transcription, processing and transport of RNA (Zeitlin et al. 1987; reviewed by Berezney, 1984; Jackson et al. 1984; Vogelstein et al. 1985; Zehnbauer & Vogelstein, 1985), steroid hormone action (reviewed by Barrack & Coffey, 1982; Diamond & Barrack, 1984; Kirsch et al. 1986; Alexander et al. 1987) and viral replication (reviewed by Berezney, 1984; Simard et al. 1986). The nuclear'matrix also seems to play a role in carcinogenesis (Berezney, 1984). It is important to note that the suggested role for the nuclear matrix in various nuclear functions is based primarily on the recovery of several relevant functional 11

Chromntin

Nucleolus

Nuclear lamina

Outer nuclear membrane

Inner nuclear membrane

Rough endoplasmic reticulum

Fig. 1. Schematic view of a cell nucleus. The nuclear envelope is composed of inner and outer nuclear membranes, which fuse at the regions where nuclear pores are situated. The nuclear lamina is localized on the nucleoplasmic site of the inner membrane.

molecules in nuclear matrix preparations, and thus comprises only circumstantial evidence. This aspect has to be borne in mind in discussing the relation of particular nuclear functions with the nuclear matrix. Although in this review we want to focus mainly on recently characterized nuclear matrix proteins, increasing evidence indicates that heterogeneous nuclear RNA (hnRNA or pre-mRNA) plays a structural role in the organization of the internal nuclear matrix (Miller et al. 1978*7; Berezney, 1980; Brasch, 1982; Gallinaro et al. 1983; Long &"Schrier, 1983; Fey et al. 1986a,b). Therefore, part of this review will be dealing with RNA processing and transport in the nucleus as well. The main thesis proposed in this respect is that RNA polymerase II transcripts, soon after the initiation of their synthesis, bind to proteins to set up a structural backbone. Among these proteins are the so-called core proteins. Subsequently, other functional molecules that are involved in RNA and DNA metabolism may be tethered to this hnRNP structure. During processing and transport the RNA remains involved in the maintenance of the integrity of the internal matrix and is not released until after the last maturation step, that is after changing its set of core proteins for proteins known to be associated with cytoplasmic mRNA. The pore complex-lamina The major structural components of the nuclear envelope are the inner and outer nuclear membranes enclosing a lumen or perinuclear space, as well as the nuclear lamina and pore complexes. The outer membrane on the cytoplasmic side appears to be continuous with the rough endoplasmic reticulum and is covered with 12 R. Verheijen et al.

ribosomes on its outermost surface. The inner membrane is smooth. The nuclear pore complexes are situated in those regions where the two membranes fuse (Franked al. 19816; Gerace, 1986). The nuclear lamina is a fibrillar meshwork of proteinaceous material, which is intercalated between the chromatin and the inner membrane of the nuclear envelope (see Fig. 1). In preparing nuclear matrices, the nuclear envelope is exposed to buffers containing non-ionic detergents, nucleases and high-molarity salt buffers. Morphologically, only the pore complexes and the nuclear lamina seem to be resistant to such treatments. This residual framework is, therefore, mostly referred to as the nuclear pore complex-lamina, and is considered to be a part of the nuclear matrix. It should be noted, however, that much of the data about the pore complexlamina have not originated from studies of nuclear matrix preparations but from investigations on isolated nuclear envelopes. The nuclear pore complexes The nuclear pore complexes are large organelles that form channels for nucleocytoplasmic transport through the nuclear envelope (reviewed by Newport & Forbes, 1987). They have also been postulated to serve as genegating organelles capable of interacting specifically with expanded (transcribable) portions of the genome (Blobel, 1985). The pore complexes are non-randomly distributed on the nuclear surface. Their three-dimensional structure has been determined by electron microscopy using nuclear envelopes from Xenopus oocytes (Unwin & Milligan, 1982). These authors found the pore complex

to be a symmetrical structure (outer diameter 120 nm) framed by two widely separated, coaxial rings. Each ring is composed of eight globular subunits, and attaches to the nuclear membranes. Connected to these rings and extending radially inwards from them along a central plane are elongated structures called spokes. These spokes appear to contact a large central spherical particle, the plug. Until 1982 none of the pore complex polypeptide constituents had been defined and characterized. In recent years, however, the nuclear envelope of the Xenopus laevis oocyte (see Fig. 2) has been shown to contain one principal polypeptide of 68K (K = 10 Mr) that appeared to be a major component present in both lamina and pore complex preparations (Stick & Krohne, 1982; Benavente et al. 1984a). Gerace et al. (1982) identified a prominent 190K nuclear pore complex glycoprotein (gpl90) in rat liver nuclear envelopes. On the basis of its biochemical characteristics, these authors have suggested that the protein is involved in anchoring the pore complex to the nuclear envelope membranes. Davis & Blobel (1986) have identified and characterized a 62K protein of the nuclear pore complex from rat liver. This protein was shown to be synthesized as a soluble cytoplasmic precursor of 61K. After incorporation of the protein into the nuclear fraction the

protein seems to be modified by addition of A'-acetylglucosamine residues. gpl90 as well as the 62K protein remain associated with the pore complex-lamina after Triton X-100 extraction at low-ionic strength. However, the interaction of both proteins with the pore complex-lamina fractions was found to be destabilized in the presence of elevated salt concentrations. Recently, Gerace and co-workers reported on the identification of eight structurally distinct pore complex proteins with common epitopes, isolated from rat liver cells (Snow et al. 1987). These polypeptides with apparent molecular weights of 210, 180, 145, 100, 63, 58, 54 and 45 (Xl0 J ) copurified with the pore complexes under various conditions of ionic strength and non-ionic detergent, and were characterized using monoclonal antibodies. All members of this group of proteins contained multiple O-linked A:-acetylglucosamine residues (see also Holt et al. 1987). Using immunoelectron microscopic techniques it was found that the proteins recognized by the monoclonal antibodies were situated on the cytoplasmic as well as on the nucleoplasmic surfaces of the nuclear membranes, but were absent from the lumen. Because of some similarities between the biochemical properties of the 63K protein and the 62K protein described by Davis & Blobel (1986), it was concluded by Snow et al. (1987)

Fig. 2. Electron micrograph showing the nuclear envelope manually isolated from oocytes of A', laevis (for details see Franke & Scheer, 1970; Scheer, 1972). In the upper picture numerous pore complexes are denoted by vertical arrows, brackets denote some of intranuclear tangles of the fibrils associated with the pore complexes. In the lower picture the arrows denote the individual annular granules on either side of the pore margin, the arrowhead points to the central granule. N, nucleoplasmic side; C, cytoplasmic side; o, outer side; i, inner side of the nuclear envelope. Note that in these cells the nuclear lamina at the inner aspect of the nuclear envelope is very thin and barely detectable. In whole-mount preparations it appears as a single layer of a loose filamentous meshwork (Scheer et al. 1976; Aebi et al. 1986). Bars, 0-1 /im. (Courtesy of Professor Dr W. W. Franke.)

The nuclear matrix

13

that these two pore complex constituents are probably identical. Berrios et al. (1983) have identified a glycoprotein associated with nuclear matrix pore complex-lamina preparations obtained from Drosophila melanogaster embryos. The molecular weight of this protein was initially estimated to be 17X103, but more recent studies have shown that it was only 2x 10 smaller than gpl90 described by Gerace et al. (1982) in rat liver. This Drosophila glycoprotein is therefore designated here as gp 188. It seems to be the homologue of the rat liver gpl90, since the two polypeptides have some biochemical properties in common (Filson et al. 1985). Antisera prepared against gpl90 were found not to cross-react with gpl88 (Filson et al. 1985). In contrast, one of the two antibodies raised against gpl88 crossreacted weakly with glycoproteins of similar molecular weight in isolated nuclear fractions from Xenopus oocytes, as well as chicken, opossum and rat liver. There was no detectable release of gpl88 from the nuclear fraction after treatment with Triton X-100 or DNase I and RNase A. Extraction of the residual nuclear material with 1 M-NaCl resulted in the apparent solubilization of approximately 10-20% of the glycoprotein, but the majority of this component resisted salt extraction (and even 5 M-urea) and was found associated with the nuclear matrix pore complexlamina. Berrios et al. (1983) have also reported on the existence of a 188K protein in the nuclear matrix pore complex-lamina fraction of Drosophila embryos, which is distinct from gpl88. This protein was identified as an ATPase/dATPase and appeared not to be glycosylated, as it was completely resistant to digestion by endoglycosidase H. It remains to be determined if the ATPase/dATPase is a real constituent of the pore complex in vivo. Nuclear matrix preparations described by many authors differ considerably with respect to the highionic strength conditions under which they are isolated. It is not known to what extent the pore complex proteins described above remain associated with such different preparations. When nuclear envelopes are exposed to rigorous extraction treatments involving solutions of high-ionic strength or non-denaturing detergents, the basic structural elements of the pore complexes are still identifiable (Franke et al. 19816). An interesting postulation in this respect is that only the components on the cytoplasmic surface would be sensitive to salt extraction (Davis & Blobel, 1986), whereas the components on the nuclear surface would remain attached to the lamina by a link that is resistant to extraction with high-salt solution. Whether or not its link to the lamina makes the nuclear part of the pore complex resistant to high-salt extraction remains to be examined. 14 R. Verheijen et al.

The nuclear lamina The nuclear lamina consists of a proteinaceous layer situated subjacent to the inner nuclear membrane (reviewed by Franke et al. 19816; Gerace & Blobel, 1982; Gerace, 1986; Krohne & Benavente, 1986; Newport & Forbes, 1987). The lamina is usually copurified together with the pore elements. The predominant polypeptides in such preparations are the lamins, proteins in the molecular weight range of 60-80 (XlO3), which are immunologically related. In mammals and avian species three main lamin proteins, i.e. lamins A, B and C, have been characterized, while at least five different lamins have been described in amphibia and one or two in certain invertebrates (reviewed by Krohne & Benavente, 1986). Immunohistochemical studies using specific antibodies to the lamin proteins have confirmed the localization of these proteins at the rim of the nucleus (see, e.g., Fig. 3). Several authors have detected lamin precursors (Gerace et. al. 1984; DagenaiseZ al. 1985; Lehnere/a/. 1986), while additional minor components that display the biochemical properties characteristic of the lamins have been described for rat and chicken by Lehner et al. (1986). This indicates that the composition of the nuclear lamina in these species is probably more complex than previously assumed. Recently, McKeon et al. (1986) and Fisher et al. (1986) have characterized the cDNA clones for human lamins A and C. From the protein sequences deduced from these cDNAs it became apparent that lamin A has an additional region of approximately 9K at its carboxyl terminus as compared to lamin C (Fisher et al. 1986). Both lamin sequences show a marked homology with the intermediate filament proteins (reviewed by Franke, 1987). In vitro translation studies performed by Laliberte et al. (1984) showed that lamins A and C are encoded for by different mRNAs, with lamin A being a precursor approximately 2K larger than mature lamin A. Using a mouse monoclonal antibody (IFA) raised against a common domain of all intermediate filament proteins, Lebel & Raymond (1987), as well as Osborn & Weber (1987), have shown that lamin B also shares some sequence homology with the intermediate filament proteins. This homology of the lamins to the intermediate filament proteins may account for the fibrillar nature of the lamina as seen, for example, in Fig. 4. Lamin B is thought to fulfil a role in anchoring the lamina to the inner membrane of the lipid bilayer, since this protein is more resistant to chemical extractions from nuclear membranes compared with lamins A and C (Gerace & Blobel, 1982; Lebel & Raymond, 1984). It also remains selectively associated with membrane vesicles after nuclear envelope disassembly during mitosis (Burke & Gerace, 1986).

Fig. 3. Peripheral localization

of the nuclear lamina in mouse P19 cells as detected by an antiserum to lamin B. The set of pictures was obtained by the use of a confocal scanning laser microscope (Brakenhoff et at. 1985), which scanned through the nucleus from bottom (1) to top (9) at 2-j.im intervals. Bar, 10/tm. (Courtesy of Dr G. J, Brakenhoff & Dr R. van Driel.)

Recent studies by Georgatos & Blobel (\987a,b) have demonstrated that lamin B also constitutes an intermediate filament attachment site at the nuclear envelope. Their approach consisted of in vitro binding studies with isolated bovine lens vimentin and avian erythrocyte nuclear membranes. Removal of lamin B from the nuclear envelope by urea extraction or blocking with anti-lamin B antibodies were found to reduce the binding of vimentin to these membranes. Other techniques, such as immunoprecipitation, rate zonal sedimentation and affinity chromatography, pointed to a specific vimentinlamin B association under;/? vitro conditions. The 6-6K carboxyl terminus of vimentin was found to be involved in this interaction, whereas the binding was positively influenced when lamin A was present. From these data Georgatos & Blobel (19876) concluded that intermediate filaments may be anchored directly to the nuclear lamina. These anchorage places are suggested to be restricted to certain distinct foci along the lamina, coinciding with nuclear pores, and not to be uniformly distributed over the nuclear surface (see also Goldman etal. 1985).

A direct association between the cytoskeletal framework and the nuclear lamina had been described by Capco et al. (1982) by using whole-mount microscopy to visualize nuclear matrices prepared from mouse 3T3 fibroblasts and HeLa cells. Their electron micrographs showed that cytoskeletal filaments were attached to the nuclear lamina. Two-dimensional gel electrophoresis revealed vimentin to be present in their nuclear matrix fractions. Several other investigators have demonstrated the presence of vimentin and cytokeratins in nuclear matrix preparations from cells grown in suspension culture (see, for example, Verheijen et al. 1986a). Another indication of the attachment of intermediate filaments to the nuclear envelope has been reported by Staufenbiel & Deppert (1982), who showed that after isolation of nuclei from cells grown in suspension culture the majority of the cytokeratin and vimentin filaments had collapsed onto the nuclear surface but still constituted a filamentous system. These collapsed filaments could be partially unfolded again by centrifugation through an isotonic buffer. The nuclear matrix 15

Fig. 4. Rotary-shadowed platinum/carbon replica of an isolated and cntical-point-dned BHKnuclear matrix preparation showing the lamina as a meshwork of anastomosing 8-10nm filaments. Cell extraction was essentially performed as described by Simard el al. (1986) (0-5% Triton X-100, 5 fig ml" 1 DNase I, 2lM-NaCl). Bar, 1-OjUm. (Courtesy of Dr V. Bibor-I lardy.)

16

R. Yerlieijen el al.

A remarkable observation in the field of intermediate filament-nuclear matrix interactions has recently been made by Carmo-Fonseca et al. (1987). These investigators isolated nuclear matrix-intermediate filament scaffolds from cultured rat ventral prostate cells and isolated rat uterine epithelial cells. Subsequently, the scaffolds were critical-point dried, platinum-carbon replicated and examined by electron microscopy. In such preparations the intermediate filaments were not seen to abut on the nuclear lamina, but rather to be looped and to follow the nuclear surface. Short, direct connections of a cross-bridge type (5-7 nm in diameter) extended laterally from the intermediate filaments and fused with the nuclear pore complexes (see Fig. 5). These cross-bridges appeared to be about 75-100 nm long in uterine epithelial cells and were shown to be associated with cytokeratin filaments, while in fibroblasts they were considerably shorter (approximately 14 nm long) and probably associated with vimentin filaments. The cross-bridges were not recognized by anti-cytokeratin antibodies. Considering the very short length of these linking structures, the authors concluded that their finding does not contradict the observations of Georgatos 8c Blobel (1987a,b). Also Fey et al. (1984a) have clearly shown interactions of intermediate filament structures with the nuclear periphery (see Fig. 6). Apart from investigations on cytoskeleton-nuclear lamina interactions, other studies have established that during interphase the lamina is in intimate contact with the peripheral chromatin (Boulikas, 1986; reviewed by Hancock & Boulikas, 1982; Hubert & Bourgeois, 1986; Hancock & Dessev, 1987). Such interactions are probably important for stabilizing or maintaining certain aspects of higher-order chromosome architecture (Lebkowski & Laemmli, 19826). In conclusion, the lamina not only determines the nuclear shape and the spatial organization of the pore complexes, but seems also to be directly involved in anchoring both the intermediate filaments and the chromosomes at the nuclear periphery during interphase.

The nucleolar residue The nucleolus is the site of synthesis and processing of pre-ribosomal RNA, and of assembly of the ribosomal proteins and ribosomal RNA into pre-ribosomal particles. Only a relatively small number of the many nuclear proteins are confined exclusively to the nucleolus and presumably play specific roles in its structure and function. One such major nucleolarspecific protein is nucleolin or C23. This phosphorylated protein (HOK/pI 5-5) is probably involved in prerRNA transcription and ribosome assembly (Bugler et

al. 1982). Small nuclear RNPs (U3 and U8 in particular), which may function in rRNA processing, are also accumulated in the nucleolus (Epstein et al. 1984; Reddy et al. 1985). Many authors have dealt with the morphological and biochemical aspects of the nucleolus (for reviews see Jordan & Cullis, 1982; Goessens, 1984; Hadjiolov, 1985; Sommerville, 1986). At the Eighth European Nucleolar Workshop in 1983 an attempt was made to standardize nucleolar nomenclature. Thus, the nucleolar matrix was defined as the residual structure left after extraction procedures to reveal the nuclear matrix (Jordan, 1984). Such nucleolar matrices retain the size and shape of the original nucleoli. However, as nucleolar residues present in nuclear matrix preparations may have been subjected to treatment with DNase I only or with a combination of DNase I and RNase A, it is necessary to extend the definition given by Jordan (1984). The term nucleolar matrix is used here to indicate the residual nucleolus in nuclear matrix preparations in which only DNase I has been used in the nuclease step. Little is known about the composition of the nucleolar remainders in nuclear matrix preparations. It is difficult to detect any morphological resemblance between the nucleolar residues found in nuclear matrices and structural components of the intact nucleolus. Aggregation and condensation of nucleolar structures, probably due to the presence of divalent cations, make interpretation of electron-microscopic images very difficult. Procedures using EDTA, however, can yield fairly decondensed nucleolar matrices. This is illustrated by the studies of Long & Ochs (1983), who prepared chromatin-depleted nuclei from Friend erythroleukaemia cells under conditions that avoid the use of high salt concentrations. These authors partially digested the DNA, and then washed the cells twice in 2mM-EDTA. Compared with the amounts in whole nuclei, the amounts retained in the resulting structures were approximately 1 % of the DNA, 65 % of the total RNA, 70% of the hnRNA, 74% of the snRNA, 29% of all protein and 2 % of the histories. Although in electron micrographs of nuclei treated in this way no morphological evidence was found for residual nucleoli, immunofluorescence studies showed that protein C23 was located in distinct centrally localized regions. Exposing these EDTA-prepared chromatindepleted nuclei to 2mM-MgCl2 resulted both in the reformation of morphologically distinct residual nucleoli and in aggregation of matrix fibrils. Similar effects have been observed in rat liver nuclei by Galcheva-Gargova et al. (1982) and also by Hubert et al. (1981), Bouvier et al. (1980) and Aaronson & Woo (1981). For some time it was not clear whether the nucleolar matrix structures and the intranuclear matrix network were composed of distinct or identical constituents. The nuclear matrix 17

Fig. 5. A. Rotary-shadowed platinum/carbon replica of an isolated, extracted and critical-point-dried rat uterine epithelial cell. Cell extraction was performed according to Fey et al. (1984a) in order to obtain nuclear matrix-intermediate filament scaffolds. Nuclear pore complexes are seen attached to intermediate filaments through short filamentous cross-bridges approximately 5 nm in diameter (arrows). B. Lateral view of the thin filaments, extending from the intermediate filaments to the nuclear matrix in a similar preparation as shown in A. The thin filaments are seen to abut on the nuclear lamina (arrows) and pore complexes (open arrow). Filaments with identical diameters link adjacent intermediate (cytokeratin) filaments (arrowheads). Bars, 0'2,i(m. (Courtesy of Dr M. Carmo-Fonseca & Dr A. Cidadao; see also reference CarmoFonseca*?/ al. (1988).)

18

R. Verheijen et al.

The first experiments that provided evidence for different protein compositions of these two structures were performed in rat liver by Berezney & Coffey (1977). Later, Todorov & Hadjiolov (1979) found five distinct protein bands with apparent molecular weights of 30, 40, 56, 70 and 82 (XlO3) enriched in the nueleolar matrix fraction. Distinct nueleolar matrix proteins from rat liver were found also upon two-dimensional gel electrophoresis by Comings & Peters (1981). They found many basic proteins specific for the nucleolus, the most prominent one being a 33K protein. The elegant studies of Franke et al. (1981a) led to the isolation of a nueleolar matrix from Xenopus oocyte nuclei comprising filaments of about 4nm diameter, which were densely coiled into higher-order fibrils of 30-40 nm diameter. This matrix was resistant to treatments with low-salt and high-salt buffers, DNases and RNases, sulphydryl agents and non-ionic detergents, and contained a single 145K/pI 6-2 protein (see also Benavente et al. 1984a,b). Furthermore, Olson & Thompson (1983) showed that a 160K polypeptide was predominantly found in the nueleolar matrix fraction from Novikoff hepatoma ascites cells prepared by digestion with DNase I and extraction with high-salt. Recent studies of Olson et al. (1986), also performed on Novikoff hepatoma cells, revealed a nueleolar matrix fraction that was enriched in polypeptides with molecular weights of 28, 37-5, 40, 70, 72, 110 and 160X103. The 110K protein was recognized by an antibody directed against protein C23 (Olson et al. 1981). About 25 % of the protein, 50 % of the RNA and less than 4 % of the DNA of untreated nucleoli were recovered in such nueleolar matrix preparations. Olson et al. (1986) also showed that pre-ribosomal RNP particles are major constituents of nueleolar matrix preparations extracted with DNase I and high-salt solutions. RNase A treatment during the DNase I digestion stage, together with the inclusion of 1 % /3-mercaptoethanol in the high-molarity salt washes, reduced the protein content to 15 % and the RNA content to about 2-5% of that of untreated nuclei. Under such conditions protein C23 and the 160K protein were also removed. The remaining polypeptides were largely represented in the 30-50K range, and electron microscopy revealed only amorphous material instead of the granular elements usually found in nueleolar matrix preparations (Olson et al. 1986). In preparing nueleolar matrices from mouse L-cells, Shiomi et al. (1986) used 50/xgmr 1 DNase I, a concentration five times higher than that used by Olson et al. (1986), and reproducibly obtained core nucleoli (the nueleolar fraction that remains after extensive DNase I action, without a high-salt extraction) with a minimum protein content. The nueleolar matrix fractions prepared by Shiomi et al. (1986) contained about 5 % of the amount of DNA present in isolated nucleoli,

16% of the RNA and less than 4 % of its original protein content, being enriched for proteins of 34K, 36K, 43K, 57K, lamins A and C (70 and 62K). Also higher molecular weight proteins, including a lOOK/pI 6-8 and a phosphorylated 160K/pI5-5 protein were found in these preparations. A portion of ribosomal spacer DNA remained tightly bound after treatment with2M-NaCl (see also Bollaef al. 1985). Shiomi et al. (1986) found the C23 protein to be quantitatively released from the nueleolar matrix by 2M-NaCl, which is in contrast to the results of Olson et al. (1986). This was verified with a specific antibody test indicating that the 110K protein was distinct from protein C23. The observations of several investigators that the nucleolus is often in close contact with the nuclear envelope (Rae & Franke, 1972; Goessens, 1979; Bouteillee/ al. 1982; Bourgeois et al. 1981, 1984) may explain the presence of lamins A and C in the nueleolar matrix preparations of Shiomi et al. (1986). Evidence for a nucleolusnuclear envelope junction in the form of a 'pedicle' or 'stalk' has been presented by Rae & Franke (1972) in mouse hepatocyte nuclei and more recently by Hubert et al. (1984), who isolated nucleoli containing nuclear shells from membrane-depleted rat liver nuclei (see also Bouviere/ al. 19856). Hubert e? al. (1984) showed the nucleoli to be anchored to the peripheral lamina by a pedicle that was continuous with an intranucleolar network. The pedicle and the network that supports the nueleolar DNA were composed mainly of nonhistone proteins insoluble in 2M-NaCl. In conclusion, the nueleolar residues in nuclear matrix preparations appear to be complex structures comprising several different elements, including many proteins not yet adequately characterized.

The internal matrix Morphological and biochemical aspects Several authors have provided evidence of the protcinaceous nature of the internal matrix, based on the fact that a fibrillar structure is observed in the nucleus when both DNA and RNA are removed by nucleases (Capco et al. 1982). Other authors (Galcheva-Gargova et al. 1982; Kaufmann & Shaper, 1984; Kaufmann et al. 1986) have observed a complete reorganization of the internal matrix structure, depending on the isolation conditions used. In this regard there is concern about aggregation or precipitation of otherwise soluble nuclear components, due to the high-salt concentrations required to extract chromatin during matrix preparation, or about cross-links formed by oxidation of sulphhydryl groups. Other studies indicated that the structural integrity of the internal network depends on the maintenance of The nuclear matrix 19

20

/?. Verheijen et al.

certain metalloprotein interactions during matrix isolation, e.g. with Ca 2+ and Cu 2 + (Lebkowski & Laemmli, 1982a,b) or with Mg 2 + (Bouvier et al. 1985a). These latter authors found the intranuclear structures in HeLa nuclear matrix preparations to be composed of residual (DNase- and salt-resistant) RNP complexes of both nucleolar and non-nucleolar origin. These intranuclear structures comprised two distinct but superimposed networks, which appeared as thin fibrillar elements (2-3 nm) and as thick fibrogranular elements of varying size. Both networks disappeared as a result of RNase digestion in the presence of low-ionic strength EDTA before extraction of nuclei in 2M-NaCl. However, the thick fibrillar elements were preserved from being eluted in 2 M-NaCl when RNase was used in buffers containing Mg 2 + . This remaining network was enriched in two proteins of 49K and 70K. From these results Bouvier et al. (1985c/) concluded that in the presence of Mg + interactions between certain RNP complexes are established, which then become able to form a salt-resistant intranuclear network. Herman et al. (1978) showed that after removal of 99 % of the chromatin in a two-step extraction procedure, both steady-state and newly synthesized hnRNA were associated with the remaining nuclear structure. Their suggestion that the integrity of the nuclear matrix is dependent on the RNA was in contrast with the conclusion of Miller et al. (1978a) that RNase treatment of the nuclear matrix does not alter the morphology of this network. Considering the results obtained by the many workers on nuclear matrix structure and composition, it will be obvious that the presence or absence of the internal network in nuclear matrix preparations depends on the experimental protocol used. The effects of divalent cations, the molarity of extraction buffers, the effect of high-salt treatment, the extent of disulphide cross-linking during preparation of the matrices, the order in which the various preparation steps are applied, the use of ( N H ^ S C ^ instead of NaCl for the extraction itself, the presence of endolytic enzymes other than proteases inhibited by phenylmethylsulphonyl fluoride (PMSF) or phenylmethylsulphonyl chloride (PMSC), and many other factors that are used for the preparation of the matrix structure have not yet Fig. 6. A. Whole-mount transmission electron micrographs of the nuclear matrix-intermediate filement (NM-IF) scaffold from a breast carcinoma cell line. The chromatin-depleted nuclear matrix (NM) is apparently in association with intermediate filaments (IF) largely consisting of cytokeratins. Note the nuclear pores (NP) present in the nuclear lamina. Bar, 0-5jum. B. Immunogold staining of intermediate filaments (IF) as described by Fey et al. (1984) in a similar preparation as in A, using anticytokeratin antibodies. Bar, 01 /*m. (Courtesy of Dr E. G. Fey.)

been studied sufficiently. Some of these technical problems have been tackled in recent papers but answers are still incomplete (Long & Ochs, 1983; Kaufmann & Shaper, 1984; Staufenbiel & Deppert, 1984; Fey et al. I984a,b, I986a,b; Mirkovitch et al. 1984; Verheijen et al. 1986a). Electron-microscopic studies performed by Pouchelet et al. (1986) have shown the existence of a welldefined network in nuclei of resting mouse lymphocytes in situ. These authors prepared nuclear matrices from formaldehyde-fixed cells. The nuclease step in such isolations was performed either with DNase only or with DNase in combination with RNase. In both types of preparations three well-defined networks were observed: the lamina, an intra-chromatin network and an inter-chromatin network. This latter structure could be superimposed on the internal network of isolated nuclear matrices. Its sensitivity to pepsin digestion was increased tenfold when the digestion was preceded by treatment with RNase. This latter finding indicates that RNA appears to be essential for the maintenance of this structure. Also Fey et al. (1986a) have shown that the morphology of the internal matrix changed drastically when RNases were used in the extraction buffers (see Fig. 7). In conclusion, the relevance of an internal matrix structure in vivo is still a matter of controversy and at present time many questions are still unanswered. It is beyond the scope of this review to compare all the different isolation procedures for the preparation of nuclear matrices or to discuss all the indications that support or deny the existence of an extranucleolar network in vivo. We merely conclude that considerable evidence indicates that when an internal matrix structure is obtained without prior treatment with RNase, hnRNA is found to be an integral constituent of this structure (Miller et al. 1978; Berezney, 1980; Brasch, 1982; Gallinaroe/ al. 1983; Long & Schrier, 1983; Fey et al. 1986a,b). A relevant question is whether the presence of an internal matrix is a general feature of cells. Using procedures similar to those published by Merman et al. (1978) and Miller et al. (1978a) a relatively stable RNAmatrix association can be found in various types of cells (Berezney, 1980; van Eekelen & van Venrooij, 1981; Brasch, 1982; Gallinaro et al. 1983; Long et al. 1979; Fey et al. \986a,b). The adult chicken erythrocyte nucleus, however, in which virtually no DNA or RNA synthesis takes place, was found to lack an internal nuclear matrix. Also, mild extraction procedures resulted in empty shells of pore complexlamina together with loose aggregates of core histones (Lafond & Woodcock, 1983). In contrast, rat liver nuclei showed a typical intranuclear salt-resistant skeleton after the same treatments. These results indicate that an internal matrix is not an obligatory nuclear The nuclear matrix

21

Fig. 7. Transmission electron micrographs of HeLa nuclear matrix preparations in unembedded resin-free sections (0-2fim thick) as described by Fey et al. (1986c;). A shows an RNP-containing nuclear matrix preparation, and B an RNP-depleted nuclear matrix. The RNP-containing nuclear matrix reveals fibres (F) that extends throughout the nucleus, forming continuous associations between nucleoli (Nu) and the nuclear lamina (L). Cytoplasmic filaments (Cy) are observed in association with the lamina. The RNP-depleted nuclear matrix displays a distortion of nuclear shape. The interior of the nuclear matrix is composed of condensed and fragmented filament aggregates (FA). The distortion of the interior by digestion with RNase suggests that RNA is an important structural component of the nucleus. Bars, l-Of/m. (Courtesy of

Dr E. G. Fey.) component and that in erythrocytes it is apparently not required for the spatial organization of chromatin. In contrast, much more active 5-day-old embryonic erythrocytes did contain an interchromatinous nuclear matrix (Lafond & Woodcock, 1983). Subsequent observations from the same group showed that an internal nuclear matrix is generated during the reactivation of chick crythrocyte nuclei in mouse L-cell cytoplasts. The nuclei enlarge and chromatin decondenses, accompanied by an influx of proteins from the host cytoplasm and the onset of RNA synthesis (Lafond & Woodcock, 1983). Recently, Woodcock & Woodcock (1986) have used the same experimental system to identify 15 major polypeptides that, after a 16-h reactivation period, had migrated into the nucleus. Five of the identified proteins in the 30-70K molecular weight range appeared to be nuclear matrix proteins; two of these had their counterparts in L-cell nuclei. During the concanavalin-A-induced stimulation of lymphocytes Setterfield et al. (1983) observed that the 22 R. Verheijen et al. nuclear volume increased up to sixfold, together with an extensive synthesis of stable interchromatinous matrix proteins. All these results suggest a correlation between the presence of nuclear matrix structures and nuclear 'activity'. In summary, the composition and organization of hnRNA as part of the granular and fibrous internal nuclear matrix structure still require more precise characterization. The results obtained to date permit the conclusion that when nuclear matrices are isolated in the absence of RNase, hnRNA can be isolated almost quantitatively as an intimate part of it. Considering the complicated composition of the nuclear matrix it will be obvious that it is not a static structure, but must display considerable dynamic activity. Heterogeneous nuclear RNP (hnRNP) particles Shortly after hnRNA has been synthesized it associates with proteins to form fibrillar ribonucleoprotein (RNP) particles and granules resembling 'beads on a

string' that extend away from the DNA-protein axis. These fibres (7 nm thick) and granules (20-25 nm diameter) are mostly referred to as the nuclear RNP network. hnRNP structures can be isolated from cell nuclei in a wide range of sedimentation values (30-250 S), depending on the isolation procedure applied (reviewed by LeStourgeon et al. 1981; Holoubek, 1984). The monomeric forms of these structures are hnRNP particles of 30-40 S. Next to an RNA fragment in the range of 125-800 nucleotides, these 40 S hnRNP particles contain a set of proteins that comprise 75-90% of its mass (Holoubek, 1984; Dreyfuss, 1986). Although non-specific binding of proteins to the RNA during the isolation of the particles has never been excluded, it is generally accepted that a distinct set of so-called core proteins is present in 40 S hnRNP particles. According to the nomenclature of Beyer et al. (1977), HeLa cells contain the following core proteins: A1(34K), A2(36K), B1(37K), B2(38K), C1(41K) and C2(43K) (reviewed by LeStourgeon et al. 1981; Holoubek, 1984; Dreyfuss, 1986). Proteins Cl and C2 appear to play a role in hnRNA processing, as a monoclonal antibody to these proteins inhibits in vitro splicing of an mRNA precursor, while depletion of these proteins from the splicing extract abolishes its capacity to splice pre-mRNA (Choi et al. 1986). Twodimensional gel separations of the core proteins have made it necessary to extend the nomenclature of Beyer et al. (1977), as has been done by Wilk<?/ al. (1985) and CeWs etal. (1986). Recently, Dreyfuss' group analysed the composition of hnRNP complexes obtained from HeLa cells by immunopurification with monoclonal antibodies. By two-dimensional gel electrophoresis they identified a set of over 24 proteins in the molecular weight range of 34-120K as consistent hnRNP components, of which the A, B and C proteins appeared to be just one subgroup. Chromatography on single-stranded DNA-agarose indicated that almost all these proteins are single-stranded nucleic acid binding proteins (G. Dreyfuss, personal communication). The A, B and C proteins are also bound to the matrix-associated hnRNA (van Eekelen & van Venrooij, 1981; Mariman et al. 1982a; Dreyfuss et al. 1984). Nuclear hnRNP particles can be isolated either by a lengthy low-salt extraction procedure or by sonication of isolated nuclei (reviewed by Holoubek, 1984). The extraction of hnRNP particles from intact nuclei is dependent on the action of nuclear RNases and on the slightly alkaline pH that is required for the release of the particles from nuclei. Since the nuclear matrix has been shown to be extremely susceptible to proteolytic activities (Miller et al. 1978a), it is likely that this procedure may also release RNP particles from the nucleus by proteolytic degradation of the nuclear structure. The sonication procedure, on the other

hand, releases RNP particles by the shearing forces employed. Faiferman & Pogo (1975) have shown that the yields of particles from disrupted nuclei are proportional to the shearing forces applied. It is evident that this procedure also destroys the delicate nuclear infrastructure. In studying the release of RNA from the nuclear matrix several authors have found that hnRNA is bound tenaciously to other matrix components and that it can be separated from the attaching structure only after disruption of the nuclear integrity (Longed al. 1979; van Eekelen & van Venrooij, 1981). In summary, it can be concluded that isolation of hnRNP particles necessarily implies the fragmentation of the internal nuclear matrix. As a consequence, depending on the procedure used, varying quantities of large hnRNP complexes and RNP core particles must have been present in nuclear matrix preparations isolated by various workers during the last decade. The complexity of the nuclear matrix and salt-resistant hnRNP structures has been compared by Gallinaro et al. (1983). Apart from small nuclear RNAs about 40 proteins in the 25-120K molecular weight range were characterized as common constituents of the nuclear matrix and the hnRNP particles. In addition, the premRNAs and maturation products present in both structures were compared. The results confirmed the similarity of the structures, strongly suggesting that pre-mRNA in the nuclear matrix and in the saltresistant complexes derived from hnRNP share a common constitutive unit (Gallinaro et al. 1983). Small nuclear RNP (snRNP) particles Eukaryotic cells contain a group of metabolically stable, capped RNAs known as U RNAs (for reviews, see Holoubek, 1984; Brunei et al. 1985). The major representatives are designated U1-U6 RNAs, while in the higher eukaryotes four minor U RNA species (U7-U10) have been described as well (Reddy et al. 1985). The U RNAs are found in discrete ribonucleoprotein particles, which all appear to play a role in the processing of pre-mRNA (reviewed by Padgett et al. 1986; Maniatis & Reed, 1987). In interphase cells the localization of U3 RNP (and probably U8 RNP) is restricted to the nucleolus, whereas the other U snRNP particles are mainly located in the nucleoplasm. The association of U RNAs with the nuclear matrix was first reported by Zieve & Penman (1976) and Miller et al. (19786), who demonstrated that, like hnRNA, small nuclear RNAs remained in the nuclear matrix after removal of nuclear membranes and chromatin. Similar observations were made by Herlan et al. (1979), Maundrell et al. (1981) and Ross et al. (1982) for Tetrahymena, duck erythroblast and chicken erythroblast nuclear matrices, respectively. Miller et al. (19786) and Gallinaro et al. (1983) found a quantitative association of all snRN As with the The nuclear matrix 23

nuclear matrix. Zieve & Penman (1976), however, characterized U2, U3, U4 and U6 RNA as being nuclear matrix-associated, whereas Ul RNA was mainly lost upon chromatin extraction. In contrast, Ciejek el al. (1982) have shown that only some of the U snRNAs were tightly associated with chicken oviduct matrices. They did not find a quantitative association nor a specific enrichment of one or more of the snRNAs. snRNPs can be released efficiently from isolated nuclei or from nuclear matrices by disintegration of the nuclear structure, for example by sonication. To a great extent the snRNPs are released in the form of 10 S snRNP particles; however, they are partly found in association with larger particles (30-60 S) containing hnRNA. Part of the snRNA is believed to be basepaired to these particles (reviewed by Padgett et al. 1986). In cross-linking experiments using 4'-aminomethyl-4,5',8-trimethylpsoralen (Pogo et al. 1982) it was shown that in chromatin-depleted nuclei snRNAs, mainly Ul and U6, can be cross-linked to hnRNA. Similar experiments on whole cells in vivo indicated that both Ul and U2 RNAs can be found base-paired to hnRNA (reviewed by Padgett et al. 1986). Although the latter experiments and the association of snRNA with hnRNP complexes suggest a close correlation between the structural organization of hnRNA and snRNA, Pogo and co-workers were able to remove snRNA from nuclear matrices by treatment with 1 % sodium deoxycholate, leaving the hnRNA in the structure (Pogo et al. 1982). This suggests that at least some of the snRNA and hnRNA are in different nuclear organizations. Sera from patients with connective-tissue diseases often contain antibodies directed against snRNP particles. Anti-Sm sera recognize the complete set of Ul, U2, U4, U5 and U6 (U1-U6) RNP particles, whereas anti-Ul RNP sera exclusively react with Ul RNPspecific proteins (Pettersson et al. 1984; Habets et al. 1985fl). Up to 12 polypeptides have now been identified as constituents of the snRNPs U1-U6, i.e. 70K, A(32K), A'(31K), B'(27K), B"(26K), B(25K), C(22K), D(16K), D'(15K), E(12K), F(11K) and G(9K) (Bringmann & Luhrmann, 1986). Proteins 70K, A and C are unique to the Ul RNP particle, whereas A' and B" are specific constituents of the U2 RNP particle. Vogclstein & Hunt (1982) showed that upon preparation of nuclear matrices of 3T3 cells a subset of the antigens recognized by anti-Sm sera was retained in the matrices. Judging from the intensities of immunofluorcsccnce, 65 % of the Sm antigens remained in the residual matrix after extraction with 2M-NaCl. Treatment of the matrices with RNase removed the Sm antigens, which implies that they are associated with 24 R. \ erheijen et al.

the matrix structure via RNA. When a human autoimmune serum containing anti-La antibodies was used as a control, immunodecoration of the matrices did not occur. This means that the La RNP particles (containing polymerase III transcripts such as pre-tRNAs and precursor forms of 7 S RNA and 5 S rRNA; Rinke & Steitz, 1982) were lost during matrix preparation. Similar experiments to immunodecorate isolated matrix structures with antibodies directed against U RNP antigens have been performed by van Eekelen el al. (1982), Spector <?/ al. (1983) and, more recently, by Reuter et al. (1984). They showed that several antigens recognized by these sera are almost quantitatively retained in nuclear matrices prepared from cultured cells. There is, however, some uncertainty as to whether these antigens can be removed by RNase treatment or not. An explanation for such discrepancies can probably be found in the fact that these human sera recognize different antigenic polypeptides in an snRNP particle. Cell fractionation studies performed by Habets et al. (19856) have shown that the Ul RNP-specific 70K antigen and the B'/B (U1-U6) RNA-associated antigens are more tightly complexed with the nuclear matrix compared with the other U RNP proteins, since substantial amounts of both the 70K and B'/B were resistant to subsequent treatments with detergents, DNase, RNase and high-salt solution (see also Verheijen et al. 19866). The other U snRNP-associated proteins, for example the A and D antigens, were readily extracted under these circumstances. Therefore, using a serum that contains a strong anti-70K activity one would expect an RNase-resistant immunofluorescence, as indeed has been found (Verheijen el al. 1986a). In contrast, a human serum that recognizes predominantly the Sm antigens will not decorate nuclear matrices as clearly after RNase treatment. The fact that some antigens, for example, the Ul RNA-specific 70K protein, are more tightly associated with the nuclear matrix compared with other U RNP proteins, suggests that such proteins can be involved in binding U RNAs to the hnRNAmatrix complex. Using immunoelectron microscopy with Sm antibodies on Novikoff hepatoma cells, Spector et al. (1983) found the snRNPs to be localized in a reticular nuclear network. This localization of snRNPs was altered when PtK2 cells had been treated with 5,6dichloro-1-jS-D-ribofuranosylbenzimidazole (DRB), which inhibits 60-75 % of the hnRNA synthesis. Drug-treated cells showed an accumulation or clumping of antigens recognized by anti-Sm antibodies in the central region of the nucleus. This effect on the distribution of snRNPs by inhibition of the hnRNA synthesis supports the assumption of a functional association between these two nuclear components (Spector etal. 1983).

Fakan et al. (1986) have applied immunoelectron microscopic techniques to mouse and Drosophila tissue-culture cells, using monoclonal antibodies directed against hnRNP core proteins or against U RNP proteins. Their studies provided direct evidence for an association of Ul RNP and possibly also of other U RNP species with extranucleolar RNA during early transcription elongation. In addition the results of these authors confirmed the presence of hnRNP proteins within the growing RNP chains in the transcription complexes. In summary, the limited number of studies performed on the interaction between snRNA or snRNA and nuclear matrix structures indicates a specific interaction between these two complexes. It will be of great interest to elucidate the nature of such an interaction and to establish its functional significance.

Nuclear actin Considerable evidence has accumulated over the past decade to show that actin is a constituent not only of the cytoplasm but also of interphase nuclei in a wide variety of cells (for a review, see Scheer et al. 1984). Actin has also been demonstrated as a major protein in manually isolated and cleaned nuclei of Amoeba and amphibian oocytes, in which concentrations of 3-4mgml~' can be found (Krohne & Franke, 1980; Gounon & Karsenti, 1981). Experiments by Manley et al. (1980) have indicated that at least part of the nuclear actin may be involved in RNA transcription. The transcription of proteincoding genes in eukaryotic systems is performed by RNA polymerase II, which in vitro requires supplementation with crude cellular extracts to initiate accurate transcription. Such cellular extracts contain multiple factors, some of them recognizing specific promotor elements such as the TATA box (Breathnach & Chambon, 1981). One of these factors has been purified and characterized as a protein that is strikingly similar to actin (Egly et al. 1984). In a completely different approach, Scheer et al. (1984) observed that injection of antibodies against actin into oocytes resulted in the cessation of transcription by RNA polymerase II, loop retraction and chromosome condensation. Moreover, even stronger inhibition was observed after injection of actin-binding proteins from different sources, such as fragmin from Physarum polycephalum and an actin modulator protein from mammalian smooth muscle. The idea that actin is involved in some way in RNA transcription is also supported by the finding that this protein is tightly associated with purified RNA polymerase II (Smith et al. 1979) and possibly involved in the regulation of poly(A) metabolism mediated by poly(A) polymerase (Schroder et al. 1982).

Additional evidence for a possible role of nuclear actin in RNA metabolism has been documented in several studies by Nakayasu & Ueda (1984, 1985, 1986). These authors have shown an interaction between pre-mRNAs and actin filaments in the nuclear matrix of mouse L-cells (Nakayasu & Ueda, 1985). In a previous study it had been shown that actin filaments are closely associated with small nuclear RNPs (Nakayasu & Ueda, 1984). Recently, the existence of two additional acidic species of actin in the nuclei of mouse L-cells were reported next to the two common /3-actins and y-actins (Nakayasu & Ueda, 1986). The most acidic actin (pi 5-1) was localized predominantly in the nuclear matrix. Other authors have also found actin as a major protein in nuclear matrix preparations (Capco et al. 1982; Staufenbiel & Deppert, 1984; Verheijen et al. 1986a; Nakayasu & Ueda, 1986). Although the possibility of contamination with cytoplasmic actin cannot be totally excluded, the observations described above justify the conclusion that actin may be a major and well-defined nuclear matrix protein that might have a defined function in RNA synthesis in vivo. Enzymes involved in DNA and RNA metabolism Enzymes involved in DNA and RNA metabolism that have been found in nuclear matrix fractions are numerous. These include DNA alpha and beta polymerases (Nishizawa et al. 1984; Smith et al. 1984; Foster & Collins, 1985), topoisomerase I (Nishizawa et al. 1984), topoisomerase II (Halligan et al. 1984; Berrios et al. 1985), RNA polymerase II (Lewis et al. 1984), poly(A) polymerase (Schroder et al. 1984), DNA methylase (Burdon et al. 1985) and DNA primase (Wood & Collins, 1986; Tubo & Berezney, 1987a,b). Vims-specific proteins It has been shown that the nuclear matrix is an important site of viral interaction (reviewed by Berezney, 1984; Simard et al. 1986). Viral DNA (see, e.g., Smith et al. 1985) and virus-specific proteins (see, e.g., Jones &Su, 1982; Bibor-Hardy et al. 1985; Khittoo et al. 1986) have been found enriched in nuclear matrix preparations. The bulk of large tumour antigen (large T) in simian virus 40 (SV40)-infected cells is present in three subnuclear locations: in the nucleoplasm, associated with the cellular chromatin and tightly bound to the nuclear matrix. Schirmbeck & Deppert (1987) have analysed the distribution of large T in lytically infected monkey cells and found that the amounts of large T associated with both chromatin and nuclear matrix increased markedly after transition from early to late phase of viral infection. The amount of nuclcoplasmic large T increased only slightly. During the course of infection large T accumulated in the chromatin and in The nuclear matrix 25

the nuclear matrix fraction, in parallel with the increase of viral DNA synthesis. Recent studies by the same group have indicated that the association of SV40 large T with the chromatin and the nuclear matrix is mediated by protein-protein interactions, rather than by sequence-specific DNA binding (Hinzpeter & Deppert, 1987).

Associated transcripts Newly synthesized SV40 RNA appears to be quantitatively associated with the nuclear matrix, while its processing and transport also appear to take place on this structure (Ben Ze'ev & Aloni, 1983; Abulafia et al. 1984). Similarly, influenza viral RNA sequences (Jackson et al. 1982) and the primary transcripts as well as the spliced RNA intermediates of adenovirus-specific genes (Mariman et al. 1982a; van Venrooij et al. 19826, 1985) have been shown to be tightly bound to the matrix structure. Rearrangements in the nuclear matrix morphology after infection with adenovirus type 2 have been demonstrated by the electron-microscopic studies of Zhonghe et al. (1987). Other studies concerning nuclear matrix-associated transcripts have been presented by Ciejek e al. (1982). RNA was isolated from oviduct nuclear matrices and analysed by hybridization to cloned probes for ovalbumin and ovomucoid mRNA. More than 95% of all of the precursors of these mRNAs, including various splicing intermediates, were associated with the matrix. Less than 50 % of the mature mRNA present in intact nuclei was recovered in the nuclear matrix. Schroder et al. (1987a) have also studied the release of total mRNA, as well as of specific high-abundance ovalbumin mRNA, from hen oviduct nuclear matrices. Their results confirmed the earlier findings of Ciejek et al. (1982), by demonstrating that ovalbumin premRNA was almost quantitatively associated with the oviduct nuclear matrix, whereas only one-third of the mature ovalbumin mRNA of whole nuclei was recovered in the nuclear matrix fraction. In addition, they showed that the binding of both pre-mRNA and matrix-bound mature mRNA displayed no difference in strength when the matrices were subjected to treatments with high-salt (3M-NaCl), urea (4M), detergent (2% Triton X-100) or EDTA (SOmM). The mature mRNA, however, was released selectively from the nuclear matrix by either ATP, AMP plus pyrophosphate, ADP or ATP analogues containing non-hydrolysable a,/3 or j6,y bonds. Whereas mRNA translocation through the nuclear pores is dependent on hydrolysis of ATP or GTP, mRNA release from the nuclear matrix apparently does not require hydrolysis of the ft,y phosphodiester bond. From these results Schroder et al. (1987a) suggested that the release of RNA might be caused by a conformational change of a nuclear matrix 26 R. Verheijen et al.

(or mRNP) component induced by ATP or its derivatives without cleavage of any high-energy bond. The hnRNA remained completely bound to the matrix in the presence of ATP. Furthermore, the release of mature mRNA by ATP could be strongly inhibited by various inhibitors of topoisomerase II, by a mechanism not yet understood. Other remarkable results were that both mature and pre-mRNA were released from the matrix structure in the presence of poly(A), ethidium bromide or the copper chelator 1,10-phenanthroline (Schroder et al. 1987a). The general conclusion reached by these authors was that nucleoplasmic RNA transport is apparently regulated not only during passage through the pore complex but also at the level of RNA release from the nuclear matrix. Although the possibility that the non-matrix-bound mRNAs in the studies of Ciejek et al. (1982) and Schroder et al. (1987a) are contaminations from cytoplasmic mRNAs cannot be completely ruled out, the data indicate that a substantial part of the processed mRNA in the nucleus is bound differently and not as tightly to the nuclear matrix as are the mRNA precursors. Since nuclear mRNAs are associated with a different set of proteins, as compared to cytoplasmic RNA molecules (van Eekelen et al. 19186; van Venrooij et al. 1982a), one would expect the selection of mature mRNA for nucleocytoplasmic transport to occur by means of exchange of hnRNP proteins (Dreyfuss, 1986) with a set of mRNA-binding proteins. Possible candidates for such proteins that could exchange with hnRNP proteins are the transport proteins described by Moffet & Webb (1983) or cytoplasmic mRNP proteins (Wagenmakers et al. 1980; Setyono & Greenberg, 1981; van Eekelen et al. 1981a; van Venrooij et al. 1982a). All these results support the concept that the nuclear matrix may be the structural site for RNA processing within the nucleus of eukaryotic cells. The nuclear matrix and RNA transport Recently, Schroder et al. (19876) have reviewed nucleoplasmic mRNA transport and discussed requirements for mRNA release. For this reason we will focus only on the fact that different RNA species have different rates of transportation. A general finding is that smaller RNAs are transported faster to the cytoplasm than the larger mRNAs. Newly synthesized globin mRNA (9 S), for example, is released into the cytoplasm about 5 min after initiation of its synthesis (Bastos & Aviv, 1977; Kinniburgh & Ross, 1979). This RNA is polyadenylated and spliced. Histone mRNA, which is non-polyadenylated, unspliced and similar in size to the globin mRNA, is released into the cytoplasm within 10 min (Adesnik & Darnell, 1972). Adenovirus pIX mRNA (about 9 S ,

polyadenylated and unspliced) reaches the cytoplasm within 4 min after the start of synthesis (Mariman et al. 19826). In contrast, the bulk of the late adenovirus mRNAs reach the cytoplasm only after 16 min (Mariman et al. 19826), a phenomenon that has been found for most cellular mRNAs as well (see, e.g., van Venrooij et al. 1975). The reason for these divergent rates of transportation between matrix-bound mRNAs is not understood. Probably most of the larger premRNAs require more complicated processing patterns, which means that the rate of transportation of mRNA depends mainly on the rate of maturation. Another possibility is that mRNAs emerging rapidly into the cytoplasm are not assembled into the usual hnRNP structures, as has been suggested by Pederson for transcripts lacking introns (Pederson, 1983). In this respect it should be mentioned that the final maturation step of mRNA, that is the binding of the typical cytoplasmic mRNA-associated proteins (Wagenmakers et al. 1980), is accompanied by the release of hnRNAassociated proteins (van Eekelen et al. 19816; van Venrooij et al. 1982a). It has also been established that mature mRNAs in the nucleus cannot be considered as integral components of the nuclear matrix, as are precursor RNAs (Ciejek et al. 1982; Schroder et al. 1987a). Behaviour of nuclear matrix components during mitosis The onset of mitosis is accompanied by an extensive rearrangement of nuclear components (see Fig. 8). As the cell approaches mitosis, the nucleolus first decreases in size and then disappears as the chromosomes condense and all RNA synthesis ceases. At prophase the lamins become highly phosphorylated (Ottaviano & Gerace, 1985), followed by dissassembling of the nuclear envelope. In immunofluorescence localization studies it has been shown that during prophase many nuclear (matrix) proteins shift from their distinct nuclear locations to a diffusely cytoplasmic distribution, excluding the condensed chromosomes (Chaly et al. 1984). At telophase this process is reversed. Such behaviour has been documented for several pore complex proteins (see, e.g., Davis & Blobel, 1986; Snow e( al. 1987), the lamins (see, e.g., Gerace, 1986; Verheijen et al. 19866), snRNP proteins (see, e.g., Reuter et al. 1985; Spector & Smith, 1986; Verheijen et al. 19866) and hnRNP proteins (see, e.g., Martin & Okamura, 1981). For some of these proteins the molecular state of association in mitotic cells has been investigated. Lamin B in Chinese hamster ovary (CHO) cells, for example, has been found to remain associated with phospholipid vesicles during mitosis, while lamins A and C are converted into a soluble form (Burke & Gerace, 1986). In a previous study, Lahiri &

Thomas (1985) found that the hnRNP core proteins in mitotic HeLa cells were not free but stably associated with high molecular weight RNA in the form of hnRNP particles that sedimented between 80 and 200 S. Recent studies by these authors (Lahiri & Thomas, 1987) showed that during mitosis at least 95 % of the total cellular snRNPs was also present in such hnRNP complexes, sedimenting at about 100 S. Other nuclear proteins, when investigated with immunocytochemical techniques, appear to be associated with the condensed chromosomes in mitosis (Chaly et al. 1984). Such behaviour has been specifically described for some nucleolar proteins (see, e.g., Pfeifle et al. 1986), topoisomerase I (see, e.g., Verheijen et al. 19866) and topoisomerase II (Earnshaw et al. 1985). The question arises as to how this group of nuclear (matrix) proteins is associated with the chromatin. The models for chromosome architecture that are in vogue have been reviewed by Earnshaw (1986). In one of these models a non-histone scaffold supposedly maintains the higher-order topological organization of DNA in mitotic chromosomes. Such scaffolds have been visualized in the electron microscope by Paulson & Laemmli (1977) after treating isolated mitotic HeLa chromosomes with dextran sulphate and heparin in order to remove the histones. Their preparations consisted of a subset of non-histone proteins attached to intact chromosomal DNA. The proteinaceous component or chromosome scaffold has been isolated from such structures after nuclease digestion and extraction of chromosomal proteins (Adolph et al. 1977; Lewis & Laemmli, 1982) and was initially proposed to be a rigid linear axial backbone in each chromatid, responsible for the morphology of metaphase chromosomes (Paulson & Laemmli, 1977). Among the scaffold proteins, Lewis & Laemmli (1982) found two prominent highmolecular weight polypeptides, Sc-1 and Sc-2, of about 170 and 135K, respectively. Using a polyclonal antibody that recognized chicken scaffold protein Sc-1, Earnshaw et al. (1985) have shown this polypeptide to be topoisomerase II (see also Gassere/ al. 1986). Other components that have been found in isolated chromosome scaffolds are centromeric proteins (Earnshaw et al. 1984). Chromosome scaffolds respond dramatically to changes in the ionic environment. Scaffolds isolated in the presence of 2 M-NaCl differ completely in appearance from those isolated at low ionic strength in the presence of dextran sulphate and heparin (Earnshaw & Laemmli, 1983), even though the protein composition of the two forms appears to be identical (Lewis & Laemmli, 1982). Exposure of isolated scaffolds to millimolar concentrations of Mg 2 + also causes a similar dramatic alteration in scaffold appearance (see Fig. 9; and references, Earnshaw & Laemmli, 1983). The nuclear matrix 27

Fig. 8. Immunofluorescent localization of various nuclear antigens in MR65 cells (human lung carcinoma) in interphase ( A l - D l ) and metaphase (A2-D2) with monoclonal antibodies directed against: A, the lamins (antibody 41CC4); B, the Ul RNP-specific 70K protein (antibody 2.73); C, the hnRNP-associated C proteins (antibody 4F4); D, a nucleolusassociated cell proliferation marker (antibody Ki-67). X1150.

'

'

Fig. 9. Electron micrographs of chromosome scaffolds prepared from highly purified HeLa mitotic chromosomes. A. Scaffold prepared at low ionic strength using a dextran sulphate/heparin lysis mix. B. Scaffold prepared at low ionic strength using a dextran sulphate/heparin lysis mix, followed by exposure to S mMMgC^. Centromere region, indicated by arrows. Bar, 1-0 jUm. (Courtesy of Dr W. C. Earnshaw; see also Earnshaw & Laemmh (1983).)

28

R. Verlieijen et al.

Hancock & Dessev (1987) have isolated chromosomes by lysis with Nonidet P40 in low ionic strength buffer without divalent cations or EDTA. These chromosomes slowly stretched to up to several times their original length while conserving an identifiable chromosome morphology (see Fig. 10). This finding is incompatible with a rigid nature of a supposed shapedetermining skeletal element. These authors further observed that each chromatid was not linear, but consisted of a spiraled fibre, many times the chromatid's length, which was unwound upon incubation of polyamine-stabilized chromosomes with deoxvcholate or diiodosalicylate (see Fig. 11). The winding of this fibre was suggested by Hancock & Dessev (1987) as being determined by protein-protein interactions between neighbouring gyres. These authors also concluded that if a skeletal element does exist in mitotic chromosomes it does not itself dictate the dimensions of the chromosome, but rather are the form and

dimensions of the skeletal element determined by the DNA associated with it. Using the same polyclonal antibody to chicken topoisomerase II as described above, Earnshaw & Heck (1985) have examined the distribution of this enzyme in intact, swollen but unextracted, chromosomes from MSB-1 chicken lymphoblastoid cells. Under the conditions used in their experiments, topoisomerase 11 was localized in numerous separate spots that appeared to be 120-200 nm across when covered with bound antibody. It was therefore concluded that these data did not provide evidence for the existence of a rigid corelike scaffold structure. In summary, despite the controversial experimental data, we think it is justifiable to conclude that, if present /// vivo, the chromosomal scaffold is not a rigid structure, but rather a component of the mitotic chromosome with dvnamic features.

Fig. 10. Mitotic chromosomes from CHO cells, prepared by lysis with 0-25% Nonidet P40 in 0-15 Msucrose/0-2M-phosphate, pH7-5. The lysate was stained with Hoechst 33258 and examined in suspension by fluorescence microscopy. The chromosomes showed a dramatically extended appearance (B,C) as compared to their initial shape (A). Magnifications in A, B and C are the same. Bar, 10;/m. (Courtesy of Dr R. Hancock; see also Hancock & Dessev (1988).)

The nuclear matrix

29

Fig. 11. Mitotic chromosomes from CIIO cells prepared in a polyamine-containing medium by lysis with Nonidet P40. Lithium diioclosalicylate was added to a final concentration of 25 mM and the chromosomes were immediately examined by fluorescent staining with Hoechst 33258, showing that chromatids are wound in helical gyres. Bar, lOjUm. (Courtesy of Dr R. Hancock; see also Hancock & Dessev (1988).) This study was supported by the Netherlands Cancer Foundation, grant no. NUKC 1984-11. The authors thank Professor Dr W. W. Franke (Heidelberg, FRG) for providing Fig. 2, Drs G. Brakenhoff and Dr R. van Driel (Amsterdam, The Netherlands) for Fig. 3, Dr V. BiborHardy (Sherbrooke/Montreal, Canada) for Fig. 4, Drs A. Cidadao and M. Carmo-Fonseca (Oeiras, Portugal) for Fig. 5, Drs E. G. Fey and S. Penman (Cambridge, USA) for Figs 6 and 7, Dr W. C. Earnshaw (Baltimore, MD, USA) for Fig. 9, and Dr R. Hancock (Quebec, Canada) for Figs 10 and 11. We also acknowledge the kind gifts of the monoclonal antibodies 41CC4 (from Dr G. Warren; Heidelberg, FRG), 2.73 (from Dr S. I loch; La Jolla, USA) and 4F4 (from Dr G. Dreyfuss; Evanston, USA).

ALEXANDER, R. B., GREENE, G. L. & BARRACK, E. R.

(1987). Estrogen receptors in the nuclear matrix: direct demonstration using monoclonal antireceptor antibody. Endocrinology 120, 1851-1857.
BARRACK, E. R. & COFFEY, D. S. (1982). Biological

properties of the nuclear matrix: steroid hormone binding. Recent Prog. Honn. Res. 38, 133-195. BASTOS, R. N. & Aviv, II. (1977). Globin RNA precursor molecules: biosynthesis and processing in crythroid cells. Cell 11, 641-650.
BENAVENTE, R., KROHNE, G., SCHMIDT-ZACHMANN, M. S.,

HOGLE, B. & FRANKE, W. W. (19846). Karyoskeletal

proteins and the organization of the amphibian oocyte nucleus. J. Cell Sci. Suppl. 1, 161-186.
BENAVENTE, R., KROHNE, G., STICK, R. & FRANKE, W. W.

References
AARONSON, R. P. & Woo, E. (1981). Organization in the cell nucleus: divalent cations modulate the distribution of condensed and diffuse chromatin. J. Cell Biol. 90, 181-186.
ABULAFIA, R., BEN-ZE'EV, A., HAY, N. & ALONI, Y.

(1984). Control of late SV40 transcription by the attenuation mechanism and transcriptionally active ternary complexes are associated with the nuclear matrix. J. moiec. Biol. 172, 467-487. ADESNIK, M. & DARNELL, J. E. (1972). Biogenesis and characterization of histone mRNA in HeLa cells. J. molec. Biol. 67, 397-406.
ADOLPH, K. W., CHENG, S. M. & LAEMMLI, U. K. (1977).

Role of nonhistone proteins in metaphase chromosomes. Cell 12, 805-816.

AEBI, U., COHN, J., BUHLE, L. & GERACE, L. (1986). The

(1984a). Electron microscopic immunolocalization of a karyoskeletal protein of molecular weight 145,000 in nucleoli and perinucleolar bodies of Xenopus laeris. Expl Cell Res. 151, 224-235. BEN-ZE'EV, A. & ALONI, Y. (1983). Processing of SV40 RNA is associated with the nuclear matrix and is not followed by the accumulation of low-molecular-wcight RNA products. Virology 125, 475-479. BEREZNEY, R. (1980). Fractionation of the nuclear matrix. I. Partial separation into matrix protein fibrils and a residual ribonucleoprotein fraction. J . Cell Biol. 85, 641-650. BEREZNEY, R. (1984). Organization and functions of the nuclear matrix. In Chmmosomal Nonhistone ProteinsStructural Associations, vol. IV (eel. L. S. Hnilica), pp. 119-180. Boca Raton, Florida: CRC Press.
BEREZNEY, R. & COFFEY, D. S. (1974). Identification of a

nuclear lamina is a meshwork of intermediate-type filaments. Nature, Land. 323, 560-564.

nuclear protein matrix. Bioclieni. biophys. Res. Coniinun. 60, 1410-1417.

30

R. Verheijen el al.

BEREZNEY, R. & COFFEY, D. S. (1977). Nuclear matrix: isolation and characterization of a framework structure from rat liver nuclei. J. Cell Biol. 73, 616-637.
BERRIOS, M., FILSON, A. J., BLOBEL, G. & FISHER, P. A.

neuroblastoma nuclei shown by confocal scanning laser microscopy. Nature, Land. 317, 748-749. BRASCH, K. (1982). Fine structure and localization of the nuclear matrix in situ. Expl Cell Res. 140, 161-171.
BREATHNACH, R. & CHAMBON, P. (1981). Organization and

(1983). A 174-kilodalton ATPase/dATPase polypeptide and a glycoprotein of apparently identical molecular weight are common but distinct components of higher eukaryotic nuclear structural protein subfractions. J. biol. Chem. 258, 13 384-13 390.
BERRIOS, M., OSHEROFF, N. & FISHER, P. A. (1985). In

expression of eukaryotic split genes coding for proteins. A. Rev. Biochem. 50, 349-383.
BRINGMANN, P. & LUHRMANN, R. (1986). Purification of

situ localization of DNA topoisomerase II, a major polypeptide component of the Drosophila nuclear matrix fraction. Proc. natn. Acad. Sci. U.S.A. 82, 4142-4146.
BEYER, A. L., CHRISTENSEN, M. E., WALKER, B. W. &

the individual snRNPs U l , U2, U5 and U4/U6 from HeLa cells and characterization of their protein constituents. EMBOJ. 5, 3509-3516.
BRUNEL, C , SRI-WIDADA, J. & JEANTEUR, P. (1985).

LESTOURGEON, W. M. (1977). Identification and characterization of the packaging proteins of core 40S hnRNP particles. Cell 11, 127-138.
BIBOR-HARDY, V., BERNARD, M. & SIMARD, R. (1985).

snRNPs and scRNPs in eukaryotic cells. In Prog, molec. subcell. Biol., vol. IX (ed. S. E. Hahn, D. J. Kopecko & W. E. G. Muller), pp. 1-52. Berlin: Springer-Verlag.
BUGLER, B., CAIZERGUES-FERRER, M., BOUCHE, G.,

BOURBON, H. & AMALRIC, F. (1982). Detection and

Nuclear matrix modifications at different stages of infection by herpes simplex virus type l.jf.gen. Virol. 66, 1095-1103. BLOBEL, G. (1985). Gene gating: a hypothesis. Proc. natn. Acad. Sci. U.S.A. 82, 8527-8529.
BOLLA, R. I., BRAATEN, D. C , SHIOMI, Y., HEBERT, M.

localization of a class of proteins immunologically related to a 100-kDa nucleolar protein. Eur. J. Biochem. 128, 475-480.
BURDON, R. H., QURESHI, M. & ADAMS, R. L. P. (1985).

B. & SCHLESSINGER, D. (1985). Localization of specific rDNA spacer sequences to the mouse L-cell nucleolar matrix. Molec. cell. Biol. 5, 1287-1294. BOULIKAS, T. (1986). Protein-protein and protein-DNA interactions in calf thymus nuclear matrix using crosslinking by ultraviolet irradiation. Biochem. Cell Biol. 64, 474-484.
BOURGEOIS, C. A., COSTAGLIOLA, D., LAQUERRIERE, F.,

Nuclear matrix-associated DNA methylase. Biochim. biophys. Ada 825, 70-79. BURKE, B. & GERACE, L. (1986). A cell-free system to study reassembly of the nuclear envelope at the end of mitosis. Cell 44, 639-652.
CAPCO, D. G., WAN, K. M. & PENMAN, S. (1982). The

nuclear matrix: three dimensional architecture and protein composition. Cell 29, 847-858.
CARMO-FONSECA, M., CIDADAO, A. J. & DAVID-FERREIRA,

BARD, F., HEMON, D. & BOUTEILLE, M. (1984). In situ

J. F. (1988). Filamentous cross-bridges link intermediate filaments to the nuclear pore complexes. Eiir.jf. Cell Biol. 45, 282-290.
CELIS, J. E., BRAVO, R., ARENSTORF, II. P. &

arrangement of non-bearing chromosomes in the interphase nucleus of Aotus trivirgatus. jf. Cell Sci. 69, 107-115.
BOURGEOIS, C. A., HEMON, D., BEAURE D'AUGERES, C ,

ROBINEAUX, R. & BOUTEILLE, M. (1981). Kinetics of

LESTOURGEON, W. M. (1986). Identification of proliferation-sensitive human proteins amongst components of the 40S hnRNP particles. FEBS Lett. 194, 101-109.
CHALY, N., BLADON, T., SETTERFIELD, G., LITTLE, J. E.,

nucleolus location within the nucleus by time-lapse microcinematography. Biol. Cell 40, 229-232.
BOUTEILLE, M., HERNANDEZ-VERDUN, D., DUPUY-COIN,

KAPLAN, J. G. & BROWN, D. L. (1984). Changes in

A. M. & BOURGEOIS, C. A. (1982). Nucleoli and nucleolar related structures in normal, infected and drugtreated cells. In The Nucleolus (ed. E. G. Jordan & C. A. Cullis), pp. 179-211. London, New York: Cambridge University Press.
BOUVIER, D., DUPUY-COIN, A.-M., BOUTEILLE, M. &

distribution of nuclear matrix antigens during the mitotic cell cycle. J. Cell Biol. 99, 661-671.
CHOI, Y. D., GRABOWSKI, P. J., SHARP, P. A. &

DREYFUSS, G. (1986). Heterogeneous nuclear ribonucleoproteins: role in RNA splicing. Science 231, 1534-1539.
CIEJEK, E. M., NORSTROM, J. L., TSAI, M.-J. &

MOENS, P. (1980). Three-dimensional electron microscopy of the nuclear matrix components of HeLa cells. Biol. Cell 39, 121-124.
BOUVIER, D., HUBERT, J., SEVE, A.-P. & BOUTEILLE, M.

O'MALLEY, B. W. (1982). Ribonucleic acid precursors are associated with the chick oviduct nuclear matrix. Biochemistry 21, 4945-4953.
COMINGS, D. E. & PETERS, K. E. (1981). Two-dimensional

(1985a). Nuclear RNA-associated proteins and their relationship to the nuclear matrix and related structures in HeLa cells. Can. J. Biochem. Cell Biol. 63, 631-643.
BOUVIER, D., HUBERT, J., SEVE, A.-P. & BOUTEILLE, M.

gel electrophoresis of nuclear particles. In The Cell Nucleus, vol. 9 (ed. H. Busch), pp. 89-118. New York: Academic Press.
DAGENAIS, A., BIBOR-HARDY, V., LALIBERTE, J.-F.,

(19856). Characterization of lamina-bound chromatin in the nuclear shell isolated from HeLa cells. Expl Cell Res. 156, 500-512.
BRAKENHOFF, G. J., VAN DER VOORT, H. T. M., VAN SPRONSEN, E. A., LINNEMANS, W. A. M. & NANNINGA,

ROYAL, A. & SIMARD, R. (1985). Detection in BHK cells of a precursor form for lamin A. Expl Cell Res. 161, 269-276.
DAVIS, L. 1. & BLOBEL, G. (1986). Identification and

N. (1985). Three-dimensional chromatin distribution in

characterization of a nuclear pore complex protein. Cell 45, 699-709.

The nuclear matrix

31

DIAMOND, D. A. & BARRACK, E. R. (1984). The

FILSON, A. J., LEWIS, A., BLOBEL, G. & FISCHER, P. A.

relationship of androgen receptor levels to androgen responsiveness in the Dunning R3327 rat prostate tumor sublines. J . Urol. 132, 821-827. DREYFUSS, G. (1986). Structure and function of nuclear and cytoplasmic ribonucleoprotein particles. A. Rev. Cell Hiol. 2, 459-498.
DREYFUSS, G., CHOI, Y. D. & ADAM, S. A. (1984).

(1985). Monoclonal antibodies prepared against the major Drosophila nuclear matrix-pore complex-lamina glycoprotein bind specifically to the nuclear envelope in situ. J. biol. Chem. 260, 3164-3172.
FISHER, D. Z., CHAUDHARY, N. & BLOBEL, G. (1986).

Characterization of heterogeneous nuclear RNA-protein complexes in vivo with monoclonal antibodies. Molec. cell. Rial. 4, 1104-1114. EARNSHAW, W. C. (1986). Mitotic chromosome structure: an update. In Cluvmosomal Proteins and Gene Expression (ed. G. R. Reeck, G. A. Goodwin & P. Puigdomenech), pp. 55-75. New York: Plenum.
EARNSHAW, W. C , HALLIGAN, B., COOKE, C. A., HECK,

cDNA sequencing of nuclear lamins A and C reveals primary and secondary structural homology to intermediate filament proteins. Proc. iiatn. Acad. Sci. U.S.A. 83, 6450-6454.
FOSTER, K. A. & COLLINS, J. M. (1985). The interrelation

M. M. S. & Liu, L. F. (1985). Topoisomerase II is a structural component of mitotic chromosome scaffolds. J. Cell Rial. 100, 1706-1715.
EARNSHAW, V V . C , HALLIGAN, N., COOKE, C. A. &

between DNA synthesis rates and DNA polymerases bound to the nuclear matrix in synchronized HeLa cells. J. biol. Chem. 260, 4229-4235. FRANKE, VV. VV. (1987). Nuclear lamins and cytoplasmic intermediate filament proteins: a growing multigene family. Cell 48, 3-4.
FRANKE, VV. VV. & SCHEER, U. (1970). The infrastructure

ROTHFIELD, N. F. (1984). The kinetochore is part of the metaphase chromosome scaffold. J. Cell Biol. 98, 352-357.
EARNSHAW, VV. C. & HECK, M. M. S. (1985). Localization

of the nuclear envelope of amphibian oocytes: a reinvestigation. I. The mature oocyte. J. Ultrastruct. Res. 30, 288-316.
FRANKE, V V . VV., KLEINSCHMIDT, J. A., SPRING, II., KROHNE, G., GRUND, C , TRENDELENBURG, M. F.,

of topoisomerase II in mitotic chromosomes. .7- Cell Biol. 100, 1716-1725.

EARNSHAW, VV. C. & LAEMMLI, U. K. (1983). Architecture

STOEHR, M. & SCHEER, U. (1981). A nucleolar skeleton of protein filaments demonstrated in amplified nucleoli of Xenopus laevis. J. Cell Biol. 90, 289-299.
FRANKE, V V . VV., SCHEER, U., KROHNE, G. & JARASCH, E.-

of metaphase chromosome scaffolds. J. Cell Biol. 96, 84-93.

EGLY, J. M., MIYAMOTO, N. G., MONCOLLIN, V. &

D. (19816). The nuclear envelope and the architecture of the nuclear periphery.^. Cell Biol. 91, 39s-50s.
GALCHEVA-GARGOVA, Z., PETROV, P. & DESSEV, G. N.

CHAMBON, P. (1984). Is actin a transcription initiation factor for RNA polymerase B? EMBOJ. 3, 2363-2371.
EPSTEIN, P., REDDY, R. & BUSCH, II. (1984). Multiple

(1982). Effect of chromatin decondensation on the intranuclear matrix. Eur.J. Cell Biol. 28, 155-159.
GALLINARO, H., PUVION, E., KISTER, L. & JACOB, M.

states of U3 RNA in Novikoff hepatoma nucleoli. Biochemistry 23, 5421-5425. FAIFERMAN, I. & POGO, A. O. (1975). Isolation of a nuclear ribonucleoprotein network that contains heterogeneous RNA and is bound to the nuclear envelope. Biochemistry 14, 3808-3816.
FAKAN, S., LESER, G. & MARTIN, T. E. (1986).

(1983). Nuclear matrix and hnRNP share a common structural constituent associated with premessenger RNA. EMBOJ. 2, 953-960.
GASSER, S. M., LAROCHE, T., FALQUET, J., DE LA TOUR,

Immunoelectron microscope visualization of nuclear ribonucleoprotein antigens within spread transcription complexes. J. Cell Biol. 103, 1153-1157.
FEY, E. G., CAPCO, D. G., KROCHMALNIC, G. & PENMAN,

E. B. & LAEMMLI, U. K. (1986). Metaphase chromosome structure. Involvement of topoisomerase II. J. molec. Biol. 186, 613-629.
GEORGATOS, S. D. & BLOBEL, G. (1987a). Two distinct

S. (19846). Epithelial structure revealed by chemical dissection and unembedded electron microscopy. J. Cell Biol. 99, 203s-208s.
FEY, E. G., KROCHMALNIC, G. & PENMAN, S. (1986).

attachment sites for vimentin along the plasma membrane and the nuclear envelope in avian erythrocytes: a basis for a vectorial assembly of intermediate filaments. J. Cell Biol. 105, 105-115.
GEORGATOS, S. D. & BLOBEL, G. (19876). Lamin B

The nonchromatin substructures of the nucleus: the ribonucleoprotein(RNP)-containing and RNP-depleted matrices analyzed by sequential fractionation and resinless section electron microscopy. J. Cell Biol. 102, 1654-1665.
FEY, E. G., ORNELLES, D. A. & PENMAN, S. (19866).

Association of RNA with the cytoskeleton and the nuclear matrix. J . Cell Sci. Suppl. 5, 99-119.
FEY, E. G., WAN, K. M. & PENMAN, S. (1984a).

constitutes an intermediate filament attachment site at the nuclear envelope. J . Cell Biol. 105, 117-125. GERACE, L. (1986). Nuclear lamina and organization of the nuclear envelope. Trends Biochem. Sci. 11, 443-446. GERACE, L. & BLOBEL, G. (1982). Nuclear lamina and organization of the nuclear envelope. Cold Spring Harbor Symp. quant. Biol. 46, 967-978.
GERACE, L., COMEAU, C. & BENSON, M. (1984).

Epithelial cytoskeletal framework and nuclear matrixintermediate filament scaffold: three-dimensional organization and protein composition. .7. Cell Biol. 98, 1973-1984.

Organization and modulation of nuclear lamina structure../. Cell Sci. Suppl. I, 137-160.
GERACE, L., OTTAVIANO, Y. & KONDOR-KOCH, C. (1982).

Identification of a major polypeptide of the nuclear pore complex. J . Cell Biol. 95, 826-837.

32

R. Verheijen et al.

GOESSENS, G. (1979). Relations between fibrillar centers and nucleol us-associated chromatin in Ehrlich tumor cells. Cell Biol. Int. Rep. 3, 337-343. GOESSENS, G. (1984). Nuleolar structure. Int. Rev. Cytol. 87, 107-158.
GOLDMAN, R., GOLDMAN, A., GREEN, K., JONES, J.,

HUBERT, J., BUREAU, J. & BOUTEILLE, M. (1984).

Anchorage of the nucleolus in the pore complex-lamina by a DNA-bearing structure masked in situ in rat liver nuclei. Biol. Cell 52, 91-102.
JACKSON, D. A., CATON, A. J., MCCREADY, S. J. & COOK,

LIESKA, N. & Y A N G , H.-Y. (1985). Intermediate

P. R. (1982). Influenza virus RNA is synthesized at fixed sites in the nucleus. Nature, Land. 296, 366-368.
JACKSON, D. A., MCCREADY, S. J. & COOK, P. R. (1984).

filaments: possible functions as cytoskeletal connecting links between the nucleus and the cell surface. Ann. N.Y. Accul. Sci. 455, 1-17. GOUNON, P. & KARSENTI, E. (1981). Involvement of contractile proteins in the changes in consistency of oocyte nucleoplasm of the newt Pleurodeles waltlii. J. Cell Biol. 88, 410-421.
HABETS, W. J., BERDEN, ] . H. M., HOCH, S. O. & VAN

Replication and transcription depend on attachment of DNA to the nuclear cage. J . Cell Sci. Suppl. I, 59-79. JONES, C. & Su, R. T. (1982). DNA polymcrase alpha from the nuclear matrix of cells infected with simian virus 40. Niicl. Acids Res. 10, 5517-5532. JORDAN, E. G. (1984). Nuleolar nomenclature..J. Cell Sci. 67, 217-220.
JORDAN, E. G. & CULLIS, C. A. (1982). The Nucleolus.

VENROOIJ, W. J. (1985fl). Further characterization and subcellular localization of Sm and Ul ribonucleoprotein antigens. Eur. J. Immun. 15, 992-997.
HABETS, W., HOET, M., BRINGMANN, P., LOHRMANN, R.

London, New York: Cambridge University Press.

KAUFMANN, S. H., FIELDS, A. P. & SHAPER, J. H. (1986).

& VAN VENROOIJ, W. (19856). Autoantibodies to ribonucleoprotein particles containing U2 small nuclear RNA. EMBOjf. 4, 1545-1550. HADJIOLOV, A. A. (1985). The Nucleolus and Ribosome Biogenesis, Cell Biology Monographs, vol. 12. New York: Springer-Verlag.
HALIJGAN, B. D., SMALL, D., VOGELSTEIN, B., HSIEH, T.-

The nuclear matrix: current concepts and unanswered questions. In Meth. Achiev. exp. Path. Xuclear Electmit Microscopy, vol. 12 (ed. G. Jasmin & R. Simard), pp. 141-171. Basel: Karger.
KAUFMANN, S. H. & SHAPER, J. H. (1984). A subset of

S. & Liu, L. E. (1984). Localization of type II DNA topoisomerase in nuclear matrix, jf. Cell Biol. 99, 128a.
HANCOCK, R. & BOULIKAS, T. (1982). Functional

non-histone nuclear proteins reversibly stabilized by the sulfhydryl cross-linking reagent tetrathionate. Expl Cell Res. 155, 477-495.
KHITTOO, G., DELORME, L., DERY, C. V., TREMBLAY, M. L., WEBER, J. M., BIBOR-HARDY, V. & SIMARD, R.

organization in the nucleus. Int. Rev. Cytol. 79, 165214. HANCOCK, R. & DESSEV, G. (1988). Nonhistone proteins and the organisation of eukaryotic DNA. In Progress in Nonhistone Protein Research, Florida: CRC Press (in press).
HERLAN, G., ECKERT, W. A., KAFFENBERGER, W. &

(1986). Role of the nuclear matrix in adenovirus maturation. Virus Res. 5, 391-403. KINNIBURGH, A. J. & Ross, J. (1979). Processing of the mouse /3-globin mRNA precursor: at least two cleavage-ligation reactions are necessary to excise the larger intervening sequence. Cell 17, 915-921.
KlRSCH, T . M . , MlLLER-DlENER, A. & LlTWACK, G .

WUNDERLICH, F. (1979). Isolation and characterization of an RNA-containing nuclear matrix from Tetrahyntena macronuclei. Biochemistry 18, 1782-1788.
HERMAN, R., WEYMOUTH, L. & PENMAN, S. (1978).

(1986). The nuclear matrix is the site of glucocorticoid receptor complex action in the nucleus. Biochem. biophys. Res. Commun. 137, 640-648.
KROHNE, G. & BENAVENTE, R. (1986). The nuclear

Heterogeneous nuclear RNA-protein fibers in chromatin-depleted nuclei. J . Cell Biol. 78, 663-674.
HINZPETER, M. & DEPPERT, W. (1987). Analysis of

biological and biochemical parameters for chromatin and nuclear matrix association of SV40 large T antigen in transformed cells. Oncogene 1, 119-129. HOLOUBEK, V. (1984). Nuclear ribonucleoproteins containing heterogeneous RNA. In Chromosomal Nonhistone Proteins-Structural Associations, vol. IV (ed. L. S. Hnilica), pp. 21-117. Boca Raton, Florida: CRC Press.
HOLT, G. D., SNOW, C. M., SENIOR, A., HALTIWANGER,

lamins: a multigene family of proteins in evolution and differentiation. Expl Cell Res. 162, 1-10. KROHNE, G. & FRANKE, W. W. (1980). A major soluble acidic protein located in nuclei of diverse vertebrate species. Expl Cell Res. 129, 167-189.
LAFOND, R. E. & WOODCOCK, C. L. F. (1983). Status of

the nuclear matrix in mature and embryonic chick erythrocyte nuclei. Expl Cell Res. 147, 31-39.
LAHIRI, D. K. & THOMAS, J. O. (1985). The fate of

R. S., GERACE, L. & HART, G. W. (1987). Nuclear pore complex glycoproteins contain cvtoplasmically disposed Olinked A'-acetylglucosamine. J. Cell Biol. 104, 1157-1164.
HUBERT, J., BOUVIER, D., ARNOULT, J. & BOUTEILLE, M.

(1981). Isolation and partial characterization of the nuclear shell of HeLa cells. Expl Cell Res. 131, 446-455.
HUBERT, J. & BOURGEOIS, C. A. (1986). The nuclear

heterogeneous nuclear ribonucleoprotein complexes during mitosis. J . biol. Chem. 260, 598-603. LAHIRI, D. K. & THOMAS, J. O. (1987). Small nuclear RNAs exist as large RNA-protein complexes during mitosis. In Abstracts of papers of the meeting on RNA processing 1987, Cold Spring Harbor, p. 107. New York: Cold Spring Harbor Press.
LALiBERTfi, J.-F., DAGENAIS, A., FILION, M., BIBORHARDY, V., SIMARD, R. & ROYAL, A. (1984).

skeleton and the spatial arrangement of chromosomes in the intcrphase nucleus of vertebrate somatic cells. Hum. Genet. 74, 1-15.

Identification of distinct messenger RNAs for nuclear lamin C and a putative precursor of nuclear lamin A. J. Cell Biol. 98, 980-985.

The nuclear matrix

33

LEBEL, S. & RAYMOND, Y. (1984). Lamin B from rat liver nuclei exists both as a lamina protein and as an intrinsic membrane protein. J. biol. Chem. 259, 2693-2696.
LEBEL, S. & RAYMOND, Y. (1987). Lamins A, B and C

Cell Nucleus, vol. 9 (ed. H. Busch), pp. 119-144. New York: Academic Press.
MAUNDRELL, K., MAXWELL, E. S., PUVION, E. &

share an epitope with the common domain of intermediate filament proteins. Expl Cell Res. 169, 560-565.
LEBKOWSKI, J. S. & LAEMMLI, U. K. (1982). Evidence

SCHERRER, K. (1981). The nuclear matrix of duck erythroblasts is associated with globin mRNA coding sequences but not with the major proteins of 40S nuclear RNP. Expl Cell Res. 136, 435-445.
MCKEON, F. D., KIRSCHER, M. W. & CAPUT, D. (1986).

for two levels of DNA folding in histone-depleted HeLa interphase nuclei. J. molec. Biol. 156, 309-324.
LEBKOWSKI, J. S. & LAEMMLI, U. K. (19826). Non-histone

Homologies in both primary and secondary structure between nuclear envelope and intermediate filament proteins. Nature, bond. 319, 463-468.
MILLER, T . E., HUANG, C.-Y. & POGO, A. O. (1978a).

proteins and long-range organization of HeLa interphase DNA. J. molec. Biol. 156, 325-344.
LEHNER, C. F., KURER, V., EPPENBERGER, H. M. & NIGG,

Rat liver nuclear skeleton and ribonucleoprotein complexes containing hnRNA. J. Cell Biol. 76, 675-691.
MILLER, T. E., HUANG, C.-Y. & POGO, A. O. (19786). Rat

E. A. (1986). The nuclear lamin protein family in higher vertebrates: identification of quantitatively minor lamin proteins by monoclonal antibodies. .7. biol. Chem. 261, 13293-13301.
LESTOURGEON, W. M., LOTHSTEIN, L., WALKER, B. W. &

liver nuclear skeleton and small molecular weight RNA species. J. Cell Biol. 76, 692-704.
MlRKOVITCH, J . , MlRAULT, M . - E . & LAEMMLI, U . K .

BEYER, A. L. (1981). The composition and general topology of RNA and protein in monomer 40S ribonucleoprotein particles. In The Cell Nucleus, vol. 9 (ed. H. Busch), pp. 49-87. New York: Academic Press. LEWIS, C. D. & LAEMMLI, U. K. (1982). High order metaphase chromosome structure: evidence for metalloprotein interactions. Cell 29, 171-181.
LEWIS, C. D., LEBKOWSKI, J. S., DALY, A. K. &

(1984). Organization of the higher-order chromatin loop: specific DNA attachment sites on nuclear scaffold. Cell 39, 223-232.
MOFFET, R. B. & WEBB, T . E. (1983). Relationship

LAEMMLI, U. K. (1984). Interphase nuclear matrix and metaphase scaffolding structures. .J. Cell Sci. Suppl. I, 103-122.
LONG, B. H., HUANG, C.-Y. & POGO, A. O. (1979).

Isolation and characterization of the nuclear matrix in Friend erythroleukemia cells: chromatin and hnRNA interactions with the nuclear matrix. Cell 18, 1079-1090. LONG, B. H. & OCHS, R. L. (1983). Nuclear matrix, hnRNA, and snRNA in Freund erythroleukemia nuclei depleted of chromatin by low ionic strength EDTA. Biol. Cell 48, 89-98. LONG, B. H. & SCHRIER, W. H. (1983). Isolation from Friend erythroleukemia cells of an RNase-sensitive nuclear matrix fibril fraction containing hnRNA and snRNA. Biol. Cell 48, 99-108. MANIATIS, T . & REED, R. (1987). The role of small nuclear ribonucleoprotein particles in pre-mRNA splicing. Nature, Land. 325, 673-678.
MANLEY, J. L., FIRE, A., CANO, A., SHARP, P. A. &

between the transport from isolated nuclei of two abundant cytoplasmic messengers and the source of a messenger RNA transport factor. Molec. Biol. Rep. 9, 227-230. NAKAYASU, H. & UEDA, K. (1984). Small nuclear RNA-protein complex anchors on the actin filaments in bovine lymphocyte nuclear matrix. Cell Struct. Funct. 9, 317-325. NAKAYASU, H. & UEDA, K. (1985). Association of rapidlylabelled RNAs with actin in nuclear matrix from mouse L5178Y cells. Expl Cell Res. 160, 319-330. NAKAYASU, H. & UEDA, K. (1986). Preferential association of acidic actin with nuclei and nuclear matrix from mouse leukemia L5178Y cells. Expl Cell Res. 163, 327-336.
NELSON, W. G., PIENTA, K. J., BARRACK, E. R. &

COFFEY, D. S. (1986). The role of the nuclear matrix in the organization and function of DNA. A. Rev. Biophys. biophys. Chem. 15, 457-475.
NEWPORT, J. W. & FORBES, D. J. (1987). The nucleus:

structure, function and dynamics. A. Rev. Biochem. 56,

535-565.
NISHIZAWA, M., TANABE, K. & TAKAHASHI, T . (1984).

GEFTER, M. L. (1980). DNA-dependent transcription of adenovirus genes in a soluble whole-cell extract. Proc.

DNA polymerases and DNA topoisomerases solubilized from nuclear matrices of regenerating rat livers. Biochem. biophys. Res. Commun. 124, 917-924.
OLSON, M. O. J., GUETZOW, K. & BUSCH, H. (1981).

naln. Acad. Sci. U.SA. 77, 3855-3859.

MARIMAN, E., HAGEBOLS, A.-M. & VAN VENROOU, W. J.

Localization of phosphoprotein C23 in nucleoli by immunological methods. Expl Cell Res. 135, 259-265.
OLSON, M. 0 . J. & THOMPSON, B. A. (1983). Distribution

(19826). On the localization and transport of specific adenoviral mRNA-sequences in the late infected HeLa cell. Nucl. Acids Res. 10, 6131-6145.
MARIMAN, E. C. M., VAN EEKELEN, C. A. G., REINDERS, R. J., BERNS, A. J. M. & VAN VENROOU, W. J. (1982a).

of proteins among chromatin components of nucleoli. Biochemistry 22, 3187-3193.

OLSON, M. O. J., WALLACE, M. O., HERRERA, A. H., MARSHALL-CARLSON, L. & HUNT, R. C. (1986).

Adenoviral heterogeneous nuclear RNA is associated with the host nuclear matrix during splicing. J. molec. Biol. 154, 103-119.
MARTIN, T . E. & OKAMURA, C. S. (1981).

Preribosomal ribonucleoprotein particles are a major component of a nucleolar matrix fraction. Biochemistry 25, 484-491.
OSBORN, M. & WEBER, K. (1987). Cytoplasmic

Immunochemistry of nuclear hnRNP complexes. In The

intermediate filament proteins and the nuclear lamins A,

34

R. Verheijen et al.

B and C share the IFA epitope. Expl Cell Res. 170, 195-203.
OTTAVIANO, Y. & GERACE, L. (1985). Phosphorylation of

nature of the nuclear pore complex material. Z. Zellforsch. mikmsk. anat. 127, 127-148.
SCHEER, U . , HlNSSEN, H . , FRANKE, W. W . & JOCKUSCH,

the nuclear lamins during interphase and mitosis. J. biol. Chem. 260, 624-632.
PADGETT, R. A., GRABOWSKI, P. J., KONARSKA, M. M.,

SEILER, S. & SHARP, P. A. (1986). Splicing of messenger RNA precursors. A. Rev. Biochem. 55, 1119-1150.
PAULSON, J. R. & LAEMMLI, U. K. (1977). The structure

B. M. (1984). Microinjection of actin-binding proteins and actin antibodies demonstrates involvement of nuclear actin in transcription of lampbrush chromosomes. Cell 39, 111-122.
SCHEER, U., KARTENBECK, J., TRENDELENBURG, M. F.,

STADLER, J. & FRANKE, W. W. (1976). Experimental

of histone-depleted metaphase chromosomes. Cell 12, 817-828. PEDERSON, T. (1983). Nuclear RNA-protein interactions and messenger RNA processing. J. Cell Biol. 97, 1321-1326. PETTERSON, I., HlNTERBERGER, M., MlMORI, T., GOTTLIEB, E. & STEITZ, J. A. (1984). The structure of mammalian small nuclear ribonucleoproteins. J. biol. Chem. 259, 5907-5914.
PFEIFLE, J., BOLLER, K. & ANDERER, F. A. (1986).

disintegration of the nuclear envelope. Evidence for pore-connecting fibrils. J. Cell Biol. 69, 1-18.
SCHIRMBECK, R. & DEPPERT, W. (1987). Specific

interaction of simian virus 40 large T antigen with cellular chromatin and nuclear matrix during the course of infection. J . Virol. 61, 3561-3569.
SCHRODER, H. C , BACHMANN, M., DIEHL-SEIFERT, B. &

MULLER, W. E. G. (19876). Transport of mRNA from nucleus to cytoplasm. Prog. Nucl. Acid Res. molec. Biol. 45, 98-142.
SCHRODER, H. C , NITZGEN, D. E., BERND, A., KURELEC, B., ZAHN, R. K., GRAMZOW, M. & MULLER, W. E. G.

Phosphoprotein ppl35 is an essential component of the nucleolus organizer region (NOR). Expl Cell Res. 162, 11-22. POGO, A. O., CORNUDELLA, L., GREBANIER, A. E.,
PROCYCK, R. & ZBRZEZNA, V. (1982). Cross-linking

experiments in nuclear matrix: nonhistone proteins to histones and snRNA to hnRNA. In The Nuclear Envelope and the Nuclear Matrix (ed. G. G. Maul), pp. 223-233. New York: Alan R. Liss, Inc.
POUCHELET, M., ANTEUNIS, A. & GANSMULLER, A. (1986).

(1984). Inhibition of nuclear envelope nucleoside triphosphate-regulated nucleocytoplasmic messenger RNA translocation by 9-jS-D-arabinofuranosyladeninc 5'triphosphate in rodent cells. Cancer Res. 44, 3812-3819.
SCHRODER, H. C , TROLLTSCH, D., FRIESE, U., BACHMANN, M. & MULLER, W. E. G. (1987C/). Mature

Correspondence of two nuclear networks observed in situ with the nuclear matrix. Biol. Cell 56, 107-112. RAE, P. M. M. & FRANKE, W. W. (1972). The interphase distribution of satellite DNA-containing heterochromatin in mouse nuclei. Chromosoma 39, 443-456.
RAZIN, S. V., CHERNOKHVOSTOV, V. V., YAROVAYA, O. V.

mRNA is selectively released from the nuclear matrix by an ATP/dATP-dependent mechanism sensitive to topoisomerase inhibitors. J . biol. Chem. 262, 8917-8925.
SCHRODER, H. C , ZAHN, R. K. & MULLER, W. E. G.

(1982). Role of actin and tubulin in the regulation of poly(A) polymerase-endoribonuclease IV complex from calf thymus. J. biol. Chem. 257, 2305-2309.
SETTERFIELD, G., HALL, R., BLADON, T., LITTLE, J. &

& GEORGIEV, G. P. (1985). Organization of the sites for DNA attachment to the nonhistone proteinaceous nuclear skeleton. In Prog. Nonhistone Protein Res., vol. II (ed. Bekhor), pp. 91-114. Boca Raton, Florida: CRC Press.
REDDY, R., HENNING, D. & BUSCH, H. (1985). Primary

and secondary structure of U8 small nuclear RNA. J. biol. Chem. 260, 10930-10935.
REUTER, R., APPEL, B., BRINGMANN, P., RINKE, J. &

KAPLAN, J. G. (1983). Changes in structure and composition of lymphocyte nuclei during mitogenic stimulation. J . Ultrastmct. Res. 82, 264-282. SETYONO, B. & GREENBERG, J. (1981). Proteins associated with poly A and other regions of mRNA and hnRNA molecules as investigated by cross-linking. Cell 24, 775-783.
SHAPER, J. H., PARDOLL, D. M., KAUFMANN, S. H., BARRACK, E. R., VOGELSTEIN, B. & COFFEY, D. S.

LUHRMANN, R. (1984). 5'-Terminal caps of snRNAs are reactive with antibodies specific for 2,2,7trimethylguanosine in whole cells and nuclear matrices. Expl Cell Res. 154, 548-560.
REUTER, R., APPEL, B., RINKE, J. & LUHRMANN, R.

(1979). The relationship of the nuclear matrix to cellular structure and function. In Adv. Enzyme Regulation, vol. 17 (ed. G. Weber), pp. 213-248. Oxford, New York: Pergamon Press.
SHIOMI, Y., POWERS, J., BOLLA, R. I., VAN NGUYEN, T. &

(1985). Localization and structure of snRNPs during mitosis. Expl Cell Res. 159, 63-79. RINKE, J. & STEITZ, J. A. (1982). Precursor molecules of both human 5S ribosomal RNA and transfer RNAs are bound by a cellular protein reactive with anti-La lupus antibodies. Cell 29, 149-159. Ross, D. A., YEN, R.-W. & CHAE, C.-B. (1982). Association of globin ribonucleic acid and its precursors with the chicken erythroblast nuclear matrix. Biochemistry 21, 764-771. SCHEER, U. (1972). The ultrastructure of the nuclear envelope of amphibian oocytes. IV. On the chemical

SCHLESSINGER, D. (1986). Proteins and RNA in mouse L-cell core nucleoli and nucleolar matrix. Biochemistn' 25, 5745-5751.
SIMARD, R., BIBOR-HARDY, V., DAGENAIS, A., BERNARD,

M. & PINARD, M. F. (1986). Role of the nuclear matrix during viral replication. Meth. Achiev. exp. Pathol. 12, 172-199.
SMITH, H. C , BEREZNEY, R., BREWSTER, J. M. & REKOSH,

D. (1985). Properties of adenoviral DNA bound to the nuclear matrix. Biochemistiy 24, 1197-1202.
SMITH, H. C , PUVION, E., BUCHHOLTZ, L. A. &

BEREZNEY, R. (1984). Spatial distribution of DNA loop

The nuclear matrix

35

attachment and replicational sites in the nuclear matrix. J.Cell Biol. 99, 1794-1802.
SMITH, S. S., KELLY, K. H. & JOCKUSCH, B. M. (1979).

adenovirus specific mRNA in the cytoplasm. I-'EBS Ix'tt. 145, 62-66.

VAN VENROOIJ, V V . J., VAN EEKELEN, C. A. G., MARIMAN,

Actin co-purifies with RNA polymerase II. Biochem. biophys. Res. Commun. 86, 161-166.
SNOW, C. M., SENIOR, A. & GERACE, L. (1987).

Monoclonal antibodies identify a group of nuclear pore complex glycoproteins.J. Cell Biol. 104, 1143-1156. SOMMERVILLE, J. (1986). Nucleolar structure and ribosome biogenesis. Trends Biochem. Sci. 11, 443-446.
SPECTOR, D. L., SCHRIER, W. H. & BUSCH, H. (1983).

E. C. M. & REINDERS, R. J. (19826). On the binding of host and viral RNA to the nuclear matrix. In The Nuclear Envelope and the Nuclear Matrix (cd. G. G. Maul), pp. 235-245. New York: Alan R. Liss.
VAN VENROOIJ, V V . J., VERHEIJEN, R. & MARIMAN, E. C.

Immunoelectron microscopic localization of snRNPs. Biol. Cell 49, 1-10.

SPECTOR, D. L. & SMITH, H. C. (1986). Redistribution of

(1985). Adenoviral hnRNA is associated with the host nuclear matrix during processing. In Viral Messenger Il\'A (ed. Y. Becker), pp. 147-163. Boston: Martinus Nijhoff.
VERHEIJEN, R., KUIJPERS, II., VOOIJS, P., VAN VENROOIJ,

VV. & RAMAEKERS, F. (1986O). Protein composition of

U-snRNPs during mitosis. Expl Cell Res. 163, 87-94.

STAUFENBIEL, M. & DEPPERT, W. (1982). Intermediate

nuclear matrix preparations from HeLa cells: an immunochemical approach..J. Cell Sci. 80, 103-122.
VERHEIJEN, R., KUIJPERS, II., VOOIJS, P., VAN VENROOIJ,

filament systems are collapsed onto the nuclear surface after isolation of nuclei from tissue culture cells. Expl Cell Res. 138, 207-214.
STAUFENBIEL, M. & DEPPERT, W. (1984). Preparation of

nuclear matrices from cultured cells: subfractionation of nuclei in situ. 7. Cell Biol. 98, 1886-1894.
STICK, R. & KROHNE, G. (1982). Immunological

localization of the major architectural protein associated with the nuclear envelope of the Xenopus laevis oocyte. Expl Cell Res. 138, 319-330.
TODOROV, I. T. & HADJIOLOV, A. A. (1979). A comparison

VV. & RAMAEKERS, F. (19866). Distribution of the 70K Ul RNA-associated protein during interphase and mitosis. J. Cell Sci. 86, 173-190. VOGELSTEIN, B. & HUNT, B. F. (1982). A subset of small nuclear ribonucleoprotein particle antigens is a component of the nuclear matrix. Biochem. biophys. Res. Commun. 105, 1224-1232.
VOGELSTEIN, B., SMALL, D., ROBINSON, S. & NELKIN, B.

of nuclear and nucleolar matrix proteins from rat liver. Cell Biol. Int. Rep. 3, 753-757.
TUBO, R. A. & BERZNEY, R. (1987a). Identification of 100

(1985). The nuclear matrix and the organization of nuclear DNA. In Prog. Nouhistone Protein Res., vol. II (ed. Bekhor), pp. 115-129. Boca Raton, Florida: CRC Press.
WAGENMAKERS, A. J. M., REINDERS, R. J. & VAN

and 150S DNA polymerase a-primase megacomplexes solubilized from the nuclear matrix of regenerating rat liver. J . biol. Chew. 262, 5857-5865. TUBO, R. A. & BEREZNEY, R. (19876). Nuclear matrixbound DNA primase to the nuclear matrix in HeLa cells. J . biol. Chew. 262, 6637-6642.
UNWIN, P. N. T. & MILLIGAN, R. A. (1982). A large

VENROOIJ, VV. J. (1980). Cross-linking of mRNA to proteins by irradiation of intact cells with ultraviolet light. Eur'.J. Biochem. 112, 323-330.
WILK, H..-E., VVERR, H., FRIEDRICH, D., KILTZ, II. II. &

particle associated with the perimeter of the nuclear pore complex. J . Cell Biol. 93, 63-75.
VAN EEKELEN, C. A. G., MARIMAN, E. C. M., REINDERS,

SCHAFER, K. P. (1985). The core proteins of 35S heterogeneous nuclear ribonucleoprotein complexes: characterization of nine different species. Eur. J. Biochem. 146, 71-81.
WOOD, S. I-I. & COLLINS, J. M. (1986). Preferential

R. J. & VAN VENROOIJ, W. J. (1981fl). Adenoviral heterogeneous nuclear RNA is associated with host cell proteins. Eur.J. Biochem. 119, 461-467.
VAN EEKELEN, C. A. G., RIEMEN, T. & VAN VENROOIJ, W.

binding of DNA primase to the nuclear matrix in HeLa cells. J . biol. Client. 261, 7119-7122.
WOODCOCK, C. L. F. & WOODCOCK, H. (1986). Nuclear

J. (19816). Specificity in the interaction of hnRNA and mRNA with proteins as revealed by in vivo crosslinking. EEBS Lett. 130, 223-226.
VAN EEKELEN, C. A. G., SALDEN, M. H. L., HABETS, W.

matrix generation during reactivation of avian erythrocyte nuclei: an analysis of the protein traffic in cybrids.J. Cell Sci. 84, 105-127.
ZEHNBAUER, B. A. & VOGELSTEIN, B. (1985). Supercoiled

J. A., VAN DE PlITTE, L. B. A. & VAN VENROOIJ, W. J. (1982). On the existence of an internal nuclear protein structure in HeLa cells. Expl Cell Res. 141, 181-190.
VAN EEKELEN, C. A. G. & VAN VENROOIJ, W. J. (1981).

loops and the organization of replication and transcription in eukaryotes. BioEssays 2, 52-54.
ZEITLIN, S., PARENT, A., SILVERSTEIN, S. & EFSTRATIADIS,

A. (1987). Pre-mRNA splicing and the nuclear matrix. Molec. cell. Biol. 7, 111-120.
ZHONGHE, Z., NICKERSON, J. A., KROCHMALNIC, G. &

hnRNA and its attachment to a nuclear protein matrix. J. Cell Biol. 88, 554-563.
VAN VENROOIJ, V V . J., GIELKENS, A. L. J., JANSSEN, A. P.

M. & BLOEMENDAL, H. (1975). Transport of messenger RNA into different classes of membrane-associated polyribosomes in Ehrlich-ascites-tumor cells. Eur. J. Biochem. 56, 229-238.
VAN VENROOIJ, V V . J., RIEMEN, T . & VAN EEKELEN,

PENMAN, S. (1987). Alterations in nuclear matrix structure after adenovirus infection. J . Viml. 61, 1007-1018. ZIEVE, G. & PENMAN, S. (1976). Small RNA species of the HeLa cell: metabolism and subcellular localization. Cell 8, 19-31. (Received 2 October 1987 - Accepted 25 January I9SS)

C. A. G. (1982ci). Host proteins are associated with

36

R. Verheijen el a\.