Portal de Periódicos da CAPES

The Tpr Protein: Linking Structure and Function in the Nuclear Interior?

1998; Elsevier BV; Volume: 63; Issue: 2 Linguagem: Inglês

10.1086/301989

1537-6605

Michael R. Paddy,

Cardiomyopathy and Myosin Studies

As the organelle housing the genome, the nucleus plays a fundamental role in the operation of the cell. A large number of essential and complex functions occur there, including DNA packaging and replication, RNA transcription, RNA processing, and mRNA transport. The number and complexity of these tasks has long led to suggestions of internal nuclear structures that organize and facilitate these functions (e.g., see Comings, 1968Comings D The rationale for an ordered arrangement of chromatin in the interphase nucleus.Am J Hum Genet. 1968; 20: 440-457PubMed Google Scholar). During the past several years, considerable new evidence has accumulated for nuclear functions occurring in discrete spatial domains (for a review, see Strouboulis and Wolffe, 1996Strouboulis J Wolffe AP Functional compartmentalization of the nucleus.J Cell Sci. 1996; 109: 1991-2000Crossref PubMed Google Scholar). These observations raise, once again, long-standing questions about whether extrachromosomal structures exist within the nucleus to spatially organize nuclear functions. Is the nucleus like a bag of chromosomal spaghetti, where nuclear functions occur only via freely diffusing factors that self-associate to form the observed nuclear domains? Or is the nucleus a very highly structured organelle where, for instance, a nucleoskeleton ties function to discrete spatial positions? Morphological and biochemical evidence has long suggested a nuclear skeleton or “nuclear matrix” that might organize nuclear functions. In this skeleton, networks of filamentous proteins provide structural continuity between the nuclear interior and the nuclear periphery. It is hypothesized that this skeleton provides binding sites for any number of functional complexes, and, indeed, many components of nuclear spatial domains have been found associated with nuclear matrix preparations (e.g., see Berezney et al., 1995Berezney R Mortillaro MJ Ma H Wei X Samarabandu J The nuclear matrix: a structural milieu for genomic function.Int Rev Cytol. 1995; 162A: 1-65PubMed Google Scholar). What has been missing is the identification of the molecular constituents of the nuclear matrix structures and the demonstration that the structures formed by these gene products play a role in nuclear function. It was against this background that my laboratory began studies of a large (262-kD) predicted filamentous protein in the nuclear interior, a protein now known as “Tpr.” Much to the surprise of my colleagues and me, both the characteristics of Tpr and recent work in the nuclear structure field combine to suggest an alternative formulation of a nuclear skeleton. Here, nuclear pore complex–associated filamentous proteins provide structural connectivity between the nuclear interior and nuclear periphery in the channels between chromosomes. This model is an attractive means of linking the major function of the nucleus—that is, gene expression—to known but poorly characterized structures in the nuclear interior. Furthermore, there is a direct path to testing this model by use of in vivo analysis tools such as genetic and in vivo imaging, which are readily availablein a metazoan model organism such as Drosophila melanogaster. Much of the controversy about internal nuclear structures that might organize function results from 2 decades of technically difficult biochemical approaches to electron micrograph (EM)–observable structures named “nuclear matrix” or “nuclear skeleton.” As first characterized by Berezney and Coffey in the mid-1970s (Berezney and Coffey, 1974Berezney R Coffey DS Identification of a nuclear protein matrix.Biochem Biophys Res Commun. 1974; 60: 1410-1417Crossref PubMed Scopus (772) Google Scholar, Berezney and Coffey, 1977Berezney R Coffey DS Nuclear matrix: isolation and characterization of a framework structure from rat liver nuclei.J Cell Biol. 1977; 73: 616-637Crossref PubMed Scopus (446) Google Scholar), nuclear matrices result from selective extraction procedures employing salts, detergents, and nucleases to produce an insoluble nuclear remnant. Many features of these remnants appear enticing. For example, whole-mount EMs can show fibers in the nuclear interior that are anastomosed with the nuclear lamina (reviewed by Nickerson and Penman, 1991Nickerson JA Penman S BioVision: microscopy in three dimensions.Semin Cell Biol. 1991; 2: 117-129PubMed Google Scholar). The internal nuclear fibers can have the width and repeat spacing expected for intermediate-filamentlike proteins (e.g., see Jackson and Cook, 1988Jackson DA Cook PR Visualization of a filamentous nucleoskeleton with a 23 nm axial repeat.EMBO J. 1988; 7: 3667-3677Crossref PubMed Scopus (154) Google Scholar). Two-dimensional protein gels of these remnants show both distinctive patterns of proteins in different cell types (e.g., see Fey and Penman, 1988Fey EG Penman S Nuclear matrix proteins reflect cell type of origin in cultured human cells.Proc Natl Acad Sci USA. 1988; 85: 121-125Crossref PubMed Scopus (275) Google Scholar) and shifting patterns of proteins during the development of cancers (e.g., see Partin et al., 1993Partin AW Getzenberg RH Carmichael MJ Vindivich D Yoo J Epstein JI Coffey DS Nuclear matrix protein patterns in human benign prostatic hyperplasia and prostate cancer.Cancer Res. 1993; 53: 744-746PubMed Google Scholar), much as one would expect if these proteins play a role in changing patterns of gene expression. However, historically the great difficulty in these biochemical approaches has been the wide divergence in the structures produced and in the proteins isolated by different groups (reviewed by Cook, 1988Cook PR The nucleoskeleton: artefact, passive framework or active site?.J Cell Sci. 1988; 90: 1-6PubMed Google Scholar). The scientific discussion has long been dominated by demonstrations of artifact in different preparation methods (e.g., see Kaufmann et al., 1986Kaufmann SH Fields AP Shaper JH The nuclear matrix: current concepts and unanswered questions.Methods Achiev Exp Pathol. 1986; 12: 141-171PubMed Google Scholar; Belgrader et al., 1991Belgrader P Siegel AJ Berezney R A comprehensive study on the isolation and characterization of the HeLa S3 nuclear matrix.J Cell Sci. 1991; 98: 281-291PubMed Google Scholar), so much so that the entire notion of any internal nuclear skeleton has often been called into question (e.g., see Cook, 1988Cook PR The nucleoskeleton: artefact, passive framework or active site?.J Cell Sci. 1988; 90: 1-6PubMed Google Scholar). This field is thus in the difficult position of having provided a strong indication of what an internal nuclear skeleton is likely to be like, while, at the same time, being unable to show unequivocally that such a structure exists and performs essential functions in the living cell. One way to unequivocally demonstrate the existence of a nuclear skeleton would be to identify specific proteins that form the expected matrix structures and that also perform demonstrable functions (e.g., via mutant genetic phenotypes). Cell biologists began on this course during the early 1990s, using traditional “reverse genetics” approaches. Surprisingly, most of the proteins identified have roles in RNA metabolism and no obvious skeletonlike characteristics. For example, van Driel's group (Mattern et al., 1996Mattern KA Humbel BM Muijsers AO de Jong L van Driel R hnRNP proteins and B23 are the major proteins of the internal nuclear matrix of HeLa S3 cells.J Cell Biochem. 1996; 62: 275-289Crossref PubMed Scopus (111) Google Scholar) recently completed a systematic study of the 21 most abundant proteins of the nuclear matrix of HeLa S3 cells. These proteins represent approximately three quarters of the mass of the matrix internal to the nuclear lamina. Of the 21 proteins, 16 are known hnRNP proteins, 1 is an abundant nucleolar protein (B23 or numatrin or nucleophosmin), and the remaining 4 are unidentified. Generally similar results have been obtained by other groups, with several SR-related proteins (e.g., see Blencowe et al., 1995Blencowe BJ Issner R Kim J Mccaw P Sharp PA New proteins related to the ser-arg family of splicing factors.RNA. 1995; 1: 852-865PubMed Google Scholar) and a hyperphosphorylated form of the RNA polymerase II large subunit (Mortillaro et al., 1996Mortillaro MJ Blencowe BJ Wei X Nakayasu H Du L Warren SL Sharp PA et al.A hyperphosphorylated form of the large subunit of RNA polymerase II is associated with splicing complexes and the nuclear matrix.Proc Natl Acad Sci USA. 1996; 93: 8253-8257Crossref PubMed Scopus (280) Google Scholar; Vincent et al., 1996Vincent M Lauriault P Dubois MF Lavoie S Bensaude O Chabot B The nuclear matrix protein p255 is a highly phosphorylated form of RNA polymerase II largest subunit which associates with spliceosomes.Nucleic Acids Res. 1996; 24: 4649-4652Crossref PubMed Scopus (77) Google Scholar) also being identified as matrix proteins. Although hnRNP proteins have long been known to be associated with the nuclear matrix, where are the filamentous proteins that should be at the core? Early candidates were the mammalian NuMA protein (reviewed by Cleveland, 1995Cleveland DW NuMA: a protein involved in nuclear structure, spindle assembly, and nuclear reformation.Trends Cell Biol. 1995; 5: 60-64Abstract Full Text PDF PubMed Scopus (55) Google Scholar) and the Nuf1p protein of budding yeast (Mirzayan et al., 1992Mirzayan C Copeland CS Snyder M The NUF1 gene encodes an essential coiled-coil related protein that is a potential component of the yeast nucleoskeleton.J Cell Biol. 1992; 116: 1319-1332Crossref PubMed Scopus (75) Google Scholar). Both of these proteins are localized to the nuclear interior and have the large coiled-coil secondary-structural motifs expected for filamentous proteins. However, the major function of both proteins appears to be in mitosis, with no obvious interphase function. In this void of candidate nuclear skeletal proteins, the Tpr protein (Byrd et al., 1994Byrd DA Sweet DJ Pante N Konstantinov KN Guan T Saphire ACS Mitchell PJ et al.Tpr, a large coiled coil protein whose amino terminus is involved in activation of oncogenic kinases, is localized to the cytoplasmic surface of the nuclear pore complex.J Cell Biol. 1994; 127: 1515-1526Crossref PubMed Scopus (97) Google Scholar; Cordes et al., 1997Cordes VC Reidenbach S Rackwitz H-R Franke WW Identification of protein p270/Tpr as a constitutive component of the nuclear pore complex-attached intranuclear filaments.J Cell Biol. 1997; 136: 1-15Crossref PubMed Scopus (184) Google Scholar; Zimowska et al., 1997Zimowska G Aris JP Paddy MR A Drosophila Tpr protein homolog is localized both in the extrachromosomal channel network and to nuclear pore complexes.J Cell Sci. 1997; 110: 927-944PubMed Google Scholar) has assumed the role of a protein to watch. Drosophila Tpr isolates exclusively with the nuclear matrix fraction of a traditional nuclear matrix preparation (Zimowska et al., 1997Zimowska G Aris JP Paddy MR A Drosophila Tpr protein homolog is localized both in the extrachromosomal channel network and to nuclear pore complexes.J Cell Sci. 1997; 110: 927-944PubMed Google Scholar). The nuclear interior-staining pattern of Drosophila Tpr is decidedly nonuniform, often having a linear or fibrous appearance when observed with immunofluorescence (Zimowska et al., 1997Zimowska G Aris JP Paddy MR A Drosophila Tpr protein homolog is localized both in the extrachromosomal channel network and to nuclear pore complexes.J Cell Sci. 1997; 110: 927-944PubMed Google Scholar). The Tpr cDNA sequence predicts two distinct structural domains: a large N-terminal domain (180 kD) strongly predicted to form a coiled-coil and an acidic C-terminus (82 kD) predicted to form a random coil (Byrd et al., 1994Byrd DA Sweet DJ Pante N Konstantinov KN Guan T Saphire ACS Mitchell PJ et al.Tpr, a large coiled coil protein whose amino terminus is involved in activation of oncogenic kinases, is localized to the cytoplasmic surface of the nuclear pore complex.J Cell Biol. 1994; 127: 1515-1526Crossref PubMed Scopus (97) Google Scholar; Zimowska et al., 1997Zimowska G Aris JP Paddy MR A Drosophila Tpr protein homolog is localized both in the extrachromosomal channel network and to nuclear pore complexes.J Cell Sci. 1997; 110: 927-944PubMed Google Scholar). The Tpr coiled-coil domain therefore may assemble into a skeletal structure, leaving the C-terminus free to interact with other macromolecules. The localization of Tpr within the nucleus also offers an unexpected twist: Drosophila Tpr is excluded from the chromosomal and nucleolar spaces but generally is found in all other regions (Zimowska et al., 1997Zimowska G Aris JP Paddy MR A Drosophila Tpr protein homolog is localized both in the extrachromosomal channel network and to nuclear pore complexes.J Cell Sci. 1997; 110: 927-944PubMed Google Scholar). Because these extrachromosomal regions are the major sites of RNA metabolism and transport within the nucleus (e.g., see Zachar et al., 1993Zachar Z Kramer J Mims IP Bingham PM Evidence for channeled diffusion of pre-mRNAs during nuclear RNA transport in metazoans.J Cell Biol. 1993; 121: 729-742Crossref PubMed Scopus (117) Google Scholar), it is worth considering a novel model of nuclear structure, in which Tpr, RNAs, and RNA-binding proteins fill the spaces between chromosomes. During the past decade, chromosome-painting methods have revealed a simple but profound fact: within the interphase nucleus, individual chromosomes occupy distinct, nonoverlapping regions (see Kurz et al., 1996Kurz A Lampel S Nickolenko JE Bradl J Benner A Zirbel RM Cremer T et al.Active and inactive genes localize preferentially in the periphery of chromosome territories.J Cell Biol. 1996; 135: 1195-1205Crossref PubMed Scopus (226) Google Scholar); that is, unlike the simplistic view from a traditional EM image, the euchromatic regions of different chromosomes do not intermingle within the nuclear interior but, rather, remain separate from each other in distinct volumes occupied by each chromosome. One immediate consequence is that the chromosomes themselves segregate the nuclear interior into a continuous region that is simply the spaces between the chromosomes. Such interchromosomal channels provide a continuous path for molecular exchange from the deep nuclear interior to pore complexes in the nuclear periphery, or vice versa (fig. 1a; also see, e.g., Bridger et al., 1998Bridger JM Herrmann H Munkel C Lichter P Identification of an interchromosomal compartment by polymerization of nuclear-targeted vimentin.J Cell Sci. 1998; 111: 1241-1253PubMed Google Scholar). For some time now, there have been demonstrations that these channels, alternately labeled the “interchromosomal channel domain” (Zirbel et al., 1993Zirbel RM Mathieu UR Kurz A Cremer T Lichter P Evidence for a nuclear compartment of transcription and splicing located at chromosome domain boundaries.Chromosome Res. 1993; 1: 93-106Crossref PubMed Scopus (194) Google Scholar) or the “extrachromosomal channel network” (Zachar et al., 1993Zachar Z Kramer J Mims IP Bingham PM Evidence for channeled diffusion of pre-mRNAs during nuclear RNA transport in metazoans.J Cell Biol. 1993; 121: 729-742Crossref PubMed Scopus (117) Google Scholar), concentrate activities required for mRNA metabolism (e.g., pre-mRNA splicing factors) and provide a nonrandom pathway for mRNA to be transported out of the nucleus (fig. 1a; for reviews, see Kramer et al., 1994Kramer J Zachar Z Bingham PM Nuclear pre-mRNA metabolism: channels and tracks.Trends Cell Biol. 1994; 4: 35-37Abstract Full Text PDF PubMed Scopus (30) Google Scholar; Razin and Gromova, 1995Razin SV Gromova II The channels model of nuclear matrix structure.BioEssays. 1995; 17: 443-450Crossref PubMed Scopus (80) Google Scholar; Strouboulis and Wolffe, 1996Strouboulis J Wolffe AP Functional compartmentalization of the nucleus.J Cell Sci. 1996; 109: 1991-2000Crossref PubMed Google Scholar). Given the initial evidence that most genes (active and inactive) appear to be localized at the edges of the chromosomal territories (Kurz et al., 1996Kurz A Lampel S Nickolenko JE Bradl J Benner A Zirbel RM Cremer T et al.Active and inactive genes localize preferentially in the periphery of chromosome territories.J Cell Biol. 1996; 135: 1195-1205Crossref PubMed Scopus (226) Google Scholar), these interchromosomal channels provide a simple means of concentrating and coordinating gene transcription, mRNA processing, and mRNA transport in the same region of the nuclear interior. Such interchromosomal channel networks need not exclude the existence of a nuclear skeleton or matrix. Indeed, Razin and Gromova, 1995Razin SV Gromova II The channels model of nuclear matrix structure.BioEssays. 1995; 17: 443-450Crossref PubMed Scopus (80) Google Scholar have proposed an alternative formulation of the nuclear matrix that places nuclear matrix proteins on the surfaces of chromosomes (as opposed to the centers of chromosomes), lining the interchromosomal channels (fig. 1b). This model makes it possible to reconcile all the available matrix data, particularly that which indicates that chromosomal DNA is organized as a series of loops tethered to a matrixlike structure. Thus, perhaps a better way to find a nuclear matrix is to seek the expected filamentous proteins in the spaces between the chromosomes. Classic EM studies of amphibian oocyte nuclei have long shown filaments extending from the inner face of nuclear pore complexes to a considerable distance into the nuclear interior (as far as 1 μm; e.g., see Franke and Scheer, 1970Franke WW Scheer U The ultrastructure of the nuclear envelope of amphibian oocytes: a reinvestigation. I. The mature oocyte.J Ultrastruct Res. 1970; 30: 288-316Crossref PubMed Scopus (97) Google Scholar). These filaments, which may stretch from pore complexes to the nucleolus, contain both the Tpr protein and another nuclear pore complex protein, Nup153 (Cordes et al., 1993Cordes VC Reidenbach S Kohler A Stuurman N van Driel R Franke WF Intranuclear filaments containing a nuclear pore complex protein.J Cell Biol. 1993; 123: 1333-1344Crossref PubMed Scopus (100) Google Scholar, Cordes et al., 1997Cordes VC Reidenbach S Rackwitz H-R Franke WW Identification of protein p270/Tpr as a constitutive component of the nuclear pore complex-attached intranuclear filaments.J Cell Biol. 1997; 136: 1-15Crossref PubMed Scopus (184) Google Scholar). Because Nup153 is implicated in nuclear export (Bastos et al., 1996Bastos R Lin A Enarson M Burke B Targeting and function in mRNA export of nuclear pore complex protein Nup153.J Cell Biol. 1996; 134: 1141-1156Crossref PubMed Scopus (165) Google Scholar), it seems likely that the long intranuclear filaments form a channel or act as tracks for movements of molecules between the nuclear interior and the pore complexes. Indeed, Ris, 1997Ris H High-resolution field-emission scanning electron microscopy of nuclear pore complex.Scanning. 1997; 19: 368-375Crossref PubMed Scopus (61) Google Scholar has observed these same Xenopus oocyte nuclei with high-resolution field-emission scanning EM and has found that the nucleoplasmic faces of pore complexes are interconnected by 50-nm channels formed from eight 6-nm filaments. Channels from several pore complexes merge to form a common channel that then runs deeper into the nuclear interior. Such channels deep in the nuclear interior could provide the filamentous proteins expected at the core of a nuclear matrix structure. A nuclear matrix model based on pore complex–associated intranuclear filaments in the interchromosomal channels is sketched in figure 1c. This sketch is intentionally vague, to emphasize the general notion rather than exhaustive details. This model is attractive because it uses filamentous structures known to exist and integrates nuclear structure with gene expression, a nuclear function of paramount importance. Furthermore, both the RNA metabolism function of most nuclear matrix proteins characterized to date and the RNase sensitivity of traditional matrix preparations are anticipated in this model. Finally, note that, because the nuclear matrix is generally observed only after DNase treatment, a matrix that exists only in the spaces between chromosomes, such as is shown in figure 1, would not be obvious in traditional preparations. Are the functions of Tpr consistent with a role in an interchromosomal channel–based nuclear matrix? Can even the more modest hypothesis of a Tpr role in nuclear transport be documented? Currently, Tpr functional studies are only beginning, and thus all the intriguing possibilities remain open. The initial functional evidence is that Tpr plays some role in intranuclear transport. As this review was being written, four groups were reporting data implicating Tpr either in the export of mRNA or in the import of proteins. Overexpression of either full-length mammalian Tpr or certain Tpr-deletion constructs in tissue-culture cells leads to accumulation of poly(A)+ RNA within the nucleus (B. Burke, personal communication). This may mean that some soluble factor required for export binds to the overexpressed Tpr and is not available for its usual function. Similarly, my laboratory has observed that overexpression of Drosophila Tpr full-length or deletion constructs in yeast, mammalian tissue-culture cells, and fly salivary glands leads to an accumulation of poly(A)+ RNA (G. Zimowska, V. Lamian, and M. R. Paddy, unpublished data). Surprisingly, however, my laboratory has yet to find conditions in which overexpression of Drosophila Tpr or of Drosophila Tpr–deletion constructs produces organismic death, aside from the trivial case of overexpressed protein blocking transport through pore complexes (G. Zimowska, V. Lamian, and M. R. Paddy, unpublished data). Similarly, Cordes has found that none of a large number of Tpr-deletion constructs leads to cell death when overexpressed in tissue-culture cells (V. Cordes, personal communication). Taken together, these results may indicate that Tpr plays only an accessory—and not an obligatory—role in nuclear export (i.e., it may facilitate but not be required for export). Alternately, it is possible that Tpr plays some other role upstream of mRNA export, a role that either remains unrecognized or is masked by the endogenous Tpr protein present in all these experiments. Clearly, there is a need to repeat these experiments in a Tpr genetic-null background. Surprisingly, Forbes's group finds that Tpr is a major physiological binding site for importin β, which, in combination with importin α, is required for import of classic nuclear localization signal–containing proteins (Shah et al., 1998Shah S Tugendreich S Forbes D Major binding sites for the nuclear import receptor are the internal nucleoporin Nup153 and the adjacent nuclear filament protein Tpr.J Cell Biol. 1998; 141: 31-49Crossref PubMed Scopus (166) Google Scholar). This result was unanticipated, because it is unclear why the nucleus would require that binding sites for importin β be at any place other than the nucleoplasmic face of the pore complex. Shah et al., 1998Shah S Tugendreich S Forbes D Major binding sites for the nuclear import receptor are the internal nucleoporin Nup153 and the adjacent nuclear filament protein Tpr.J Cell Biol. 1998; 141: 31-49Crossref PubMed Scopus (166) Google Scholar have hypothesized that Tpr participates either in binding of importin α/β after release of the nuclear localization signal–containing protein or else in the recycling of importin β back to the cytoplasm. Thus, Tpr may participate in nuclear trafficking in both directions. If Tpr acts in mRNA export, it touches on a classic debate: Does mRNA export from the nucleus occur along a defined path or track, or does it occur by simple diffusion (contrast the review of Xing and Lawrence, 1993Xing Y Lawrence JB Nuclear RNA tracks: structural basis for transcription and splicing?.Trends Cell Biol. 1993; 3: 346-353Abstract Full Text PDF PubMed Scopus (47) Google Scholar with that of Kramer et al., 1994Kramer J Zachar Z Bingham PM Nuclear pre-mRNA metabolism: channels and tracks.Trends Cell Biol. 1994; 4: 35-37Abstract Full Text PDF PubMed Scopus (30) Google Scholar)? Most theoretical and experimental studies suggest that mRNA transport occurs by simple diffusion (e.g., see Zachar et al., 1993Zachar Z Kramer J Mims IP Bingham PM Evidence for channeled diffusion of pre-mRNAs during nuclear RNA transport in metazoans.J Cell Biol. 1993; 121: 729-742Crossref PubMed Scopus (117) Google Scholar; Politz et al., 1998Politz JC Browne ES Wolf DE Pederson T Intranuclear diffusion and hybridization state of oligonucleotides measured by fluorescence correlation spectroscopy in living cells.Proc Natl Acad Sci USA. 1998; 95: 6043-6048Crossref PubMed Scopus (229) Google Scholar). However, Ris's recent images of pore complex–associated intranuclear channels evoke earlier work suggesting a defined path or track for RNA export (Xing and Lawrence, 1993Xing Y Lawrence JB Nuclear RNA tracks: structural basis for transcription and splicing?.Trends Cell Biol. 1993; 3: 346-353Abstract Full Text PDF PubMed Scopus (47) Google Scholar). If Tpr were part of a filamentous track or path for mRNA export, one would expect Tpr-containing intranuclear filaments to run continuously from pore complexes to deep into the nuclear interior. Currently, there is no evidence for this. Indeed, Drosophila appears unique in the extent to which Tpr is localized throughout the nuclear interior, not merely adjacent to pore complexes. Furthermore, the preliminary data from my laboratory suggest that, at certain developmental stages, Tpr is found in a form other than the intranuclear filaments. In some stages of larval salivary-gland development, Tpr is localized to small spherical structures that often border the nucleolus (Zimowska et al., 1997Zimowska G Aris JP Paddy MR A Drosophila Tpr protein homolog is localized both in the extrachromosomal channel network and to nuclear pore complexes.J Cell Sci. 1997; 110: 927-944PubMed Google Scholar; G. Zimowska and M. R. Paddy, unpublished data). This may represent a storage form of newly synthesized Tpr, Tpr complexed with pre-mRNA, or Tpr cycling on and off pore-associated intranuclear filaments. If these preliminary observations are confirmed, a simple, static intranuclear-filament structure for Tpr will not suffice. A dynamic nuclear matrix has often been proposed (e.g., see Mattern et al., 1996Mattern KA Humbel BM Muijsers AO de Jong L van Driel R hnRNP proteins and B23 are the major proteins of the internal nuclear matrix of HeLa S3 cells.J Cell Biochem. 1996; 62: 275-289Crossref PubMed Scopus (111) Google Scholar), and the true structure of a dynamic matrix might never be made clear in traditional static (fixed-sample) structural methods. In vivo imaging of either Tpr green-fluorescence protein chimeras or Tpr tagged with photoactivatable fluorescence probes should provide a direct path toward examination of any dynamic Tpr structures or any mRNA transport along an intranuclear filament. Although descriptive evidence continues to mount that many nuclear functions are organized in discrete spatial domains, the existence of a nuclear skeleton or matrix to direct this organization remains unproved. What I have suggested here is that the broad outlines of a nuclear matrix already may be evident if we relax our expectations of what a nuclear matrix must be (i.e., a rigid nuclear skeleton) and, instead, focus on two structures already known to exist—interchromosomal channels and pore complex–associated intranuclear filaments. Such a matrix model unifies transcription at the surfaces of chromosomes, mRNA metabolism, and nuclear transport, through common structural elements. Of course, at the heart of any nuclear matrix model must be the long-awaited filamentous proteins, to provide the expected skeletal functions. The fact that the Tpr protein, with its predicted large (∼180-kD) coiled-coil filamentous domain, has been localized to the pore-associated intranuclear filaments provides hope that, at last, one of these filamentous proteins has been isolated. Tpr now must run the gauntlet of molecular functional and structural analysis, to see whether this hope can be realized.