General organization of protein in HeLa 40S nuclear ribonucleoprotein particles

hnRNP A1 in RNA metabolism regulation and as a potential therapeutic target

Abnormal RNA metabolism, regulated by various RNA binding proteins, can have functional consequences for multiple diseases. Heterogeneous nuclear ribonucleoprotein A1 (hnRNP A1) is an important RNA binding protein, that regulates various RNA metabolic processes, including transcription, alternative splicing of pre-mRNA, translation, miRNA processing and mRNA stability. As a potent splicing factor, hnRNP A1 can regulate multiple splicing events, including itself, collaborating with other cooperative or antagonistical splicing factors by binding to splicing sites and regulatory elements in exons or introns. hnRNP A1 can modulate gene transcription by directly interacting with promoters or indirectly impacting Pol II activities. Moreover, by interacting with the internal ribosome entry site (IRES) or 3′-UTR of mRNAs, hnRNP A1 can affect mRNA translation. hnRNP A1 can alter the stability of mRNAs by binding to specific locations of 3′-UTR, miRNAs biogenesis and Nonsense-mediated mRNA decay (NMD) pathway. In this review, we conclude the selective sites where hnRNP A1 binds to RNA and DNA, and the co-regulatory factors that interact with hnRNP A1. Given the dysregulation of hnRNP A1 in diverse diseases, especially in cancers and neurodegeneration diseases, targeting hnRNP A1 for therapeutic treatment is extremely promising. Therefore, this review also provides the small-molecule drugs, biomedicines and novel strategies targeting hnRNP A1 for therapeutic purposes.

40S hnRNP particles are a novel class of nuclear biomolecular condensates

Heterogenous nuclear ribonucleoproteins (hnRNPs) are abundant proteins implicated in various steps of RNA processing that assemble on nuclear RNA into larger complexes termed 40S hnRNP particles. Despite their initial discovery 55 years ago, our understanding of these intriguing macromolecular assemblies remains limited. Here, we report the biochemical purification of native 40S hnRNP particles and the determination of their complete protein composition by label-free quantitative mass spectrometry, identifying A-group and C-group hnRNPs as the major protein constituents. Isolated 40S hnRNP particles dissociate upon RNA digestion and can be reconstituted in vitro on defined RNAs in the presence of the individual protein components, demonstrating a scaffolding role for RNA in nucleating particle formation. Finally, we revealed their nanometer scale, condensate-like nature, promoted by intrinsically disordered regions of A-group hnRNPs. Collectively, we identify nuclear 40S hnRNP particles as novel dynamic biomolecular condensates.

An analysis of the role of HnRNP C dysregulation in cancers

Heterogeneous nuclear ribonucleoproteins C (HnRNP C) is part of the hnRNP family of RNA-binding proteins. The relationship between hnRNP C and cancers has been extensively studied, and dysregulation of hnRNP C has been found in many cancers. According to existing public data, hnRNP C could promote the maturation of new heterogeneous nuclear RNAs (hnRNA s, also referred to as pre-mRNAs) into mRNAs and could stabilize mRNAs, controlling their translation. This paper reviews the regulation and dysregulation of hnRNP C in cancers. It interacts with some cancer genes and other biological molecules, such as microRNAs (miRNAs), long noncoding RNAs (lncRNAs), and double-stranded RNAs (dsRNAs). Even directly binds to them. The effects of hnRNP C on biological processes such as alternative cleavage and polyadenylation (APA) and N6-methyladenosine (m6A) modification differ among cancers. Its main function is regulating stability and level of translation of cancer genes, and the hnRNP C is regarded as a candidate biomarker and might be valuable for prognosis evaluation.

Heterogeneous Nuclear Ribonucleoproteins Involved in the Functioning of Telomeres in Malignant Cells

Heterogeneous nuclear ribonucleoproteins (hnRNPs) are structurally and functionally distinct proteins containing specific domains and motifs that enable the proteins to bind certain nucleotide sequences, particularly those found in human telomeres. In human malignant cells (HMCs), hnRNP-A1-the most studied hnRNP-is an abundant multifunctional protein that interacts with telomeric DNA and affects telomerase function. In addition, it is believed that other hnRNPs in HMCs may also be involved in the maintenance of telomere length. Accordingly, these proteins are considered possible participants in the processes associated with HMC immortalization. In our review, we discuss the results of studies on different hnRNPs that may be crucial to solving molecular oncological problems and relevant to further investigations of these proteins in HMCs.

Galectin-3 interacts with components of the nuclear ribonucleoprotein complex

Background

The multifunctional β-galactoside-binding protein galectin-3 is found in many distinct subcellular compartments including the cell nucleus. Expression and distribution of galectin-3 between the cell nucleus and the cytosol changes during cell differentiation and cancer development. Nuclear functions of galectin-3 and how they contribute to tumorigenesis are not understood.

Methods

In order to identify nuclear galectin-3 interaction partners, we used affinity chromatography and co-immunoprecipitation. Spatial proximity in the nucleus was assessed by immunofluorescence and proximity ligation assay. We also investigated the function of galectin-3 on mRNA-export by fluorescence in situ hybridization and on mRNA-processing by RNA-sequencing.

Results

The heterogeneous ribonucleoprotein particle component hnRNPA2B1 was identified as a novel galectin-3 binding protein that associates with the lectin in a lactose-dependent manner in the cell nucleus. Specific individual depletion of galectin-3 does not affect the mRNA distribution between cytoplasm and nucleus. A significant alteration of this distribution was observed after combined depletion of galectin-1 and −3. However, silencing of galectin-3 was sufficient to alter the splicing patterns of several genes.

Conclusions

Galectin-3 and hnRNPA2B1 interact as members of the early splicing machinery. Galectin-3 and −1 have redundant functions in mRNA transport and at least in part in mRNA splicing. RNA-sequencing data points to a specific function of the hnRNPA2B1/galectin-3 interaction in the processing of transcripts coding for the nuclear oncoprotein SET.

Electronic supplementary material

The online version of this article (doi:10.1186/s12885-016-2546-0) contains supplementary material, which is available to authorized users.

Evidence for cooperative tandem binding of hnRNP C RRMs in mRNA processing

The human hnRNP C is a ubiquitous cellular protein involved in mRNA maturation. Recently, we have shown that this protein specifically recognizes uridine (U) pentamers through its single RNA recognition motif (RRM). However, a large fraction of natural RNA targets of hnRNP C consists of much longer contiguous uridine stretches. To understand how these extended sites are recognized, we studied the binding of the RRM to U-tracts of 8-11 bases. In vivo investigation of internal translation activation of unr (upstream of N-ras) mRNA indicates that the conservation of the entire hnRNP C binding site, UC(U)8, is required for hnRNP C-dependent IRES activation. The assays further suggest a synergistic interplay between hnRNP C monomers, dependent on the protein's ability to oligomerize. In vitro spectroscopic and thermodynamic analyses show that isolated RRMs bind to (U)11 oligomers as dimers. Structural modeling of a ternary double-RRM/RNA complex indicates additionally that two RRM copies can be accommodated on the canonical sequence UC(U)8. The proposed tandem RRM binding is in very good agreement with the transcriptome-wide recognition of extended U-tracts by full-length hnRNP C, which displays a cross-linking pattern consistent with a positively cooperative RRM dimer binding model.

Heterogeneous nuclear ribonucleoprotein A1 in health and neurodegenerative disease: from structural insights to post-transcriptional regulatory roles

Heterogeneous nuclear ribonucleoproteins (hnRNPs) are a family of conserved nuclear proteins that associate with nascent RNA polymerase II transcripts to yield hnRNP particles, playing key roles in mRNA metabolism, DNA-related functions and microRNA biogenesis. HnRNPs accompany transcripts from stages of transcriptional regulation through splicing and post-transcriptional regulation, and are believed to affect the majority of expressed genes in mammals. Most hnRNP mRNA transcripts undergo alternative splicing and post-translational modifications, to yield a remarkable diversity of proteins with numerous functional elements that work in concert in their multiple functions. Therefore, mis-regulation of hnRNPs leads to different maladies. Here, we focus on the role of one of the best-known members of this protein family, hnRNP A1 in RNA metabolism, and address recent works that note its multileveled involvement in several neurodegenerative disorders. Initially discovered as a DNA binding protein, hnRNP A1 includes two RNA recognition motifs, and post-translational modifications of these and other regions in this multifunctional protein alter both its nuclear pore shuttling properties and its RNA interactions and affect transcription, mRNA splicing and microRNA biogenesis. HnRNP A1 plays several key roles in neuronal functioning and its depletion, either due to debilitated cholinergic neurotransmission or under autoimmune reactions causes drastic changes in RNA metabolism. Consequently, hnRNP A1 decline contributes to the severity of symptoms in several neurodegenerative diseases, including Alzheimer's disease (AD), spinal muscular atrophy (SMA), fronto-temporal lobar degeneration (FTLD), amyotrophic lateral sclerosis (ALS), multiple sclerosis (MS), hereditary spastic paraparesis (HSP) and HTLV-I associated myelopathy/tropical spastic paraparesis (HAM/TSP). At the translational level, these properties of hnRNP A1 led to massive research efforts aimed at developing RNA-targeted therapeutic tools such as splicing-modulating oligonucleotides with promising pharmaceutical potential. HnRNP A1 thus presents an intriguing example for the complexity and importance of heteronuclear ribonucleoproteins in health and disease. This article is part of a Special Issue entitled 'RNA and splicing regulation in neurodegeneration'.

HuR-hnRNP interactions and the effect of cellular stress

The heterogeneous nuclear ribonucleoproteins (hnRNPs) constitute an important group of RNA-binding proteins (RBPs) that play an active role in post-transcriptional gene regulation. Here, we focus on representative members of the hnRNP group of RBPs, namely hnRNP A1 and hnRNP C1/C2, which participate mainly in RNA splicing, as well as on HuR, a prototype of the AU-rich element-binding proteins (ARE-BP), which has an established role in regulating the stability and translation of target mRNAs. HuR and most hnRNPs are primarily localized in the nucleoplasm, and they can shuttle between the nucleus and the cytoplasm. Herein, we have extended our recently reported findings on the ability of HuR to associate with the immunopurified from mammalian cell extracts hnRNP and mRNP complexes by the application of an anti-HuR antibody that selects HuR-RNP complexes. We find that the protein components precipitated by the anti-HuR antibody are very similar to the hnRNP-HuR complexes reported previously. The in vivo association of HuR and hnRNP proteins is examined in the presence and the absence of thermal stress by confocal microscopy of intact cells and by in situ nuclear matrix preparation. We find notable heat-induced changes of HuR and of hnRNP A1, which exit the nucleus and co-localize to large cytoplasmic foci that represent heat-induced stress granules. The functional implications of HuR-hnRNP interactions in stressed and unstressed cells are discussed.

ELAV multimerizes on conserved AU4-6 motifs important for ewg splicing regulation

ELAV is a gene-specific regulator of alternative pre-mRNA processing in Drosophila neurons. Since ELAV/Hu proteins preferentially bind to AU-rich regions that are generally abundant in introns and untranslated regions, it has not been clear how gene specificity is achieved. Here we used a combination of in vitro biochemical experiments together with phylogenetic comparisons and in vivo analysis of Drosophila transgenes to study ELAV binding to the last ewg intron and splicing regulation. In vitro binding studies of ELAV show that ELAV multimerizes on the ewg binding site and forms a defined and saturable complex. Further, sizing of the ELAV-RNA complex and a series of titration experiments indicate that ELAV forms a dodecameric complex on 135 nucleotides in the last ewg intron. Analysis of the substrate RNA requirements for ELAV binding and complex formation indicates that a series of AU(4-6) motifs spread over the entire binding site are important, but not a strictly defined sequence element. The importance of AU(4-6) motifs, but not spacing between them, is further supported by evolutionary conservation in several melanogaster species subgroups. Finally, using transgenes we demonstrate in fly neurons that ELAV-mediated regulation of ewg intron 6 splicing requires several AU(4-6) motifs and that introduction of spacer sequence between conserved AU(4-6) motifs has a minimal effect on splicing. Collectively, our results suggest that ELAV multimerization and binding to multiple AU(4-6) motifs contribute to target RNA recognition and processing in a complex cellular environment.

A proteomic study of the arabidopsis nuclear matrix

The eukaryotic nucleus has been proposed to be organized by two interdependent nucleoprotein structures, the DNA-based chromatin and the RNA-dependent nuclear matrix. The functional composition and molecular organization of the second component have not yet been resolved. Here, we describe the isolation of the nuclear matrix from the model plant Arabidopsis, its initial characterization by confocal and electron microscopy, and the identification of 36 proteins by mass spectrometry. Electron microscopy of resinless samples confirmed a structure very similar to that described for the animal nuclear matrix. Two-dimensional gel electrophoresis resolved approximately 300 protein spots. Proteins were identified in batches by ESI tandem mass spectrometry after resolution by 1D SDS-PAGE. Among the identified proteins were a number of demonstrated or predicted Arabidopsis homologs of nucleolar proteins such as IMP4, Nop56, Nop58, fibrillarins, nucleolin, as well as ribosomal components and a putative histone deacetylase. Others included homologs of eEF-1, HSP/HSC70, and DnaJ, which have also been identified in the nucleolus or nuclear matrix of human cells, as well as a number of novel proteins with unknown function. This study is the first proteomic approach towards the characterization of a higher plant nuclear matrix. It demonstrates the striking similarities both in structure and protein composition of the operationally defined nuclear matrix across kingdoms whose unicellular ancestors have separated more than one billion years ago.

Cytoplasmic intermediate filaments are stably associated with nuclear matrices and potentially modulate their DNA-binding function

The tight association of cytoplasmic intermediate filaments (cIFs) with the nucleus and the isolation of crosslinkage products of vimentin with genomic DNA fragments, including nuclear matrix attachment regions (MARs) from proliferating fibroblasts, point to a participation of cIFs in nuclear activities. To test the possibility that cIFs are complementary nuclear matrix elements, the nuclei of a series of cultured cells were subjected to the Li-diiodosalicylate (LIS) extraction protocol developed for the preparation of nuclear matrices and analyzed by immunofluorescence microscopy and immunoblotting with antibodies directed against lamin B and cIF proteins. When nuclei released from hypotonically swollen L929 suspension cells in the presence of digitonin or Triton X-100 were exposed to such strong shearing forces that a considerable number were totally disrupted, a thin, discontinuous layer of vimentin IFs remained tenaciously adhering to still intact nuclei, in apparent coalignment with the nuclear lamina. Even in broken nuclei, the distribution of vimentin followed that of lamin B in areas where the lamina still appeared intact. The same retention of vimentin together with desmin and glial IFs was observed on the nuclei isolated from differentiating C2C12 myoblast and U333 glioma cells, respectively. Nuclei from epithelial cells shed their residual perinuclear IF layers as coherent cytoskeletal ghosts, except for small fractions of vimentin and cytokeratin IFs, which remained in a dot-to cap-like arrangement on the nuclear surface, in apparent codistribution with lamin B. LIS extraction did not bring about a reduction in the cIF protein contents of such nuclei upon their transformation into nuclear matrices. Moreover, in whole mount preparations of mouse embryo fibroblasts, DNA/chromatin emerging from nuclei during LIS extraction mechanically and chemically cleaned the nuclear surface and perinuclear area from loosely anchored cytoplasmic material with the production of broad, IF-free annular spaces, but left substantial fractions of the vimentin IFs in tight association with the nuclear surface. Accordingly, double-immunogold electron microscopy of fixed and permeabilized fibroblasts disclosed a close neighborhood of vimentin IFs and lamin B, with a minimal distance between the nanogold particles of ca. 30 nm. These data indicate an extremely solid interconnection of cIFs with structural elements of the nuclear matrix, and make them, together with their susceptibility to crosslinkage to MARs and other genomic DNA sequences under native conditions, complementary or even integral constituents of the karyoskeleton.

An antiparallel four-helix bundle orients the high-affinity RNA binding sites in hnRNP C: a mechanism for RNA chaperonin activity

Previous studies have shown that the C protein of 40 S hnRNP complexes contains a leucine-zipper domain, residues 180-207, and that a 40 residue highly basic domain, immediately preceding the zipper, is responsible for almost all of the free energy of RNA binding to C protein. Because this domain arrangement is like that seen in the bZIP transcription factors it has been termed the bZIP-like-motif or bZLM. We report here that the zipper domain drives C protein oligomerization through its spontaneous assembly into an anti-parallel four-helix bundle approximately 50 A in length. The anti-parallel nature of the four-helix bundle positions the tetramer's four high-affinity RNA binding domains at opposing ends of a rigid core formed by the helix bundle. This domain topology is ideally suited to accommodate and direct a double wrapping of RNA around the tetramer and is fully consistent with C protein's ability to bind and order 230 nt lengths of pre-mRNA through a highly cooperative RNA binding mode. We have used a novel sequence-specific 13C/15N labeling strategy and multidimensional NMR spectroscopy to define the anti-parallel orientation of the four-helix bundle and its molecular dimensions. In vitro reconstitution and hydrodynamic studies on native C protein, on several C protein fragments, and on various synthetic peptides, are consistent with the proposed model and indicate that C protein's canonical RNA recognition motifs probably function in tetramer-tetramer interactions during 40 S hnRNP assembly.

Heterogeneous nuclear ribonucleoproteins C1 and C2 associate with the RNA component of human telomerase

Here we demonstrate that heterogeneous nuclear ribonucleoproteins (hnRNPs) C1 and C2 can associate directly with the integral RNA component of mammalian telomerase. The binding site for hnRNPs C1 and C2 maps to a 6-base uridylate tract located directly 5' to the template region in the human telomerase RNA (TR) and a 4-base uridylate tract directly 3' to the template in the mouse TR. Telomerase activity is precipitated with antibodies specific to hnRNPs C1 and C2 from cells expressing wild-type human TR but not a variant of the human TR lacking the hnRNPs C1 and C2 binding site, indicating that hnRNPs C1 and C2 require the 6-base uridylate tract within the human TR to associate with the telomerase holoenzyme. In addition, we demonstrate that binding of hnRNPs C1 and C2 to telomerase correlates with the ability of telomerase to access the telomere. Although correlative, these data do suggest that the binding of hnRNPs C1 and C2 to telomerase may be important for the ability of telomerase to function on telomeres. The C proteins of the hnRNP particle are also capable of colocalizing with telomere binding proteins, suggesting that the C proteins may associate with telomeres in vivo. Therefore, human telomerase is capable of associating with core members of the hnRNP family of RNA binding proteins through a direct and sequence-specific interaction with the human TR. This is also the first account describing the precise mapping of a sequence in the human TR that is required to associate with an auxiliary component of the human telomerase holoenzyme.

Nuclear matrix-like filaments and fibrogranular complexes form through the rearrangement of specific nuclear ribonucleoproteins

The behavior of nuclear pre-mRNA-binding proteins after their nuclease and/or salt-induced release from RNA was investigated. After RNase digestion or salt extraction, two proteins that initially exist as tetramers (A2)(3)B1 in isolated heterogeneous nuclear ribonucleoprotein (hnRNP) complexes quantitatively reassociated to form regular helical filaments ranging in length from 100 nm to >10 microm. In highly magnified preparations prepared for scanning transmission electron microscopy, single filaments have diameters near 18 nm. In conventional negatively stained preparations viewed at low magnification, the diameters of the thinnest filaments range from 7 to 10 nm. At protein concentrations of >0.1 mg/ml, the filaments rapidly aggregated to form thicker filamentous networks that look like the fibrogranular structures termed the "nuclear matrix." Like the residual material seen in nuclear matrix preparations, the hnRNP filaments were insoluble in 2 M NaCl. Filament formation is associated with, and may be dependent on, disulfide bridge formation between the hnRNP proteins. The reducing agent 2-mercaptoethanol significantly attenuates filament assembly, and the residual material that forms is ultrastructurally distinct from the 7- to 10-nm fibers. In addition to the protein rearrangement leading to filament formation, nearly one-third of the protein present in chromatin-clarified nuclear extracts was converted to salt-insoluble material within 1 min of digestion with RNase. These observations are consistent with the possibility that the residual material termed the nuclear matrix may be enriched in, if not formed by, denatured proteins that function in pre-mRNA packaging, processing, and transport.

Half a century of "the nuclear matrix"

A cell fraction that would today be termed "the nuclear matrix" was first described and patented in 1948 by Russian investigators. In 1974 this fraction was rediscovered and promoted as a fundamental organizing principle of eukaryotic gene expression. Yet, convincing evidence for this functional role of the nuclear matrix has been elusive and has recently been further challenged. What do we really know about the nonchromatin elements (if any) of internal nuclear structure? Are there objective reasons (as opposed to thinly veiled disdain) to question experiments that use harsh nuclear extraction steps and precipitation-prone conditions? Are the known biophysical properties of the nucleoplasm in vivo consistent with the existence of an extensive network of anastomosing filaments coursing dendritically throughout the interchromatin space? To what extent may the genome itself contribute information for its own quarternary structure in the interphase nucleus? These questions and recent work that bears on the mystique of the nuclear matrix are addressed in this essay. The degree to which gene expression literally depends on nonchromatin nuclear structure as a facilitating organizational format remains an intriguing but unsolved issue in eukaryotic cell biology, and considerable skepticism continues to surround the nuclear matrix fraction as an accurate representation of the in vivo situation.

Analysis of RNA-protein interactions of mouse liver cytochrome P4502A5 mRNA

In our previous studies we have identified a 37/39 kDa, pyrazole-inducible, cytochrome P4502A5 (CYP2A5) mRNA binding protein and provided evidence that it may play a role in the stabilization and processing of the RNA [Geneste, Rafalli and Lang (1996) Biochem. J. 313, 1029-1037; Thulke-Gross, Hergenhahn, Tilloy-Ellul, Lang and Bartsch (1998) Biochem. J. 331, 473-481]. Details of the RNA-protein interactions are, however, not known. In this report we have performed an analysis of the interaction between the CYP2A5 mRNA and the 37/39 kDa protein. With UV-cross linking experiments, using RNA probes corresponding to various parts of the CYP2A5 mRNA, and with antisense oligonucleotides complementary to certain areas of the 3'-untranslated region (3'UTR), we could map the primary binding site to the tip of a 71 nt hair-pin loop at the 3'-UTR. This analysis also showed that the protein may have more than one site of interaction with the RNA and/or that, within the binding region, there could be more than one protein molecule binding to the RNA. Analysis of the probable conformations of the various probes used in the UV cross-linking experiments, in combination with the estimated binding affinities of the protein to the different probes, suggests that important factors in the high-affinity binding are the UAG triplet flanked by GA-rich sequences at the tip of the hair-pin loop, in addition to the conformation of the loop itself. Within the binding region, similarities with known binding sites of heterogeneous nuclear ribonucleoprotein (hnRNP) A1 in other RNA molecules were revealed by sequence alignment analysis. Moreover, competition experiments with an oligoribonucleotide corresponding to a known high-affinity binding site of hnRNP A1, and immunoprecipitation of the UV cross-linked 37/39 kDa complex showed that the protein binding to the CYP2A5 mRNA could be hnRNP A1 or its close analogue. It was also shown that the 37/39 kDa protein binds with less affinity to CYP2A4 mRNA than to CYP2A5 mRNA. This is in accordance with experiments characterizing the binding site, since these two otherwise highly homologous genes are kown to have a three nucleotide difference within the region important for the high binding affinity. Since the response of CYP2A4 to pyrazole is known to be weak, as compared with CYP2A5, this observation provides further evidence for a regulatory role of the 37/39 kDa protein in CYP2A5 mRNA metabolism.

Oligonucleotide binding specificities of the hnRNP C protein tetramer

Through the use of various non-equilibrium RNA binding techniques, the C protein tetramer of mammalian40S hnRNP particles has been characterized previously as a poly(U) binding protein with specificity for the pyrimidine-rich sequences that often precede 3' intron-exon junctions. C protein has also been characterized as a sequence-independent RNA chaperonin that is distributed along nascent transcripts through cooperative binding and as a protein ruler that defines the length of RNA packaged in 40S monoparticles. In this study fluorescence spectroscopy was used to monitor C protein-oligonucleotide binding in a competition binding assay under equilibrium conditions. Twenty nucleotide substrates corresponding to polypyrimidine tracts from IVS1 of the adenovirus-2 major late transcript, the adenovirus-2 oncoprotein E1A 3' splice site, IVS2 of human alpha-tropomyosin, the consensus polypyrimidine tract for U2AF65, AUUUA repeats and r(U)20were used as competitors. A 20 nt beta-globin intronic sequence and a randomly generated oligo were used as competitor controls. These studies reveal that native C protein possesses no enhanced affinity for uridine-rich oligonucleotides, but they confirm the enhanced affinity of C protein for an oligonucleotide identified as a high affinity substrate through selection and amplification. Evidence that the affinity of C protein for the winner sequence is due primarily to its unique structure or to a unique context is seen in its retained substrate affinity when contiguous uridines are replaced with contiguous guanosines.

Thinking about a nuclear matrix

The possible existence in eukaryotic cells of an internal, non-chromatin nuclear structural framework that facilitates gene readout as a set of spatially concerted reactions has become a popular but controversial theater of investigation. This article endeavors to present a circumspect review of the nuclear matrix concept as we presently know it, framed around two contrasting hypotheses: (1) that an internal nuclear framework actively enhances gene expression (in much the same way the cytoskeleton mediates cell locomotion, mitosis and intracellular vesicular traffic) versus (2) that the interphase chromosomes have fixed, inherited positions and that the DNA replication, transcripton and RNA processing machinery diffusionally arrives at sites of gene readout, with some aspects of nuclear structure thus being more a result than a cause of gene expression. On balance, the available information suggests that interactions among various gene expression machines may contribute to isolated nuclear matrix preparations. Some components of isolated nuclear matrix preparations may also reflect induced or reconfigured protein-protein associations. The protein characterization and ultrastructural analysis of the isolated nuclear matrix has advanced significantly in recent years, although controversies remain. Important new clues are now coming in from promising contemporary lines of research that report on nuclear structure in living cells.

Growth-dependent and growth-independent translation of messengers for heterogeneous nuclear ribonucleoproteins

The hnRNP A1 transcript has a relatively short 5'- untranslated region (UTR) starting with a pyrimidine tract similar to that of mRNAs encoded by the TOP [terminal oligo(pyrimidine)] genes in vertebrates. Such genes code for ribosomal proteins and for other proteins directly or indirectly involved in the production and function of the translation apparatus. As expected from the role of the pyrimidine tract in the translational regulation of TOP mRNAs, the A1 mRNA is more efficiently loaded onto polysomes in growing than in resting cells. On the other hand, a less stringent regulation with respect to that of other TOP mRNAs is observed, partially due to the presence of multiple transcription start sites within the pyrimidine tract, where transcripts with shorter TOP sequences are less sensitive to regulation. Thus, from the point of view of structural features and translation behaviour the A1 mRNA can be included in the class of TOP genes, suggesting a possible role of A1 in translation. Interestingly, a TOP-like behaviour was observed for hnRNP I mRNA but not for hnRNP C1/C2 and A2/B1 mRNAs, indicating the existence of two classes of hnRNPs with different translational regulation.

hnRNP proteins and B23 are the major proteins of the internal nuclear matrix of HeLa S3 cells

The nuclear matrix is the structure that persists after removal of chromatin and loosely bound components from the nucleus. It consists of a peripheral lamina-pore complex and an intricate internal fibrogranular structure. Little is known about the molecular structure of this proteinaceous internal network. Our aim is to identify the major proteins of the internal nuclear matrix of HeLa 53 cells. To this end, a cell fraction containing the internal fibrogranular structure was compared with one from which this structure had been selectively dissociated. Protein compositions were quantitatively analyzed after high-resolution two-dimensional gel electrophoresis. We have identified the 21 most abundant polypeptides that are present exclusively in the internal nuclear matrix. Sixteen of these proteins are heterogeneous nuclear ribonucleoprotein (hnRNP) proteins. B23 (numatrin) is another abundant protein of the internal nuclear matrix. Our results show that most of the quantitatively major polypeptides of the internal nuclear matrix are proteins involved in RNA metabolism, including packaging and transport of RNA.

A novel methyltransferase (Hmt1p) modifies poly(A)+-RNA-binding proteins

RNA-binding proteins play many essential roles in the metabolism of nuclear pre-mRNA. As such, they demonstrate a myriad of dynamic behaviors and modifications. In particular, heterogeneous nuclear ribonucleoproteins (hnRNPs) contain the bulk of methylated arginine residues in eukaryotic cells. We have identified the first eukaryotic hnRNP-specific methyltransferase via a genetic screen for proteins that interact with an abundant poly(A)+-RNA-binding protein termed Npl3p. We have previously shown that npl3-1 mutants are temperature sensitive for growth and defective for export of mRNA from the nucleus. New mutants in interacting genes were isolated by their failure to survive in the presence of the npl3-1 allele. Four alleles of the same gene were identified in this manner. Cloning of the cognate gene revealed an encoded protein with similarity to methyltransferases that was termed HMT1 for hnRNP methyltransferase. HMT1 is not required for normal cell viability except when NPL3 is also defective. The Hmt1 protein is located in the nucleus. We demonstrate that Npl3p is methylated by Hmt1p both in vivo and in vitro. These findings now allow further exploration of the function of this previously uncharacterized class of enzymes.

Proteins C1 and C2 of heterogeneous nuclear ribonucleoprotein complexes bind RNA in a highly cooperative fashion: support for their contiguous deposition on pre-mRNA during transcription

Proteins C1 and C2 together comprise about one-third the protein mass of mammalian core 40S heterogeneous nuclear ribonucleoprotein particles (40S hnRNP) and exist as heterotetramers of (C1)3C2. On the basis of nonequilibrium binding studies, it has been suggested that the C proteins specifically bind oligo(U)- and poly(U)-rich sequences, and preferentially associate with uridine-rich regions near the 3' termini of many introns. We describe here a more quantitative characterization of the equilibrium binding properties of native and recombinant C protein to homoribopolymers using fluorescence spectroscopy. Like C protein from HeLa cells, the recombinant proteins spontaneously oligomerize to form tetramers with the same hydrodynamic properties as native protein. Near-stoichiometric binding titrations of the fluorescent homoribopolymer polyethenoadenosine (poly[r(epsilon A)]) with recombinant (C1)4 and (C2)4 homotetramers along with competition binding assays with poly(A) and poly(C) indicate that the binding site size (n) is between 150 and 230 nucleotides. This site size range is in close agreement with that previously determined for native C protein through hydrodynamic and ultrastructural studies (approximately 230 nucleotides). (C1)4 and (C2)4 bind poly(G) with intrinsic affinities (Ki) of 10(9) M-1, which are a hundredfold higher than their affinities for poly(U). In opposition to reports that C protein does not bind poly(A) and poly(C), we find that the C proteins bind these substrates with moderate Ki, but with high cooperativity (omega). The overall affinity (K omega) for the binding of both proteins to poly(A) and poly(C) is 10-fold higher (> 10(8) but < 10(9) M-1) than their affinities for poly(U). The highly cooperative binding of C protein to these substrates provides a mechanistic basis for the distribution of C protein along the length of nucleic acid substrates.

Chromatin domains and prediction of MAR sequences

Polynuceosomes are constrained into loops or domains and are insulated from the effects of chromatin structure and torsional strain from flanking domains by the cross-complexation of matrix-attached regions (MARs) and matrix proteins. MARs or SARs have an average size of 500 bp, are spaced about every 30 kb, and are control elements maintaining independent realms of gene activity. A fraction of MARs may cohabit with core origin replication (ORIs) and another fraction might cohabit with transcriptional enhancers. DNA replication, transcription, repair, splicing, and recombination seem to take place on the nuclear matrix. Classical AT-rich MARs have been proposed to anchor the core enhancers and core origins complexed with low abundancy transcription factors to the nuclear matrix via the cooperative binding to MARs of abundant classical matrix proteins (topoisomerase II, histone H1, lamins, SP120, ARBP, SATB1); this creates a unique nuclear microenvironment rich in regulatory proteins able to sustain transcription, replication, repair, and recombination. Theoretical searches and experimental data strongly support a model of activation of MARs and ORIs by transcription factors. A set of 21 characteristics are deduced or proposed for MAR/ORI sequences including their enrichment in inverted repeats, AT tracts, DNA unwinding elements, replication initiator protein sites, homooligonucleotide repeats (i.e., AAA, TTT, CCC), curved DNA, DNase I-hypersensitive sites, nucleosome-free stretches, polypurine stretches, and motifs with a potential for left-handed and triplex structures. We are establishing Banks of ORI and MAR sequences and have undertaken a large project of sequencing a large number of MARs in an effort to determine classes of DNA sequences in these regulatory elements and to understand their role at the origins of replication and transcriptional enhancers.

In vivo and in vitro arginine methylation of RNA-binding proteins

Heterogenous nuclear ribonucleoproteins (hnRNPs) bind pre-mRNAs and facilitate their processing into mRNAs. Many of the hnRNPs undergo extensive posttranslational modifications including methylation on arginine residues. hnRNPs contain about 65% of the total NG,NG-dimethylarginine found in the cell nucleus. The role of this modification is not known. Here we identify the hnRNPs that are methylated in HeLa cells and demonstrate that most of the pre-mRNA-binding proteins receive this modification. Using recombinant human hnRNP A1 as a substrate, we have partially purified and characterized a protein-arginine N-methyltransferase specific for hnRNPs from HeLa cells. This methyltransferase can methylate the same subset of hnRNPs in vitro as are methylated in vivo. Furthermore, it can also methylate other RNA-binding proteins that contain the RGG motif RNA-binding domain. This activity is evolutionarily conserved from lower eukaryotes to mammals, suggesting that methylation has a significant role in the function of RNA-binding proteins.

The C-protein tetramer binds 230 to 240 nucleotides of pre-mRNA and nucleates the assembly of 40S heterogeneous nuclear ribonucleoprotein particles

A series of in vitro protein-RNA binding studies using purified native (C1)3C2 and (A2)3B1 tetramers, total soluble heterogeneous nuclear ribonucleoprotein (hnRNP), and pre-mRNA molecules differing in length and sequence have revealed that a single C-protein tetramer has an RNA site size of 230 to 240 nucleotides (nt). Two tetramers bind twice this RNA length, and three tetramers fold monoparticle lengths of RNA (700 nt) into a unique 19S triangular complex. In the absence of this unique structure, the basic A- and B-group proteins bind RNA to form several different artifactual structures which are not present in preparations of native hnRNP and which do not function in hnRNP assembly. Three (A2)3B1 tetramers bind the 19S complex to form a 35S assembly intermediate. Following UV irradiation to immobilize the C proteins on the packaged RNA, the 19S triangular complex is recovered as a remnant structure from both native and reconstituted hnRNP particles. C protein-RNA complexes composed of three, six, or nine tetramers (one, two, or three triangular complexes) nucleate the stoichiometric assembly of monomer, dimer, and trimer hnRNP particles. The binding of C-protein tetramers to RNAs longer than 230 nt is through a self-cooperative combinatorial mode. RNA packaged in the 19S complex and in 40S hnRNP particles is efficiently spliced in vitro. These findings demonstrate that formation of the triangular C protein-RNA complex is an obligate first event in the in vitro and probably the in vivo assembly the 40S hnRNP core particle, and they provide insight into the mechanism through which the core proteins package 700-nt increments of RNA. These findings also demonstrate that unless excluded by other factors, the C proteins are likely to be located along the length of nascent transcripts.

The C-protein tetramer binds 230 to 240 nucleotides of pre-mRNA and nucleates the assembly of 40S heterogeneous nuclear ribonucleoprotein particles

A series of in vitro protein-RNA binding studies using purified native (C1)3C2 and (A2)3B1 tetramers, total soluble heterogeneous nuclear ribonucleoprotein (hnRNP), and pre-mRNA molecules differing in length and sequence have revealed that a single C-protein tetramer has an RNA site size of 230 to 240 nucleotides (nt). Two tetramers bind twice this RNA length, and three tetramers fold monoparticle lengths of RNA (700 nt) into a unique 19S triangular complex. In the absence of this unique structure, the basic A- and B-group proteins bind RNA to form several different artifactual structures which are not present in preparations of native hnRNP and which do not function in hnRNP assembly. Three (A2)3B1 tetramers bind the 19S complex to form a 35S assembly intermediate. Following UV irradiation to immobilize the C proteins on the packaged RNA, the 19S triangular complex is recovered as a remnant structure from both native and reconstituted hnRNP particles. C protein-RNA complexes composed of three, six, or nine tetramers (one, two, or three triangular complexes) nucleate the stoichiometric assembly of monomer, dimer, and trimer hnRNP particles. The binding of C-protein tetramers to RNAs longer than 230 nt is through a self-cooperative combinatorial mode. RNA packaged in the 19S complex and in 40S hnRNP particles is efficiently spliced in vitro. These findings demonstrate that formation of the triangular C protein-RNA complex is an obligate first event in the in vitro and probably the in vivo assembly the 40S hnRNP core particle, and they provide insight into the mechanism through which the core proteins package 700-nt increments of RNA. These findings also demonstrate that unless excluded by other factors, the C proteins are likely to be located along the length of nascent transcripts.

hnRNP A2/B1 binds specifically to single stranded vertebrate telomeric repeat TTAGGGn

We have previously isolated a protein from mouse liver nuclei that specifically binds to single stranded (TTAGGG)n repeats. TTAGGG is the telomeric repeats of mammals and we therefore named the new protein single stranded telomere binding protein (sTBP). Further studies now identify sTBP as heterogeneous nuclear ribonucleoprotein (hnRNP) A2/B1 on the basis of amino acid sequence determination and antibody reactivity. A2 and B1 form a major part of the protein component of hnRNP particles and are abundant nuclear proteins. Unexpectedly, A2/B1 has a high specificity for binding to the RNA equivalent of TTAGGG, UUAGGG, but under the same conditions does not appear to have a strong affinity for a number of other RNA species.

Demonstration of a 7-nm RNP fiber as the basic structural element in a premessenger RNP particle

Balbiani ring granules in Chironomus salivary glands represent premessenger ribonucleoprotein (RNP) particles, each containing a 35- to 40-kb message for a secretory polypeptide. Their gross structure can be described as an RNP ribbon bent into a toroid. We now demonstrate that an unfolded, thin RNP fiber is observed after low salt treatment of isolated Balbiani ring granules. Moreover, the thin RNP fiber, 7 nm in diameter, can be revealed as the main structural element in Balbiani ring granules studied in situ in 3-D with electron microscope tomography. It is proposed that the thin RNP fiber consists of a premessenger RNA molecule coiled around a filamentous core of polymeric proteins, which has functional implications for processes such as assembly of RNP, intranuclear degradation of RNA, and delivery of RNA through the nuclear pores.

The NUF1 gene encodes an essential coiled-coil related protein that is a potential component of the yeast nucleoskeleton

In an attempt to identify structural components of the yeast nucleus, subcellular fractions of yeast nuclei were prepared and used as immunogens to generate complex polyclonal antibodies. One such serum was used to screen a yeast genomic lambda gt11 expression library. A clone encoding a gene called NUF1 (for nuclear filament-related) was identified and extensively characterized. Antibodies to NUF1 fusion proteins were generated, and affinity-purified antibodies were used for immunoblot analysis and indirect immunofluorescence localization. The NUF1 protein is 110 kD in molecular mass and localizes to the yeast nucleus in small granular patches. Intranuclear staining is present in cells at all stages of the cell cycle. The NUF1 protein of yeast is tightly associated with the nucleus; it was not removed by extraction of nuclei with nonionic detergent or salt, or treatment with RNAse and DNAse. Sequence analysis of the NUF1 gene predicts a protein 945 amino acids in length that contains three domains: a large 627 residue central domain predicted to form a coiled-coil structure flanked by nonhelical amino-terminal and carboxy-terminal regions. Disruption of the NUF1 gene indicates that it is necessary for yeast cell growth. These results indicate that NUF1 encodes an essential coiled-coil protein within the yeast nucleus; we speculate that NUF1 is a component of the yeast nucleoskeleton. In addition, immunofluorescence results indicate that mammalian cells contain a NUF1-related nuclear protein. These data in conjunction with those in the accompanying manuscript (Yang et al., 1992) lead to the hypothesis that an internal coiled-coil filamentous system may be a general structural component of the eukaryotic nucleus.

Isolation of hnRNP complexes from Drosophila melanogaster

Nascent RNA polymerase II transcripts, heterogeneous nuclear RNAs (hnRNAs), become associated with nuclear proteins (hnRNP Proteins), and their processing into mRNAs takes place in these hnRNP complexes. hnRNP complexes have previously been purified from vertebrate cells. Here we report the isolation of hnRNP complexes from an invertebrate organism, the fruitfly Drosophila melanogaster. Candidate hnRNP proteins were purified from D. melanogaster embryos by ssDNA affinity chromatography, and mAbs were produced to many of the major proteins. Genuine hnRNP proteins were identified by several criteria, including nucleoplasmic localization, association with nascent transcripts, crosslinking to poly(A)-containing RNA in living cells, and amino acid sequence. In addition, mAbs that cross-react between the fruitfly and human hnRNP proteins were obtained. Most importantly, using hnRNP-specific mAbs we have purified the hnRNP complexes from D. melanogaster cells. These RNAase-sensitive complexes contain at least 10 major proteins designated hrps, the most abundant proteins having apparent molecular masses of 36, 38, 39, 40, 44, 48, 54, 62, 70, and 75 kD. cDNAs and complete sequences for several of these proteins have been obtained and are presented in the accompanying paper (Matunis, E. L., M. J. Matunis, and G. Dreyfuss. 1992. J. Cell Biol. 116:257-269). The purification of D. melanogaster hnRNP complexes will facilitate genetic and cytological studies on the function of hnRNA-binding proteins and on the posttranscriptional regulation of gene expression.

Compartmentalization within the nucleus: discovery of a novel subnuclear region

Antibodies to a set of structurally related autoantigens (p23-25) bind to a previously uncharacterized, large structural domain in the nucleus of a variety of human cell types. This subnuclear domain is visible by phase contrast alone as a region of decreased density after several different fixation protocols. The morphology of this region changes dramatically during the cell cycle and we have given it the name PIKA (for polymorphic interphase karyosomal association) based on preliminary evidence that the PIKA proteins may be associated with chromatin. The function of the PIKA is not yet known, but our immunolocalization data indicate that it is unlikely to be associated with regions of ongoing DNA replication, heterogeneous nuclear RNA storage, or mRNA processing. The discovery of the PIKA provides evidence supporting an emerging model of nuclear structure. It now appears that the nucleus is organized into distinct domains which include not only the nucleolus, but also previously unidentified regions such as the PIKAs. Furthermore, structural rearrangements undergone by the nucleolus and the PIKAs may be indicative of a broad tendency for nuclear organization to change in a cell cycle-specific fashion.

The core proteins A2 and B1 exist as (A2)3B1 tetramers in 40S nuclear ribonucleoprotein particles

The six "core" proteins of HeLa cell 40S nuclear ribonucleoprotein particles (hnRNP particles) package 700-nucleotide lengths of pre-mRNA into a repeating array of regular particles. We have previously shown that the C proteins exist as anisotropic tetramers of (C1)3C2 in 40S hnRNP particles and that each particle probably contains three such tetramers. We report here that proteins A2 and B1 also exist in monoparticles as (A2)3B1 tetramers and that each monoparticle contains at least three such tetramers. Proteins A2 and B1 dissociate from isolated monoparticles as a stable tetramer upon nuclease digestion. In low-salt gradients, the tetramers sediment at 6.8S, which is consistent with a mass of 145 kDa. In 200 mM salt, the concentration which dissociates these proteins from RNA, only 4.2S dimers exist in solution. Tetramers of (A2)3B1 possess the ability to package multiples of 700 nucleotides of RNA in vitro into an array of regular, 22.5-nm 43S particles. Unlike the in vitro assembly of intact 40S hnRNP, the (A2)3B1 tetramers assemble by means of a highly cooperative process. These findings indicate that the (A2)3B1 tetramers play a major role in hnRNP assembly and they further support the contention that 40S monoparticles are regular structures composed of three copies of three different tetramers, i.e., 3[(A1)3B2, (A2)3B1, (C1)3C2].

The core proteins A2 and B1 exist as (A2)3B1 tetramers in 40S nuclear ribonucleoprotein particles

The six "core" proteins of HeLa cell 40S nuclear ribonucleoprotein particles (hnRNP particles) package 700-nucleotide lengths of pre-mRNA into a repeating array of regular particles. We have previously shown that the C proteins exist as anisotropic tetramers of (C1)3C2 in 40S hnRNP particles and that each particle probably contains three such tetramers. We report here that proteins A2 and B1 also exist in monoparticles as (A2)3B1 tetramers and that each monoparticle contains at least three such tetramers. Proteins A2 and B1 dissociate from isolated monoparticles as a stable tetramer upon nuclease digestion. In low-salt gradients, the tetramers sediment at 6.8S, which is consistent with a mass of 145 kDa. In 200 mM salt, the concentration which dissociates these proteins from RNA, only 4.2S dimers exist in solution. Tetramers of (A2)3B1 possess the ability to package multiples of 700 nucleotides of RNA in vitro into an array of regular, 22.5-nm 43S particles. Unlike the in vitro assembly of intact 40S hnRNP, the (A2)3B1 tetramers assemble by means of a highly cooperative process. These findings indicate that the (A2)3B1 tetramers play a major role in hnRNP assembly and they further support the contention that 40S monoparticles are regular structures composed of three copies of three different tetramers, i.e., 3[(A1)3B2, (A2)3B1, (C1)3C2].

The Drosophila Hrb87F gene encodes a new member of the A and B hnRNP protein group

Nascent premessenger RNA transcripts are packaged into heterogeneous nuclear ribonucleoprotein (hnRNP) complexes containing specific nuclear proteins, the hnRNP proteins. The A and B group proteins constitute a major class of small basic proteins found in mammalian hnRNP complexes. We have previously characterized the Drosophila melanogaster Hrb98DE gene, which is alternatively spliced to encode four protein isoforms closely related to the A and B proteins. We report here that the Drosophila genome contains a family of genes related to the Hrb98DE gene. One member of the family, Hrb87F, is very homologous to Hrb98DE in both sequence and structure. The Hrb87F transcripts (1.7 and 2.2 kb) utilize two alternative polyadenylation sites, are abundant in ovaries and early embryos, and are present in lesser amounts throughout development. In one wildtype strain of Drosophila there is a naturally-occurring polymorphism in this gene due to the insertion of a 412 transposable element in the 3' untranslated region. The larger transcript is not produced in these files and thus is not required for viability. Sequence identities among the Drosophila Hrb proteins and the vertebrate A and B hnRNP proteins suggest that these proteins may form a distinct subfamily within the larger family of related RNA binding proteins.

A uridylate tract mediates efficient heterogeneous nuclear ribonucleoprotein C protein-RNA cross-linking and functionally substitutes for the downstream element of the polyadenylation signal

Every RNA added to an in vitro polyadenylation extract became stably associated with both the heterogeneous nuclear ribonucleoprotein (hnRNP) A and C proteins, as assayed by immunoprecipitation analysis using specific monoclonal antibodies. UV-cross-linking analysis, however, which assays the specific spatial relationship of certain amino acids and RNA bases, indicated that the hnRNP C proteins, but not the A proteins, were associated with downstream sequences of the simian virus 40 late polyadenylation signal in a sequence-mediated manner. A tract of five consecutive uridylate residues was required for this interaction. The insertion of a five-base U tract into a pGEM4 polylinker-derived transcript was sufficient to direct sequence-specific cross-linking of the C proteins to RNA. Finally, the five-base uridylate tract restored efficient in vitro processing to several independent poly(A) signals in which it substituted for downstream element sequences. The role of the downstream element in polyadenylation efficiency, therefore, may be mediated by sequence-directed alignment or phasing of an hnRNP complex.

A uridylate tract mediates efficient heterogeneous nuclear ribonucleoprotein C protein-RNA cross-linking and functionally substitutes for the downstream element of the polyadenylation signal

Every RNA added to an in vitro polyadenylation extract became stably associated with both the heterogeneous nuclear ribonucleoprotein (hnRNP) A and C proteins, as assayed by immunoprecipitation analysis using specific monoclonal antibodies. UV-cross-linking analysis, however, which assays the specific spatial relationship of certain amino acids and RNA bases, indicated that the hnRNP C proteins, but not the A proteins, were associated with downstream sequences of the simian virus 40 late polyadenylation signal in a sequence-mediated manner. A tract of five consecutive uridylate residues was required for this interaction. The insertion of a five-base U tract into a pGEM4 polylinker-derived transcript was sufficient to direct sequence-specific cross-linking of the C proteins to RNA. Finally, the five-base uridylate tract restored efficient in vitro processing to several independent poly(A) signals in which it substituted for downstream element sequences. The role of the downstream element in polyadenylation efficiency, therefore, may be mediated by sequence-directed alignment or phasing of an hnRNP complex.

Attachment of DNA to the nucleoskeleton of HeLa cells examined using physiological conditions

Although it is widely believed that eukaryotic DNA is looped by attachment to a nucleoskeleton, there is controversy about its composition and which sequences are attached to it. As most nuclear derivatives are isolated using unphysiological conditions, the criticism that attachments seen in vitro are generated artifactually has been difficult to rebut. Therefore we have re-investigated attachments of chromatin to the skeleton using physiological conditions. HeLa cells are encapsulated in agarose microbeads and lysed using Triton in a 'physiological' buffer. Then, most chromatin can be electroeluted after treatment with a restriction enzyme to leave some at the base of the loops still attached. Analysis of the size and amounts of these residual fragments indicates that the loops are 80-90kbp long. The residual fragments are stably attached, with about 1kbp of each fragment protected from nuclease attack. This is very much longer than a typical protein-binding site of 10-20bp.

Vaults. II. Ribonucleoprotein structures are highly conserved among higher and lower eukaryotes

Vaults are cytoplasmic ribonucleoprotein structures that display a complex morphology reminiscent of the multiple arches which form cathedral vaults, hence their name. Previous studies on rat liver vaults (Kedersha, N. L., and L. H. Rome. 1986. J. Cell Biol. 103:699-709) have established that their composition is unlike that of any known class of RNA-containing particles in that they contain multiple copies of a unique small RNA and more than 50 copies of a single polypeptide of 104,000 Mr. We now report on the isolation of vaults from numerous species and show that vaults appear to be ubiquitous among eukaryotes, including mammals, amphibians (Rana catesbeiana and Xenopus laevis), avians (Gallus Gallus), and the lower eukaryote Dictyostelium discoideum. Electron microscopy reveals that vaults purified from these diverse species are similar both in their dimensions and morphology. The vaults from these various species are also similar in their polypeptide composition; each being composed of a major polypeptide with an approximate mass of 100 kD and several minor polypeptides with molecular masses similar to those seen in the rat. Antibodies raised against rat vaults recognize the major vault protein of all species including Dictyostelium. Vaults therefore appear to be strongly conserved and broadly distributed, suggesting that their function is essential to eukaryotic cells.

Core filaments of the nuclear matrix

The nuclear matrix is concealed by a much larger mass of chromatin, which can be removed selectively by digesting nuclei with DNase I followed by elution of chromatin with 0.25 M ammonium sulfate. This mild procedure removes chromatin almost completely and preserves nuclear matrix morphology. The complete nuclear matrix consists of a nuclear lamina with an interior matrix composed of thick, polymorphic fibers and large masses that resemble remnant nucleoli. Further extraction of the nuclear matrices of HeLa or MCF-7 cells with 2 M sodium chloride uncovered a network of core filaments. A few dark masses remained enmeshed in the filament network and may be remnants of the nuclear matrix thick fibers and nucleoli. The highly branched core filaments had diameters of 9 and 13 nm measured relative to the intermediate filaments. They may serve as the core structure around which the matrix is constructed. The core filaments retained 70% of nuclear RNA. This RNA consisted both of ribosomal RNA precursors and of very high molecular weight hnRNA with a modal size of 20 kb. Treatment with RNase A removed the core filaments. When 2 M sodium chloride was used directly to remove chromatin after DNase I digestion without a preceding 0.25 M ammonium sulfate extraction, the core filaments were not revealed. Instead, the nuclear interior was filled with amorphous masses that may cover the filaments. This reflected a requirement for a stepwise increase in ionic strength because gradual addition of sodium chloride to a final concentration of 2 M without an 0.25 M ammonium sulfate extraction uncovered core filaments.

Multiple splicing factors are released from endogenous complexes during in vitro pre-mRNA splicing

Pre-mRNA splicing occurs in a macromolecular complex called the spliceosome. Efforts to isolate spliceosomes from in vitro splicing reactions have been hampered by the presence of endogenous complexes that copurify with de novo spliceosomes formed on added pre-mRNA. We have found that removal of these large complexes from nuclear extracts prevents the splicing of exogenously added pre-mRNA. We therefore examined these complexes for the presence of splicing factors and proteins known or thought to be involved in RNA splicing. These fast-sedimenting structures were found to contain multiple small nuclear ribonucleoproteins (snRNPs) and a fragmented heterogeneous nuclear ribonucleoprotein complex. At least two splicing factors other than the snRNPs were also associated with these large structures. Upon incubation with ATP, these splicing factors as well as U1 and U2 snRNPs were released from these complexes. The presence of multiple splicing factors suggests that these complexes may be endogenous spliceosomes released from nuclei during preparation of splicing extracts. The removal of these structures from extracts that had been preincubated with ATP yielded a splicing extract devoid of large structures. This extract should prove useful in the fractionation of splicing factors and the isolation of native spliceosomes formed on exogenously added pre-mRNA.

Multiple splicing factors are released from endogenous complexes during in vitro pre-mRNA splicing

Pre-mRNA splicing occurs in a macromolecular complex called the spliceosome. Efforts to isolate spliceosomes from in vitro splicing reactions have been hampered by the presence of endogenous complexes that copurify with de novo spliceosomes formed on added pre-mRNA. We have found that removal of these large complexes from nuclear extracts prevents the splicing of exogenously added pre-mRNA. We therefore examined these complexes for the presence of splicing factors and proteins known or thought to be involved in RNA splicing. These fast-sedimenting structures were found to contain multiple small nuclear ribonucleoproteins (snRNPs) and a fragmented heterogeneous nuclear ribonucleoprotein complex. At least two splicing factors other than the snRNPs were also associated with these large structures. Upon incubation with ATP, these splicing factors as well as U1 and U2 snRNPs were released from these complexes. The presence of multiple splicing factors suggests that these complexes may be endogenous spliceosomes released from nuclei during preparation of splicing extracts. The removal of these structures from extracts that had been preincubated with ATP yielded a splicing extract devoid of large structures. This extract should prove useful in the fractionation of splicing factors and the isolation of native spliceosomes formed on exogenously added pre-mRNA.

The nucleoskeleton and the topology of transcription

Transcription is conventionally believed to occur by passage of a mobile polymerase along a fixed template. Evidence for this model is derived almost entirely from material prepared using hypotonic salt concentrations. Studies on subnuclear structures isolated using hypertonic conditions, and more recently using conditions closer to the physiological, suggest an alternative. Transcription occurs as the template moves past a polymerase attached to a nucleoskeleton; this skeleton is the active site of transcription. Evidence for the two models is summarised. Much of it is consistent with the polymerase being attached and not freely diffusible. Some consequences of such a model are discussed.