Georg Sauermann

Proto-oncoprotein tls/fus is associated to the nuclear matrix and complexed with splicing factors ptb, srm160, and sr proteins

TLS/FUS is a nucleic acid-binding protein whose N-terminal half functions as a transcriptional activator domain in fusion oncoproteins found in human leukemias and liposarcomas. Previous reports have suggested a role for TLS/FUS in transcription and splicing processes. Here we report the association of TLS/FUS with the nuclear matrix and investigate its role in splicing. Splicing of two pre-mRNAs was inhibited in a TLS/FUS-immunodepleted extract and could only be partly restored by addition of recombinant TLS/FUS or/and SR proteins, known interaction partners of TLS/FUS. The subsequent analysis of TLS/FUS immunoprecipitates revealed that, in addition to the SR proteins SC35 and SRp75, the splicing factor PTB (hnRNPI) and the splicing coactivator SRm160 are complexed with TLS/FUS, thus explaining the inability to restore splicing completely. Coimmunolocalization confirmed the nuclear matrix association and interaction of TLS/FUS with PTB, SR proteins, and SRm160. Our results suggest that the matrix protein TLS/FUS plays a role in spliceosome assembly.

Proteome analysis of nuclear matrix proteins during apoptotic chromatin condensation

The nuclear matrix (NM) is considered a proteinaceous scaffold spatially organizing the interphase nucleus, the integrity of which is affected during apoptosis. Caspase-mediated degradation of NM proteins, such as nuclear lamins, precedes apoptotic chromatin condensation (ACC). Nevertheless, other NM proteins remain unaffected, which most likely maintain a remaining nuclear structure devoid of chromatin. We, therefore, screened various types of apoptotic cells for changes of the nuclear matrix proteome during the process of apoptotic ACC. Expectedly, we observed fundamental alterations of known chromatin-associated proteins, comprising both degradation and translocation to the cytosol. Importantly, a consistent set of abundant NM proteins, some (e.g. hNMP 200) of which displaying structural features, remained unaffected during apoptosis and might therefore represent constituents of an elementary scaffold. In addition, proteins involved in DNA replication and DNA repair were found accumulated in the NM fraction before cells became irreversibly committed to ACC, a time point characterized in detail by inhibitor studies with orthovanadate. In general, protein alterations of a consistent set of NM proteins (67 of which were identified), were reproducibly detectable in Fas-induced Jurkat cells, in UV-light treated U937 cells and also in staurosporine-treated HeLa cells. Our data indicate that substantial alterations of proteins linking chromatin to an elementary nuclear protein scaffold might play an intriguing role for the process of ACC.

Similarity between nuclear matrix proteins of various cells revealed by an improved isolation method

Comparative analysis of nuclear matrix proteins by two‐dimensional electrophoresis may be greatly impaired by copurifying cytoskeletal proteins. The present data show that the bulk of adhering cytofilaments may mechanically be removed by shearing of nuclei pretreated with vanadyl ribonucleoside complexes. Potential mechanisms of action not based on ribonuclease inhibition are discussed. To individually preserve the integrity of nuclear structures, we developed protocols for the preparation of nuclear matrices from three categories of cells, namely leukocytes, cultured cells, and tissue cells. As exemplified with material from human lymphocytes, cultured amniotic cells, and liver tissue cells, the resulting patterns of nuclear matrix proteins appeared quite similar. Approximately 300 spots were shared among the cell types. Forty‐nine of these were identified, 21 comprising heterogeneous nuclear ribonucleoproteins. Heterogeneous nuclear ribonucleoproteins L and nuclear lamin B2 isoforms were identified by amino acid sequencing and mass spectrometry. However, individually expressed proteins, such as the proliferating cell nuclear antigen, also pertained following application of the protocols. Thus, enhanced resolution and comparability of proteins improve systematic analyses of nuclear matrix proteins from various cellular sources. J. Cell. Biochem. 71:363–374, 1998.

Antibodies to nuclear lamins in autoimmune liver disease

Antibodies to nuclear lamins were detected in sera of patients with autoimmune liver disease. In indirect immunofluorescence tests, these sera revealed staining of the nuclear periphery. Using isolated nuclei, nuclear matrices, nuclear lamina-pore complexes, and chromatographically purified lamins as antigen source, the nuclear lamins A, B, and C were identified as reactive antigens in immunoblotting experiments. The lamins were also identified by 2-D gel electrophoresis. Antibodies to nuclear lamins occurred in 12 of 16 cases of active lupoid hepatitis, but not in 35 patients with the disease in remission. However, only 3 of 37 sera of patients with primary biliary cirrhosis contained anti-lamin antibodies. Autoimmune liver disease sera reacted preferentially with lamins A/C and less frequently with lamin B or lamins A/B/C.

Differential nuclear localization and nuclear matrix association of the splicing factors PSF and PTB

A monoclonal antibody raised against nuclear matrix proteins detected a protein of basic pI in human nuclear matrix protein samples of various cellular origin. The ubiquitously occurring (common) nuclear matrix protein was identified as splicing factor PSF (PTB associated splicing factor). The interaction between the splicing factors PSF and PTB/hnRNP I was confirmed by co‐immunoprecipitation from nuclear salt extracts. However, the nuclear localization of PSF and PTB and their distribution in subnuclear fractions differed markedly. Isolated nuclear matrices contained the bulk of PSF, but only minor amounts of PTB. In confocal microscopy both proteins appeared in speckles, the majority of which did not co‐localize. Removing a large fraction of the soluble PTB structures by salt extraction revealed some colocalization of the more stable PTB fraction with PSF. These PTB/PSF complexes as well as the observed PSF‐PTB interaction may reflect the previously reported presence of PTB and PSF in spliceosomal complexes during RNA processing. The present data, however, point to different cellular distribution and nuclear matrix association of the majority of PSF and PTB. J. Cell. Biochem. 76:559–566, 2000.

Nuclear matrix proteins specific for subtypes of human hematopoietic cells

Nuclear matrices were prepared from isolated subtypes of human hematopoietic cells and from cultured leukemia cells. The nuclear matrix proteins were analyzed by high‐resolution two‐dimensional gel electrophoresis and computer‐assisted image analysis. While more than 200 protein spots were shared among the cells, about 50 distinct spots were found characteristic for individual cells or groups of related cells. This allowed to differentiate between hematopoietic cells and nonhematopoietic cells, lymphocytes and myeloid cells, monocytes, neutrophils, and promyelocytic leukemia cells. B and T lymphocytes could not be differentiated. Myeloid cells with their polymorph nuclei were characterized by the presence of 13 and by the absence of seven distinct spots, as well as by low concentrations of nuclear lamins and of heterogeneous nuclear ribonucleoproteins. Neutrophils with multilobular nuclei displayed six additional spots, while lacking 18 nuclear matrix protein spots. The nuclear matrix of proliferating cells showed three distinct spots in addition to proliferating cell nuclear antigen, increased concentrations of numatrin (B23), and heterogeneous nuclear ribonucleoproteins. The described cell‐specific nuclear matrix proteins may represent new markers for hematopoietic cells. J. Cell. Biochem. 72:470–482, 1999.

Transport of beta-globin mRNA from nuclei of murine friend erythroleukemia cells. Reversible inhibition of transport by the oxidizing sulfhydryl reagent o-iodosobenzoate

An in vitro assay system for analysis of beta-globin mRNA transport is described. Nuclei isolated from murine Friend erythroleukemia cells induced to synthesize globin mRNA, were incubated in micro-assays. By electrophoresis and hybridization analysis, released 9-S beta-globin mRNA was shown to be undegraded. After direct blotting, the released mRNA was quantified by hybridization with a labeled plasmid containing a beta-globin DNA restriction fragment. The inducibility of beta-globin mRNA transport corresponded to that previously reported for the release of rapidly labeled RNA in other assay systems. In contrast to the ineffectiveness of high concentrations of the sulfhydryl reagent iodoacetate, low concentrations of the oxidizing sulfhydryl reagent, o-iodosobenzoate, inhibited the release of beta-globin mRNA from nuclei of erythroleukemia cells, as well as the release of rapidly labeled RNA from rat liver nuclei. The inhibitory effect of the oxidizing agent on beta-globin mRNA transport could be reversed by postincubation of the nuclei with the reducing agent, dithiothreitol. The potential role of disulfide bond formation on RNA transport is discussed.

Identification of adp-ribosylated histones by the combined use of high-performance liquid chromatography and electrophoresis

Reversed-phase high-performance liquid chromatography (HPLC) was employed for analysing mono- and oligo(ADP-ribosyl)ated histones. Under the chromatographic conditions described, the ADP-ribosylated histones showed similar retention times to the unmodified histones, although the molecular weight and the charge of the proteins are significantly altered by their modification. The simultaneous elution of unmodified and labelled modified histones was detected by two types of gel electrophoresis and by autoradiography. In addition, the HPLC fractions did not display overlapping ladders of the multiply modified histones, as is commonly seen in one-dimensional electrophoretic analyses of unfractionated material. Hence individual bands could be unambiguously assigned. After in vitro labelling of isolated rat liver nuclei, the following ADP-ribosylated and unmodified histones were identified by HPLC and gel electrophoresis: histone H1(0), four histone H1 subfractions, histone H2A.1, histone H2A.2, oxidized histone H2A.2, histone H2A.X, histone H2A.Z, histone H2B, three histone H3 variants and histone H4.

Poly(ADP-ribose) effectively competes with DNA for histone H4 binding.

The effect of poly(ADP-ribose) on DNA-histone H4 interaction was studied using a nitrocellulose filter binding assay. Poly-(ADP-ribose) was found to form poly(ADP-ribose)-histone H4 complexes at physiological salt concentrations. The homopolymer effectively competed with DNA for histone H4 binding. Poly(ADP-ribose) was also capable of displacing DNA from preformed DNA-histone H4 complexes. Our hypothesis is that poly(ADP-ribose), locally and transiently formed at the site of DNA damage, causes dissociation of DNA from the nucleosome particle or nucleosome unfolding.

Nuclear columns Kinetics of RNA synthesis and release in isolated rat liver nuclei

By continuous perfusion of columns containing isolated immobilized rat liver nuclei with media containing labeled RNA precursors, the in vitro synthesis and release of RNA was studied. The combined reaction of synthesis and release could be adjusted to proceed at a constant rate. The reaction rate responded to variation of termperature, ionic conditions, nucleoside triphosphate concentration and to the addition of RNA polymerase inhibitors. During 60 min perfusion approximately equal amounts of radioactive low molecular weight RNA and of ribonucleoproteins were released. Pulse-chase experiments showed that the low molecular weight RNA was synthesized throughout the perfusion and released immediately after formation. The ribonucleoproteins were primarly labeled during the first period of perfusion and were gradually released. Synthesis of RNA contained in the ribonucleoproteins was inhibited by low alpha-amanitin concentrations, indicating that it was catalyzed by RNA polymerase II. The in vitro labeled ribonucleoproteins exhibited properties of the stable nuclear particles which can be extracted from isolated nuclei after rapid in vivo labeling of RNA. They had a buoyant density of 1.41--1.43 in CsCl, were partially unstable in 1% deoxycholate, but stable in 0.1% deoxycholate, in 100 mM NaCl and in 10 mM EDTA. Due to the dilution by the perfusion medium, the ribonucleoproteins sedimented with a peak at 22--27 S, and not at 30--45 S. The RNA synthesized in the immobilized nuclei was not degraded during the perfusion. Less than 20% was gradually released, whereby the 20--30 S peak zone was reduced. While the properties of the in vitro labeled ribonucleoproteins and of rapidly in vivo labeled ribonucleoproteins were the same, the kinetics of their release differed.