Protein C23 Research Articles

Nucleolin in Neurons of the Human Substantia Nigra

Using immunocytochemistry and confocal laser microscopy, the distribution of nucleolin (protein C23) in neurons of the human substantia nigra was investigated. Fragments of the human midbrain (n = 8) with the verified lack of neurodegeneration were used in the study. The material was fixed in zinc ethanol formaldehyde, a special fixative that provides high preservation of antigenic determinants. New morphological data on the distribution of nucleolin in neurons of the human substantia nigra were obtained. Nucleolin was found in the nucleolus and nucleoplasm of dopaminergic neurons but not in the cytoplasm or cell surface. In the nucleolus and nucleoplasm, protein C23 is distributed irregularly in the form of some accumulations (clumps) with fuzzy borders. Moreover, nucleolin distribution varied in structurally similar neurons containing cytoplasmic neuromelanin. The Marinesco bodies typically not exhibiting nucleolin expression were identified in the nucleus, in addition to nucleoli, by confocal laser microscopy with simultaneous use of transmitted light detector. The findings can contribute to elucidation of the role of nucleolin in the specific functions of normal dopaminergic neurons in the human substantia nigra and assessment of probable involvement of C23 in the development of nucleolar stress and neurodegeneration.

Retraction Note to: Nuclear Protein C23 on the Cell Surface Plays an Important Role in Activation of CXCR4 Signaling in Glioblastoma.

The chemokine receptor CXCR4 and its ligand stromal cell-derived factor 1 (SDF-1) plays an important role in tumor progression and are associated with angiogenesis. Meanwhile, the implications of C23 in multiple signaling pathways have been also investigated. However, the effects of C23 on CXCR4 pathway in glioblastoma are not fully characterized. In the present study, C23 and CXCR4 of U87 cell line were inhibited by anti-C23 and anti-CXCR4 antibodies, respectively; and then C23 and CXCR4 siRNAs were used to knock down endogenous C23 and CXCR4, respectively. In addition, MTT assay was also introduced. Our data showed that either anti-C23 or anti-CXCR4 antibodies efficaciously repressed the phosphorylation levels of ERK (p < 0.000) and AKT (p < 0.000) compared with SDF-1 alone and control. As expected, either C23 or CXCR4 siRNAs indeed resulted in C23 and CXCR4 knockdown and further suppressed the expression of p-ERK and p-AKT. Most importantly, immunoprecipitation revealed C23 interacted with CXCR4 once U87 was exposed to SDF-1 treatment. In addition, MTT assay identified that C23 or CXCR4 siRNAs could obviously decreased cell proliferation capacity (p = 0.002). In conclusion, our results suggest that C23 plays a crucial role in activation of SDF-1-induced ERK and PI3K/AKT pathways via interacting with CXCR4. Furthermore, C23 could be recommended as an important element in glioblastoma development and a new target for glioblastoma treatment.

Involvement of nuclear protein C23 in activation of EGFR signaling in cervical cancer.

Nuclear protein C23 and epidermal growth factor receptor (EGFR) are reported to be correlated with cervical cancer (CC). However the correlations between C23 and EGFR were rarely reported. Here, this study explored the effects of C23 in activation of EGFR signaling pathway. In our study, immunohistochemistry was used to identify the expression of C23 or EGFR in CC tissues. The level of the phosphorylated EGFR was observed by western blot, and cell invasion capacity was detected by Transwell assay. In this study, we found that C23 and EGFR were highly expressed in cervical cancer tissues, while C23 on the cell surface mainly expressed in CC tissues with lymph node metastasis, and was correlated to EGFR statistically. In vitro, western blot showed that either anti-C23 or anti-EGFR antibodies can inhibit the phosphorlation of EGFR with significant differences (p < 0.01). Besides, based on Transwell assay, the number of membrane-invading cells was reduced significantly in anti-C23 group, and no significant difference was found compared with anti-EGFR treatment (p > 0.05). In conclusion, C23 on the cell surface may be a kind of indispensable component in activation of EGFR signaling, by which C23 can participate in the growth and invasion of tumors. C23 antagonists may provide a new field for cervical cancer therapy.

C23 protein meditates bone morphogenetic protein-2-mediated EMT via up-regulation of Erk1/2 and Akt in gastric cancer.

In our previous study, the epithelial-to-mesenchymal transition (EMT) has been identified to be involved in gastric cancer progression. Notably, nuclear protein C23 and bone morphogenetic protein-2 (BMP2) have been linked into EMT. However, the specific mechanisms underlying BMP2 pathway-mediated EMT are not still unraveled. In this study, we adopted immunohistochemistry and immunoblotting to determine the expression of C23 and BMP2 receptor II (BMPR-II) in 90 gastric cancer samples and cell lines. Subsequently, relevant cell lines were selected to be treated with si-C23 or si-BMPRII and the detection of in vitro assay. Our results revealed that both C23 and BMPRII were aberrantly and constitutively expressed in gastric cancer specimens and cell lines, whose expression was positively associated with metastasis, stage and differentiation, and portended poor survival outcome of gastric cancer patients. In vitro assay validated the increased expression of p-Erk1/2, p-Akt, vimentin, N-cadherin, and MMP2 in BMP2-stimulated MGC803 cells, which was in a dose-dependent manner. By contrast, si-C23 treatment attenuated the BMP2-stimulated expression of p-Erk1/2, p-Akt, vimentin, N-cadherin, and MMP2. Also, the treatment of either si-C23 or si-BMPRII decreased the ability of migration and invasion of MGC803 cells. In conclusion, C23 mediates BMP2-induced EMT progression via the up-regulation of Erk1/2 and Akt signaling pathway in gastric cancer, which indicated both C23 and BMPRII pathway could be recommended as prospective targets or biomarkers to antagonize the progression of gastric cancer.

Nuclear Protein C23 on the Cell Surface Plays an Important Role in Activation of CXCR4 Signaling in Glioblastoma.

The chemokine receptor CXCR4 and its ligand stromal cell-derived factor 1 (SDF-1) plays an important role in tumor progression and are associated with angiogenesis. Meanwhile, the implications of C23 in multiple signaling pathways have been also investigated. However, the effects of C23 on CXCR4 pathway in glioblastoma are not fully characterized. In the present study, C23 and CXCR4 of U87 cell line were inhibited by anti-C23 and anti-CXCR4 antibodies, respectively; and then C23 and CXCR4 siRNAs were used to knock down endogenous C23 and CXCR4, respectively. In addition, MTT assay was also introduced. Our data showed that either anti-C23 or anti-CXCR4 antibodies efficaciously repressed the phosphorylation levels of ERK (p < 0.000) and AKT (p < 0.000) compared with SDF-1 alone and control. As expected, either C23 or CXCR4 siRNAs indeed resulted in C23 and CXCR4 knockdown and further suppressed the expression of p-ERK and p-AKT. Most importantly, immunoprecipitation revealed C23 interacted with CXCR4 once U87 was exposed to SDF-1 treatment. In addition, MTT assay identified that C23 or CXCR4 siRNAs could obviously decreased cell proliferation capacity (p = 0.002). In conclusion, our results suggest that C23 plays a crucial role in activation of SDF-1-induced ERK and PI3K/AKT pathways via interacting with CXCR4. Furthermore, C23 could be recommended as an important element in glioblastoma development and a new target for glioblastoma treatment.

Cell surface protein C23 affects EGF-EGFR induced activation of ERK and PI3K-AKT pathways.

The epidermal growth factor (EGF) pathway has been reported as canonical causes in cancer development. Meanwhile, the involvement of C23 in multiple signaling pathways has been also investigated (Lv et al., 2014). However, the effect of C23 on EGF pathway in glioblastoma is not fully characterized. In the present study, C23 and the epidermal growth factor receptor (EGFR) of U251 cell line were inhibited by C23 and EGFR antibodies, respectively; and then C23 and EGFR siRNAs were used to knock down endogenous C23 and EGFR, respectively. In addition, soft-agar and MTT assay were also introduced. Compared with control, either C23 or EGFR antibodies efficiently repressed the phosphorylation levels of ERK1/2 (p<0.000) and AKT (p<0.000). Similarly, either C23 or EGFR siRNAs indeed resulted in C23 and EGFR knockdown, and further suppressed the expression of p-ERK1/2 and p-AKT. Most importantly, immunoprecipitation revealed C23 interacted with EGFR once U251 was exposed to EGF treatment. In addition, the MTT and soft-agar assay also identified that C23 or EGFR siRNAs could obviously affected cell growth (p=0.004) and invasiveness, as cell viability and colony formation decreased markedly. Our results suggest that C23 plays a crucial role in activation of EGF-induced ERK and PI3K-AKT pathways via interacting with EGFR; furthermore, C23 could be indicative of an important factor in glioblastoma development and a useful target for glioblastoma treatment.

Nucleolar changes in bovine nucleotransferred embryos.

This study focused on nucleolar changes in bovine embryos reconstructed from enucleated mature oocytes fused with blastomeres of morulae or with cultured, serum unstarved bovine fetal skin fibroblasts (embryonic vs. somatic cloning). The nucleotransferred (NT) embryos were collected and fixed at time intervals of 1-2 h (early 1-cell stage), 10-15 h (late 1-cell stage), 22-24 h (2-cell stage), 37-38 h (4-cell stage), 40-41 h (early 8-cell stage), 47-48 h (late 8-cell stage), and 55 h (16-cell stage) after fusion. Immunocytochemistry by light and electron microscopy was used for structure-function characterization of nucleolar components. Antibodies against RNA, protein B23, protein C23, and fibrillarin were applied. In addition, DNA was localized by the terminal deoxynucleotidyl transferase (TdT) technique, and the functional organization of chromatin was determined with the nick-translation immunogold approach. The results show that fully reticulated (active) nucleoli observed in donor cells immediately before fusion as well as in the early 1-cell stage after fusion were progressively transformed into nucleolar bodies displaying decreasing numbers of vacuoles from the 2- to 4-cell stage in both types of reconstructed embryos. At the late 8-cell stage, morphological signs of resuming nucleolar activity were detected. Numerous new small vacuoles appeared, and chromatin blocks reassociated with the nucleolar body. During this period, nick-translation technique revealed numerous active DNA sites in the periphery of chromatin blocks associated with the nucleolar body. Fully reticulated nucleoli were again observed as early as the 16-cell stage of embryonic cloned embryos. In comparison, the embryos obtained by fetal cloning displayed a lower tendency to develop, mainly during the first cell cycle and during the period of presumed reactivation. Correlatively, the changes in nucleolar morphology (desegregation and rebuilding) were at least delayed in many somatic NT embryos in comparison with the embryonic NT group. It is concluded that complete reprogramming of rRNA gene expression is part of the general nuclear reprogramming necessary for development after NT.

In vivo interaction of nucleophosmin/B23 and protein C23 during cell cycle progression in HeLa cells

By using the cross-linking reagent, DSP, efforts were made to identify the protein(s) that interact with nucleophosmin/B23. A cross-linked protein complex at molecular weight of about 140 kDa was recognized by both nucleophosmin/B23 and protein C23 MAbs. Both C23 and nucleophosmin/B23 could be detected from the cross-linked complex immunoprecipitated by C23 MAb. The association between nucleophosmin/B23 and protein C23 while being observed at interphase and cytokinesis, was not detected in prometaphase and metaphase cells. Interactions of nucleophosmin/B23 with C23 not only could be found in cells in which nucleophosmin/B23 and C23 were both mainly localized to the nucleolus, but also in cells in which nucleophosmin/B23 and C23 had translocated from the nucleolus to the nucleoplasm during actinomycin D-induced cell growth inhibition. The purified recombinant GST-B23 being phosphorylated by prometaphase cell extracts (nocodazole-arrested cells) or cdc2 kinase could still be co-immunoprecipitated with C23. Consequently, the fact that nucleophosmin/B23 did not interact with C23 during mitosis could not be explained simply by mitotic phosphorylation of nucleophosmin/B23. Our findings suggest some possibilities for further elucidation of the actions of nucleophosmin/B23 and protein C23 in cell cycle progression and cell growth.

Nucleologenesis in the cleaving bovine embryo: immunocytochemical aspects.

In vivo nucleologenesis was studied in bovine embryos by electron microscopic immunogold labelling of DNA, RNA, protein C23 and protein B23. We have used the classification of Kopecný et al. (1989b) and Kopecný (1990) dividing nucleologenesis in four steps: compact nucleolar precursor body (NPB), monovacuolated NPB, NPB containing secondary vacuoles and fully reticulated nucleolus. These different features of early bovine embryo nucleologenesis were mainly observed during the eight-cell stage. In the first step of nucleolar development, the association of compact NPB with DNA structures was observed. DNA was also labelled in some small secondary vacuoles appearing during the third developmental step. From the second step onward, the labelling of protein C23 was observed in the compact fibrillar network of the NPB. Protein B23 started to be labelled in the compact fibrillar mass at the third step. RNA labelling was also observed for the first time in NPB containing secondary vacuoles. Labelled RNA was located in the peripheral region of compact fibrillar mass as well as along the border of the vacuoles. In the reticulated nucleolus, the dense fibrillar component was found to contain both proteins and RNA.

C23 interacts with B23, a putative nucleolar-localization-signal-binding protein.

The human protein C23 (nucleolin) is a major nucleolar protein. Its interactions with other proteins were studied with the two-hybrid system which identified nucleolar protein B23 (nucleophosmin) as being associated with C23. Both proteins were co-immunoprecipitated from HeLa cell nuclear extract by either monoclonal anti-C23 or monoclonal anti-B23. Binding studies utilizing deletion mutants indicated that the binding of C23 and B23 involves specific motifs. In addition to an approximately 46-amino-acid-binding domain in B23 (amino acids 194-239), amino acids 540-628 of C23 were required for binding; this region of C23 is required for the nucleolar localization. In addition, nucleolar protein p120 was also found to be co-immunoprecipitated with B23. A fragment of p120 containing a functional nucleolar localization signal bound to the truncated binding domain of B23, as did C23. These results suggest that the interaction of C23 and B23 may represent a nucleolar-targeting mechanism in which B23 acts as a nucleolar-localization signal-binding protein.

Limited proteolysis as a probe of the conformation and nucleic acid binding regions of nucleolin

Nucleolin, also called protein C23, is a RNA-associated protein implicated in the early stages of ribosome assembly. To study the general conformation and map the nucleic acid binding regions, rat nucleolin was subjected to limited proteolysis using trypsin and chymotrypsin in the presence or absence of poly(G). The cleavage sites were classified according to their locations in the three putative domains: the highly polar amino-terminal domain, the central nucleic acid binding domain, which contains four 90-residue repeats, and the carboxyl-terminal domain, which is rich is glycine, dimethylarginine, and phenylalanine. The most labile sites were found in basic segments of the amino-terminal domain. This region was stabilized by Mg2+. At low enzyme concentrations, cleavage by trypsin or chymotrypsin in the amino-terminal domain was enhanced by poly(G). Trypsin produced a relatively stable 48-kDa fragment containing the central and carboxyl-terminal domains. The enhanced cleavage suggests that binding of nucleic acid by the central domain alters the conformation of the amino-terminal domain, exposing sites to proteolytic cleavage. At moderate enzyme concentrations, the 48-kDa fragment was protected by poly(G) against tryptic digestion. At the highest enzyme concentrations, both enzymes cleaved near the boundaries between repeats 2, 3, and 4 with some sites protected by poly(G), suggesting that the repeats themselves form compact units. The carboxyl-terminal domain was resistant to trypsin but was cleaved by chymotrypsin either in the presence or in the absence of poly(G), indicating exposure of some phenylalanines in this region. These studies provide a general picture of the topology of nucleolin and suggest that the nucleic acid binding region communicates with the amino-terminal domain.

Simultaneous immunoelectron microscopic visualization of protein B23 and C23 distribution in the HeLa cell nucleolus.

The intranucleolar distribution of phosphoproteins B23 and C23 was visualized simultaneously by post-embedding immunoelectron microscopy in HeLa cell nucleoli, using specific antibodies. The data show that proteins B23 and C23 co-localize to the same nucleolar compartments, i.e., the dense fibrillar component and the granular component. Neither of the two antibodies is significantly associated with the fibrillar centers in these cells, although the fibrillar centers appear positive after silver staining. These findings suggest that other unidentified components must be responsible for the silver staining observed in the fibrillar centers of interphase nucleoli. The results are discussed in the light of previously reported data obtained by preembedding immunolabeling techniques and by silver staining, which both suggested a localization of protein C23 inside the fibrillar centers.

Effects of anti-C23 (Nucleolin) antibody on transcription of ribosomal DNA in Chironomus salivary gland cells

Protein C23 (also called nucleolin or 100-kDa nucleolar protein) is a major nucleolar phosphoprotein involved in ribosome biogenesis. To determine the effects of protein C23 on preribosomal RNA (pre-rRNA) synthesis anti-C23 antiserum was microinjected into nuclei of Chironomus tentans salivary glands. Transcription was measured by incubation of the glands with 32P-labeled RNA precorsors followed by microdissection of nucleoli, RNA extraction, and electrophoretic analyses. Injection of the anti-C23 antibody caused a 2- to 3.5-fold stimulation of 32P incorporation into 38S pre-rRNA. No stimulation was observed in salivary glands injected with preimmune serum or antiserum preabsorbed with protein C23. The stimulatory effect was selective for pre-rRNA as indicated by the lack of stimulation of 32P incorporation into extranucleolar RNA. Injection of the antiserum produced little or no effect on pre-RNA processing as measured by the relative amounts of 32P-labeled intermediate cleavage products of pre-rRNA in stimulated versus control glands. When protein extracts of Chironomus tentans salivary gland nuclei were probed on Western blots with anti-C23 antibody the predominant cross-reacting species was a 110-kDa polypeptide which had an electrophoretic mobility similar to that of protein C23. These results suggest that protein C23 not only is involved in ribosome assembly but also plays a role in regulating the transcription of the preribosomal RNA.

CDNA and deduced primary structure of rat protein B23, a nucleolar protein containing highly conserved sequences.

Protein B23 (Mr/pI = 38,000/5.1) is a major RNA-associated nucleolar phosphoprotein which contains highly acidic segments and has a high affinity for silver ions. Using synthetic oligonucleotides as probes cloned cDNAs encoding protein B23 were isolated and characterized. One of the cDNAs, obtained from a rat brain library, contained an insert of 1232 base pairs of DNA encoding a polypeptide of 292 amino acid residues. Segments of the protein sequence were confirmed by partial sequencing of CNBr fragments from rat hepatoma protein B23. The protein contains a methionine-rich amino-terminal sequence and two highly acidic segments in the center of the sequence. The first acidic segment, in which 11 of the 13 residues are acidic, begins at residue 120 and contains a major phosphorylation site. In the second segment (residues 159-187) there are four copies of the sequence Asp-Asp-Glu, and all but two of the 29 residues have acidic side chains. When the sequence of the rat protein was compared with available sequences from other species a high degree of conservation was found; the 77-residue carboxyl-terminal sequence is identical with that of human protein B23 (Chan, P. K., Chan, W.-Y., Yung, B. Y. M., Cook, R. G., Aldrich, M. B., Ku, D., Goldknopf, I. L., and Busch, H. (1986) J. Biol. Chem. 261, 14335-24341), and about 63% of the residues are identical when the rat B23 sequence is compared with protein N038 from Xenopus laevis (Schmidt-Zachmann, M. S., Hügle-Dörr, B., and Franke, W. (1987) EMBO J. 6, 1881-1890). Except for the presence of highly acidic regions no significant similarities were found with protein C23 (nucleolin), the other major nucleolar protein.

Studies on satellite nucleoli in rat hepatocytes and Novikoff hepatoma cells.

The present study was undertaken to provide information on the presence and frequency of satellite nucleoli in cells with increased nucleolar biosynthetic activity. The number of hepatocytes containing satellite nucleoli was analyzed in rat liver, regenerating liver 18 h after partial hepatectomy and in Novikoff hepatoma ascites cells. In comparison with hepatocytes of normal liver, the number of both stimulated hepatocytes and malignant hepatoma cells containing satellite nucleoli was significantly reduced. The results also indicated that whereas most satellite nucleoli contain protein C23, a smaller percentage contain protein B23.

Primary structure of human nuclear ribonucleoprotein particle C proteins: conservation of sequence and domain structures in heterogeneous nuclear RNA, mRNA, and pre-rRNA-binding proteins.

In the eucaryotic nucleus, heterogeneous nuclear RNAs exist in a complex with a specific set of proteins to form heterogeneous nuclear ribonucleoprotein particles (hnRNPs). The C proteins, C1 and C2, are major constituents of hnRNPs and appear to play a role in RNA splicing as suggested by antibody inhibition and immunodepletion experiments. With the use of a previously described partial cDNA clone as a hybridization probe, full-length cDNAs for the human C proteins were isolated. All of the cDNAs isolated hybridized to two poly(A)+ RNAs of 1.9 and 1.4 kilobases (kb). DNA sequencing of a cDNA clone for the 1.9-kb mRNA (pHC12) revealed a single open reading frame of 290 amino acids coding for a protein of 31,931 daltons and two polyadenylation signals, AAUAAA, approximately 400 base pairs apart in the 3' untranslated region of the mRNA. DNA sequencing of a clone corresponding to the 1.4-kb mRNA (pHC5) indicated that the sequence of this mRNA is identical to that of the 1.9-kb mRNA up to the first polyadenylation signal which it uses. Both mRNAs therefore have the same coding capacity and are probably transcribed from a single gene. Translation in vitro of the 1.9-kb mRNA selected by hybridization with a 3'-end subfragment of pHC12 demonstrated that it by itself can direct the synthesis of both C1 and C2. The difference between the C1 and C2 proteins which results in their electrophoretic separation is not known, but most likely one of them is generated from the other posttranslationally. Since several hnRNP proteins appeared by sodium dodecyl sulfate-polyacrylamide gel electrophoresis as multiple antigenically related polypeptides, this raises the possibility that some of these other groups of hnRNP proteins are also each produced from a single mRNA. The predicted amino acid sequence of the protein indicates that it is composed of two distinct domains: an amino terminus that contains what we have recently described as a RNP consensus sequence, which is the putative RNA-binding site, and a carboxy terminus that is very negatively charged, contains no aromatic amino acids or prolines, and contains a putative nucleoside triphosphate-binding fold, as well as a phosphorylation site for casein kinase type II. The RNP consensus sequence was also found in the yeast poly(A)-binding protein (PABP), the heterogeneous nuclear RNA-binding proteins A1 and A2, and the pre-rRNA binding protein C23. All of these proteins are also composed of at least two distinct domains: an amino terminus, which possesses one or more RNP consensus sequences, and a carboxy terminus, which is unique to each protein, being very acidic in the C proteins and rich in glycine in A1, and C23 and rich in proline in the poly(A)-binding protein. These findings suggest that the amino terminus of these proteins possesses a highly conserved RNA-binding domain, whereas the carboxy terminus contains a region essential to the unique function and interactions of each of the RNA-binding proteins.

Primary structure of human nuclear ribonucleoprotein particle C proteins: conservation of sequence and domain structures in heterogeneous nuclear RNA, mRNA, and pre-rRNA-binding proteins.

In the eucaryotic nucleus, heterogeneous nuclear RNAs exist in a complex with a specific set of proteins to form heterogeneous nuclear ribonucleoprotein particles (hnRNPs). The C proteins, C1 and C2, are major constituents of hnRNPs and appear to play a role in RNA splicing as suggested by antibody inhibition and immunodepletion experiments. With the use of a previously described partial cDNA clone as a hybridization probe, full-length cDNAs for the human C proteins were isolated. All of the cDNAs isolated hybridized to two poly(A)+ RNAs of 1.9 and 1.4 kilobases (kb). DNA sequencing of a cDNA clone for the 1.9-kb mRNA (pHC12) revealed a single open reading frame of 290 amino acids coding for a protein of 31,931 daltons and two polyadenylation signals, AAUAAA, approximately 400 base pairs apart in the 3' untranslated region of the mRNA. DNA sequencing of a clone corresponding to the 1.4-kb mRNA (pHC5) indicated that the sequence of this mRNA is identical to that of the 1.9-kb mRNA up to the first polyadenylation signal which it uses. Both mRNAs therefore have the same coding capacity and are probably transcribed from a single gene. Translation in vitro of the 1.9-kb mRNA selected by hybridization with a 3'-end subfragment of pHC12 demonstrated that it by itself can direct the synthesis of both C1 and C2. The difference between the C1 and C2 proteins which results in their electrophoretic separation is not known, but most likely one of them is generated from the other posttranslationally. Since several hnRNP proteins appeared by sodium dodecyl sulfate-polyacrylamide gel electrophoresis as multiple antigenically related polypeptides, this raises the possibility that some of these other groups of hnRNP proteins are also each produced from a single mRNA. The predicted amino acid sequence of the protein indicates that it is composed of two distinct domains: an amino terminus that contains what we have recently described as a RNP consensus sequence, which is the putative RNA-binding site, and a carboxy terminus that is very negatively charged, contains no aromatic amino acids or prolines, and contains a putative nucleoside triphosphate-binding fold, as well as a phosphorylation site for casein kinase type II. The RNP consensus sequence was also found in the yeast poly(A)-binding protein (PABP), the heterogeneous nuclear RNA-binding proteins A1 and A2, and the pre-rRNA binding protein C23. All of these proteins are also composed of at least two distinct domains: an amino terminus, which possesses one or more RNP consensus sequences, and a carboxy terminus, which is unique to each protein, being very acidic in the C proteins and rich in glycine in A1, and C23 and rich in proline in the poly(A)-binding protein. These findings suggest that the amino terminus of these proteins possesses a highly conserved RNA-binding domain, whereas the carboxy terminus contains a region essential to the unique function and interactions of each of the RNA-binding proteins.

Association of protein C23 with rapidly labeled nucleolar RNA.

The association of nucleolar phosphoprotein C23 with preribosomal ribonucleoprotein (RNP) particles was examined in Novikoff hepatoma nucleoli. RNA was labeled with [3H]uridine for various times in cell suspensions, and RNP particles were extracted from isolated nucleoli and fractionated by sucrose gradient ultracentrifugation. The majority of protein C23 cosedimented with fractions containing rapidly labeled RNA (RL fraction). To determine whether there was a direct association of RNA with protein C23, the RL fraction was exposed to ultraviolet (UV) light (254 nm) for short periods of time. After 2 min of exposure there was a 50% decrease in C23 as measured by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) analyses, with no significant further decrease at longer times. When UV-treated fractions were subjected to phenol/chloroform extractions, as much as 30% of the labeled RNA was found in the phenol (protein) layer, indicating that RNA became cross-linked to protein. Similarly, there was an increase in protein C23 extracted into the water layer after irradiation. By SDS-PAGE analyses the cross-linked species migrated more slowly than protein C23, appearing as a smear detected either by [3H]uridine radioactivity or by anti-C23 antibody. With anti-C23 antibodies, up to 25% of the labeled RNA was precipitated from the RL fraction. Dot-blot hybridizations, using cloned rDNA fragments as probes, indicated that the RNA in the RL fraction and the immunoprecipitated RNA contained sequences from 18S and 28S ribosomal RNA.(ABSTRACT TRUNCATED AT 250 WORDS)

Proteins and RNA in mouse L cell core nucleoli and nucleolar matrix

When intact nucleoli were prepared in the presence of enough leupeptin and phenylmethanesulfonyl fluoride to inhibit protease action, electrophoretic patterns of their constituent proteins were reproducible and very similar for L, HeLa, CHO, and rat hepatoma cells. "Core nucleoli", defined as that nucleolar fraction which remains after extensive DNase I action, had a protein composition similar to that of crude intact nucleoli, but were enriched for snRNA U3. Core nucleolar proteins included all of the histones, ribosomal proteins, and phosphorylated proteins with mobilities corresponding to 110 (protein C23) and 160 kilodaltons (kDa). The presence of protein C23 and of lamins A and C in nucleoli and core nucleoli was further verified by reaction with specific antibodies after one- or two-dimensional electrophoresis. A class of higher molecular weight proteins, ranging from 70 to greater than 200 kDa by mobility, was observed. It included at least 25 specific proteins, almost all of them highly acidic (pI less than 3.5). Treatment of core nucleoli with ethylenediaminetetraacetic acid/hypotonic buffer solubilized 30-35% of the small and large molecular weight proteins. In contrast, washing core nucleoli with 2 M NaCl selectively released U3 snRNA, 95% of the ribosomal RNA, and about half of the proteins, including C23 and most of the histones, ribosomal proteins, and other lower molecular weight proteins. The fraction remaining insoluble, "nucleolar matrix", was enriched for proteins of 34 and 57 kDa, lamins A and C, and most higher molecular weight proteins, as well as a portion of ribosomal spacer DNA.

Protein and cDNA sequence of a glycine-rich, dimethylarginine-containing region located near the carboxyl-terminal end of nucleolin (C23 and 100 kDa).

By a combination of protein chemistry and recombinant DNA methods a glycine-rich region was found to be located near the carboxyl terminus of the nucleolar specific phosphoprotein, nucleolin, from Novikoff hepatoma (protein C23) and Chinese hamster ovary cells (100-kDa nucleolar protein). A sequence of 192 amino acid residues was derived from partial sequences of cyanogen bromide and N-bromosuccinimide fragments of protein C23 and deduced protein sequence from Chinese hamster ovary cell 100-kDa cDNA sequences. The 66 residues sequenced by protein methods were identical to the corresponding residues deduced by DNA sequencing. The multiple residues of NG,NG-dimethylarginine (DMA) contained in the nucleolin polypeptide were found to be limited to a segment of less than 10 kDa near the carboxyl-terminal end of the protein. This segment also contained internally repeated sequences (e.g. 7 copies of the sequence Gly-Gly-Arg-Gly-Gly were found) which were unrelated to sequences closer to the amino-terminal end. Most arginine residues in this region were surrounded by 2 or 3 glycine residues and were relatively close in sequence to phenylalanine residues.