Direct Interaction of the Spinal Muscular Atrophy Disease Protein SMN with the Small Nucleolar RNA-associated Protein Fibrillarin
2001; Elsevier BV; Volume: 276; Issue: 42 Linguagem: Inglês
10.1074/jbc.m106161200
ISSN1083-351X
AutoresKevin W. Jones, Karen Gorzynski, Chadwick M. Hales, Utz Fischer, Farah Badbanchi, Rebecca M. Terns, Michael P. Terns,
Tópico(s)RNA Research and Splicing
ResumoDisruption of the survival motor neuron (SMN) gene leads to selective loss of spinal motor neurons, resulting in the fatal human neurodegenerative disorder spinal muscular atrophy (SMA). SMN has been shown to function in spliceosomal small nuclear ribonucleoprotein (snRNP) biogenesis and pre-mRNA splicing. We have demonstrated that SMN also interacts with fibrillarin, a highly conserved nucleolar protein that is associated with all Box C/D small nucleolar RNAs and functions in processing and modification of rRNA. Fibrillarin and SMN co-immunoprecipitate from HeLa cell extracts indicating that the proteins exist as a complex in vivo. Furthermore, in vitro binding studies indicate that the interaction between SMN and fibrillarin is direct and salt-stable. We show that the glycine/arginine-rich domain of fibrillarin is necessary and sufficient for SMN binding and that the region of SMN encoded by exon 3, including the Tudor domain, mediates the binding of fibrillarin. Tudor domain missense mutations, including one found in an SMA patient, impair the interaction between SMN and fibrillarin (as well as the common snRNP protein SmB). Our results suggest a function for SMN in small nucleolar RNP biogenesis (akin to its known role as an snRNP assembly factor) and reveal a potential link between small nucleolar RNP biogenesis and SMA. Disruption of the survival motor neuron (SMN) gene leads to selective loss of spinal motor neurons, resulting in the fatal human neurodegenerative disorder spinal muscular atrophy (SMA). SMN has been shown to function in spliceosomal small nuclear ribonucleoprotein (snRNP) biogenesis and pre-mRNA splicing. We have demonstrated that SMN also interacts with fibrillarin, a highly conserved nucleolar protein that is associated with all Box C/D small nucleolar RNAs and functions in processing and modification of rRNA. Fibrillarin and SMN co-immunoprecipitate from HeLa cell extracts indicating that the proteins exist as a complex in vivo. Furthermore, in vitro binding studies indicate that the interaction between SMN and fibrillarin is direct and salt-stable. We show that the glycine/arginine-rich domain of fibrillarin is necessary and sufficient for SMN binding and that the region of SMN encoded by exon 3, including the Tudor domain, mediates the binding of fibrillarin. Tudor domain missense mutations, including one found in an SMA patient, impair the interaction between SMN and fibrillarin (as well as the common snRNP protein SmB). Our results suggest a function for SMN in small nucleolar RNP biogenesis (akin to its known role as an snRNP assembly factor) and reveal a potential link between small nucleolar RNP biogenesis and SMA. small nucleolar survival motor neuron spinal muscular atrophy ribonucleoprotein glutathione S-transferase glycine/arginine-rich Fibrillarin is one of four proteins known to interact selectively with all Box C/D family small nucleolar (sno)1 RNAs (1Lafontaine D.L. Tollervey D. RNA (New York). 1999; 5: 455-467Crossref PubMed Scopus (139) Google Scholar, 2Lafontaine D.L. Tollervey D. Mol. Cell. Biol. 2000; 20: 2650-2659Crossref PubMed Scopus (128) Google Scholar, 3Watkins N.J. Segault V. Charpentier B. Nottrott S. Fabrizio P. Bachi A. Wilm M. Rosbash M. Branlant C. Luhrmann R. Cell. 2000; 103: 457-466Abstract Full Text Full Text PDF PubMed Scopus (284) Google Scholar, 4Wu P. Brockenbrough J.S. Metcalfe A.C. Chen S. Aris J.P. J. Biol. Chem. 1998; 273: 16453-16463Abstract Full Text Full Text PDF PubMed Scopus (118) Google Scholar, 5Tyc K. Steitz J.A. EMBO J. 1989; 8: 3113-3119Crossref PubMed Scopus (303) Google Scholar, 6Schimmang T. Tollervey D. Kern H. Frank R. Hurt E. EMBO J. 1989; 8: 4015-4024Crossref PubMed Scopus (236) Google Scholar). In eukaryotes, large numbers of snoRNAs direct cleavages and specific nucleotide modifications of pre-rRNA that are required for ribosome biogenesis (7Filipowicz W. Pelczar P. Pogacic V. Dragon F. Acta Biochim. Pol. 1999; 46: 377-389Crossref PubMed Scopus (34) Google Scholar, 8Maxwell E.S. Fournier M.J. Annu. Rev. Biochem. 1995; 64: 897-934Crossref PubMed Scopus (529) Google Scholar, 9Weinstein L.B. Steitz J.A. Curr. Opin. Cell Biol. 1999; 11: 378-384Crossref PubMed Scopus (248) Google Scholar, 10Tollervey D. Kiss T. Curr. Opin. Cell Biol. 1997; 9: 337-342Crossref PubMed Scopus (369) Google Scholar, 11Terns, M. P., and Terns, R. M. (2001) Gene Expr.,in press.Google Scholar). Most Box C/D family snoRNAs guide the site-specific 2′-O-methylation of rRNA (12Bachellerie J.-P. Cavaille J. Grosjean H. Benne B. Modification and Editing of RNA. American Society for Microbiology, Washington, D. C.1998: 255-272Google Scholar, 13Kiss-Laszlo Z. Henry Y. Bachellerie J.P. Caizergues-Ferrer M. Kiss T. Cell. 1996; 85: 1077-1088Abstract Full Text Full Text PDF PubMed Scopus (648) Google Scholar). snoRNAs function as RNA-protein complexes known as small nucleolar ribonucleoprotein particles (snoRNPs) (11Terns, M. P., and Terns, R. M. (2001) Gene Expr.,in press.Google Scholar, 14Venema J. Tollervey D. Annu. Rev. Genet. 1999; 33: 261-311Crossref PubMed Scopus (649) Google Scholar). Fibrillarin has sequence and structural homology to known methyltransferases (15Niewmierzycka A. Clarke S. J. Biol. Chem. 1999; 274: 814-824Abstract Full Text Full Text PDF PubMed Scopus (190) Google Scholar, 16Wang H. Boisvert D. Kim K.K. Kim R. Kim S.H. EMBO J. 2000; 19: 317-323Crossref PubMed Scopus (143) Google Scholar). This observation, coupled with functional studies in yeast (17Tollervey D. Lehtonen H. Jansen R. Kern H. Hurt E.C. Cell. 1993; 72: 443-457Abstract Full Text PDF PubMed Scopus (403) Google Scholar), has led to the hypothesis that fibrillarin is the catalytic factor in Box C/D snoRNA-directed 2′-O-methylation of ribosomal RNA. The survival motor neuron (SMN) protein is linked with one of the most common inheritable causes of childhood mortality, spinal muscular atrophy (SMA) (18Czeizel A. Hamula J. J. Med. Genet. 1989; 26: 761-763Crossref PubMed Scopus (72) Google Scholar, 19Pearn J. J. Med. Genet. 1978; 15: 409-413Crossref PubMed Scopus (405) Google Scholar, 20Pearn J.H. J. Med. Genet. 1973; 10: 260-265Crossref PubMed Scopus (140) Google Scholar, 21Roberts D.F. Chavez J. Court S.D. Arch. Dis. Child. 1970; 45: 33-38Crossref PubMed Scopus (135) Google Scholar). The SMN1 gene is deleted or mutated in patients with SMA (22Lefebvre S. Burglen L. Reboullet S. Clermont O. Burlet P. Viollet L. Benichou B. Cruaud C. Millasseau P. Zeviani M. Le Paslier D. Frezal J. Cohen D. Weissenbach J. Munnich A. Melki J. Cell. 1995; 80: 155-165Abstract Full Text PDF PubMed Scopus (2895) Google Scholar) resulting in loss of spinal motor neurons accompanied by progressive muscular atrophy. SMN has been implicated in an array of cellular pathways. Antibody inhibition experiments have demonstrated that SMN is required for the biogenesis of spliceosomal small nuclear RNPs (snRNPs) (23Buhler D. Raker V. Luhrmann R. Fischer U. Hum. Mol. Genet. 1999; 8: 2351-2357Crossref PubMed Scopus (217) Google Scholar, 24Fischer U. Liu Q. Dreyfuss G. Cell. 1997; 90: 1023-1029Abstract Full Text Full Text PDF PubMed Scopus (549) Google Scholar), and studies using in vitro splicing systems and yeast mutants have demonstrated a key role for SMN in pre-mRNA splicing (25Meister G. Buhler D. Laggerbauer B. Zobawa M. Lottspeich F. Fischer U. Hum. Mol. Genet. 2000; 9: 1977-1986Crossref PubMed Scopus (121) Google Scholar, 26Pellizzoni L. Kataoka N. Charroux B. Dreyfuss G. Cell. 1998; 95: 615-624Abstract Full Text Full Text PDF PubMed Scopus (461) Google Scholar, 27Hannus S. Buhler D. Romano M. Seraphin B. Fischer U. Hum. Mol. Genet. 2000; 9: 663-674Crossref PubMed Scopus (63) Google Scholar). SMN has also been implicated in regulation of gene expression at the transcriptional level (28Strasswimmer J. Lorson C.L. Breiding D.E. Chen J.J. Le T. Burghes A.H. Androphy E.J. Hum. Mol. Genet. 1999; 8: 1219-1226Crossref PubMed Scopus (93) Google Scholar, 29Williams B.Y. Hamilton S.L. Sarkar H.K. FEBS Lett. 2000; 470: 207-210Crossref PubMed Scopus (54) Google Scholar), in the assembly of the polymerase II transcription machinery (30Pellizzoni L. Charroux B. Rappsilber J. Mann M. Dreyfuss G. J. Cell Biol. 2001; 152: 75-85Crossref PubMed Scopus (197) Google Scholar), and as a neuron-specific anti-apoptotic factor (31Sato K. Eguchi Y. Kodama T.S. Tsujimoto Y. Cell Death Differ. 2000; 7: 374-383Crossref PubMed Scopus (23) Google Scholar, 32Iwahashi H. Eguchi Y. Yasuhara N. Hanafusa T. Matsuzawa Y. Tsujimoto Y. Nature. 1997; 390: 413-417Crossref PubMed Scopus (175) Google Scholar, 33Kerr D.A. Nery J.P. Traystman R.J. Chau B.N. Hardwick J.M. Proc. Natl. Acad. Sci. U. S. A. 2000; 97: 13312-13327Crossref PubMed Scopus (87) Google Scholar). To learn more about snoRNP biogenesis and structure, we performed a yeast two-hybrid screen for proteins that interact withXenopus fibrillarin. We identified the survival motor neuron (SMN) gene multiple times in screens of both Xenopus and human cDNA libraries. Fibrillarin had been detected previously in a two-hybrid screen using SMN (34Liu Q. Dreyfuss G. EMBO J. 1996; 15: 3555-3565Crossref PubMed Scopus (634) Google Scholar). We have now demonstrated an in vivo interaction between SMN and fibrillarin, and we show that the two proteins interact directly in vitro. We have mapped the domains of each protein responsible for the interaction, and we found that SMN interacts with fibrillarin and the SmB snRNP protein via the same domain. Our findings suggest a function for SMN in the biogenesis and/or function of snoRNPs (similar to its established role with snRNPs (23Buhler D. Raker V. Luhrmann R. Fischer U. Hum. Mol. Genet. 1999; 8: 2351-2357Crossref PubMed Scopus (217) Google Scholar, 24Fischer U. Liu Q. Dreyfuss G. Cell. 1997; 90: 1023-1029Abstract Full Text Full Text PDF PubMed Scopus (549) Google Scholar, 25Meister G. Buhler D. Laggerbauer B. Zobawa M. Lottspeich F. Fischer U. Hum. Mol. Genet. 2000; 9: 1977-1986Crossref PubMed Scopus (121) Google Scholar, 26Pellizzoni L. Kataoka N. Charroux B. Dreyfuss G. Cell. 1998; 95: 615-624Abstract Full Text Full Text PDF PubMed Scopus (461) Google Scholar, 35Liu Q. Fischer U. Wang F. Dreyfuss G. Cell. 1997; 90: 1013-1021Abstract Full Text Full Text PDF PubMed Scopus (529) Google Scholar, 36Pellizzoni L. Charroux B. Dreyfuss G. Proc. Natl. Acad. Sci. U. S. A. 1999; 96: 11167-11172Crossref PubMed Scopus (216) Google Scholar)). The yeast strain PJ69-4A (MATa, trp1-901, leu2-3, 112, ura3-52, his3-200, gal4Δ, gal80Δ, GAL2-ADE2, LYS2::GAL1-HIS3, met2::GAL7-lacZ, generously provided by Phil James, University of Wisconsin) was used for both two-hybrid screens. Xenopus fibrillarin (37Lapeyre B. Mariottini P. Mathieu C. Ferrer P. Amaldi F. Amalric F. Caizergues-Ferrer M. Mol. Cell. Biol. 1990; 10: 430-434Crossref PubMed Scopus (106) Google Scholar) in pGBT9 was used to screen both a Xenopus (CLONTECH) and HeLa cell (generously provided by William Marzluff, University of North Carolina) cDNA GAL4 fusion library. Transformants were screened on media deficient in histidine or adenine and were tested for β-galactosidase activity using a liquid colorimetric assay (CLONTECH). Library constructs were recovered from yeast that expressed all three reporter genes and sequenced. Reporter gene expression was confirmed with recovered constructs following co-transformation specifically with the fibrillarin bait construct. HeLa S3 cells (ATCC) were cultured by standard procedures. Preparation of nucleolar extracts was essentially as described previously (38Szekely A.M. Chen Y.H. Zhang C. Oshima J. Weissman S.M. Proc. Natl. Acad. Sci. U. S. A. 2000; 97: 11365-11370Crossref PubMed Scopus (102) Google Scholar, 39Jordan P. Mannervik M. Tora L. Carmo-Fonseca M. J. Cell Biol. 1996; 133: 225-234Crossref PubMed Scopus (131) Google Scholar) with a modified lysis buffer (10 mm Tris-HCl (pH 7.5), 10 mm KCl, 2 mm MgCl2, 0.05% Triton X-100, 1 mmdithiothreitol, protease inhibitor mixture (Roche Molecular Biochemicals)). Approximately two confluent, 10-cm HeLa cultures were used for each immunoprecipitation. Nonidet P-40 (Sigma) was added to nucleolar extract to a final concentration of 0.65%. IgG from the hybridoma cell line SP2/0 (courtesy of Serafin Pinol-Roma, Mount Sinai School of Medicine) and the monoclonal anti-fibrillarin antibody 72B9 (40Reimer G. Pollard K.M. Penning C.A. Ochs R.L. Lischwe M.A. Busch H. Tan E.M. Arthritis & Rheum. 1987; 30: 793-800Crossref PubMed Scopus (221) Google Scholar) were coupled to protein-A-Sepharose (Sigma) equilibrated in IP buffer (20 mm Tris (pH 7.5), 300 mm NaCl, 1 mm EDTA, 0.65% Nonidet P-40). The coupled protein A-Sepharose was incubated with nucleolar extracts (or an equivalent amount of IP buffer) for 2 h at 4 °C, and washed 5 times for 5 min with 1 ml of cold IP buffer. Immunoblotting was carried out with the anti-SMN monoclonal antibody 7B10 (27Hannus S. Buhler D. Romano M. Seraphin B. Fischer U. Hum. Mol. Genet. 2000; 9: 663-674Crossref PubMed Scopus (63) Google Scholar) at 1:1000. Detection was by ECL (Amersham Pharmacia Biotech). Full-length cDNA clones encoding Xenopusfibrillarin (37Lapeyre B. Mariottini P. Mathieu C. Ferrer P. Amaldi F. Amalric F. Caizergues-Ferrer M. Mol. Cell. Biol. 1990; 10: 430-434Crossref PubMed Scopus (106) Google Scholar) and SMN (obtained in our two-hybrid screens) as well as human U1A (generously provided by Walther van Venrooij, University of Nijmegen) were subcloned into the glutathioneS-transferase (GST) fusion expression vector, pZEX, in which the multiple cloning site of pGEX-2T (Amersham Pharmacia Biotech) was modified to shift the reading frame. 2Z. Paroush, unpublished data. Proteins were expressed in Escherichia coli strain BL21 (room temperature induction, 5 h, 1 mmisopropyl-1-thio-β-d-galactopyranoside) and purified using glutathione-agarose (Pierce). His6-, GST-, and zz (two IgG-binding domains)-tagged human SMN constructs were prepared as described previously (23Buhler D. Raker V. Luhrmann R. Fischer U. Hum. Mol. Genet. 1999; 8: 2351-2357Crossref PubMed Scopus (217) Google Scholar). His- and zz-tagged proteins were purified by Ni2+ chelation (Qiagen) and IgG-Sepharose (Amersham Pharmacia Biotech) affinity chromatography, respectively. The SMN-Tudor construct 83/173 was a kind gift of Dr. Michael Sattler (see Ref. 41Selenko P. Sprangers R. Stier G. Buhler D. Fischer U. Sattler M. Nat. Struct. Biol. 2001; 8: 27-31Crossref PubMed Scopus (261) Google Scholar). SMN-Tudor 83/173 constructs with single amino acid replacements were generated by recombinant polymerase chain reaction. The polymerase chain reaction products were cloned via NcoI/KpnI into a modified pET24d expression vector (Novagen) containing an N-terminal His6-GST tag followed by a TEV protease cleavage site (48Samarsky D.A. Fournier M.J. Singer R.H. Bertrand E. EMBO J. 1998; 17: 3747-3757Crossref PubMed Scopus (156) Google Scholar) and sequenced. Expression and purification was carried out as described (25Meister G. Buhler D. Laggerbauer B. Zobawa M. Lottspeich F. Fischer U. Hum. Mol. Genet. 2000; 9: 1977-1986Crossref PubMed Scopus (121) Google Scholar, 48Samarsky D.A. Fournier M.J. Singer R.H. Bertrand E. EMBO J. 1998; 17: 3747-3757Crossref PubMed Scopus (156) Google Scholar). In vitro translation was carried out using a rabbit reticulocyte lysate, coupled in vitro transcription and translation kit (Promega), and [35S]methionine (Amersham Pharmacia Biotech). Xenopus fibrillarin and SMN and human U1A were translated from an SP64TEN vector (42Klein P.S. Melton D.A. Science. 1994; 265: 803-806Crossref PubMed Scopus (59) Google Scholar) that we re-engineered to include a His6, N-terminal fusion tag and a Kozak sequence. SmB and Sip1 were translated from pET vectors as described previously (23Buhler D. Raker V. Luhrmann R. Fischer U. Hum. Mol. Genet. 1999; 8: 2351-2357Crossref PubMed Scopus (217) Google Scholar). For binding assays involving two tagged recombinant proteins, E. coli cell pellets were resuspended in Ipp-250T (50 mm Tris (pH 7.5), 250 mm NaCl, 0.05% Triton X-100), mixed, and sonicated. Extracts were centrifuged at 20,000 × g for 30 min, and the supernatants were incubated with 200 µl of glutathione-agarose (Pierce), 1–2 h at 4 °C. Pellets were washed 5 times with 1 ml of Ipp-250T and resuspended in SDS sample buffer. Adjustments based on Coomassie staining were made to obtain equivalent loading of tagged proteins. Co-purified proteins were analyzed by immunoblotting with anti-SMN monoclonal antibody 7B10. For binding assays with in vitro translated proteins, similar amounts (based on Coomassie staining) of GST (or zz-)-tagged proteins were coupled to the appropriate resin equilibrated in Ipp-250T. In vitro translated proteins were added (1–3 µl) and incubated for 1 h at 4 °C in a final volume of 500 µl. Following five 1-ml washes with Ipp-250T, the bound proteins were eluted with SDS sample buffer and analyzed by fluorography. The data shown in Fig. 6 were quantitated by densitometric scanning of three separate experiments using a Bio-Rad Fluor-S multi-imager with Quantity 1 software. Yeast two-hybrid screens were performed to identify proteins that interact with the core Box C/D snoRNP protein fibrillarin. Separate screens were performed with full-lengthXenopus fibrillarin against a Xenopus laevis and a human (HeLa) cDNA. The screen of the Xenopus library resulted in isolation of 7 SMN clones from a bank of 4.8 × 106 transformants. These clones all coded for either full-length Xenopus SMN or a protein lacking only the first 10 amino acids. The amino acid sequence of Xenopus SMN isolated in our screen is shown in Fig.1B. When Xenopusfibrillarin was used to screen a human (HeLa cell) cDNA library, the human SMN gene was isolated 21 times among the 1.49 × 107 transformants screened. Reporter gene activation was only observed for strains containing both the fibrillarin (XFib-BD) and SMN (Xenopus XSMN-AD or human HSMN-AD) constructs (Fig.1A); the fibrillarin or SMN constructs alone did not activate reporter gene expression (Fig. 1A). The reciprocal two-hybrid interaction has also been detected using full-length human SMN in a two-hybrid screen of a human cDNA library (34Liu Q. Dreyfuss G. EMBO J. 1996; 15: 3555-3565Crossref PubMed Scopus (634) Google Scholar). The interaction of fibrillarin and SMN in the yeast two-hybrid system suggested the possibility that the proteins interact directly or indirectly in vivo. We have investigated whether fibrillarin interacts with SMN in vivoin co-immunoprecipitation experiments using HeLa cell extracts. Nuclear extracts enriched in nucleolar material were prepared from HeLa cells. Immunoblot analysis showed that essentially 100% of cellular fibrillarin and ∼5% of cellular SMN were present in the nucleolar-enriched extracts (data not shown). The antibody 72B9 (40Reimer G. Pollard K.M. Penning C.A. Ochs R.L. Lischwe M.A. Busch H. Tan E.M. Arthritis & Rheum. 1987; 30: 793-800Crossref PubMed Scopus (221) Google Scholar) was used to immunoprecipitate fibrillarin from HeLa cell extracts. Co-precipitation of SMN was assessed by immunoblotting with monoclonal antibody 7B10 against SMN (27Hannus S. Buhler D. Romano M. Seraphin B. Fischer U. Hum. Mol. Genet. 2000; 9: 663-674Crossref PubMed Scopus (63) Google Scholar). SMN was co-immunoprecipitated from the HeLa extracts by the anti-fibrillarin antibody (Fig.2A) showing that fibrillarin and SMN exist together in a complex in vivo. A mock immunoprecipitation performed without HeLa extract reveals that the other bands are contributed by the antibody (Fig. 2A). Non-immune IgG (SP2/0) failed to co-immunoprecipitate SMN (Fig.2A). These results demonstrate that fibrillarin and SMN are associated in vivo in HeLa cells. To determine whether fibrillarin and SMN interact directly or indirectly (i.e. as parts of a complex bridged by other components), the interaction of recombinant fibrillarin and SMN was tested in vitro. Recombinant proteins were expressed inE. coli; extracts were mixed; GST-tagged fibrillarin was purified; and co-purification of recombinant SMN was analyzed by immunoblotting with antibody 7B10. SMN specifically co-purified with GST-tagged fibrillarin but not with glutathione-agarose, GST tag alone, or GST-U1A (U1 snRNA-binding protein) (Fig. 2B). These results indicate that the interaction between fibrillarin and SMN is direct and is not mediated by other eukaryotic protein components. The strength and specificity of the interaction between SMN and fibrillarin were further characterized using tagged recombinant proteins andin vitro translated 35S-labeled proteins. GST-tagged SMN (either Xenopus or human SMN, results were not distinguishable) or GST-tagged fibrillarin were expressed inE. coli and coupled to glutathione-agarose. The protein-coupled beads were incubated with in vitrotranslated 35S-labeled proteins and washed extensively. Bound labeled proteins were analyzed by SDS-polyacrylamide gel electrophoresis and fluorography. The fibrillarin/SMN interaction occurred in a reciprocal manner, with GST-tagged SMN specifically binding fibrillarin, and GST-tagged fibrillarin specifically binding SMN (Fig. 3A). GST-tagged SMN interacted with itself, the snRNP-protein SmB, and SIP1/Gemin2, as demonstrated previously (23Buhler D. Raker V. Luhrmann R. Fischer U. Hum. Mol. Genet. 1999; 8: 2351-2357Crossref PubMed Scopus (217) Google Scholar, 34Liu Q. Dreyfuss G. EMBO J. 1996; 15: 3555-3565Crossref PubMed Scopus (634) Google Scholar, 35Liu Q. Fischer U. Wang F. Dreyfuss G. Cell. 1997; 90: 1013-1021Abstract Full Text Full Text PDF PubMed Scopus (529) Google Scholar, 43Lorson C.L. Strasswimmer J. Yao J.M. Baleja J.D. Hahnen E. Wirth B. Le T. Burghes A.H. Androphy E.J. Nat. Genet. 1998; 19: 63-66Crossref PubMed Scopus (404) Google Scholar), but not with GST tag alone or the U1-binding protein U1A (Fig. 3A and data not shown). We show here that fibrillarin also self-associates (as suggested previously (16Wang H. Boisvert D. Kim K.K. Kim R. Kim S.H. EMBO J. 2000; 19: 317-323Crossref PubMed Scopus (143) Google Scholar)) but did not interact with GST tag alone, SmB, SIP1/Gemin2, or the U1A protein (Fig. 3A and data not shown). Approximately equivalent amounts of GST-tagged proteins were used in these assays, as judged by Coomassie-stained gels (Fig.3A). To characterize further the interaction between SMN and fibrillarin, the assay was carried out in the presence of increasing salt concentrations (Fig. 3B). The interaction between tagged SMN and in vitro translated fibrillarin was observed in the presence of up to 1 m NaCl. Similarly, the interaction between GST-tagged fibrillarin and in vitro translated SMN was stable to 1 m NaCl washes (data not shown). These results indicate that hydrophobic interactions are important for the association of fibrillarin and SMN. Eukaryotic fibrillarin consists of two major domains (Fig.4A). The glycine/arginine-rich (GAR) domain consists of ∼83 amino acids at the N terminus ofXenopus fibrillarin. Based on x-ray structural data and sequence conservation, the rest of the protein (amino acids 83–323) is similar to known SAM-dependent methyltransferases (15Niewmierzycka A. Clarke S. J. Biol. Chem. 1999; 274: 814-824Abstract Full Text Full Text PDF PubMed Scopus (190) Google Scholar, 16Wang H. Boisvert D. Kim K.K. Kim R. Kim S.H. EMBO J. 2000; 19: 317-323Crossref PubMed Scopus (143) Google Scholar). To determine the domains of fibrillarin that mediate its interaction with SMN, four GST-tagged fibrillarin truncation mutants (Fig.4A) were expressed in E. coli, coupled to glutathione-agarose, and incubated with full-length, in vitro translated, 35S-labeled SMN. Binding was assessed following SDS-polyacrylamide gel electrophoresis and fluorography. Deletion of the GAR domain of fibrillarin resulted in a loss of SMN binding (Fig. 4B). Furthermore, the GAR domain alone interacted with SMN to an extent similar to full-length fibrillarin (Fig. 4B). These results indicate that the GAR domain of fibrillarin is both necessary and sufficient for association with SMN in vitro. To elucidate the domain of SMN responsible for its interaction with fibrillarin, we tested eight truncation mutants of SMN (Fig.5A). Full-length, in vitro translated fibrillarin was incubated with each of the tagged SMN truncation mutants. It was recently shown that the conserved YG box of SMN (encoded by exons VI and VII) mediates the interaction of SMN with other proteins rich in glycine and arginine (Sm and Lsm proteins) and that the interaction is disrupted by a point mutation found in some SMA patients, Y272C (36Pellizzoni L. Charroux B. Dreyfuss G. Proc. Natl. Acad. Sci. U. S. A. 1999; 96: 11167-11172Crossref PubMed Scopus (216) Google Scholar, 44Friesen W.J. Dreyfuss G. J. Biol. Chem. 2000; 275: 26370-26375Abstract Full Text Full Text PDF PubMed Scopus (119) Google Scholar). However, the C-terminal half of SMN including the YG box (159/294) failed to interact with fibrillarin (as did this same region containing the Y272C point mutation) (Fig.5B). On the other hand, the N-terminal half of SMN (1/160) supported interaction with fibrillarin comparable to that observed with full-length SMN (Fig. 5B). Fibrillarin failed to associate with fragments consisting of only the first 30 or 60 amino acids of SMN (1/30 and 1/60), but some binding of fibrillarin was observed with amino acids 1–90 of SMN (1/90, Fig. 5B). Fibrillarin binding comparable to that observed with full-length SMN was found with amino acids 90–294 or 90–160 of SMN (90/294and 90/160, Fig. 5B). Comparable amounts of the SMN truncation mutants were used in the binding assays (Fig.5B). Our results indicate that fibrillarin binding is mediated by amino acids 60–160 of SMN (encoded by exons IIb and III), with amino acids 90–160, comprising the Tudor domain (45Mushegian A.R. Bassett Jr., D.E. Boguski M.S. Bork P. Koonin E.V. Proc. Natl. Acad. Sci. U. S. A. 1997; 94: 5831-5836Crossref PubMed Scopus (214) Google Scholar, 46Ponting C.P. Trends Biochem. Sci. 1997; 22: 51-52Abstract Full Text PDF PubMed Scopus (176) Google Scholar), being of primary importance for the interaction. The Tudor domain of SMN has also been found to interact with specific Sm proteins (common components of snRNPs) (23Buhler D. Raker V. Luhrmann R. Fischer U. Hum. Mol. Genet. 1999; 8: 2351-2357Crossref PubMed Scopus (217) Google Scholar, 41Selenko P. Sprangers R. Stier G. Buhler D. Fischer U. Sattler M. Nat. Struct. Biol. 2001; 8: 27-31Crossref PubMed Scopus (261) Google Scholar). Recently, the NMR structure of the Tudor domain of SMN and its association with Sm proteins have been characterized (41Selenko P. Sprangers R. Stier G. Buhler D. Fischer U. Sattler M. Nat. Struct. Biol. 2001; 8: 27-31Crossref PubMed Scopus (261) Google Scholar). Based upon the NMR structural analysis, we generated 9 missense mutations in the Tudor domain of SMN (amino acids 83–173, Fig. 1B), and we tested their effect on association with fibrillarin. We found that amino acids Glu-104, Asp-105, Gln-136, and Leu-142 are important for fibrillarin association with the Tudor domain of SMN (Fig.6A). Mutation of these residues (E104K, D105K, Q136A, and L142A) reduced binding by 70–80%. Alteration of other amino acids within or near the Tudor domain binding pocket (Lys-97, Arg-133, Glu-135, Asn-137, and Ser-143) did not significantly alter fibrillarin binding (Fig. 6A). The same pattern of sensitivity to the mutations was observed for interaction of the Tudor domain with the core snRNP protein SmB, although the degree of the effects was greater for SmB binding than for fibrillarin (Fig.6A). Our results indicate that SMN interacts with core components of snoRNPs and snRNPs via similar mechanisms involving the Tudor domain. Spinal muscular atrophy is correlated with deletions and/or mutations within the SMN1 gene (22Lefebvre S. Burglen L. Reboullet S. Clermont O. Burlet P. Viollet L. Benichou B. Cruaud C. Millasseau P. Zeviani M. Le Paslier D. Frezal J. Cohen D. Weissenbach J. Munnich A. Melki J. Cell. 1995; 80: 155-165Abstract Full Text PDF PubMed Scopus (2895) Google Scholar). One point mutation found in SMA patients, E134K, maps to the Tudor domain of SMN and was shown to abolish binding to Sm proteins in vitro (23Buhler D. Raker V. Luhrmann R. Fischer U. Hum. Mol. Genet. 1999; 8: 2351-2357Crossref PubMed Scopus (217) Google Scholar, 41Selenko P. Sprangers R. Stier G. Buhler D. Fischer U. Sattler M. Nat. Struct. Biol. 2001; 8: 27-31Crossref PubMed Scopus (261) Google Scholar). We tested the effect of E134K on the interaction of the N-terminal half of SMN (1/160) with fibrillarin, and we found that the mutation decreased binding by ∼50% (Fig. 6B). As expected, the SmB protein also failed to interact efficiently with the E134K mutant (Fig.6B). The effect of the SMA-associated E134K point mutation on interaction with fibrillarin reveals a potential link between snoRNP biogenesis or function and spinal muscular atrophy. In this work, we have identified and characterized an interaction between the core Box C/D snoRNP protein fibrillarin and the spinal muscular atrophy-linked protein SMN. This finding is of potential importance to our understanding of snoRNP biogenesis as well as the etiology of spinal muscular atrophy. SMN interacts with fundamental components of multiple nuclear RNA-protein complexes as follows: spliceosomal snRNPs (23Buhler D. Raker V. Luhrmann R. Fischer U. Hum. Mol. Genet. 1999; 8: 2351-2357Crossref PubMed Scopus (217) Google Scholar, 35Liu Q. Fischer U. Wang F. Dreyfuss G. Cell. 1997; 90: 1013-1021Abstract Full Text Full Text PDF PubMed Scopus (529) Google Scholar, 36Pellizzoni L. Charroux B. Dreyfuss G. Proc. Natl. Acad. Sci. U. S. A. 1999; 96: 11167-11172Crossref PubMed Scopus (216) Google Scholar, 41Selenko P. Sprangers R. Stier G. Buhler D. Fische
Referência(s)