|
|
Proc Natl Acad Sci U S A. 2007 June 5; 104(23): 9753–9757. Published online 2007 May 29. doi: 10.1073/pnas.0701720104. | PMCID: PMC1887542 |
Copyright © 2007 by The National Academy of Sciences of the USA Genetics Functional redundancy of worm spliceosomal proteins U1A and U2B″ Tassa Saldi,*† Carol Wilusz,*‡ Margaret MacMorris,*† and Thomas Blumenthal*†§ *Department of Biochemistry and Molecular Genetics, University of Colorado School of Medicine, Aurora, CO 80045; and †Department of Molecular, Cellular, and Developmental Biology, University of Colorado, Boulder, CO 80309 Received February 23, 2007. In Caenorhabditis elegans, the small nuclear ribonucleoprotein (snRNP)-associated proteins U1A and U2B″ are ≈50% identical to each other, and neither bears signature characteristics of mammalian U1A or U2B″ or the single Drosophila homolog, SNF. We show here that the genes that encode these proteins (rnp-2 and rnp-3) are cotranscribed in an operon, and that ribonucleoprotein RNP-2 is U1 snRNP-associated (U1A) whereas RNP-3 is U2 snRNP-associated (U2B″). U2B″ interacts with U2 even in the absence of another U2 snRNP protein, U2A′. Like U1A and U2B″ from yeast, plants, and vertebrates, worm U1A and U2B″ are more similar to each other than they are to other U1A or U2B″ proteins, respectively. Even though U1A and U2B″ interact with different snRNPs, they are functionally redundant; knockout of both is required for a lethal phenotype. Interestingly, U1A associates with U2 RNA when U2B″ is deleted. Thus, the two members of this gene family normally function as components of different snRNPs but apparently remain capable of performing the function of the other. Redundancy results from the fact that one protein can substitute for the other, even though it normally does not. Keywords: Caenorhabditis elegans, evolution, RNA processing, splicing Removal of introns from pre-mRNAs requires the spliceosome, a complex macromolecular machine containing small nuclear ribonucleoproteins (snRNPs) U1, U2, U4, U5, and U6, each of which has a single RNA and several proteins ( 1, 2). The U1 snRNP binds to the 5′ splice site, and U2 snRNP binds to the branch point near the 3′ splice site to facilitate the first step in splicing. U1 and U2 snRNPs have protein components that bind the snRNA through RNA recognition motifs (RRMs) ( 3– 5). Two of the most extensively studied RRM proteins are U1A and U2B″, components of the U1 and U2 snRNPs, respectively. These paralogous proteins each contain two RRMs separated by a short variable sequence (except for yeast U2B″, which contains only a single RRM). Homologs of U1A and U2B″ have been identified in numerous organisms, including vertebrates, yeast, and plants ( 6– 12). Although these proteins are important for prespliceosome formation in yeast, little is known about how they function ( 13). In Drosophila melanogaster a single member of this protein family, SNF, interacts with both U1 and U2 ( 6), but only the SNF/U1 interaction is necessary for splicing ( 14). These two proteins bind to their snRNA targets through very similar sequences in all phyla, excluding yeast ( 6). U1A interacts with stem/loop II of U1 RNA, and U2B″ binds to loop IV of U2 RNA ( 15– 17). Whereas U1A binds RNA independently, U2B″ requires association with an unrelated protein, U2A′, to efficiently bind RNA in yeast, plants, vertebrates, and flies ( 6, 9, 13, 16). In vitro studies with either vertebrate or plant proteins have shown that in the absence of U2A′, the affinity of U2B″ for U2 stem/loop IV is severely reduced, and binding is nonspecific. Furthermore, the interaction of U2A′ with U2B″ is RNA-independent, and U2A′ alone has little or no affinity for RNA in vertebrates ( 17– 19). The striking similarity of U1A and U2B″ prompted detailed in vitro studies with human U1A and U2B″ to identify amino acids critical for differential association with U1 or U2, as well as U2B″ residues important for U2A′ binding () ( 16, 17, 20– 22). Structures of RNA–protein complexes have been solved for these associations ( 18, 23). With the exception of yeast, the amino acids identified as critical for binding to either U1 or U2 are conserved ( 6). An analysis of the Caenorhabditis elegans genome reveals two members of the U1A/U2B″ family. Interestingly, these two genes are contained in a single operon. Because neither one of them has sequence characteristics that distinguish U1A from U2B″, we performed experiments to determine which gene encodes which protein. We identified the genes encoding U1A and U2B″ by associations of the proteins with U1 and U2 RNAs. We showed that, in worms, the two proteins are genetically redundant. Even though they are normally bound to different snRNPs, U1A can bind to the U2 snRNP when U2B″ is deleted. We also determined that U2B″ can bind to the U2 snRNP without the aid of the C. elegans U2A′ homolog. Our results demonstrate that these proteins are redundant, because one can substitute for the other even though it does not normally do so. rnp-2 and rnp-3 Are Genetically Redundant. The rnp-2 and rnp-3 genes are the first two genes in a C. elegans operon, CEOP4116 ( A). As is typical of worm operons, the genes are separated by only 100 bp and are expressed from a single promoter. Both proteins lack residues previously reported to specify identity as either U1A or U2B″, although they are clearly members of that gene family ( B). Neither RNP-2 nor RNP-3 contain the glutamic acid residue reported to facilitate association with U2A′. Both proteins lack the QIL and VALKT amino acid sequence motifs reported to be critical for interaction with U1 or U2 RNAs, respectively ( 16, 17, 20). RNP-2 and RNP-3 are 50% identical, and the two proteins are equally similar overall to U1A and U2B″ from other phyla (values range from 52% to 55% identity for comparisons of RNP-3 and RNP-2 with vertebrate U1A and U2B″ and fly SNF on the basis of a clustal alignment of vertebrate, fly, and worm sequences). Because neither protein possesses the defining characteristics of either U1A or U2B″, they might each be functioning redundantly on both snRNPs, like Drosophila SNF does. Thus, there is no clear prediction whether knockout or knockdown of these proteins would result in a phenotype. Remarkably, a strain carrying an rnp-3 deletion mutation lacking the central exon that includes both RRMs is viable. Because there is no deletion allele of rnp-2, we used RNAi to reduce RNP-2 levels. RNAi of rnp-2 showed no phenotype, and the rnp-3 deletion reduced viability only slightly ( Table 1). | Table 1.rnp-3(ok1424)/rnp-2(RNAi) double mutant has a more severe phenotype than either knockout alone |
To test for redundancy of RNP-2 and RNP-3, the levels of both proteins were reduced by RNAi of rnp-2 in the rnp-3 deletion ( ok1424) strain. The double knockdown/knockout decreased embryonic viability to 52% compared with rnp-3( ok1424) (86%) ( Table 1). Mortality during larval stages reduced total viability to 30% in the double. Because lowering rnp-2 by RNAi is unlikely to have completely eliminated the protein, the residual viability of the double may be due to incomplete reduction of RNP-2. These data demonstrate that worms can live without RNP-3, but when RNP-2 levels are also reduced, worms are disadvantaged or inviable, indicating genetic redundancy. rnp-2 and rnp-3 Encode U1A and U2B″, Respectively. To determine the molecular identities of RNP-2 and RNP-3, we analyzed a transgenic strain (BL1240) expressing Myc-RNP-2 and Flag-RNP-3 under endogenous promoter control. The transgenes were initially tested in the rnp-3 deletion strain to encourage the use of Flag-RNP-3 in functional snRNP. Antibodies that recognize each epitope tag were used in immunoprecipitation experiments with embryonic extract, and the associated RNA was analyzed by Northern blotting. Anti-Myc immunoprecipitated U1 RNA, but not U2 (, lane 6). In contrast, anti-Flag immunoprecipitated U2 RNA, but not U1 (, lane 4). As expected, antibody against the Sm proteins, found on most snRNPs, immunoprecipitated both U1 and U2 from both the tagged and untagged strains ( lanes 1 and 2). No snRNAs were pulled down in extract from the nontransgenic strain with either anti-Myc or anti-Flag. The rnp-3 deletion did not affect the specificity of the tagged proteins, because essentially identical results to those shown in were obtained with extract from the transgenic strain (BL1223) containing a wild-type allele of rnp-3 (data not shown). We conclude that RNP-2 associates exclusively with U1 RNA and therefore encodes U1A, whereas RNP-3 only associates with U2 RNA and encodes U2B″. Deletion of rnp-3 Allows U1A to Bind to U2 snRNP. RNP-2 and RNP-3 act genetically as though they are redundant, yet they interact with different snRNPs. Do U1A and U2B″ retain the capacity to function on the other snRNP when one of the proteins is missing? To address this question we created a transgenic strain (BL1262) expressing only Myc-tagged RNP-2 (U1A) in an rnp-3 deletion background (missing U2B″), and immunoprecipitated extract from this strain with anti-Myc antibodies. When the U2B″ protein was absent, Myc-tagged U1A associated with both U1 and U2 RNAs (, lane 7). Thus, even though U1A and U2B″ are functionally distinct, acting as components of different snRNPs, U1A can bind to U2 RNA when U2B″ is eliminated. U2B″ and U2A′ Deletions Have Different Phenotypes in C. elegans. In yeast, U2A′ and U2B″ mutations have similar phenotypes, and the double-deletion strain had an identical phenotype to the two single-deletion strains ( 13). However, in worms the U2A′ and U2B″ mutated strains behaved quite differently. Analysis of the phenotypes at 15°C resulting from the U2A′ deletion allele, sap-1( ok1221), revealed a variety of defects including embryonic lethality, slow growth and sterility ( A). (The sterility may result from the fact that ok1221 deletes the 3′ end of the 3′ UTR of H20J04.6, other alleles of which cause sterility, along with the first exon of sap-1.) Although the phenotype of U2B″ deletion [ rnp-3( ok1424)] is essentially wild-type at this temperature, there is an increase in lethality and slow growth associated with the rnp-3 deletion when the strain is maintained at 25°C. We also found that when the sap-1 deletion was grown at 20°C or when the rnp-3 deletion was grown at 25°C, levels of U2 snRNA were moderately (≈25%) reduced (data not shown). To investigate the requirement of U2A′ for U2B″ to function, a double mutant deleted for both proteins was tested for a phenotype greater than the additive effects of the single deletions. If U2B″ requires the binding of U2A′ to function within the U2 snRNP, U2A′ deletion should be epistatic to U2B″ deletion; it should not make any difference whether U2B″ is deleted if U2A′ is not present. However, the doubly deleted strain has a dramatic reduction in embryonic and total viability as well as fertility, a considerably more severe phenotype than deletion of U2A′ alone (A). These data demonstrate that U2B″ can perform an important function in the absence of U2A′. Furthermore, they demonstrate that U2A′ has a function independent of U2B″, because the phenotype of the double deletion is much more severe than the phenotype of the U2B″ deletion alone. Does U2B″ Require Association with Its Binding Partner, U2A′, to Bind Efficiently to U2 RNA? To determine whether U2B″ can associate with the U2 snRNP in the absence of U2A′, a transgenic strain (BL1260) expressing Flag-U2B″ was tested in the U2A′ deletion background. We immunoprecipitated extract from this strain along with extract from a strain containing U2A′ (BL1240) by using anti-Flag antibodies. In both the presence and absence of U2A′, Flag-tagged U2B″ associated with U2 RNA (B, lanes 2 and 3). Anti-Flag did not immunoprecipitate either snRNA in the untagged strain (B, lane 1), and anti-Sm immunoprecipitated both U1 and U2 in all three extracts. These data demonstrate that in vivo U2B″ can associate with U2 RNA without its presumed binding partner, U2A′. We have washed the immunoprecipitation pellets at salt concentrations up to 500 mM with no loss of U2B″ binding from either the wild type or the U2A′ deletion strain (data not shown). The absence of U2A′ also does not affect the exclusive specificity of U2B″ for U2. No association is observed between U2B″ and U1 RNA (B). The specificity of U1A for U1 was also not affected (data not shown). Thus, the absence of U2A′ does not result in promiscuous binding of either protein. We have identified the C. elegans U1A homolog as RNP-2 and the U2B″ homolog as RNP-3, on the basis of association with their corresponding snRNAs. Worm U1A and U2B″ are genetically redundant and, whereas they usually bind exclusively to the appropriate snRNA, U1A is capable of binding U2 RNA when U2B″ is absent. Protein and RNA Sequences Required for Specific Association. In vitro studies of human U1A and U2B″ have identified residues that confer specificity for RNA binding, and in most organisms examined (excluding yeast), these residues are conserved. Interestingly, these sequences are missing from the worm proteins. Nonetheless, the proteins are capable of differential snRNP association. This shows that these residues are not crucial for exclusive association in all phyla. A similar result was obtained by Polycarpou et al. ( 6), who noted that, whereas the yeast U2B″ homolog is widely divergent from other U2B″ proteins from other organisms, it also associates specifically with U2 RNA. These results imply that sequences and/or structural motifs beyond those characterized as essential contribute to the specificity. To investigate how U1A is capable of binding U2 when U2B″ is absent, we examined the sequence of C. elegans U2 RNA in comparison to U1 and U2 RNA from other phyla (). Interestingly, the adenosine in stem–loop IV of U2, reported to be critical for U2B″/U2A′ binding to U2, is not present in C. elegans U2 stem–loop IV. Furthermore, the cytosine in U1 stem/loop II postulated to facilitate U1A association with U1 is present in the equivalent position in worm U2 stem-loop IV ( 15). Hence, we postulate that the divergent sequence of C. elegans U2 may contribute to its ability to associate with both U2B″ and U1A. The Role of U2A′. Vertebrate, insect, and plant U2B″s require association with U2A′ to efficiently bind U2. Here we have shown that in vivo worm U2B″ associates exclusively with U2 RNA even when U2A′ is deleted (). This is in contrast to human U2B″, which has equal affinity for both U1 and U2 when U2A′ is absent ( 22). Furthermore, because deletions of U2B″ and U2A′ have distinct phenotypes, these two proteins are likely to have independent functions in splicing or in some other process in C. elegans. Because the double knockout is more severe than the additive effects of single deletions, U2B″ and U2A′ may be performing partially redundant functions in the U2 snRNP. Genetic Redundancy Without Molecular Equivalence. A motivation for previous studies has been the striking similarity between U1A and U2B″ proteins, as well as their corresponding binding sites on U1 and U2 RNAs. This observation has been the impetus for studies on the sequence-specific interactions that make them distinct. Interestingly, here we have demonstrated that worm U1A and U2B″ are not completely functionally independent. These proteins are genetically redundant, and when U2B″ is absent, U1A associates with U2 RNA (). Although we could only demonstrate aberrant binding of U1A due to lack of a U1A deletion, the genetic data suggest that U2B″ may be capable of carrying out the function of U1A on U2 snRNP. These results demonstrate an interesting type of redundancy in which redundant function depends on absence of one or the other protein. The Surprising Phylogeny of U1A and U2B″. Once we had identified the worm U1A and U2B″ proteins, we compared them with the other U1A and U2B″ family members. At a gross level, the two worm proteins are ≈50% identical to each other, as well as to U1A and U2B″ from vertebrates and to fly SNF. Because this analysis did not reveal anything about the branching order, we investigated the evolutionary tree that includes nematodes along with several homologs from other phyla (data not shown). This analysis is based on informative positions and can provide information about branching order when simple percent identity does not. Since the last time the U1A/U2B″ family was analyzed ( 6), many more family members have been sequenced in genome projects. A tree showing the alignment of U1A and U2B″ family members is available from the TreeFam project ( www.treefam.org/cgi-bin/TFinfo.pl?ac=TF500212) and is also shown in Wormbase ( www.wormbase.org/db/gene/gene?name=WBGene00004385;class=Gene). Several interesting features emerge. All vertebrate U1As cluster with other vertebrate U1As. Vertebrate U2B″s cluster with other U2B″s, suggesting that U1A and U2B″ duplicated at or below the base of this lineage. Many species have only a single U1A/U2B″, including all of the insects, Ciona intestinalis (a primitive chordate), and sea urchin. In yeast, worms, plants and vertebrates, there are two U1A and U2B″ family members, but in each phylum the U1A is more closely related to the U2B″ from that phylum than it is to U1As from other phyla. This observation is consistent with at least two hypotheses. First, the duplication that created the family has occurred at least four times with subsequent divergence to specialize for interaction with either U1 or U2 snRNP. The observation that U1A and U2B″ are adjacent genes expressed from a single promoter in worms is most consistent with this idea. The second idea is that U1A and U2B″ have been coevolving in each lineage, with selection pressure to keep them similar to each other. This idea was favored by Polycarpou et al. ( 6), because it seems more parsimonious. If this idea is correct, it suggests U1A and U2B″ may interact with the same unidentified molecule to perform their roles in splicing, exerting selective pressure for these proteins to remain similar. In this scenario, the two proteins would undergo parallel changes to conserve association with this hypothesized common target as it changes over time. Furthermore, redundancy itself may provide pressure to remain similar. It has been shown that some gene pairs arising from duplication maintain redundancy beyond the time frame necessary for divergence and specialization, presumably because of selective pressure to remain redundant ( 24). We have demonstrated here that worm U1A and U2B″ are redundant. Therefore, the U1A/U2B″ gene pair may be a specific example of evolution selecting for redundancy (and consequently similar sequence), allowing splicing to occur even in the absence of one or the other protein. Construction of Tagged rnp-2/rnp-3 Constructs. A 5.15 KB SacI/BamHI fragment of the cosmid K08D10 (containing rnp-2 and rnp-3) was subcloned into pGEM3Z (pK08d10.5SB). A PCR fragment was synthesized to insert the Myc-tag (MEQKLISEEDL) at the 5′ end adjacent to the ATG of rnp-2 and another PCR fragment to insert the Flag-tag (MEYKEEEEK) at the 5′ end of rnp-3. The PCR fragments were subcloned into pK08d10.5SB, by replacing 5′ fragments of each gene. The Myc-tag was inserted by ligating a SacII/AvaI fragment to produce pK08d10.myc. That intermediate was digested with BsaA1 and BspE1 to clone in the fragment containing the Flag-rnp-3 5′ end to produce pK08d10.mf. To create the plasmid with only Myc-rnp-2, we excised a 2-kb fragment from pK08d10.myc by restriction digest beginning 180 bp after the start of rnp-3 and ending in the 3′ MCS of pGEM3Z. The remaining fragment was then religated to create pK08d10.myc2 containing Myc-rnp-2 and deleted for the majority of rnp-3. Injection and Integration of Plasmids. Plasmids were injected at ≈100 ng/μl into N2 or MT2597 [ rol-6( n1178) ( 25) pK08d10.mf was coinjected with pRF5 ( rol-6( su1006)] into MT2597 to produce BL1213. BL1213 was gamma-irradiated (3800R) to select two integrated transgenic lines, BL1223 and BL1224, containing tagged versions of the rnp-2 rnp-3 operon. To create BL1261, N2 worms were coinjected with pK08d10.myc2 and pRF5, the strain was UV-irradiated (300 J/m 2) to integrate the extrachromosomal array. The selected strain after outcrossing (BL1261) contained an array that transmitted at a higher frequency than the original. Strains. Maintenance and growth of worms was as described in ref. 26. Strains RB1308 and VC789 were provided by the Gene Knockout Consortium (University of British Columbia, Vancouver, Canada). Strains containing a transgenic construct and deletions were made by crossing. The presence of the homozygous deletion alleles rnp-3( ok1424) and sap-1( ok1221) were detected by both phenotype and PCR analysis. Primers used to detect rnp-3( ok1424) were as follows: rnp-3ex1L, 5′TTCAGGTTAAAATGG3′; rnp-3ex2L, 5′AACTCAAAACGCTCTCTTC3′; and rnp-3ex3R, 5′-CCTTGGTGTTTGGCATCC3′. Primers used to detect sap-1( ok1221) were as follows: intleft, 5′TTGTCTTCATTTTCGTTGCG3′; intright, 5′ACGGGGAACCTTTGAAAAAT3′; and ex1right 5′CTGAATTTTTCGCCAACTTGCCG3′. The homozygous sa p-1 worms (BL1263) were isolated from VC789 by picking individuals lacking the gfp-marked mIn1 balancer chromosome. The following strains were used or created during the course of this research: strain BL1223, genotype rol-6( n1178) inIs144 (pK08d10.mf; pRF5); strain BL1224, genotype rol-6( n1178) inIs145 (pK08d10.mf; pRF5); strain BL1240, genotype inIs144 (pK08d10.mf; pRF5); rnp-3( ok1424) IV; strain RB1308, genotype rnp-3( ok1424) IV; VC789, genotype sap-1( ok1221)/ mIn1[mIs14 dpy10 ( e128)] II; strain BL1260, genotype inIs145 (pK08d10.mf; pRF5); sap-1( ok1221) II; strain BL1261, genotype inEx786 (pK08d10.myc2; pRF5); strain BL1262, genotype inEx786 (pK08d10.myc2; pRF5); rnp-3( ok1424) IV; and strain BL1263, genotype sap-1( ok1221) II. Extracts. Extract was made from either embryos or mixed-stage worms grown in liquid culture with NA22 or on NGM plates with NA22. Embryonic extract was created from BL1240, and all other extracts were generated using mixed-stage worms. Worms were broken by bead beating (Minibead Beater; Biospec Products, Bartlesville, OK) in homogenization buffer [10 mM KCl/1.5 mM MgCl2/1 mM DTT/10 mM Tris·HCl (pH 8.0)/50 mM Sucrose/0.05% Nonidet P-40/1 mM PMSF/Roche (Indianapolis, IN) complete protease inhibitors] with a 1:1 ratio of 0.1 mm glass beads to worm/embryo volume. Three cycles of 30 seconds at 5,000 rpm was used and breakage verified microscopically. Extract was either flash frozen in liquid nitrogen or used immediately for immunoprecipitations. Immunoprecipitations. Protein A Sepharose (Amersham, Piscataway, NJ) and anti-Flag or anti-Myc agarose affinity gel (Sigma, St. Louis, MO) were prepared by washing in water and IP buffer [20 mM Tris (pH 8.0)/100 mM KCl/0.2 mM EDTA/0.1% Nonidet P-40/0.1 mM DTT/1× complete protease inhibitor (Roche)]. Antibody was bound to protein A Sepharose and blocked with 0.1% nonfat dry milk, and excess unbound antibody was removed by washing with IP buffer. Extract was added to beads (40–100 μl per reaction) with RNase inhibitors, RNasin or RNaseOUT (20 units per sample) in a total volume of 0.5 ml IP buffer. Each tube was rocked at 4°C for 2 h, then supernatant collected and the beads washed 3 to 4 times with IP buffer. RNA was isolated from beads by proteinase K treatment, phenol/chloroform extraction, ethanol precipitation, resuspension in formamide dye loading buffer and electrophoresis on 12% acrylamide, and 7 M urea denaturing gels. Gels were blotted to Hybond N (Amersham), UV-crossed-linked (Hoefer, San Francisco, CA), and dried. Northern Blot Analysis. Northern blots were probed according to standard procedures ( 27). Membranes were prehybridized in 6× SSC, 0.2% SDS, and 10× Denhardt's at 65°C for at least one hour and hybridized overnight at 30°C in 6× SSC, 0.2% SDS, 5× Denhardt's with 32P-labeled oligonucleotides. Oligonucleotides used as probes for hybridization included a U1 mix (U1h, 5′GCCACACAGGTCAGCCCAT3′; U1e, 5′AGCCCGCACCAAAAGTGC3′; U1pe, 5′ATAACCCCCAGCCAGGTAAG3′) and a U2 mix (U2–1, 5′GGGCCGAGCCCGGCAGCAGTGC3′; U2–2, 5′CCTTTATTACACTCATTCG3′). RNAi. Double-stranded RNA was prepared by transcribing PCR fragments generated using T7 and T3 primers (T3, 5′ATTAACCCTCACTAAAGGGA3′; T7, 5′TAATACGACTCACTATAG3′) from cDNA for K08D10.3 cloned into pGEM. Single-stranded transcripts were run on 1% agarose gels to estimate concentration. Approximately equal concentrations of T7 and T3 transcribed RNA were mixed, annealed (37°C for 30 min), and used for injection. Individual N2 or RB1308 mutant worms were injected, allowed to recover overnight, and then placed on individual plates to monitor their progeny. Individual P0's were moved to new plates each day to facilitate counts of embryos and their subsequent hatching and survival to adulthood. We thank our colleagues David Bentley, Ravinder Singh, and David Pollock for helpful suggestions and for critical reading of the manuscript. Strains VC789 and RB1308 were provided by the Caenorhabditis Genetics Center (University of Minnesota, Minneapolis, MN), which is funded by the National Institutes Health National Center for Research Resources. This work was supported by National Institute of General Medical Sciences Research Grant GM42432. | | RRM | RNA recognition motif | | | snRNP | small nuclear ribonucleoprotein. |
1. Staley JP, Guthrie C. Cell. 1998;92:315–326. [PubMed] 2. Jurica MS, Moore MJ. Mol Cell. 2003;12:5–14. [PubMed] 3. Uhlenbeck OC, Pardi A, Feigon J. Cell. 1997;90:833–840. [PubMed] 4. Nagai K. Curr Opin Struct Biol. 1996;6:53–61. [PubMed] 5. Burd CG, Dreyfuss G. Science. 1994;265:615–621. [PubMed] 6. Polycarpou-Schwarz M, Gunderson SI, Kandels-Lewis S, Seraphin B, Mattaj IW. RNA. 1996;2:11–23. [PubMed] 7. Tang J, Abovich N, Rosbash M. Mol Cell Biol. 1996;16:2787–2795. [PubMed] 8. Liao XC, Tang J, Rosbash M. Genes Dev. 1993;7:419–428. [PubMed] 9. Simpson GG, Vaux P, Clark G, Waugh R, Beggs JD, Brown JW. Nucleic Acids Res. 1991;19:5213–5217. [PubMed] 10. Simpson GG, Clark GP, Rothnie HM, Boelens W, van Venrooij W, Brown JW. EMBO J. 1995;14:4540–4550. [PubMed] 11. Sillekens PT, Habets WJ, Beijer RP, van Venrooij WJ. EMBO J. 1987;6:3841–3848. [PubMed] 12. Bennett MM, Baron MA, Craft J. Nucleic Acids Res. 1993;21:4404. [PubMed] 13. Caspary F, Seraphin B. EMBO J. 1998;17:6348–6358. [PubMed] 14. Stitzinger SM, Conrad TR, Zachlin AM, Salz HK. RNA. 1999;5:1440–1450. [PubMed] 15. Scherly D, Boelens W, van Venrooij WJ, Dathan NA, Hamm J, Mattaj IW. EMBO J. 1989;8:4163–4170. [PubMed] 16. Scherly D, Boelens W, Dathan NA, van Venrooij WJ, Mattaj IW. Nature. 1990;345:502–506. [PubMed] 17. Bentley RC, Keene JD. Mol Cell Biol. 1991;11:1829–1839. [PubMed] 18. Price SR, Evans PR, Nagai K. Nature. 1998;394:645–650. [PubMed] 19. Boelens W, Scherly D, Beijer RP, Jansen EJ, Dathan NA, Mattaj IW, van Venrooij WJ. Nucleic Acids Res. 1991;19:455–460. [PubMed] 20. Scherly D, Dathan NA, Boelens W, van Venrooij WJ, Mattaj IW. EMBO J. 1990;9:3675–3681. [PubMed] 21. Tsai DE, Harper DS, Keene JD. Nucleic Acids Res. 1991;19:4931–4936. [PubMed] 22. Rimmele ME, Belasco JG. RNA. 1998;4:1386–1396. [PubMed] 23. Nagai K, et al. Nature. 1990;348:515–520. [PubMed] 24. Nagai K, Oubridge C, Jessen TH, Li J, Evans PR. Genome Biol. 2006;7:R69. [PubMed] 25. Mello CC, Kramer JM, Stinchcomb D, Ambros V. EMBO J. 1991;10:3959–3970. [PubMed] 26. Brenner S. Genetics. 1974;77:71–94. [PubMed] 27. Conrad R, Liou RF, Blumenthal T. EMBO J. 1993;12:1249–1255. [PubMed]
|
This article has been 
