RNA unwinding in U4/U6 snrnps requires ATP hydrolysis and the DEIH-box splicing factor Brr2 Pratima L. Raghunathan and Christine Guthrie

Transcription

1 Research Paper 847 RNA unwinding in U4/U6 snrnps requires ATP hydrolysis and the DEIH-box splicing factor Brr2 Pratima L. Raghunathan and Christine Guthrie Background: The dynamic rearrangements of RNA structure which occur during pre-mrna splicing are thought to be mediated by members of the DExD/H-box family of RNA-dependent ATPases. Although three DExD/H-box splicing factors have recently been shown to unwind synthetic RNA duplexes in purified systems, in no case has the natural biological substrate been identified. A duplex RNA target of particular interest is the extensive base-pairing interaction between U4 and U6 small nuclear RNAs. Because these helices must be disrupted to activate the spliceosome for catalysis, this rearrangement is believed to be tightly regulated in vivo. Results: We have immunopurified Brr2, a DEIH-box ATPase, in a native complex containing U1, U2, U5 and duplex U4/U6 small nuclear ribonucleoprotein particles (snrnps). Addition of hydrolyzable ATP to this complex results in the disruption of U4/U6 base-pairing, and the release of free U4 and U6 snrnps. A mutation in the helicase-like domain of Brr2 (brr2-1) prevents these RNA rearrangements. Notably, U4/U6 dissociation and release occur in the absence of exogenously added pre-mrna. Address: Department of Biochemistry and Biophysics, University of California, San Francisco, California , USA. Correspondence: Christine Guthrie Received: 1 May 1998 Revised: 4 June 1998 Accepted: 4 June 1998 Published: 29 June , 8: Ltd ISSN Conclusions: Disruption of U4/U6 base-pairing in native snrnps requires ATP hydrolysis and Brr2. This is the first assignment of a DExD/H-box splicing factor to a specific biological unwinding event. The unwinding function of Brr2 can be antagonized by the annealing activity of Prp24. We propose the existence of a dynamic cycle, uncoupled from splicing, that interconverts free and base-paired U4/U6 snrnps. Background Introns are removed from pre-mrnas by spliceosomes, large ribonucleoprotein complexes consisting of five small nuclear ribonucleoprotein particles (snrnps; U1, U2, U4, U5 and U6) and a large cast of proteins that coalesce on premrna [1 3]. The spliceosome executes the two chemical steps of the splicing reaction using active sites which consist, at least in part, of intricately structured RNAs [4,5]. The progress of spliceosome assembly is marked by the sequential formation and dissolution of RNA duplexes [6,7]. Base pairs form first between U1 snrnp and the 5 splice site, and then between U2 snrnp and the intron branchpoint sequence. Next, U5 snrnp and the duplex U4/U6 snrnp join the spliceosome together in a single U4/U6.U5 triple snrnp particle. Within the U4/U6 snrnp, U4 and U6 small nuclear RNAs (snrnas) engage in extensive intermolecular base-pairing; in yeast, the remarkably stable U4/U6 snrna duplex melts in vitro at 55 C [8]. Shortly after U4/U6.U5 joins the pre-mrna, a major rearrangement occurs: U4 is destabilized from the spliceosome, and critical residues of U6 snrna are juxtaposed with the 5 splice site and with U2 snrna. These newly created U2/U6 and U6/5 splice site helices are thought to contribute to the first-step chemical reaction center of the spliceosome [5,6]. Construction of this catalytic core demands the precise positioning of RNA helices in time and space. A key unanswered question is how the spliceosome coordinates these RNA rearrangements. For example, because U4/U6 basepairing and the U2/U6 interaction are mutually exclusive, disruption of the U4/U6 duplex is believed to activate the spliceosome for catalysis [9]. This transition has been predicted to be tightly regulated [9] but the mechanism remains unknown. Destabilization of U4 from the spliceosome requires ATP in vitro, suggesting that unwinding and/or displacement are active processes [10,11]. Nevertheless, the only two factors functionally implicated in this event Prp19 and the U4 snrnp protein Prp4 do not contain ATPase motifs [12,13]. Seven yeast splicing factors exhibit homology to DNA and RNA helicases of Superfamily II, and are therefore attractive candidates for governing helical exchanges on the spliceosome [14]. These proteins belong to the DExD/Hbox family of RNA-dependent ATPases [14]. Six DExD/H-box splicing factors Prp5, Brr2, Prp2, Prp16, Prp22 and Prp43 are known to be required at distinct steps in the splicing pathway, and five Prp5, Prp2, Prp16, Prp22 and Brr2 have been shown to possess

2 848, Vol 8 No 15 RNA-dependent ATPase activity [15 21]. Recently, Prp16, Prp22 and U5 200K (the mammalian homolog of Brr2) have been shown to unwind synthetic RNA helices in purified systems; in fact, all three of these proteins can disrupt in vitro synthesized U4/U6 duplexes [22 25]. These enzymes display no apparent sequence specificity in the in vitro reactions, however, and have not been linked to defined RNA-unwinding events on the spliceosome. Thus, it remains to be determined whether these DExH/D-box factors unwind specific RNA or RNP substrates in the splicing pathway. This work focuses on the RNA rearrangements governed by the yeast DEIH-box protein Brr2 [26] (also known as Snu246 [27], Slt22 [20] and Rss1 [28]). The yeast mutant brr2-1 was identified by virtue of its cold-sensitive premrna splicing defect in vivo [26]. Independent analyses revealed that Brr2 is likely to be an RNA-dependent ATPase [20] within the U4/U6.U5 snrnp [27]. These observations prompted the hypothesis that Brr2 might disrupt U4/U6 base-pairing on the spliceosome [26,27]. In support of this idea, Laggerbauer et al. [25] have shown that the human homolog of Brr2 can unwind yeast U4/U6 duplexes that have been prepared by annealing purified in vitro transcribed RNAs. Here, we use the cold-sensitive mutant brr2-1 to test the hypothesis that Brr2 is required to disrupt base-pairing in native U4/U6 snrnps. The brr2-1 mutation was mapped to a conserved motif in the helicase-like DEIH-box domain. We have immunopurified both Brr2 and Brr2-1 in large complexes containing U1, U2, U5 and duplex U4/U6 snrnps. When the purified Brr2 complex was incubated with hydrolyzable ATP, U4/U6 base-pairing was disrupted, and free U4 and free U6 snrnps were released. In contrast, snrnp complexes that harbor Brr2-1 failed to undergo these ATP-dependent RNA rearrangements. We conclude that disruption of native U4/U6 snrnps requires ATP hydrolysis and Brr2 function. Interestingly, dissociation of U4/U6 snrnps occurred in the absence of added pre-mrna, indicating that U4/U6 unwinding is not restricted to the spliceosome. In conjunction with previous results [29], these data support the existence of a dynamic cycle that regulates U4/U6 base-pairing, independent of splicing. Results Cold-sensitivity of brr2-1 is conferred by a mutation in the first helicase-like domain The Brr2 protein, like its human homolog, U5 200K [27], is unique among the DExD/H-box splicing factors because it has two predicted ATPase domains (Figure 1). To determine whether the cold-sensitive brr2-1 lesion resided in these regions, we transferred the chromosomal brr2-1 allele by gap repair [30] to a plasmid bearing the wild-type BRR2 gene. A single mutation within the first DExH-box domain was sufficient to Figure 1 (a) Brr2 DEXH domain I DEXH domain II (b) DEXH domain I [ DEXH domain II [ Family consensus Brr2 U5 200K APTGSGKT APTGAGKT Brr2 SGKGTGKT U5 200K APTGSGKT DEXH box DEAH box DEAD box E610 A-TG-GKT GETGSGKT A-TG-GKT G in Brr2-1 PLKALV TPEKWD PMRSLV GD GD TPEKWD PMRLWQ PSGEKI GN TPVQFE GE TPEKWD P-KAL- PRRVAA PTRELA G in Brr2-1 GD T----- GY TDG--- GG TPGR-- DEIH DEIH DDAH DEVH DE-H DEAH DEAD SAT SAT AWGVNLP AWGVNLP SNC SSS CSAFACK CWGMNVA SAT SAT SAT --G-N-- --G---- ARG-D-- QMLGRAGR QMLGRAGR EMVGLASG QMVGHANR QM-GRAGR QR-GRAGR HRIGR-GR The Brr2-1 mutation lies in a highly conserved motif in the first helicaselike domain of Brr2 protein. (a) Schematic representation of the Brr2 protein, which contains two DExH-box ATPase domains. The position of the E610 G mutation (in the single-letter amino-acid code) in Brr2-1 is indicated. (b) The alignment shows eight signature motifs from each of the two DExH-box domains of Saccharomyces cerevisiae Brr2 and the human U5 200K homolog [27], compared with the consensus sequences of the DExH, DEAH and DEAD-box families [27,57]. Dashes denote non-conserved residues in the consensus sequences. The second helicase-like domain of Brr2 diverges more from the DExHbox consensus sequence than the first [27]. The brr2-1 allele contains an A to G transition at nucleotide 2417, which converts the glutamic acid at residue 610 in DExH-box domain I to a glycine (E610 G). confer cold-sensitivity when substituted into the wildtype gene (data not shown). The brr2-1 mutation, A2417 G, converts a glutamic acid to a glycine at amino-acid residue 610, which lies within a putative helicase-like motif that is absolutely conserved between the yeast and human homologs (Figure 1). Other DExD/H-box proteins with mutations in the motifs depicted in Figure 1 display impaired ATPase and/or unwindase activities [31] and splicing deficits [32]. We therefore tested the prediction that the mutant Brr2-1 protein would prevent U4/U6 disruption. ATP-dependent displacement of free U4 snrnp is blocked in brr2-1 extracts We first sought to implicate Brr2 in snrnp rearrangements using ammonium sulfate precipitates of whole-cell extracts known as fraction I. Unlike the situation in whole extracts, U4/U6.U5 snrnps in fraction I have been observed to dissociate into individual snrnps in the presence of ATP [10]. Dissociation of U4/U6.U5 appears to be mechanistically similar to the disruption of U4/U6 on the spliceosome, as both are ATP-dependent processes that appear to separate the U4/U6 helices. We reasoned that if U4/U6 helix displacement requires Brr2 function, then triple snrnp dissociation should be blocked in fraction I from brr2-1 cells. Addition of ATP to wild-type fraction I caused the appearance of a free U4 snrnp species (Figure 2a, lanes 1,2). ATPγS could not substitute for ATP (Figure 2a, lane 3), suggesting that ATP hydrolysis is required for U4 release. In contrast, U4/U6.U5 snrnps from brr2-1 fraction I did not separate

3 Research Paper U4/U6 disruption requires Brr2 and ATP Raghunathan and Guthrie 849 Figure 2 The snrnp populations and ATP-dependent dynamics are abnormal in extracts from brr2-1 cells. (a) ATP-dependent displacement of U4 is blocked in the snrnp-rich fraction I from brr2-1 cells. Fraction I prepared from wildtype BRR2 (lanes 1 3) or mutant brr2-1 cells (lanes 4 6) was incubated in the presence or absence of ATP or ATPγS. The snrnps were resolved on a native gel and subjected to northern analysis with a U4 snrna probe. Sequential stripping and reprobing confirmed the identity of U4/U6.U5 (data not shown); the mobility of this complex appears to change when ATP is present see also (b). ATP hydrolysis is necessary for the generation of free U4 in BRR2 fraction I (compare lanes 2 and 3), but no free U4 is produced in brr2-1 fraction I (lane 5). (b) The brr2-1 extract contains an aberrant snrnp distribution. Whole-cell extract from wild-type BRR2 or mutant brr2-1 cells was incubated with or without ATP, electrophoresed on a native gel to resolve snrnps, and subjected to northern analysis with the indicated snrna probes. (a) Fraction I ATP U4/U6.U5 Free U4 BRR2 brr2-1 + γs + γs Lane U4 probe U4/U6.U5 U4/U6 Free snrnp species are indicated with thin bars to the left of each panel; the U2 U4/U6.U5 multi-snrnp species is (b) Probe U4 U6 U5 U2 Extract BRR2 brr2-1 BRR2 brr2-1 BRR2 brr2-1 BRR2 brr2-1 ATP Lane indicated with a thick bar to the right of each panel. The U4/U6.U5, U4/U6 and U5 snrnps are less abundant in the brr2-1 extract. into free snrnps in the presence of ATP (Figure 2a, lanes 4,5). Therefore, snrnps from brr2-1 cells fail to undergo the ATP-dependent conformational rearrangement that is normally detected in wild-type snrnps. We next examined whether similar abnormalities could be detected in brr2-1 whole-cell extracts. When an extract derived from wild-type cells was visualized on a native gel, U4/U6.U5 and free U6 snrnps were found to be abundant (Figure 2b, lanes 1,5,9). Diffusely migrating U4/U6 snrnps were more apparent when ATP was added, because ATP stimulates U4/U6.U5 disintegration and Prp24-dependent snrnp recycling (Figure 2b, lanes 2,6; [29]). In the extract derived from brr2-1 cells, by contrast, no U4/U6 snrnps were observed in the presence of ATP, and the overall level of U4/U6.U5 snrnps was very low (Figure 2b, lanes 3,4,7,8,11,12). Free U5 and free U2 snrnps were also diminished in the mutant extract compared with the wild-type extract (Figure 2b, compare lanes 9,10 with 11,12; 13,14 with 15,16; thin bars). These deficits appear to be significant, because comparable amounts of free U6 snrnp were detected in wild-type and mutant extracts (equivalent amounts of protein were loaded). Also, a very slowmigrating species that hybridized to U4, U6, U5 and U2 was noticeable in mutant extracts, but not in wild-type extracts (Figure 2b, lanes 3,4,7,8,11,12,15,16; thick bars). In summary, the brr2-1 extract has an aberrant snrnp distribution, with few U4/U6.U5 snrnps and no U4/U6 snrnps. As with the snrnps in brr2-1 fraction I, snrnps in brr2-1 whole-cell extract are minimally changed by the addition of ATP. Brr2 and Brr2-1 associate with snrnp complexes containing U1, U2, U4, U5 and U6 snrnas Although Brr2 is part of U4/U6.U5 snrnps [27], the mutant Brr2-1 protein might fail to associate with triple snrnps, and thus prevent U4 displacement indirectly. To test this possibility, we generated yeast strains that expressed Brr2 and Brr2-1 as fusion proteins with the polyoma epitope tag (Pya) [33]. The gene fusions BRR2 Pya and brr2-1 Pya behaved identically to BRR2 and brr2-1 in complementing the BRR2::LEU2 gene disruption (data not shown). Wild-type and mutant Pyatagged proteins were immunoprecipitated from whole-cell extracts at different salt concentrations, and the bound snrnas were monitored by northern analysis. Brr2 Pya associated with U2, U4, U5 and U6 snrnas at NaCl concentrations between 50 mm and 250 mm; U1 snrna also precipitated at 50 mm NaCl (Figure 3, lanes 5 8). Importantly, the same spectrum of snrnas co-precipitated with Pya-tagged Brr2-1 at NaCl concentrations between 50 mm and 350 mm (Figure 3, lanes 10 13). These interactions are specific, because no snrnas were precipitated from extracts derived from cells expressing untagged protein (Figure 3, lanes 1 3), or tagged protein when antibody was omitted (Figure 3, lanes 4,9), or when competitor peptide expressing the polyoma epitope was added (Figure 3, lanes 14,15). Thus, Brr2-1, like Brr2, efficiently associates with U2, U4, U5 and U6 snrnas in a salt-resistant manner. Our results are in general agreement with Lauber et al. [27], who report that Brr2 is a component of the U4/U6.U5 snrnp. Moreover, the reproducible association of Brr2 with U2

4 850, Vol 8 No 15 Figure 3 Extract Antibody NaCl (mm) + Peptide BRR2 BRR2 Pya brr2-1 Pya WT brr WT brr2 Brr2 and Brr2-1 co-immunoprecipitate U2, U4, U5 and U6 snrnas in a salt-resistant manner. Extracts from cells expressing BRR2, or the gene fusions BRR2 Pya or brr2-1 Pya, were incubated with protein G Sepharose beads coupled to anti-polyoma antibodies (lanes 1 3,5 8,10 15), or to protein G Sepharose alone (lanes 4,9) in the presence of the indicated concentration of salt. After washing, the bound RNAs were isolated, resolved on a denaturing gel, and subjected to northern analysis for U1, U2, U4, U5 and U6 snrnas; U5 snrna exists in two different forms, U5 long (U5L) and U5 short (U5S). Peptide encoding the polyoma epitope competed for immunoprecipitation of all snrnas, demonstrating specificity (lanes 14,15). Lanes 16 and 17 contain deproteinized extract equivalent to one-tenth of each immunoprecipitation sample in lanes 4 8 (WT) and lanes 9 13 (brr2). Compared with wild-type cells (lane 16), brr2-1 cells appear to have lower levels of all snrnas (lane 17). As a percentage of wild-type snrna levels, brr2-1 contains 45% U2, 78% U1, 86% U5S, 64% U5L, 45% U4 and 67% U6. and U1 snrnas is consistent with evidence of yeast twohybrid interactions [34] between Brr2 and Cus1, a likely component of the yeast U2 snrnp [35], and between Brr2 and the U1 70 kda snrnp protein Snp1 [34]. To determine whether Brr2 and the snrnps are present in a single large complex, we sedimented wild-type and mutant extracts on 15 40% glycerol gradients, and performed immunoprecipitations from alternate fractions. Brr2, like Brr2-1, co-immunoprecipitated all five snrnas from the same peak on the gradient (data not shown). Thus, both Brr2 and Brr2-1 associate with large snrnp complexes in whole-cell extract in the absence of exogenously added pre-mrna. Brr2 and ATP are required to disrupt U4/U6 base-pairing We next tested whether these Brr2-bound snrnp complexes undergo rearrangements in the presence of ATP. We developed an assay to monitor U4/U6 disruption in snrnp complexes that were specifically associated with Brr2 or Brr2-1. Brr2 Pya and Brr2-1 Pya snrnp complexes U2 U1 U5L U5S U4 Lane U6 from whole-cell extracts were immunopurified on beads, washed extensively, and incubated in the presence of buffer with or without ATP. These supernatants were separated from the beads, and RNAs were extracted. The snrnas in each sample were analyzed on a denaturing gel by northern blotting. When isolated Brr2 snrnp complexes were exposed to ATP, U4 and U6 snrnas were physically released into the supernatant (Figure 4a, compare lanes 5,6 and 7,8). Most of the associated U2 and U5 snrnas appeared to remain with Brr2 on the beads. The non-hydrolyzable ATP analogs ATPγS and AMP-PNP did not support this discharge of U4 and U6, suggesting that restructuring of the Brr2 particle requires ATP hydrolysis by an intrinsic component (data not shown). In contrast, complexes containing Brr2-1 failed to undergo this ATP-dependent snrnp rearrangement: in the presence of ATP, all snrnas remained associated with Brr2-1 Pya on the beads (Figure 4a, lanes 11 14). Thus, ATP causes U4 and U6 (and a small fraction of U2) to separate from the Brr2-associated snrnp complex. Moreover, the displacement of U4 and U6 snrnas from this snrnp complex requires ATP hydrolysis and wild-type Brr2. An important question is whether ATP disrupts U4/U6 base-pairing in these Brr2 and Brr2-1 complexes. The deproteinized RNA samples from the previous experiment were subjected to non-denaturing gel electrophoresis to resolve duplex U4/U6 RNA from the free species [36]. In the absence of ATP, mostly duplex U4/U6 was found to have associated with Brr2 Pya on the beads (Figure 4b, lane 3). However, ATP addition resulted in the disruption of these U4/U6 helices, and substantial amounts of unpaired U4 and U6 RNAs were released into the supernatant (Figure 4b, lanes 5,6). Notably, the U4/U6 duplex remained intact within the Brr2-1 complex in the presence of ATP (Figure 4b, lanes 9 12). This is direct evidence that ATP alters base-pairing status in native U4/U6 snrnps. Furthermore, these results demonstrate conclusively that disruption of RNA basepairing in duplex U4/U6 snrnps requires ATP and wildtype Brr2 function. Discussion Disruption of U4/U6 base-pairing in native snrnps requires ATP and Brr2 From our current knowledge of the mechanism of premrna splicing, numerous RNA duplexes must be rearranged (that is, unwound and subsequently reannealed) during each spliceosome cycle in vivo [7]. The fundamental problem is to understand how these RNA transactions are effected and regulated. A long-standing hypothesis has been that the DExD/H-box factors promote RNA unwinding in vivo. Yet it has not been possible to engineer an assay to capture a specific RNA-dependent ATPase in the act of unwinding a defined spliceosomal substrate. Consequently, available information

5 Research Paper U4/U6 disruption requires Brr2 and ATP Raghunathan and Guthrie 851 (a) Extract BRR2 BRR2 Pya brr2-1 Pya Antibody ATP B S B S B S B S B S B S B S (b) Extract Antibody ATP U4 probe U6 probe Lane BRR2 Pya brr2-1 Pya B S B S B S B S B S B S BRR2 brr2 1 Lane U2 U1 U5L U5S U BRR2 brr2 1 U4/U6 duplex Free U4 U4/U6 duplex Free U6 is limited to the study of DExD/H-box proteins with synthetic RNA duplexes. Many DExD/H-box proteins with no known roles in splicing can dissociate synthetic RNA duplexes in vitro; these include eif-4a [37], RNA helicase A [38], and vaccinia virus RNA helicase [39]. Only very recently have any DExH-box splicing factors been reported to have strand-displacement activity [22 25]. In these non-physiological enzymatic assays, diverse DExD/H-box proteins share common characteristics: they unwind artificial substrates with no apparent sequence specificity, and they dissociate naked RNA duplexes rather than RNA protein complexes. Indeed, the synthetic U4/U6 duplex prepared from in vitro transcribed RNAs can be unwound by three different DExH-box splicing factors, including the human homolog of Brr2 [22,23,25]. Thus, two important questions remain unanswered: do the observed in vitro unwinding activities accurately reflect the physiological functions of these proteins and, if so, what are the identities of the specific RNA targets for these enzymes in vivo? Our strategy has been to devise functional assays to characterize the disruption of a natural helical substrate, the duplex U4/U6 snrnp (Figures 2a and 4a,b). We have provided direct evidence that ATP influences the base-pairing U6 Figure 4 Brr2 and ATP are required to disrupt the U4/U6 snrna helices and release the free snrnas. (a) The Brr2 snrnp complex physically releases U4 and U6 snrnas in response to ATP, but the Brr2-1 snrnp complex does not. The snrnp complexes associated with Brr2 Pya (lanes 3 8) and Brr2-1 Pya (lanes 9 14) were bound to anti-polyoma antibodies coupled to protein G Sepharose. After extensive washing, the matrices were incubated with splicing buffer in the absence of ATP or containing 2 mm ATP. The beads (B) were then separated from the supernatants (S), and the RNAs were isolated. Northern analysis to detect the snrnas was performed as described for Figure 3. As specificity controls, immunoprecipitations were performed using extracts derived from cells expressing BRR2 (lanes 1,2) or the tagged versions of BRR2 or brr2-1 in the presence of protein G Sepharose alone (lanes 3,4,9,10). Lanes 15 and 16 contain deproteinized extract equivalent to one-fortieth of each immunoprecipitation sample in lanes 3 8 (BRR2) and lanes 9 14 (brr2-1). The Brr2 snrnp complex contains 1%, 0.3%, 12%, 15%, 25% and 7% of total cellular U2, U1, U5L, U5S, U4 and U6 snrnas, respectively. Thus Brr2 associates with much more of the cellular U4 and U6, compared with the small fraction of cellular U2. The Brr2-1 snrnp complex contains 4%, 1%, 19%, 24%, 24% and 3% of total cellular U2, U1, U5L, U5S, U4 and U6 snrnas, respectively. (b) The U4/U6 snrna duplex is disrupted in the presence of ATP in the Brr2 snrnp complex, but not in the Brr2-1 snrnp complex. RNA samples from (a) were hybridized in solution to labeled oligos specific for U4 (upper panel) or U6 (lower panel) snrnas, and electrophoresed on non-denaturing gels to distinguish slow-migrating U4/U6 duplex from the free snrnas [36]. Lanes 13 and 14 contain deproteinized extract equivalent to one-eightieth of each immunoprecipitation sample in lanes 1 6 (BRR2) and lanes 7 12 (brr2-1). In the Brr2 complexes exposed to ATP, the U4/U6 helices are separated, and free U4 and free U6 snrna are released into the supernatant (lanes 5,6). The Brr2-1 complexes fail to disrupt U4/U6 in response to ATP (lanes 11,12). status of authentic U4/U6 snrnps (Figure 4b). Furthermore, a mutation in a conserved helicase motif of Brr2 (Figure 1) was found to prevent the ATP-dependent dissociation of the U4/U6 helices (Figure 4a,b). Therefore, we conclude that functional Brr2 is required to disrupt native U4/U6 snrnps. Although we cannot exclude the formal possibility that Brr2 is an upstream regulator of an intervening ATPase, our findings are complemented by those of Laggerbauer et al. [25], who demonstrated that the human homolog of Brr2 directly unwinds synthetic U4/U6 RNA duplexes in a purified system. The significance of our results is different, because we have linked a naturally occurring RNA-unwinding event to a specific DExD/Hbox protein. As discussed by Laggerbauer et al. [25], an interesting and important question is whether U5 200K dissociates U4/U6 on the spliceosome. We attempted to test this hypothesis for Brr2, reasoning that in vitro splicing in the brr2-1 extract should be inhibited before the first chemical step because of a defect in U4 release. In fact, the brr2-1 extract did not support in vitro splicing (data not shown), though for an unexpected reason. In the mutant extract, we observed low levels of U4/U6.U5 snrnps (Figure 2b) and inefficient assembly of U4, U6 and U5 snrnas onto pre-mrna (data not shown). We believe that the brr2-1

6 852, Vol 8 No 15 extract contains insufficient U4/U6.U5 snrnps that are competent for spliceosome assembly. This apparent upstream requirement for Brr2 in U4/U6.U5 loading does not rule out a subsequent role in U4/U6 dissociation. Given that Brr2 is required to disrupt U4/U6 base-pairing in snrnps, Brr2 is likely to perform an analogous function on spliceosomes. Definitive proof of this hypothesis is still needed, however. Brr2 and Brr2-1 associate with large multi-snrnp complexes in the absence of exogenous RNA Unexpectedly, we found that Brr2 and Brr2-1 exist in large, salt-stable snrnp complexes in the absence of added premrna (Figures 2b and 3 and data not shown). Several reports have previously described co-immunoprecipitation of U2, U4, U5 and U6 snrnas with antibodies to individual snrnp proteins, but these interactions have been interpreted as background or non-specific binding because of the low to moderate salt concentrations employed [40 44]. We observed, however, that these snrnps co-sediment in Figure 5 Brr2 ATP U4 U4 U2 U4/U6.U5 Brr2 U2 U5 Brr2 Free U4 Free U6 Antagonistic regulation of U4/U6 base-pairing by the DEIH-box ATPase Brr2 and the RNA-annealing protein Prp24. Brr2 is a component of a multi-snrnp complex (U2 U4/U6.U5 Brr2). ATP hydrolysis by Brr2 triggers the disruption of the U4/U6 helices and the displacement of free U4 and U6 snrnps from the U2 U5 Brr2 complex. The unpaired U4 and U6 snrnps are reannealed by Prp24, reversing the unwinding mediated by Brr2. The reformed U4/U6 snrnp may then be reincorporated into higher-order snrnp complexes. This dynamic snrnp cycle is proposed to be independent of pre-mrna splicing. U6 U6 Prp24 large complexes on glycerol gradients in association with Brr2 and Brr2-1 (data not shown). Moreover, we have detected slowly migrating complexes in the brr2-1 extract that are stable to electrophoresis on native snrnp gels (Figure 2b). One hypothesis that can account for these observations is that the Brr2-1 complexes represent endogenous spliceosomes that have been stalled by the failure to disrupt U4/U6. The accumulation of these dead-end complexes would thus prevent recycling and regeneration of U4/U6.U5 snrnps in the brr2-1 extract, and consequently inhibit de novo splicing (Figure 2b and data not shown). A different hypothesis is necessary to explain the behavior of the wild-type Brr2 snrnp complex, which is unanticipated from two standpoints. First, U4 destabilization has been predicted to be strictly regulated in the splicing pathway [9], yet U4/U6 dissociation within the Brr2 snrnp complex does not require the addition of premrna. Second, ATP induces free U4 and free U6 to separate from each other as well as from Brr2, U5 and U2 snrnps (Figure 4a). In contrast, the spliceosome does not liberate U6 snrnp along with U4. These observations suggest that ATP hydrolysis by Brr2 governs dissociation of the U2 U4/U6.U5 snrnp complex in a manner that is not necessarily coupled to splicing. Prp24 antagonizes Brr2 in an apparent snrnp cycle We have recently exploited the discovery of these Brr2- mediated snrnp rearrangements to advance understanding of Prp24, an RNA-binding protein and spliceosomal recycling factor [29]. Prp24 has been proposed to anneal U4 and U6 snrnps in vivo [45], but an explicit test of this hypothesis has been hampered by the inability to isolate the transient free U4 snrnp. By adding ATP to the immobilized Brr2 complex, we were able to purify free U4 and U6 snrnps away from other species. We have now demonstrated that these unpaired U4 and U6 snrnps can be rapidly reannealed by Prp24 [29]. Interestingly, the annealing activity of Prp24 is markedly more efficient with the Brr2-generated RNP substrates than with deproteinized U4 and U6 snrnas [29]. Thus, Prp24 and Brr2 appear to be antagonistic regulators of U4/U6 base-pairing: Brr2 directs ATP-dependent dissociation of U4/U6 snrnps, whereas Prp24 restores the disrupted duplex. Notably, both of these events occur in the absence of added pre-mrna, indicating that ATP-driven rearrangements of U4/U6 snrnps can occur independently of pre-mrna splicing. Taken together, these results support the existence of a dynamic snrnp cycle in which Brr2 and Prp24 antagonistically control U4/U6 base-pairing in vivo (Figure 5). The following snrnp rearrangements are proposed to be uncoupled from splicing. Brr2 hydrolyzes ATP in the multi-snrnp complex, triggering the disruption of the U4/U6 helices and the displacement of free U4 and U6 snrnps from the U2 U5 Brr2 complex. Next, Prp24 reanneals the unpaired U4 and U6 snrnps, reversing the unwinding mediated by

7 Research Paper U4/U6 disruption requires Brr2 and ATP Raghunathan and Guthrie 853 Brr2. The reformed U4/U6 snrnp may then be reincorporated into higher-order snrnp complexes (that is, U4/U6.U5 and U2 U4/U6.U5). What might be the biological function of this putative snrnp cycle? Conceivably, the dynamic interconversion between duplex and free snrnps affords the opportunity for rapid regulation of spliceosome assembly. The cellular splicing machinery may be finely attuned to the availability of splicing-competent snrnps containing duplex U4/U6. Such a mechanism may operate in HeLa cells during heat shock, when inactivation of a U4/U6.U5 snrnp assembly factor results in inhibition of pre-mrna splicing [46]. Alternatively, these snrnp dynamics might mimic spliceosomal events. For example, ATP-dependent U4/U6 dissociation might provide a proofreading step to ensure proper U4/U6.U5 positioning on the spliceosome. In this scenario, constitutive U4/U6 disruption by Brr2 might reflect a nonproductive discard pathway that improves the accuracy of concurrent helical exchanges (for example, the formation of U2/U6 and U6/5 splice site helices) [7]. Another DExD/Hbox splicing factor with RNA unwinding activity [22], Prp16, has been proposed to govern fidelity of branchpoint recognition through a similar scheme [47]. Conclusions RNA-dependent ATPases of the DExD/H-box family have long been hypothesized to promote RNA unwinding in vivo. Despite evidence that many DExD/H-box proteins can carry out strand displacement reactions on synthetic RNA duplexes in vitro, in no case has a biological substrate been identified. Using the cold-sensitive allele brr2-1, which carries a mutation in the first of two DExHbox domains, we have shown that Brr2 function is required for the ATP-dependent disruption of U4/U6 duplex in native snrnps. Because the observed U4/U6 unwinding does not appear to be coupled to splicing, we propose that U4/U6 base-pairing may be regulated outside the context of the spliceosome. Brr2 appears to govern U4/U6 disruption, whereas Prp24 restores the duplex U4/U6 snrnp. These findings provide evidence that the antagonistic actions of a DEIH-box ATPase and an RNAannealing protein control the interconversion between free and base-paired U4/U6 snrnps. Materials and methods Yeast methods and strains Yeast genetic manipulations were performed using standard methods [48]. The four yeast strains employed in this work were created by plasmid shuffling in PRY118 (MAT a brr2::leu2 ade2 lys2 his3 ura3 leu2 pse360-brr2). PRY118 was a haploid segregant of the heterozygous BRR2/brr2::LEU2 diploid strain YSN404 transformed with psn108 [26]. Therefore, these strains differ only in their plasmid-borne BRR2 genes (Table 1). Plasmid construction Standard molecular biological techniques were used [49] in the HIS3 CEN ARS vector pse362 (= pun90; [50]) or in pbluescript KS (Stratagene). The 7 kb SacI BRR2 and brr2-1 fragments were inserted into pse362 to create ppr130 and ppr133, respectively. To place the polyoma epitope tag at the carboxyl terminus of each of these genes, we used the Bluescript subclone psn125, which contains an 860 bp SalI fragment encompassing the final 653 bp of BRR2 (S.M. Noble and C.G., unpublished). A NotI site was introduced by PCR immediately upstream of the stop codon to generate ppr149; the amplified BRR2 sequence was confirmed by sequencing. Annealed, kinased oligos encoding the polyoma epitope were inserted into the NotI site (5 -GGCCGCATGGAATATATGCCAATGGAAATGGAATATATGC- CAATGGAAGGC-3 ; 5 -GGCCGCCTTCCATTGGCATATATTCCA- TTTCCATTGGCATATATTCCATGC-3 ). The resulting SalI fragment with a BRR2 Pya carboxy-terminal fusion was swapped into ppr130 and ppr133 to create plasmids ppr150 and ppr151, respectively. Identification of the brr2-1 mutation The brr2-1 allele was recovered by gap repair onto plasmid psn123, which contains the BRR2 gene on the 2µ URA3 vector prs426 (S.M. Noble and C.G., unpublished; [51]). Briefly, this plasmid was linearized with BglII (which excises a 2.5 kb fragment spanning the first putative helicase domain of BRR2) and transformed into the brr2-1 strain YSN405 [26]. Plasmids were rescued from Ura +, cold-sensitive transformants. One plasmid (ppr137) did not complement the cold-sensitive brr2-1 lesion when transformed into YSN405; this plasmid was used to create ppr133. The 2.5 kb BglII fragment from ppr137 was subcloned into pbluescript KS + (Stratagene) for sequencing. Within this construct, the 447 bp ApaLI StuI fragment was fully sequenced on both strands, and only the A2417 G mutation was discovered. This fragment was replaced into ppr130, and found to confer coldsensitivity after plasmid shuffling in PRY118. Extract preparation Yeast whole-cell extracts were prepared from PRY122 (untagged BRR2), PRY123 (untagged brr2-1), PRY132 (BRR2 Pya), and PRY135 (brr2-1 Pya) using the liquid nitrogen method [52] with modifications [53]. Fraction I was prepared from PRY122 and PRY123 as described [54]. Gel analysis of snrnps Native snrnp gel analysis without heparin was modified from [11]. Samples (5 µl) contained 40% extract, 2 mm ATP (or ATPγS), 2.5 mm MgCl 2, 3% PEG 8000, 60 mm potassium phosphate ph 7, and 1 mm spermidine unless otherwise stated. Reactions were incubated 30 min at o C, and loaded on a pre-run, 4% polyacrylamide (80:1) gel ( cm) made up in TGM buffer (50 mm Tris base, 50 mm glycine, 2 mm MgCl 2 ). Electrophoresis was for 6 7 h at 160 V in TGM buffer at 4 C, and the gel was electroblotted onto Hybond-N membrane at V for h in 0.1 M sodium phosphate, ph 6.5. The blot was sequentially probed with kinased oligos complementary to the snrnas [55]. Immunoprecipitations Protein G Sepharose beads were coupled to anti-polyoma antibodies as described [33] and washed with NET50 buffer (50 mm Tris ph 7.4, Table 1 Yeast strains used in this study. Strain BRR2 genotype Plasmid PRY122 BRR2 ppr130 PRY123 brr2-1 ppr133 PRY132 BRR2 Pya ppr150 PRY135 brr2-1 Pya ppr151

8 854, Vol 8 No % Nonidet-P40, 50 mm NaCl). For co-immunoprecipitations of snrnps, 20 µl whole-cell extract, 200 µl NET with the indicated NaCl concentration, and 10 µl settled beads were mixed on a rocking platform for 1 h at 4 C. Some samples contained 1 mm competitor peptide encoding the polyoma epitope (EYMPME) to demonstrate specificity. The beads were washed three times with 500 µl NET50, and the associated RNAs were isolated by phenol chloroform extraction at 4 C followed by ethanol precipitation. RNA samples were separated on 6% polyacrylamide 7 M urea denaturing gels and subjected to northern analysis as described [55]. RNA levels from northern blots of denaturing gels were quantitated with a phosphorimager (Molecular Dynamics). Assay for snrnp release The snrnps were immunoprecipitated with Brr2 Pya or Brr2-1 Pya as described above. The beads were washed three times with 500 µl NET50, and then incubated with 30 µl splicing buffer ± ATP (2 mm ATP, 2.5 mm MgCl 2, 3% PEG, 60 mm potassium phosphate ph 7.2, 40% buffer D [56]) for 5 min at C. After centrifugation, the supernatants were carefully collected and the beads were washed once in 500 µl NET50. RNAs were extracted and split into three portions. To determine snrna content, one-half of each sample was electrophoresed on a denaturing gel and subjected to northern analysis [55]. To determine base-pairing status of U4 and U6 snrnas, one-quarter of each sample was hybridized in solution to an oligo probe for either U4 or U6 snrnas, and electrophoresed on a non-denaturing gel [36]. Acknowledgements We thank Bernhard Laggerbauer and Reinhard Lührmann for communicating results before publication; Cathy Collins, Guillaume Chanfreau and Jon Staley for discussions; Sean Burgess, Cathy Collins, Amy Kistler, Hiten Madhani, Stephen Rader, Jon Staley and Yan Wang for critical comments on the manuscript; Stephen Rader and Cathy Collins for generous sharing of reagents; Kurt Zingler for skillful manipulation of figures; and Lucita Esperas, Carol Pudlow and Heli Roiha for exceptional technical assistance. P.L.R. was a Howard Hughes Medical Institute predoctoral fellow. C.G. is an American Cancer Society Research Professor of Molecular Genetics. This work was supported by a grant from the NIH. References 1. Rymond B, Rosbash M: Yeast pre-mrna splicing. In The Molecular and Cellular Biology of the Yeast Saccharomyces. Cold Spring Harbor, New York: Cold Spring Harbor Laboratory Press; 1992: Moore MJ, Query CC, Sharp PA: Splicing of precursor to mrna by the spliceosome. In The RNA World. Cold Spring Harbor, New York: Cold Spring Harbor Laboratory Press; 1993: Will CL, Luhrmann R: Protein functions in pre-mrna splicing. Curr Opin Cell Biol 1997, 9: Nilsen TW: RNA-RNA interactions in the spliceosome: unraveling the ties that bind. Cell 1994, 78: Ares M, Weiser B: Rearrangement of snrna structure during assembly and function of the spliceosome. Prog Nucleic Acid Res Mol Biol 1995, 50: Madhani HD, Guthrie C: Dynamic RNA-RNA interactions in the spliceosome. Annu Rev Genet 1994, 28: Staley JP, Guthrie C: Mechanical devices of the spliceosome: motors, clocks, springs and things. Cell 1998, 92: Brow DA, Guthrie C: Spliceosomal U6 RNA is remarkably conserved from yeast to mammals. Nature 1988, 334: Madhani H, Guthrie C: A novel base-pairing interaction between U2 and U6 snrnas suggests a mechanism for the catalytic activation of the spliceosome. Cell 1992, 71: Cheng S-C, Abelson J: Spliceosome assembly in yeast. Genes Dev 1987, 1: Konarska MM, Sharp PA: Interactions between small nuclear ribonucleoprotein particles in formation of spliceosomes. Cell 1987, 49: Tarn W-Y, Lee K-R, Cheng S-C: Yeast precursor mrna processing protein PRP19 associates with the spliceosome concomitant with or just after dissociation of U4 small nuclear RNA. Proc Natl Acad Sci USA 1993, 90: Ayadi L, Miller M, Banroques J: Mutations within the yeast U4/U6 snrnp protein Prp4 affect a late stage of spliceosome assembly. RNA 1997, 3: Gorbalenya AE, Koonin EV: Helicases: amino acid sequence comparisons and structure-function relationships. Curr Opin Struct Biol 1993, 3: O Day CL, Dalbadie-McFarland G, Abelson J: The Saccharomyces cerevisiae Prp5 protein has RNA-dependent ATPase activity with specificity for U2 small nuclear RNA. J Biol Chem 1996, 271: Kim S-H, Smith J, Claude A, Lin R-J: The purified yeast pre-mrna splicing factor PRP2 is an RNA-dependent NTPase. EMBO J 1992, 11: Kim S-H, Lin R-J: Pre-mRNA splicing within an assembled yeast spliceosome requires an RNA-dependent ATPase and ATP hydrolysis. Proc Natl Acad Sci USA 1993, 90: Schwer B, Guthrie C: PRP16 is an RNA-dependent ATPase that interacts transiently with the spliceosome. Nature 1991, 349: Company M, Arenas J, Abelson J: Requirement of the RNA helicaselike protein PRP22 for release of messenger RNA from spliceosomes. Nature 1991, 349: Xu D, Nouraini S, Field D, Tang S-J, Friesen JD: An RNA-dependent ATPase associated with U2/U6 snrnas in pre-mrna splicing. Nature 1996, 381: Arenas JE, Abelson JN: Prp43: an RNA helicase-like factor involved in spliceosome disassembly. Proc Natl Acad Sci USA 1997, 94: Wang Y, Wagner JDO, Guthrie C: Prp16, a DEAH-box splicing factor, unwinds RNA in vitro. Curr Biol 1998, 8: Wagner JDO, Jankowsky E, Company M, Pyle AM, Abelson JN: The DEAH-box protein Prp22 is an ATPase that mediates ATPdependent mrna release from the spliceosome and unwinds RNA duplexes. EMBO J 1998, 17: Schwer B, Gross CH: Prp22, a DExH-box RNA helicase, plays two distinct roles in yeast pre-mrna splicing. EMBO J 1998, 17: Laggerbauer B, Achsel T, Lührmann R: The human 200kD DEXHbox U5 snrnp protein unwinds U4/U6 RNA duplices in vitro. Proc Natl Acad Sci USA 1998, 95: Noble SM, Guthrie C: Identification of novel genes required for yeast pre-mrna splicing by means of cold-sensitive mutations. Genetics 1996, 143: Lauber J, Fabrizio P, Teigelkamp S, Lane WS, Hartmann E, Luhrmann R: The HeLa 200 kda U5 snrnp-specific protein and its homologue in Saccharomyces cerevisiae are members of the DEXH-box protein family of putative RNA helicases. EMBO J 1996, 15: Lin J, Rossi JJ: Identification and characterization of yeast mutants that overcome an experimentally introduced block to splicing at the 3 splice site. RNA 1996, 2: Raghunathan PL, Guthrie C: A spliceosomal recycling factor that reanneals U4 and U6 small nuclear ribonucleoprotein particles. Science 1998, 279: Rothstein R: Targeting, disruption, replacement, and allele rescue: integrative DNA transformation in yeast. In Methods in Enzymology: Guide to Yeast Genetics and Molecular Biology. San Diego: Academic Press; 1991: Pause A, Sonenberg N: Mutational analysis of a DEAD box RNA helicase: the mammalian translation initiation factor eif-4a. EMBO J 1992, 11: Plumpton M, McGarvey M, Beggs JD: A dominant negative mutation in the conserved RNA helicase motif SAT causes splicing factor Prp2 to stall in spliceosomes. EMBO J 1994, 13: Schneider KR, Smith RL, O Shea EK: Phosphate-regulated inactivation of the kinase PHO80-PHO85 by the CDK inhibitor PHO81. Science 1994, 266: Fromont-Racine M, Rain J-C, Legrain P: Toward a functional analysis of the yeast genome through exhaustive two-hybrid screens. Nat Genet 1997, 16: Wells SE, Neville M, Haynes M, Wang J, Igel H, Ares M Jr: CUS1, a suppressor of cold-sensitive U2 snrna mutations, is a novel yeast splicing factor homologous to human SAP145. Genes Dev 1996, 10: Li Z, Brow DA: A rapid assay for quantitative detection of specific RNAs. Nucleic Acids Res 1993, 21: Rozen F, Edery I, Meerovitch K, Dever TE, Merrick WC, Sonenberg N: Bidirectional RNA helicase activity of eucaryotic translation initiation factors 4A and 4F. Mol Cell Biol 1990, 10:

9 Research Paper U4/U6 disruption requires Brr2 and ATP Raghunathan and Guthrie Lee C-G, Hurwitz J: A new RNA helicase isolated from HeLa cells that catalytically translocates in the 3 to 5 direction. J Biol Chem 1992, 267: Shuman S: Vaccinia virus RNA helicase: an essential enzyme related to the DE-H family of RNA-dependent NTPases. Proc Natl Acad Sci USA 1992, 89: Banroques J, Abelson JN: PRP4: a protein of the yeast U4/U6 small nuclear ribonucleoprotein particle. Mol Cell Biol 1989, 9: Galisson F, Legrain P: The biochemical defects of prp4-1 and prp6-1 yeast splicing mutants reveal that the PRP6 protein is required for the accumulation of the [U4/U6.U5] tri-snrnp. Nucleic Acids Res 1993, 21: Ruby SW, Chang T-H, Abelson J: Four yeast spliceosomal proteins (PRP5, PRP9, PRP11, and PRP21) interact to promote U2 snrnp binding to pre-mrna. Genes Dev 1993, 7: Arenas JE, Abelson JN: The Saccharomyces cerevisiae PRP21 gene product is an integral component of the prespliceosome. Proc Natl Acad Sci USA 1993, 90: Horowitz DS, Abelson J: A U5 small nuclear ribonucleoprotein particle protein involved only in the second step of pre-mrna splicing in Saccharomyces cerevisiae. Mol Cell Biol 1993, 13: Shannon KW, Guthrie C: Suppressors of a U4 snrna mutation define a novel U6 snrnp protein with RNA-binding motifs. Genes Dev 1991, 5: Utans U, Behrens S-E, Lührmann R, Kole R, Krämer A: A splicing factor that is inactivated during in vivo heat shock is functionally equivalent to the [U4/U6.U5] triple snrnp-specific proteins. Genes Dev 1992, 6: Burgess SM, Guthrie C: A mechanism to enhance mrna splicing fidelity: the RNA-dependent ATPase Prp16 governs usage of a discard pathway for aberrant lariat intermediates. Cell 1993, 73: Guthrie C, Fink GR: Guide to yeast genetics and molecular biology. In Methods in Enzymology. San Diego: Academic Press; 1991:3-21, Sambrook J, Fritsch E, Maniatis T: Molecular Cloning: A Laboratory Manual. 2nd edn. Plainview, New York: Cold Spring Harbor Laboratory Press; Elledge SJ, Davis RW: A family of versatile centromeric vectors for use in the sectoring-shuffle mutagenesis assay in Saccharomyces cerevisiae. Gene 1988, 70: Sikorski RS, Hieter P: A system of shuttle vectors and yeast host strains designed for efficient manipulation of DNA in Saccharomyces cerevisiae. Genetics 1989, 122: Umen JG, Guthrie C: A novel role for a U5 snrnp protein in 3 splice site selection. Genes Dev 1995, 9: Ansari A, Schwer B: SLU7 and a novel activity, SSF1, act during the PRP16-dependent step of yeast pre-mrna splicing. EMBO J 1995, 14: Cheng S-C, Abelson J: Fractionation and characterization of a yeast mrna splicing extract. Proc Natl Acad Sci USA 1986, 83: Noble SM, Guthrie C: Transcriptional pulse-chase analysis reveals a role for a novel snrnp-associated protein in the manufacture of spliceosomal snrnps. EMBO J 1996, 15: Lin R-J, Newman AJ, Cheng S-C, Abelson J: Yeast mrna splicing in vitro. J Biol Chem 1985, 260: Schmid SR, Linder P: D-E-A-D protein family of putative RNA helicases. Mol Microbiol 1992, 6: Because operates a Continuous Publication System for Research Papers, this paper has been published on the internet before being printed. The paper can be accessed from for further information, see the explanation on the contents page.