Link of NTR-Mediated Spliceosome Disassembly with DEAH-Box ATPases Prp2, Prp16, and Prp22

Abstract

The DEAH-box ATPase Prp43 is required for disassembly of the spliceosome after the completion of splicing or after the discard of the spliceosome due to a splicing defect. Prp43 associates with Ntr1 and Ntr2 to form the NTR complex and is recruited to the spliceosome via the interaction of Ntr2 and U5 component Brr2. Ntr2 alone can bind to U5 and to the spliceosome. To understand how NTR might mediate the disassembly of spliceosome intermediates, we arrested the spliceosome at various stages of the assembly pathway and assessed its susceptibility to disassembly. We found that NTR could catalyze the disassembly of affinity-purified spliceosomes arrested specifically after the ATP-dependent action of DEAH-box ATPase Prp2, Prp16, or Prp22 but not at steps before the action of these ATPases or upon their binding to the spliceosome. These results link spliceosome disassembly to the functioning of splicing ATPases. Analysis of the binding of Ntr2 to each splicing complex has revealed that the presence of Prp16 and Slu7, which also interact with Brr2, has a negative impact on Ntr2 binding. Our study provides insights into the mechanism by which NTR can be recruited to the spliceosome to mediate the disassembly of spliceosome intermediates when the spliceosome pathway is retarded, while disassembly is prevented in normal reactions.

INTRODUCTION

Introns are removed from precursor mRNAs (pre-mRNAs) via two transesterification reactions. The reactions take place on a large ribonucleoprotein complex called the spliceosome, which is composed of five small nuclear RNAs (snRNAs) and numerous protein factors. These factors bind to the pre-mRNA in a sequential manner to assemble the spliceosome into a functional complex for catalysis (for reviews, see references 1 to 4). After completion of the splicing reaction, the mature message is released and the spliceosome is disassembled to recycle its components.

Extensive structural rearrangement of the spliceosome, including exchange of RNA base-pairing and protein components (1, 2, 4–9), is associated with each step of the spliceosome assembly process. DEXD/H-box RNA helicases have been proposed to mediate structural changes of the spliceosome in distinct steps (10–15). In the budding yeast Saccharomyces cerevisiae, eight DEXD/H-box proteins are required for splicing. Prp5 and Sub2 are involved in early steps of spliceosome assembly to facilitate the formation of the prespliceosome (16–18). Prp28 and Brr2 are required in releasing U1 and U4, respectively, for the activation of the spliceosome (11, 15). Prp2 and Prp16 are required for the catalytic steps, and their activities are associated with the release of U2 components SF3a and SF3b (SF3a/b) and step-one factors Yju2 and Cwc25, respectively (19–24). After the completion of splicing, Prp22 is required for the release of mature mRNA and Prp43 for the disassembly of the spliceosome to recycle spliceosomal components (25–28). Although some of these proteins have been shown to unwind the RNA duplex in vitro, none show substrate specificity. Nevertheless, the DEXD/H-box proteins are regulated to function at precise stages of the splicing pathway. It has been shown that the RNA-unwinding activity of Brr2 is stimulated by a carboxy-terminal fragment of Prp8 (29), while Prp43 is activated by an amino-terminal fragment of Ntr1 containing the G-patch domain (30), suggesting that the function of the DEXD/H-box proteins might be regulated by associated splicing factors. The U5 component Snu114 has also been shown to serve as a signal-dependent switch to regulate spliceosome dynamics via regulation of Brr2 activity (32). Posttranslational modification of other splicing factors or of the DEXD/H-box proteins themselves has also been implicated in regulating the function of these helicases during spliceosome assembly (31, 33–35).

A protein complex termed NTR has been shown previously to mediate the disassembly of the spliceosome after the release of mature mRNA (36). NTR, which contains Ntr1 (also called Spp382), Ntr2, and the DEAH-box RNA helicase Prp43, interacts with U5 snRNP in a dynamic manner via the interaction of Ntr2 and U5 component Brr2. Such interactions may mediate the recruitment of NTR to the spliceosome by U5 snRNP (37). Since U5 is associated with the spliceosome early in the assembly pathway, the recruitment of NTR by U5 can occur early before the completion of the reaction. In support of this notion, Prp43 has recently been shown to be required for the disassembly of discarded spliceosome intermediates (38, 39). Nevertheless, NTR normally does not function until the mature message is released from the spliceosome after the completion of splicing, suggesting that the interaction of NTR with U5 is temporally regulated.

We investigated how NTR can mediate the disassembly of spliceosome intermediates and how the disassembly is prevented under normal splicing conditions. We found that NTR could specifically catalyze the disassembly of the spliceosome arrested at steps after the action of DEAH-box RNA helicase Prp2, Prp16, or Prp22 but not at steps before the action of these proteins or upon their binding. The actions of Prp2, Prp16, and Prp22 are known to be associated with the release of SF3a/b, Yju2/Cwc25, and Prp22/Prp18/Slu7, respectively (21, 23, 24, 46). Since these factors bind to the catalytic center of the spliceosome, near the branch site or the 3′ splice site, at each specific stage, the susceptibility of the spliceosome to disassembly might require the removal of these factors from the catalytic center. Analysis of Ntr2 binding revealed that splicing complexes exhibit differential binding affinity for Ntr2, generally in accordance with their susceptibility to disassembly, but not all complexes that bind Ntr2 well are susceptible. Our results demonstrate that NTR can be recruited to the spliceosome only at defined stages, and only when specific components that bind to the catalytic center are removed is the spliceosome susceptible to disassembly. Furthermore, the binding of Ntr2 to spliceosome intermediates is competitively inhibited by the presence of Prp16 and Slu7. Thus, our results provide a mechanism for prevention of the disassembly of spliceosome intermediates under normal conditions.

MATERIALS AND METHODS

Yeast strains.

The yeast strains used were BJ2168 (mata prc1 prb1 pep4 leu2 trp1 ura3), YSCC5 (mata prc1 prb1 pep4 leu2 trp1 ura3 YJU2-V5), YSCC10 (mata prc1 prb1 pep4 leu2 trp1 ura3 dbr1::LEU2), YSCC16 (mata prc1 prb1 pep4 leu2 trp1 ura3 URA3::GAL.PRP16), YSCC701 (mata prc1 prb1 pep4 leu2 trp1 ura3 SLU7-V5), SS304 (mata ura3 trp1 his3 ade2 prp2-1), and prp22 (mata ura3 trp1 his3 ade2 prp22-1).

Antibodies and reagents.

The anti-V5 antibody was purchased from Serotec Inc. The antibody against hemagglutinin (HA) was produced by immunizing mice with a keyhole limpet hemocyanin (KLH)-conjugated HA peptide (unpublished data). Anti-Prp16, anti-Prp22, anti-Slu7, anti-Ntr2, and anti-Cwc25 polyclonal antibodies were produced by immunizing rabbits with recombinant Prp16, Prp22, Slu7, Ntr2, and Cwc25 proteins, respectively. Dinucleotides 4-thio-UpG and UpG were purchased from Dharmacon. RNasin and SP6 RNA polymerase are from Promega, nuclease P1 from Sigma, and RNA ligase 1 from New England Biolabs. Protein A-Sepharose and the Ni-nitrilotriacetic acid (NTA) affinity column were obtained from Amersham Biosciences and Qiagen, respectively.

Oligonucleotides.

The following oligonucleotides were used: A1 (CTTCATCACCAACGTAG), A8 (TGTTAGTACATGAGAC), A9 (GCAACAAAAAGAATGAAGCAATCGTACATGAGACTTAGTAACA), S6 (TACGATTTAGGTGACAC), P2-6 (CTTTAAAATTGCTTATATCATTAGCAACAATGAACGCAAAAAA), P2-7 (TTTTTTGCGTTCATTGTTGCTAATGATATAAGCAATTTTAAAG), P16-5 (GATAAATATTCGTGTGTTATTATTGCTGAAGCTCATGAAAGGTCATTAAAT), and P16-6 (ATTTAATGACCTTTCATGAGCTTCAGCAATAATAACACACGAATATTTATC).

Purification of Prp2, prp2-S378L, prp16, prp16-D473A, prp22, prp22-S635A, Ntr2-HA, and Slu7-V5.

The PRP2 gene encoding a protein tagged with 4 copies of the V5 epitope at the amino terminus (41) and 2 copies of the HA epitope at the carboxyl terminus was cloned into pET15b. A serine-to-leucine mutation at position 378 was introduced by site-directed mutagenesis using primers P2-6 and P2-7. Both the wild-type protein and the prp2-S378L mutant recombinant protein were purified by consecutive chromatography on a nickel affinity column and anti-HA antibody-conjugated protein A-Sepharose beads. To purify Prp2 from yeast, extracts were prepared from a strain harboring plasmid pYES-HIS-PRP2 (a gift from R.-J. Lin) and were fractionated by chromatography on a nickel affinity column. The PRP16 gene encoding a protein tagged with 3 copies of HA at the carboxyl terminus was subcloned into pET28b for expression of the recombinant protein. An aspartate-to-alanine mutation at position 473 was introduced by site-directed mutagenesis using primers P16-5 and P16-6. Both the wild-type protein and the prp16-D473A mutant recombinant protein were purified as described for Prp2. Prp22 and the prp22-S635A mutant protein were purified according to the method of Tseng and Cheng (42). The NTR2 gene encoding a protein tagged with HA at the carboxyl terminus and the SLU7 gene encoding a protein tagged with V5 at the amino terminus were individually cloned into plasmid pSUMO for the expression of SUMO fusion proteins. After purification by chromatography on a nickel affinity column, the fusion proteins were treated with SUMO protease to remove SUMO.

Depletion of Cwc25, Slu7, Prp16, Prp22, and NTR.

To immunodeplete a factor from yeast extract, 25 mg of protein A-Sepharose was swollen in 1 ml of NET-2 buffer (50 mM Tris-HCl [pH 7.5], 150 mM NaCl, 0.05% NP-40) and was conjugated with a specific antibody. For Cwc25 depletion, 200 μl of a polyclonal anti-Cwc25 antibody was conjugated to 100 μl of protein A-Sepharose and used for depletion of 200 μl of extracts. For Prp16 depletion, 100 μl of a polyclonal anti-Prp16 antibody was used. For Ntr1 depletion, 50 μl of a polyclonal anti-Ntr1 antibody was used. For Slu7 depletion, 100 μl of a polyclonal anti-Slu7 antibody was used. For codepletion of Slu7 and Prp22, 100 μl of an anti-Slu7 antibody and 2.6 μg of a purified anti-Prp22 antibody were used. Extracts were incubated with antibody-conjugated protein A-Sepharose at 4°C for 1 h, and supernatants were collected as depleted extracts.

Arrest of spliceosome intermediates.

For the precatalytic spliceosome, splicing was carried out in heat-inactivated prp2-1 mutant extracts. For the Prp2-associated spliceosome, splicing was carried out in heat-inactivated prp2-1 mutant extracts, and the reaction mixture was depleted of ATP by incubation at 25°C for 5 min following the addition of 10 mM glucose. Then 100 ng of recombinant wild-type Prp2 or prp2-S378L, or 28 ng of Prp2 purified from yeast, was added to each 10 μl of the reaction mixture. For the post-Prp2 spliceosome, splicing was carried out in Yju2-depleted extracts as described previously (41). For the pre-Prp16 spliceosome, splicing was carried out in Prp16-depleted extracts. For the Prp16-associated spliceosome, splicing was carried out in Prp16-depleted extracts, and the reaction mixture was depleted of ATP by incubation at 25°C for 5 min following the addition of 10 mM glucose. Then 100 ng of recombinant wild-type Prp16 or prp16-D473A, or 0.5 μl of Prp16-overexpressing yeast extracts, was added to each 10 μl of the reaction mixture. To arrest the Prp22-associated spliceosome containing splicing intermediates, splicing was carried out with the 3′ splice site mutant ACAC pre-mRNA. To arrest the postcatalytic spliceosome, splicing was carried out in the presence of 50 to 200 ng of recombinant prp22-S635A for each 10 μl of the reaction mixture (42). To arrest the terminal-stage spliceosome, splicing was carried out in NTR-depleted extracts by using an anti-Ntr1 antibody as described previously (36).

Spliceosome disassembly assays.

Spliceosome disassembly assays were performed according to the method of Tsai et al. (36) with slight modifications. The splicing reaction was carried out under normal conditions by using actin precursor mRNA as the substrate, and the spliceosome was isolated by precipitation of 10 μl of the reaction mixture with 1 μl of an anti-Ntc20 antibody conjugated to 10 μl of protein A-Sepharose. The isolated spliceosome was then incubated under splicing conditions at 25°C for 20 min with 30 μl of a buffer containing 8 mM HEPES (pH 7.9), 60 mM KPO₄ (pH 7.0), 20 mM NaCl, 0.08 mM EDTA, 1 mM spermidine, 3% polyethylene glycol (PEG) 8000, 8% glycerol, 2 mM ATP, and 4 mM MgCl₂, with or without 5 μl of purified NTR. After centrifugation, supernatant and pellet fractions were collected for RNA analysis. Sedimentation analysis on glycerol gradients was performed as described by Tsai et al. (36).

Ntr2 binding assays.

To assay for Ntr2 binding, splicing reactions were carried out as described above to arrest the spliceosome at specific stages. To each 10 μl of the reaction mixture, 0, 20, 50, 100, 200, or 500 ng recombinant Ntr2-HA was then added. The reaction mixture was then precipitated with 1 μl of an anti-Ntc20 antibody or 80 μl of an anti-HA antibody. To assay for competition between Ntr2 and Prp16, splicing was carried out in Prp16-depleted extracts. Premixed recombinant prp16-D473A-HA and Ntr2-HA were then added to reaction mixtures at prp16-D473A-HA/Ntr2-HA molar ratios of 0, 0.04, 0.1, 0.2, and 0.4, with a final concentration of 0.5 μM for Ntr2-HA, and the mixtures were subjected to immunoprecipitation with an anti-Ntr2 or anti-Prp16 antibody. To assay for competition between Ntr2 and Slu7, splicing reactions were carried out with ACAC pre-mRNA in extracts immunodepleted of Slu7 and metabolically depleted of Ntr2. Premixed recombinant Slu7-V5 and Ntr2-HA were then added to reaction mixtures at Slu7-V5/Ntr2-HA molar ratios of 0, 0.04, 0.1, 0.2, and 0.4, with a final concentration of 1 μM for Ntr2-HA, and the mixtures were subjected to immunoprecipitation with an anti-Ntr2 or anti-V5 antibody. After immunoprecipitation, RNA was extracted from each reaction product, analyzed by gel electrophoresis, and quantified by using a phosphorimager.

RESULTS

NTR mediates the disassembly of spliceosome intermediates.

Under normal splicing conditions, NTR mediates spliceosome disassembly after mature mRNA is released. We first asked whether the spliceosome could be disassembled before mRNA release by using extracts prepared from the prp22-1 mutant. Prp22 is required both for the release of mRNA from the spliceosome after the completion of splicing and for the second catalytic reaction (26, 40). In order to isolate spliceosomes for the disassembly assay, we used antibodies against Ntc20 to precipitate the spliceosome, since Ntc20 binds stably to the spliceosome at various stages and remains associated with the spliceosome until the completion of the reaction (43). The precipitated spliceosomes were then incubated with NTR in the presence of ATP to determine whether substrate RNAs were released from beads (36). A scheme for the assay procedure is shown in Fig. 1A. The results of control experiments using an anti-Ntr1 antibody for the depletion of NTR are shown for the disassembly assay in the presence (Fig. 1B, lanes 5 and 6) or absence (lanes 3 and 4) of NTR (36). Immunoprecipitation with an anti-Ntc20 antibody revealed that splicing in heat-inactivated prp22-1 extracts results in the accumulation of spliced products as well as more splicing intermediates on the spliceosome (Fig. 1B, lane 8), suggesting that both mRNA release and splicing reactions were affected in the prp22-1 mutant extract. When NTR was added to the precipitated spliceosome in order to assay for disassembly, approximately 40 to 60% of all species of RNA was found in the supernatant (Fig. 1B, lanes 11 and 12). Smaller amounts of RNA were found in the supernatant after incubation in the absence of NTR (lanes 9 and 10). The dissociated materials were analyzed by sedimentation on 10-to-30% glycerol gradients (Fig. 1C), and the RNA contents of each gradient fraction were further analyzed by gel electrophoresis (Fig. 1D). While the RNA that was dissociated from beads after incubation in the absence of NTR (Fig. 1D, center) consisted primarily of intact spliceosome (fractions 5 to 9), the majority of the RNA that was dissociated in the presence of NTR (Fig. 1D, top) cosedimented with RNA extracted from the splicing reaction mixture (Fig. 1D, bottom) near the top of the gradient (fractions 12 to 15), indicating that RNA was dissociated due to spliceosome disassembly (36). These results suggest that NTR can also mediate the disassembly of the spliceosome containing pre-mRNA or splicing intermediates, and they raise the question whether NTR can associate with the spliceosome at multiple stages if spliceosome assembly is retarded. To determine whether all splicing complexes of assembly intermediates are susceptible to NTR-mediated disassembly, we blocked spliceosome assembly at different stages and isolated various splicing complexes by immunoprecipitation with an anti-Ntc20 antibody for disassembly assays.

Fig 1.

Disassembly of the spliceosome formed in prp22-1 mutant extracts. (A) Scheme showing the procedure of the disassembly assay. (B) Spliceosomes formed in NTR-depleted (lanes 1 to 6) or heat-inactivated prp22-1 mutant (lanes 7 to 12) extracts were precipitated with an anti-Ntc20 antibody (lanes 2 and 8), and the pellets were incubated in the presence (lanes 5, 6, 11, and 12) or absence (lanes 3, 4, 9, and 10) of NTR. (C) After incubation, the supernatant fractions from heat-inactivated prp22-1 mutant extracts (corresponding to panel B, lanes 10 and 12), as well as total RNA isolated from the splicing reaction mixture, were fractionated on 10-to-30% glycerol gradients. (D) RNA was extracted from each gradient fraction and was analyzed on denaturing polyacrylamide gels. IP, immunoprecipitation; dNTR, NTR-depleted extracts; Δprp22, heat-inactivated prp22-1 mutant extracts; R, reaction; T, total precipitates; P, pellet; S, supernatant.

Besides inhibiting mRNA release, the prp22-1 mutant displayed a general splicing defect, as evidenced by concomitant accumulation of a higher level of splicing intermediates. To specifically block mRNA release without affecting other splicing steps, we used a dominant negative mutant of Prp22, prp22-S635A. The S635A mutant is defective in RNA unwinding activity and cannot catalyze mRNA release but is functional for the second catalytic step and remains stably associated with the spliceosome (44). Splicing in the presence of recombinant prp22-S635A protein resulted in the accumulation of a large amount of lariat intron (Fig. 2A, lane 3), and the spliceosome formed was resistant to disassembly upon incubation with NTR (Fig. 2B, lanes 9 to 12). This result suggests either that the dissociation of mRNA is a prerequisite for the susceptibility of the spliceosome to NTR or that the presence of the Prp22 protein prevents NTR association. Considering that the mRNA-containing spliceosome formed in heat-inactivated prp22-1 mutant extracts was susceptible to disassembly, it is less likely that the association of mRNA prevented disassembly.

Fig 2.

Binding of Prp22 and Slu7 prevented the disassembly of the spliceosome. (A) Splicing in the presence of recombinant Prp22 (lane 2) or prp22-S635A (lane 3). (B) Spliceosomes formed in NTR-depleted extracts (lanes 1 to 6) or in the presence of prp22-S635A (lanes 7 to 12) were isolated by precipitation with an anti-Ntc20 antibody for disassembly assays, performed by incubation in the presence (lanes 5, 6, 11, and 12) or absence (lanes 3, 4, 9, and 10) of NTR. (C) Spliceosomes formed in mock-depleted (lanes 1 to 7) or Prp22- and Slu7-codepleted extracts (lanes 8 to 14) using ACAC pre-mRNA were isolated by precipitation with an anti-Ntc20 antibody for disassembly assays, performed by incubation in the presence (lanes 6, 7, 13, and 14) or absence (lanes 4, 5, 11, and 12) of NTR. The reaction mixtures were also precipitated with an anti-Prp22 antibody to reveal the presence (lane 2) or absence (lane 9) of Prp22. R, reaction; T, total precipitates; P, pellet; S, supernatant; α-22, anti-Prp22 antibody; α-Ntc20, anti-Ntc20 antibody. (D) Immunoblot of extracts depleted of Slu7 (lane 2), Prp22 (lane 3), or both Prp22 and Slu7 (lane 4). (E) The splicing reaction was carried out in mock-treated, Prp22/Slu7-depleted, or Prp22/Slu7/NTR-depleted dbr1Δ extracts with ACAC pre-mRNA, and the reaction mixtures were fractionated on 10-to-30% glycerol gradients. RNA was isolated from gradient fractions and was analyzed on 8% polyacrylamide–8 M urea gels. T, total reaction mixture.

Binding of Prp22 and Slu7 prevented NTR-mediated spliceosome disassembly.

To determine whether the presence of Prp22 would inhibit NTR-mediated spliceosome disassembly, the 3′ splice site mutant ACAC pre-mRNA (45) was used for the splicing reaction. The spliceosome formed with ACAC pre-mRNA accumulates splicing intermediates, with Prp22 retained on the spliceosome (42). Incubation of the affinity-purified spliceosome with NTR resulted in the disassembly of a very small amount of the spliceosome containing splicing intermediates (Fig. 2C, lanes 3 to 7). We then depleted the extract of Prp22 in order to determine whether the disassembly efficiency could be enhanced. Since stable association of Prp22 with the spliceosome requires Slu7 (46), Slu7 was codepleted to minimize the binding of residual Prp22. An immunoblot of extracts depleted of one or both proteins is shown in Fig. 2D. Nearly 30% of the spliceosome formed in the Prp22- and Slu7-codepleted extract was disassembled (Fig. 2C, lanes 10 to 14), indicating that the presence of Prp22 and Slu7 on the spliceosome indeed prevented NTR-mediated spliceosome disassembly. In the absence of Prp22 and Slu7, the spliceosome is susceptible to NTR even before the completion of the splicing reaction.

Notably, the amount of the spliceosome precipitated by the anti-Ntc20 antibody from Prp22- and Slu7-depleted extracts was consistently much lower than that from mock-treated extracts (Fig. 2C, compare lanes 3 and 10). We reasoned that a fraction of the spliceosome might have been disassembled during the splicing reaction prior to immunoprecipitation. Hence, we performed a sedimentation analysis to determine whether released splicing intermediates were present in the reaction mixture. To prevent degradation of the released lariat intermediate, we used extracts prepared from a debranchase-deficient (dbr1Δ) strain for this experiment. Splicing reactions were carried out with ACAC pre-mRNA in dbr1Δ extracts, and the reaction mixtures were subjected to glycerol gradient sedimentation (Fig. 2E). In mock-treated extracts, the majority of substrate RNA was associated with the spliceosome in fractions 5 and 6, with a small peak in lighter fractions 11 and 12. The lighter peak increased, with a decrease in the spliceosome peak, when the extract was depleted of Prp22 and Slu7 but did not increase significantly when NTR was also depleted to prevent disassembly, suggesting that the lighter peak was the product of the disassembled spliceosome. This was confirmed by RNA analysis of gradient fractions by gel electrophoresis, which showed that fractions 11 and 12 contained pre-mRNA and lariat intermediate, while fractions 13 and 14 contained the excised exon 1. These results indicate that the ACAC spliceosome is more susceptible to disassembly in the absence of Prp22 and Slu7.

Link of Prp2 and Prp16 to spliceosome disassembly.

The susceptibility of the ACAC spliceosome to disassembly confirmed that spliceosome intermediates could undergo disassembly. We then systematically analyzed different spliceosome intermediates to see whether all complexes or only specific complexes are susceptible to NTR-mediated disassembly. Each catalytic step of the splicing reaction involves an ATP-dependent structural change of the spliceosome, which requires DEAH-box RNA helicases Prp2 and Prp16, respectively, and an ATP-independent catalytic reaction, which requires several other proteins (19, 20, 41, 46–48). We first asked whether the spliceosome was susceptible to disassembly when arrested before the action of Prp16. The spliceosome formed in Prp16-depleted extracts accumulated splicing intermediates (Fig. 3A, lane 1), which could be isolated by precipitation with an anti-Ntc20 antibody (lane 2). Incubation of the spliceosome with NTR did not significantly promote disassembly (Fig. 3A, lanes 2 to 6), suggesting that the spliceosome that was arrested before the Prp16 step was not susceptible to disassembly. We then tested whether the spliceosome was susceptible to disassembly after the binding of Prp16. The ATPase mutant of Prp16, prp16-D473A, which can bind to the spliceosome but cannot dissociate from it, was used for the formation of the Prp16-associated spliceosome (49) (see Fig. S1A in the supplemental material). Splicing reactions were carried out in Prp16-depleted extracts, and the reaction mixtures were exhausted of ATP by incubation in the presence of glucose before the addition of Prp16. Subsequently, the spliceosome was precipitated with an anti-Ntc20 antibody for disassembly assays. Under these conditions, Prp16 was retained on the spliceosome. However, upon incubation with NTR and ATP for disassembly assays, Prp16 could hydrolyze ATP to promote a structural change of the spliceosome and could be released from the spliceosome (see Fig. S1B in the supplemental material) together with Yju2 and Cwc25. In contrast, the D473A mutant protein would be retained on the spliceosome (see Fig. S1C in the supplemental material). These results show that while the presence of the prp16-D473A mutant protein prevented the disassembly of the spliceosome (Fig. 3A, lanes 12 to 16), wild-type Prp16 promoted a structural change of the spliceosome into a form susceptible to NTR-mediated disassembly (lanes 7 to 11). The resulting structure of the post-Prp16 spliceosome is presumably like that formed in Prp22- and Slu7-codepleted extracts (see the scheme in Fig. 5A).

Fig 3.

Disassembly of spliceosomes formed after the actions of Prp16 and Prp2. (A) The splicing reaction was carried out in Prp16-depleted extracts. Following the depletion of ATP, either no Prp16 (lanes 1 to 6), wild-type Prp16 (lanes 7 to 11), or prp16-D473A (lanes 12 to 16) was added to the reaction mixture, and the spliceosome was isolated by precipitation with an anti-Ntc20 antibody for disassembly assays, performed by incubation in the presence (lanes 5, 6, 10, 11, 15, and 16) or absence (lanes 3, 4, 8, 9, 13, and 14) of NTR. (B) The splicing reaction was carried out in heat-inactivated prp2-1 mutant extracts. Following the depletion of ATP, either no Prp2 (lanes 1 to 6), wild-type Prp2 (lanes 7 to 11), or prp2-S378L (lanes 12 to 16) was added to the reaction mixture, and the spliceosome was isolated for disassembly assays as described for panel A. R, reaction; T, total precipitates; P, pellet; S, supernatant.

Fig 5.

Differential affinity of Ntr2 for spliceosomes arrested at various stages. (A) The splicing reaction was carried out under the following conditions: splicing in heat-inactivated prp2-1 extracts (a), splicing in heat-inactivated prp2-1 extracts, followed first by ATP depletion and then by the addition of purified yeast Prp2 (b), splicing in Yju2-depleted (dYju2) extracts (c), splicing in Prp16-depleted extracts (d), splicing in Prp16-depleted extracts, followed first by ATP depletion and then by the addition of Prp16-overexpressing yeast extracts (e), splicing in Slu7-depleted extracts (f), splicing with ACAC pre-mRNA (g), splicing in the presence of recombinant prp22-S635A (h), or splicing in Ntr1-depleted extracts (i). The reaction mixtures (lanes 1) were precipitated with an anti-Ntc20 antibody (lanes 2) or an anti-HA antibody either without the addition of Ntr2-HA (lanes 3) or following the addition of 20 ng (lanes 4), 50 ng (lanes 5), 100 ng (lanes 6), 200 ng (lanes 7), or 500 ng (lanes 8) of recombinant Ntr2-HA. (B) The results shown in panel A were quantified using a phosphorimager, and the percentage of the spliceosome precipitated by the anti-Ntc20 antibody that was also precipitated by the anti-HA antibody was plotted as a function of the amount of recombinant Ntr2-HA added. (C) Splicing reactions were carried out in mock- or Prp16-depleted extracts, and the reaction mixture was depleted of ATP by the addition of 10 mM glucose followed by incubation at 25°C for 5 min. Upon the addition of recombinant Ntr2 (lanes 3 and 7) or an affinity-purified NTR complex (lanes 4 and 8), the reaction mixtures were precipitated with an anti-Ntr2 antibody (lanes 2 to 4 and 6 to 8).

To determine whether precatalytic spliceosomes are also susceptible to NTR, we formed the spliceosome in heat-inactivated prp2-1 mutant extracts and precipitated the spliceosome with an anti-Ntc20 antibody (Fig. 3B, lane 2). The spliceosome that was formed before the action of Prp2 was not disassembled when incubated with NTR (Fig. 3B, lanes 2 to 6). We then tested whether the precatalytic spliceosome that was formed with Prp2 associated or after the action of Prp2 was susceptible to disassembly by performing an experiment similar to that described above for Fig. 3A. Splicing was carried out in heat-inactivated prp2-1 extracts, and the reaction mixtures were depleted of ATP before the addition of recombinant Prp2 to allow Prp2 binding. When the spliceosome that was precipitated with an anti-Ntc20 antibody was incubated with NTR, ca. 40% of the spliceosome was disassembled (Fig. 3B, lanes 7 to 11), presumably due to a prior structural change of the spliceosome catalyzed by Prp2 in the presence of ATP. This result suggests that the spliceosome is transformed into a state that is susceptible to disassembly after the action of Prp2. The dominant negative mutant protein prp2-S378L, which carries a mutation in the helicase motif, can bind to the spliceosome but cannot dissociate from it (50). When prp2-S378L was used, the spliceosome was resistant to disassembly (Fig. 3B, lanes 12 to 16), indicating that the Prp2-associated spliceosome is not susceptible to disassembly but that the action of Prp2 is required to render the spliceosome susceptible.

Taken together, our results demonstrate that the NTR complex can catalyze the disassembly of the spliceosome arrested specifically after the action of Prp2 or Prp16 but not at steps before their actions or upon their binding to the spliceosome. To exclude the possibility that the presence of substrate RNAs in the supernatant fractions in the disassembly assay was due to nonspecific dissociation of the spliceosome from beads during incubation, we showed that the release of substrate RNAs depends on functional Prp43 by using the NTR complex reconstituted with Prp43 ATPase mutant prp43-T123A (27, 37), which did not promote the release of substrate RNA from beads (see Fig. S2 in the supplemental material). We further analyzed the supernatant fraction from the post-Prp16 spliceosome on glycerol gradients to reveal that the released RNAs sedimented as naked RNAs, indicative of a disassembled spliceosome (see Fig. S3 in the supplemental material).

Removal of Cwc25 is required for disassembly of the precatalytic spliceosome.

It is interesting that NTR mediates spliceosome disassembly only after the action of DEAH-box proteins. The functions of Prp2 and Prp16 have recently been shown to be associated with the release of SF3a/b and Yju2/Cwc25, respectively (21, 23, 24). Both SF3b and Cwc25 are implicated in binding to the spliceosome at the catalytic center. SF3b binds directly to the branch site at early steps of spliceosome assembly and is released upon the action of Prp2 (21, 24). Cwc25 binds to the spliceosome after the release of SF3a/b to promote the first catalytic reaction and is then released from the spliceosome upon the action of Prp16 (23, 47). The binding of Cwc25 to the spliceosome is inhibited by mutations at the branch point, suggesting that Cwc25 might bind to the pre-mRNA near the branch site (23). This is supported by site-specific photo-cross-linking of Cwc25 to pre-mRNA carrying a single 4-thiouridine (4sU) residue 3 bases downstream of the branch point (see Fig. S4 in the supplemental material). In this context, the spliceosome may be susceptible to disassembly only when specific factors binding to the catalytic center are removed. We tested whether the spliceosome assembled in Cwc25-depleted extracts is susceptible to disassembly. Cwc25 is recruited to the spliceosome immediately before the first catalytic reaction. Spliceosomes formed in Cwc25-depleted extracts contain all the other factors required to promote the reaction and can be chased to yield splicing intermediates upon the addition of recombinant Cwc25 (47). When NTR was added to the spliceosome isolated by precipitation with an anti-Ntc20 antibody, nearly half of the precatalytic spliceosome was dissociated from beads (Fig. 4A, lanes 5 and 6). The spliceosome formed in Cwc25-depleted extracts is at the post-Prp2 stage, at which SF3b has been removed from the branch site. To exclude the possibility that the fraction of the spliceosome susceptible to disassembly represents that without Yju2 bound, we specifically isolated a Yju2-containing spliceosome for disassembly assays. The spliceosome formed in Cwc25-depleted Yju2-V5 extracts was precipitated with an anti-V5 antibody and was incubated with NTR in order to examine disassembly (Fig. 4B). More than half of the spliceosome was dissociated from beads (Fig. 4B, lanes 5 and 6), indicating that the association of Yju2 did not prevent disassembly. Together, these results suggest that disassembly of the precatalytic spliceosome requires only the preclusion of Cwc25 binding.

Fig 4.

NTR can disassemble the spliceosome in the absence of Cwc25. (A) The splicing reaction was carried out in Cwc25-depleted extracts, and the spliceosome was isolated for disassembly assays by precipitation of the reaction mixture (lane 1) with an anti-Ntc20 antibody (lane 2), followed by incubation in the presence (lanes 5 and 6) or absence (lanes 3 and 4) of NTR. (B) The splicing reaction was carried out in Cwc25-depleted Yju2-V5 extracts, and the spliceosome was isolated for disassembly assays by precipitation of the reaction mixture (lane 1) with an anti-V5 antibody (lane 2), followed by incubation in the presence (lanes 5 and 6) or absence (lanes 3 and 4) of NTR. (C) The splicing reaction was carried out with brC pre-mRNA, and the spliceosome was isolated for disassembly assays by precipitation of the reaction mixture (lane 1) with an anti-Ntc20 antibody (lane 2), followed by incubation in the presence (lanes 5 and 6) or absence (lanes 3 and 4) of NTR. R, reaction; T, total precipitates; P, pellet; S, supernatant.

We have shown previously that stable association of Cwc25 with the spliceosome is greatly affected by mutations at the branch point sequence and that in the absence of ATP hydrolysis, the binding of Prp16 stabilizes the association of Cwc25 to promote the reaction (23). In the presence of ATP, Prp16 promotes the dissociation of Cwc25 from the spliceosome to prevent splicing upon ATP hydrolysis. Conceivably, the spliceosome assembled on branch point mutant pre-mRNA, particularly in the absence of Prp16, is susceptible to disassembly due to poor binding of Cwc25. Indeed, when the spliceosome that was formed on pre-mRNA carrying an A-to-C mutation at the branch point (brC) was precipitated with an anti-Ntc20 antibody, ca. 40% was dissociated from beads after incubation with NTR (Fig. 4C, lanes 5 and 6). This further finding confirms that prevention of the binding of Cwc25 to the spliceosome is critical for NTR-mediated disassembly of the precatalytic spliceosome.

Differential affinity of Ntr2 for different spliceosome intermediates.

We next addressed the question of whether the susceptibility of splicing complexes to disassembly correlates with their ability to recruit NTR. We have shown previously that Ntr2 mediates the recruitment of NTR to the spliceosome via its interaction with Brr2 and that Ntr2 alone can bind the spliceosome (37). The capacity of splicing complexes for Ntr2 binding may reflect their ability to recruit NTR. To assay for Ntr2 binding of different splicing complexes, splicing was carried out under various conditions to arrest the spliceosome at different stages. Increasing amounts of recombinant Ntr2-HA were then added to the reaction mixtures, followed by precipitation with an anti-HA or anti-Ntc20 antibody (Fig. 5A). For precatalytic steps, spliceosomes were arrested by using heat-inactivated prp2-1 mutant extracts (Fig. 5Aa), by the addition of Prp2, purified from Prp2-overexpressing extracts, to the splicing reaction mixture after ATP depletion (Fig. 5Ab), or by the depletion of Yju2 from extracts (Fig. 5Ac). For poststep 1, spliceosomes were arrested by the depletion of Prp16 (Fig. 5Ad), by the addition of diluted Prp16-overexpressing extracts, which overproduce Prp16 around 20-fold (see Fig. S5 in the supplemental material), to the splicing reaction mixture after ATP depletion (Fig. 5Ae), by the depletion of Slu7 (Fig. 5Af), or by using ACAC pre-mRNA (Fig. 5Ag). For postcatalytic spliceosomes, splicing was performed in the presence of prp22-S635A (Fig. 5Ah). For terminal-stage spliceosomes, splicing was carried out in Ntr1-depleted extracts (Fig. 5Ai). The amounts of Ntr2-bound spliceosomes were quantified by a phosphorimager and were normalized to those of Ntc20-bound spliceosomes (Fig. 5A, lanes 2) after subtraction of the amount of RNA precipitated in the absence of Ntr2-HA (lanes 3). The percentages of the Ntr2-bound spliceosome were then plotted against the amounts of Ntr2-HA added (Fig. 5B).

It is noteworthy that the percentage of the Ntr2-bound spliceosome may be slightly underestimated due to the presence of endogenous Ntr2, which is estimated at approximately 8 ng in each reaction mixture, except for Ntr1-depleted extracts (Fig. 5Ai), in which Ntr2 was codepleted (36). Nevertheless, the results show that the splicing complexes can be classified into three groups based on their affinities for Ntr2. The terminal-stage spliceosome, which is the original substrate of the NTR complex in normal splicing reactions, bound Ntr2 with the highest affinity (indicated by a black circle), and approximately 30% or 40% of the Ntc20-bound spliceosome contained Ntr2-HA when 20 ng or 200 ng, respectively, of Ntr2-HA was added (Fig. 5Ai). Spliceosome intermediates susceptible to disassembly (post-Prp2 [Fig. 5Ac] and post-Prp16 [Fig. 5Af]) bound Ntr2 with intermediate affinity (indicated by black squares), with approximately 20% of the Ntc20-bound spliceosome containing Ntr2-HA when 200 ng of Ntr2-HA was added. These splicing complexes can also bind the NTR complex, as revealed by the binding of exogenously added mutant prp43-T123A to the spliceosome formed in Yju2-depleted or Slu7-depleted extracts (see Fig. S6 in the supplemental material). In contrast, only barely detectable amounts of the spliceosome were precipitated in the presence of 20 ng of Ntr2, and less than 10% of the Ntc20-bound spliceosome contained Ntr2-HA, with as much as 500 ng of Ntr2 added for pre-Prp2 (Fig. 5Aa), Prp2-bound (Fig. 5Ab), Prp16-bound (Fig. 5Ae), and Prp22-bound (Fig. 5Ag and h) spliceosomes, none of which are susceptible to disassembly. To exclude the possibility that the HA epitope of Ntr2-HA was not as accessible to the antibody in these spliceosomes as in the others, poor binding of Ntr2 to these spliceosomes was further confirmed by their inability to efficiently recruit prp43-T123A to the spliceosome (see Fig. S7 in the supplemental material). As an exception, the pre-Prp16 spliceosome, which retained Yju2 and Cwc25 and was not susceptible to disassembly, bound Ntr2 with intermediate affinity (Fig. 5Ad). Furthermore, the purified NTR complex could also bind to the pre-Prp16 spliceosome in the absence of ATP (Fig. 5C, lane 8). These results suggest that splicing complexes that do not bind Ntr2 well are generally not susceptible to disassembly, but not all of the complexes that are able to bind NTR well are susceptible. Taken together, our results reveal two important structural features of the spliceosome that is susceptible to disassembly: (i) the ability to bind Ntr2 with good affinity and (ii) the removal of specific proteins that bind at the catalytic center.

Competitive inhibition of Ntr2 binding by Prp16 and Slu7.

The results discussed above show that NTR can mediate the disassembly of the spliceosome when it is arrested at the post-Prp2 or post-Prp16 stage, yet NTR needs to be precluded in order to avoid the disassembly of spliceosome intermediates under normal splicing reactions. Conceivably, splicing factors from various steps of the pathway may compete with Ntr2 to prevent its binding to the spliceosome. Brr2 has been shown, by yeast two-hybrid analysis, to interact with Prp2, Prp16, and Slu7 via the second helicase domain (51). As with Ntr2, such interactions might play roles in recruiting these proteins to the spliceosome during the catalytic steps. We first used two-hybrid assays to determine whether Ntr2 also interacts with the second helicase domain of Brr2 (Fig. 6A). The Brr2 protein is roughly divided into five regions carrying the amino-terminal (N), first helicase (H1), middle (M), second helicase (H2), and carboxyl-terminal (C) segments. Figure 6A shows that the N-terminal half of the protein (amino acid residues 1 to 1369), comprising the N, H1, and M segments, or a fragment containing the C-terminal sequence alone (amino acid residues 1713 to 2163), does not interact with Ntr2. However, the carboxy-terminal half of the protein (amino acid residues 1209 to 2163) comprising H2 and C interacts with Ntr2. Thus, Ntr2 also interacts with the second helicase domain of Brr2, and its binding to the spliceosome may be prevented by Prp2, Prp16, and Slu7 during the progression of the spliceosome pathway.

Fig 6.

Competition of Prp16 and Slu7 with Ntr2 for spliceosome binding. (A) Yeast two-hybrid assays showing the region of Brr2 interacting with Ntr2. N, amino terminus; H1, 1st helicase domain; M, middle region; H2, 2nd helicase domain; C, carboxy terminus; BD, binding domain; AD, activation domain; V, vector; FL, full length. (B) Splicing reactions were carried out in Prp16-depleted extracts. To the reaction mixture were added premixed recombinant Ntr2 and prp16-D473A with a final concentration of 500 nM for Ntr2 (lanes 1 to 15) and a final concentration of 0 nM (lanes 1 to 3), 20 nM (lanes 4 to 6), 50 nM (lanes 7 to 9), 100 nM (lanes 10 to 12), or 200 nM (lanes 13 to 15) for prp16-D473A. The mixtures were immunoprecipitated with an anti-Prp16 antibody (lanes 2, 5, 8, 11, and 14) or an anti-Ntr2 antibody (lanes 3, 6, 9, 12, and 15). RXN, 1/10 of reaction mixture. (C) The results shown in panel B were quantified using a phosphorimager, and the percentage of Ntr2 that remained bound to the spliceosome in the presence of prp16-D473A was plotted against the ratio of the amount of prp16-D473A to that of Ntr2. (D) Splicing reactions were carried out with ACAC pre-mRNA in extracts depleted of Slu7 in vitro and metabolically depleted of Ntr2. To the reaction mixture were added premixed recombinant Ntr2 and Slu7-V5 with a final concentration of 1,000 nM for Ntr2 (lanes 1 to 15) and a final concentration of 0 nM (lanes 1 to 3), 40 nM (lanes 4 to 6), 100 nM (lanes 7 to 9), 200 nM (lanes 10 to 12), or 400 nM (lanes 13 to 15) for Slu7-V5. The mixtures were immunoprecipitated with an anti-V5 antibody (lanes 2, 5, 8, 11, and 14) or an anti-Ntr2 antibody (lanes 3, 6, 9, 12, and 15). (E) The results shown in panel D were quantified using a phosphorimager, and the percentage of Ntr2 that remained bound to the spliceosome in the presence of Slu7-V5 was plotted against the ratio of the amount of Slu7-V5 to that of Ntr2.

To determine whether the spliceosome binding of Ntr2 can be compromised by Prp16, we performed competition assays. Splicing was carried out in Prp16-depleted extracts, and premixed recombinant Ntr2 and prp16-D473A were then added to the reaction mixtures at different ratios with a fixed Ntr2 concentration of 500 nM (Fig. 6B). The Ntr2-bound spliceosome was isolated by precipitation with an anti-Ntr2 antibody, and the percentage of Ntr2 that remained bound in the presence of prp16-D473A was plotted against the ratio of the amount of prp16-D473A to that of Ntr2. Figure 6C shows that the amount of Ntr2-bound spliceosome was reduced by more than 40% in the presence of 20 nM prp16-D473A (at a ratio of 0.04) and by 80% at 100 nM prp16-D473A (at a ratio of 0.2), indicating that Prp16 binds better than Ntr2 to the pre-Prp16 spliceosome and can compete with Ntr2 for binding to remove NTR from the spliceosome.

A similar competition experiment was performed with Slu7. Splicing reactions were carried out in Slu7-depleted extracts by using ACAC pre-mRNA. To prevent disassembly of the spliceosome in the absence of Slu7, extracts metabolically depleted of Ntr2 were used for this experiment (37). Premixed recombinant Ntr2 at a final concentration of 1 μM and various amounts of Slu7-V5 protein were added to the reaction mixtures. The Ntr2- and Slu7-bound spliceosome was then precipitated by an anti-Ntr2 and an anti-V5 antibody, respectively (Fig. 6D), and the percentage of Ntr2 that remained bound in the presence of Slu7 was plotted against the ratio of the amount of Slu7 to that of Ntr2 (Fig. 6E). Figure 6E shows that the amount of Ntr2-bound spliceosome was reduced by 25% in the presence of 40 nM Slu7 (at a ratio of 0.04), indicating that Slu7 also binds better than Ntr2 to the post-Prp16 spliceosome. The amount of Ntr2-bound spliceosome was reduced by approximately 50% at 400 nM Slu7 (at a ratio of 0.4). Further increases in the amount of Slu7 did not prevent more Ntr2 binding to the spliceosome (data not shown). Since stable binding of Slu7 to the spliceosome is facilitated by the presence of Prp22 (46), it is possible that the amount of endogenous Prp22 is not sufficient to support Slu7 binding when Slu7 is added in large amounts. Alternatively, Prp22-dependent rejection of the spliceosome might prevent further inhibition of Ntr2 binding by Slu7. Furthermore, the pre-Prp16 spliceosome appeared to accumulate in larger amounts in the absence of Slu7 and accounted for a fraction of the spliceosome whose binding of Ntr2 was resistant to inhibition by Slu7, as the addition of Prp16 was also able to inhibit Ntr2 binding under such conditions (see Fig. S8 in the supplemental material). Taken together, our results show that in the presence of Prp16 or Slu7, binding of Ntr2 to the pre-Prp16 or post-Prp16 spliceosome is competitively inhibited, providing a mechanism to prevent NTR from disassembling spliceosome intermediates under normal splicing conditions.

DISCUSSION

DEXD/H-box proteins Prp43 and Brr2 have been demonstrated to play roles in spliceosome disassembly (25, 27, 32, 36). Prp43 associates with Ntr1 and Ntr2 to form the NTR complex, which mediates disassembly. Ntr1 interacts with Prp43 via its G-patch domain (36), which also stimulates the helicase activity of Prp43 (30), and with Ntr2 via its middle region. Ntr2 mediates the binding of NTR to the spliceosome and can bind itself to the spliceosome via the interaction with Brr2 (37). Mutations in Brr2 have also been shown to impede the dissociation of lariat-intron and the separation of U2 and U6 from the Prp43-associated spliceosome (32). In view of the fact that U5 is associated with the spliceosome early during spliceosome assembly, it is possible that NTR can be recruited to mediate the disassembly of the spliceosome at early steps of the pathway. Several genetic studies support this notion. NTR1 (also named SPP382) was also identified as a suppressor of the prp38-1 mutation. Prp38 is a component of yeast tri-snRNP and is required for the activation of the spliceosome (52). Several mutations in PRP43 affecting ATPase activity have also been shown to suppress the growth defect of the prp38-1 mutation with efficiencies inversely proportionate to the measured ATPase activities (53), suggesting that reducing the activity of Prp43 could partially compensate for impaired spliceosome assembly. Recently, it has been further demonstrated that Prp43 promotes the discarding of spliceosome intermediates, cooperating with the function of Prp16 and Prp22 in proofreading the splicing reaction (38, 39). By systematic in vitro analysis, we show here that NTR can mediate the disassembly of spliceosome intermediates, but only at defined stages of the pathway. NTR can function after the ATP-dependent action of each DEAH-box protein—Prp2, Prp16, and Prp22—but not prior to their action or while they are associated with the spliceosome (Fig. 5A). This observation indicates that NTR is functionally linked to these DEAH-box proteins.

Disassembly of the spliceosome requires prior removal of proteins binding to the catalytic center.

The action of Prp2, Prp16, and Prp22 is associated with the release or destabilization of spliceosomal components binding at the catalytic center at each specific stage. SF3b components bind to the branch site during spliceosome assembly, and prior to catalysis, they are destabilized or lost from the spliceosome depending on the wash conditions, indicating that the mode of their interaction with the spliceosome is changed upon the action of Prp2 (21, 22, 24). Prior to the action of Prp2, when SF3a/b are still tightly associated, the spliceosome is not susceptible to disassembly, and only after SF3a/b are destabilized does the spliceosome become susceptible. Destabilization of SF3a/b allows the binding of Cwc25 to promote the first catalytic reaction. In the absence of Cwc25, the stalled precatalytic spliceosome is susceptible to disassembly, and the presence of another step-one factor, Yju2, does not prevent the disassembly of the spliceosome. After lariat formation, the action of Prp16 is required to remove Yju2 and Cwc25 from the spliceosome so as to prepare for the second catalytic reaction. With Yju2 and Cwc25 associated, the spliceosome is not susceptible to disassembly. The release of Yju2 and Cwc25 allows the step-two factors, Slu7, Prp18, and Prp22, to bind to the substrate RNA at the 3′ splice site to promote the second reaction. In the absence of Slu7 and Prp22, the stalled spliceosome is also susceptible to disassembly, and stable association of Prp22 and Slu7 prevents spliceosome disassembly. After exon ligation, Prp22 catalyzes the release of mRNA and is itself dissociated from the spliceosome together with Slu7 and Prp18. At this terminal stage, the second-step factors are removed from the catalytic center, and the spliceosome is readily disassembled. Taken together, these results strongly suggest that specific factors binding at the catalytic center at each catalytic step have to be removed for the spliceosome to be susceptible to NTR-mediated disassembly. How the binding of these proteins prevents disassembly is not known. RNA base pairings form the framework of the catalytic center of the spliceosome but are stabilized by the binding of protein factors. At each catalytic step, destabilization of these proteins presumably allows more structural dynamics in the catalytic center of the spliceosome so that the positioning of splice sites can be adjusted. It is possible that such dynamics may also create a fragile environment in which the spliceosome is susceptible to disassembly.

Stable binding of NTR to the spliceosome is required but not sufficient for mediating disassembly.

On examining the correlations between the susceptibilities of different splicing complexes to disassembly and their affinities for NTR binding, we found that splicing complexes can be classified into three groups based on their affinities for Ntr2. As expected, the lariat-intron-associated spliceosome, as the authentic substrate for NTR under normal splicing conditions, binds Ntr2 with a much higher affinity than any of the other complexes analyzed. Spliceosomes formed in Yju2-depleted or Slu7-depleted extracts, corresponding to post-Prp2 and post-Prp16 spliceosomes, respectively, also bind Ntr2 but with a lower affinity. All the complexes that are not susceptible to disassembly bind Ntr2 poorly, except for the pre-Prp16 spliceosome, which binds Ntr2 with an affinity comparable to that of post-Prp2 and post-Prp16 spliceosomes. These results suggest that distinct structural features of the spliceosome may determine its affinity for Ntr2. Alternatively, the binding site of Ntr2 may be blocked by other spliceosomal components to preclude premature disassembly of functional spliceosomes. Although complexes that hardly bind Ntr2 are consistently not susceptible to disassembly, not all complexes that bind Ntr2 with intermediate affinity are susceptible to disassembly. For example, in the case of the pre-Prp16 spliceosome, the binding of Cwc25 precludes the disassembly of the spliceosome.

It is worth noting that despite its ability to access spliceosome intermediates stalled at specific stages, NTR has to be excluded from the splicing pathway before the completion of the reaction to avoid premature disassembly of functional spliceosomes during the splicing reaction. One way to exclude NTR is through competitive binding of specific splicing factors to the same site of the spliceosome. We show that Ntr2 interacts with Brr2 through the second helicase domain, which has been shown by two-hybrid assays to interact with Prp2, Prp16, and Slu7 also (51). It is likely that Brr2 serves as the platform for recruiting these factors at various stages of the spliceosome pathway. Conceivably, competitive interaction of Prp2, Prp16, or Slu7 may block the interaction of Ntr2 with Brr2 to prevent NTR recruitment. In agreement with this notion, the binding of Ntr2 to pre-Prp16 and post-Prp16 spliceosomes is greatly reduced by the presence of Prp16 or Slu7, respectively. Furthermore, the Prp2-, Prp16-, and Slu7-bound spliceosomes all have very low capacities for Ntr2 binding, but the disassembly-resistant pre-Prp16 spliceosome can bind Ntr2 much better with none of these Brr2-interacting factors bound. These results support the model of competitive exclusion of NTR from the spliceosome under normal splicing conditions.

For the post-Prp2 spliceosome, the binding of Cwc25 to the spliceosome is expected to be more favorable than that of Ntr2 in preventing premature disassembly. Nevertheless, Cwc25 was not observed to compete for Ntr2 binding to the purified post-Prp2 spliceosome (data not shown). Nor does Cwc25 interact with Brr2 by yeast two-hybrid assays (data not shown). The mechanism underlying the prevention of NTR-mediated disassembly of the post-Prp2 spliceosome is unknown. We speculate that Cwc25 might be kinetically favored by the post-Prp2 spliceosome, and upon binding to the spliceosome, Cwc25 may rapidly promote the first reaction and remain stably bound to the spliceosome to prevent NTR-mediated disassembly.

Linkage of spliceosome disassembly with proofreading.

DEXD/H-box ATPases have been implicated in proofreading pre-mRNA splicing at multiple steps (54–56). Prp16 was first found to enhance the fidelity of branch site recognition, since mutations in PRP16 increase the accumulation of aberrant branch site intermediates (54). A kinetic proofreading mechanism was proposed for the role of Prp16 in fidelity control by coupling ATP hydrolysis with a discard pathway to remove incorrect RNA splicing intermediates from the normal splicing pathway (54, 57), but not until recently was a detailed molecular mechanism of the proofreading process uncovered (23, 38, 58). Using a modified U6 with sulfur substituting for a nonbridging oxygen at position U80, which binds Mg²⁺ in the first catalytic reaction (59), Koodathingal et al. showed that Prp16 mediates the rejection of the slow spliceosome before 5′ splice site cleavage and that Prp43 is required for the disassembly of the rejected spliceosome (38). We showed previously that Prp16 mediates the release of Yju2 and Cwc25 after the first catalytic reactions under normal splicing conditions but could do so before the reaction when catalysis is slow due to mutations at the branch point. This result suggests that Prp16 may function in proofreading by removing Cwc25 to prevent the catalytic reaction (23). Here we show further that NTR can mediate the disassembly of precatalytic spliceosomes devoid of Cwc25. Together, these results suggest that a slow spliceosome is likely destined for disassembly after Prp16-mediated rejection through Cwc25 removal.

Similarly, Prp22 has been shown to repress the splicing of aberrant intermediates and to enhance splicing fidelity by mediating the rejection of the spliceosome in an ATP-dependent manner (55), and the rejected intermediates have been shown to be discarded via Prp43 (39). The 3′ splice site mutant pre-mRNA ACAC formed spliceosomes that accumulate splicing intermediates. Prp22, Slu7, and Prp18 remain bound to such spliceosomes but cannot promote the second catalytic reaction (46). We show that NTR cannot catalyze the disassembly of the purified ACAC spliceosome but that prior depletion of Prp22 and Slu7 from the extract rendered it susceptible to disassembly, suggesting that Prp22 and Slu7 protect the spliceosome from disassembly. In this view, Prp22 may mediate the rejection of the spliceosome by hydrolyzing ATP; dissociation of Prp22 from the spliceosome together with Slu7 and Prp18 then allows NTR to access the spliceosome to elicit disassembly. In agreement with this notion, disassembly of the ACAC spliceosome in the absence of Prp22 and Slu7 was observed in the splicing reaction without prior purification of the spliceosome. We found that the amount of the spliceosome containing splicing intermediates was consistently smaller in Prp22- and Slu7-depleted extracts than in control extracts. When splicing reaction mixtures were analyzed by glycerol gradient sedimentation, free lariat intermediate was detected in the reaction mixture from Prp22- and Slu7-depleted extracts, indicating that a portion of the spliceosome was disassembled during the splicing reaction.

Prp5 has also been proposed to proofread the U2-branch site pairing in an early step of the spliceosome pathway (56), but the fate of the rejected spliceosome is not known. Since Prp5 acts early before the binding of tri-snRNP, it is questionable that NTR can be recruited to the spliceosome to mediate disassembly at this stage. Still, it will be interesting to follow the fate of the rejected spliceosome mediated by Prp5.