Saccharomyces cerevisiae PRP17-null mutants are temperature-sensitive for growth. In vitro splicing with extracts lacking Prp17 are kinetically slow for the first step of splicing and are arrested for the second step at temperatures greater than 34 °C. In the present study we show that these stalled spliceosomes are compromised for an essential conformational switch that is triggered by Prp16 helicase. These results suggest a plausible mechanistic basis for the second-step arrest in prp17Δ extracts and support a role for Prp17 in conjunction with Prp16. To understand the association of Prp17 with spliceosomes we used a functional epitope-tagged protein in co-immunoprecipitation experiments. Examination of co-precipitated snRNAs (small nuclear RNAs) show that Prp17 interacts with U2, U5 and U6 snRNPs (small nuclear ribonucleoproteins) but it is not a core component of any one snRNP. Prp17 association with in-vitro-assembled spliceosome complexes on actin pre-mRNAs was also investigated. Although the U5 snRNP proteins Prp8 and Snu114 are found in early pre-spliceosomes that contain all five snRNPs, Prp17 is not detectable at this step; however, Prp17 is present in the subsequent pre-catalytic A1 complex, containing unspliced pre-mRNA, formed after the dissociation of U4 snRNP. Thus Prp17 joins the spliceosome prior to both catalytic reactions. Our results indicate continued interactions in catalytic spliceosomes that contain reaction intermediates and in post-splicing complexes containing the lariat intron. These Prp17–spliceosome association analyses provide a biochemical basis for the delayed first step in prp17Δ and explain the previously known multiple genetic interactions between Prp17, factors of the Prp19-complex [NTC (nineteen complex)], functional elements in U2 and U5 snRNAs and other second-step splicing factors.
- pre-mRNA splicing
- Saccharomyces cerevisiae
- second-step splicing
Pre-mRNA splicing requires an ordered assembly of various spliceosome components on an intron-containing transcript. The process involves precise, temporally ordered and regulated interactions among the five spliceosomal snRNAs (small nuclear RNAs), U1, U2, U4/U6 and U5, and also interactions with the pre-mRNA which leads to intron excision and exon ligation. The chemical reactions are a set of two transesterifications, the catalytic centre for which is formed by the snRNA components of the spliceosome. An important role for protein factors is to direct RNA structural rearrangements in the spliceosome which are critical for catalysis . The complement of protein factors in the Saccharomyces cerevisiae spliceosome is ∼100 . Several factors are components of snRNP (small nuclear ribonucleoprotein) particles, whereas others have independent associations either directly with the nascent transcript or mediated through protein–protein interactions.
Six yeast protein factors function at the second step of splicing. They are Prp8, Prp16, Prp18, Slu7, Prp22 and Prp17 [3–9]. Prp8 is an integral component of the U5 snRNP and assembles on the spliceosome as a part of the U5 particle [10,11]. Besides its essential role in spliceosome assembly and the first catalytic reaction, Prp8 contributes to 3′ splice-site recognition during the second step . Prp16, an RNA-dependent helicase, associates transiently with the spliceosome after the first catalytic reaction and induces a critical conformational switch, detectable as a protection of the 3′ splice-site region [3,4]. Slu7 and Prp18 function at an ATP-independent step, subsequent to this Prp16 action [12,13]. Like Prp16, both Prp18 and Slu7 are in close proximity to the 3′ splice-site and Slu7 can be cross-linked with this substrate region [14,15]. Another factor positioned near the 3′ splice-site is the DEAH-box helicase, Prp22, whose second-step functions are ATP-independent . Subsequently, the helicase activity of Prp22 is required for the release of mRNA from the spliceosome.
Prp17 is essential for cell viability only at elevated temperatures [5,13]. The second-step function of Prp17 is at, or around, the step of Prp16 action . Unlike other second-step factors, Prp17 cannot be directly cross-linked to the 3′ splice-site, but its presence is needed for the efficient cross-linking of Prp8 and Slu7 [14,15]. Except for Prp8, a U5 snRNP protein, the other yeast and mammalian second-step factors are not integral components of any snRNP. They associate and dissociate from the spliceosome at distinct steps related to their functional requirements. The precise second-step function of Prp17, its snRNP interactions if any, and the stage of spliceosome assembly at which it associates with the spliceosome are not known.
In the present study, we have assessed Prp17 second-step functions and spliceosomal associations. Our results indicate that the arrest of the second reaction at higher reaction temperatures, in the absence of Prp17, is at least in part due to compromised Prp16 function. Utilizing a functional, polyoma- and poly-histidine-tagged Prp17p we have deciphered the snRNP associations and the likely entry and exit points of Prp17 during the spliceosome cycle. We find that Prp17 interacts with the U2, U5 and U6 snRNPs in assembling catalytic spliceosomes. The early association of Prp17 with the pre-catalytic A1 spliceosome formed upon the dissociation of U4 snRNP explains the previously reported genetic interactions of Prp17 with the NTC (nineteen complex; Prp19-complex). Together these results support a role for Prp17 in the early steps of spliceosome assembly and explain the multiple genetic interactions previously known between Prp17 and other spliceosome components.
Yeast strains and plasmids
Splicing extracts were prepared from the following strains: prp17ΔBJ2168 , SC261.8 , wild-type BJ2168 and prp17-ΔBJ2168 transformed with the expression construct p17NHP424. The latter was created by cloning the Prp17 open reading frame (lacking the first four amino acids) as a ClaI fragment into the AccI site of the N-terminal-epitope-tagging vector, pNHP424 (PSmD3, TRP1, 2μ origin) . Prp17 is thus expressed from the heterologous SmD3 promoter as a translational fusion with the N-terminal poly-histidine and polyoma middle T-antigen tags.
Yeast splicing extracts, in vitro transcription and pre-mRNA splicing reactions
Yeast splicing extracts were prepared as described in . In vitro transcription and splicing reactions were performed as in . [α-32P]UTP-labelled wild-type and 3′ splice-site mutant C303/305 actin transcripts (where/indicates the splice site) were generated by SP6 polymerase using EcoRI- and HindIII-linearized pSP65 actin and pSP65C303/305 plasmids respectively . Biotinylated actin substrates were prepared by including 25 μM biotin11-UTP (NEN) in addition to [α-32P]UTP during transcription . The radiolabelled 750 nucleotide intronless non-specific RNA was transcribed with T3 RNA polymerase on an SspI linearized pBS(KS+) template (Stratagene). Splicing reactions were resolved on 7% (19:1) denaturing polyacrylamide gels. Reactions with 5 mM EDTA or various ATP concentrations [21,22] were carried out as indicated. Quantification of the pre-mRNA substrate, splicing intermediates and products was performed using phosphorimaging.
RNaseH depletion of U snRNAs and Northern blot analysis
Oligonucleotide-directed RNaseH-based depletion of U2, U5 or U6 snRNAs from splicing extracts was carried out in ∼25 μl reaction mixtures as described in  for U2,  for U5 and  for U6 snRNA. U2 and U6 snRNAs were ablated from extracts of prp17Δ containing tagged Prp17 protein. Briefly, 20 μM of U2 oligo-SRU2 and U6 oligo, oligo d1, were added and the extracted aliquots were incubated at 35 °C for 30 min. For effective U5 depletion, extracts from SC261.8, expressing a modified U5 snRNA, were used . The oligos U5-1 and U5-2 were added; incubation at 35 °C for 45 min caused U5 ablation. Northern blot analysis of U snRNAs, in mock and snRNA-depleted extracts, was carried out with PCR-labelled DNA probes for U1, U2, U4, U5 and U6 snRNAs .
The Prp16-helicase-dependent 3′ splice-site protection assay was carried out with oligoA . Splicing reactions with the wild-type actin substrate were carried out at the indicated temperatures for 30 min either with or without the addition of oligoA in each 5 μl reaction mixture. Independently, 5 μl reaction mixtures with the mutant C303/305 actin substrate were performed at 34 °C and 37 °C again with or without the addition of HPLC-purified oligoA added 15 min after the start of these reactions, following which the reaction was continued for 10 min.
Immunoprecipitation of snRNPs from splicing extracts and in-vitro-assembled splicing complexes: affinity purification of spliceosomes
Anti-polyoma monoclonal antibodies  or anti-Prp17 polyclonal antibodies  were immobilized on protein-G–sepharose 4B beads (GE Healthcare). For immunoprecipitation of U snRNAs in splicing extracts, 10 μl of extract was added to 10 μl of the specified bead-immobilized antiserum. For immunoprecipitation of U snRNAs or of substrate RNA species from in vitro reactions, 20 μl aliquots were added to an equal volume of bead-immobilized antiserum. Binding was carried out in NET-150 buffer [50 mM Tris/HCl (pH 7.4), 150 mM NaCl and 0.05% Nonidet P40] for 1 h at 4 °C. Direct load of the extracts alone or of in vitro splicing reactions were usually one-fourth to onesixth of the amount used for immunoprecipitation.
For analysis of proteins in spliceosomes with biotinylated substrate RNA, splicing reactions containing 300 μl of splicing extracts were assembled on 483 ng of biotinylated C303/305 mutant actin pre-mRNA. As a control, the same substrate RNA was prepared, but was non-biotinylated. After 25 min at 22 °C, 80 μl of streptavidin–sepharose beads (GE Healthcare) in 750 μl of NET-150 buffer was added. Binding was allowed for 1.5 h at 4 °C and bead-bound spliceosomes were washed extensively with NET-150 buffer. The bead-bound proteins were extracted and analysed by Western blot analysis. All reaction mixtures were pre-incubated at 22 °C for 15 min prior to addition of the specified precursor substrate RNA.
Anti-polyoma antibodies (1:2000 dilution) were used to detect Prp17 in Western blots. Anti-Prp8.6 (kindly provided by Dr Jean Beggs, Wellcome Trust Centre for Cell Biology, University of Edinburgh, Edinburgh, Scotland, U.K.) and anti-Snu114 (kindly provided by Dr Patrizia Fabrizio, Max Planck Institute for Biophysical Chemistry, Department of Cellular Biochemistry, Göttingen, Germany) antibodies were used in Western blot analysis to detect Prp8 and Snu114 respectively.
A temperature-dependent requirement for Prp17 in pre-mRNA splicing
Prp17 is a non-essential splicing factor as cells with null alleles of prp17 are viable, although slow growing, at temperatures below 33 °C; however, this protein is essential for cell survival at temperatures at or above 33 °C . The accumulation of splicing intermediates in these cells indicated a second-step splicing defect as the likely cause for the lethality. To investigate the correlation between the splicing defect and the growth phenotype of prp17Δ, we have performed in vitro splicing reactions on wild-type actin pre-mRNA at five different temperatures, 15 °C, 22 °C, 30 °C, 34 °C and 37 °C. The reactions used extracts from wild-type (Supplementary Figure S1A at http://www.BiochemJ.org/bj/416/bj4160365add.htm) or extracts from prp17::LEU2 cells that were grown at 23 °C (Supplementary Figure S1B). Comparison of the reaction kinetics  over this temperature range showed that in the absence of Prp17, a sharp decrease in reaction kinetics occurred between 30 °C and 34 °C, which was further reduced when the reaction temperature was raised from 34 °C to 37 °C. Wild-type extracts showed only a very marginal reduction in reaction kinetics at 30–34 °C (Supplementary Figures S1C and S1D). This suggests that the absence of Prp17, and not the higher reaction temperature, is responsible for the compromised kinetics of the second step in reactions with the mutant extract.
Previous studies on the ATP requirement during the second-step reaction defined two stages within this step. Prp17 is thought to function at or around the time of Prp16 action in the early ATP-dependent stage [13,26]. Prp16, an ATP-dependent helicase, induces a change in spliceosome conformation, which can be detected by a protection of the 3′ splice-site in the lariat intermediate from RNaseH nuclease . We have examined the status of the arrested prp17Δ spliceosome, using this protection assay, to understand its nature. We reasoned that if Prp17 arrests the second step by compromising Prp16 function, there would be an increased accessibility of the 3′ splice-site to oligo-directed RNaseH cleavage. Splicing reactions using wild-type actin pre-mRNA and either prp17Δ or wild-type extracts were carried out. We examined whether the 3′ splice-site in the lariat intermediates formed in these reactions were protected from RNaseH cleavage. To trigger this cleavage we allowed the formation of RNA–DNA hybrids by supplementing wild-type and prp17Δ reactions with a DNA oligonucleotide (oligoA) complementary to sequences 16 nucleotides upstream of the 3′ splice-site of the RNA substrate. Only if the 3′ splice site is accessible would the RNA–DNA hybrid be formed and then cleaved by endogenous RNaseH present in yeast splicing extracts. The cleaved lariat intermediates would migrate faster than the intact lariat intermediate. In the pre-mRNA substrate the 3′ splice-site is entirely accessible to the oligoA-directed cleavage. The cleavage of the pre-mRNA thus serves as a positive control to ensure comparable RNaseH activity in the two extracts. The appearance of the RNaseH-cleavage product, from the lariat intermediate in reactions with prp17Δ (Figure 1A, lanes 1–6, arrow on the right-hand side) was compared with the results from reactions with wild-type extracts (Figure 1A, lanes 7–12, arrow on the right-hand side). The reactions were performed at 15 °C (lanes 2 and 8), 22 °C (lanes 3 and 9) or 30 °C (lanes 4 and 10). In reactions carried out at temperatures up to 30 °C, the appearance of the cleaved lariat intermediate was the same in prp17Δ or wild-type extracts. At 34 °C (lanes 5 and 11) and 37 °C (lanes 6 and 12), the lack of Prp17 caused a second-step defect and therefore lariat intermediates accumulated (compare lanes 5 and 6 with lanes 11 and 12). The 3′ splice-sites of these accumulated lariat intermediates were partially accessible to RNaseH cleavage (lane 5) as some cleaved lariat intermediates were generated. To quantitatively assess the extent of Prp16-mediated remodelling in the arrested prp17Δ spliceosomes, we carried out a more sensitive experiment using the C303/305 3′ splice-site mutant (CAG/AG to CAC/AC) in an actin pre-mRNA substrate . This mutation in the substrate arrests the second step, therefore catalytic spliceosomes accumulate but, importantly, it does not affect the Prp16-triggered remodelling of the 3′ splice-site . The efficiency of spliceosome remodelling in these stalled C303/305 lariat intermediates was compared in parallel reactions performed with wild-type and prp17Δ extracts (Figure 1B). The reactions were carried out at 34 °C (Figure 1B, lanes 1–6) and at 37 °C (Figure 1B, lanes 7–12). As before, the oligoA-directed cleavage of the input pre-mRNA served as an internal control to compare the two extracts for their RNaseH activity. In reactions with wild-type extracts at 34 °C, we observed only a very marginal increase in the cleaved lariat intermediate upon addition of oligoA at two different concentrations (Figure 1B, lanes 2 and 3, and Figure 1C, set 1). The same is seen even at 37 °C (Figure 1B, lanes 8 and 9). We infer that the majority of the wild-type spliceosomes have undergone the Prp16-dependent 3′ splice-site restructuring and are refractory to RNaseH cleavage at the 3′ splice site. In contrast, in reactions with prp17Δ extracts performed at 34 °C, we observed an increased amount of the cleaved lariat intermediate, showing greater accessibility of the 3′ splice-site for oligoA-directed cleavage (Figure 1B, compare lanes 2 and 3 with lanes 5 and 6, and Figure 1C, sets 1 and 2). This is notable since the cleavage of the pre-mRNA is comparable in wild-type and prp17Δ reactions (Figure 1D). Similarly, reactions at 37 °C with prp17Δ extracts showed increased levels of cleaved lariat intermediate (Figure 1B, compare lanes 8 and 9 with lanes 11 and 12). Thus at higher reaction temperatures the arrested prp17Δ spliceosomes are compromised for remodelling of the 3′ splice-site. We therefore infer that Prp17 improves the efficiency of Prp16-dependent second-step functions at higher reaction temperatures or it acts to support a complex suitable for Prp16 action.
Prp17 interacts with the U2, U5 and U6 snRNPs
Prp17p is a 52 kDa protein with a highly conserved C-terminal domain with seven WD-40 repeat motifs . The conformation of the C-terminal WD-40 motifs is critical to Prp17 function [28,29]. To investigate functional spliceosomal interactions of Prp17, we constructed an N-terminal double-tagged (polyoma middle T-antigen and poly-histidine tags) protein that completely rescued the prp17::LEU2 temperature-sensitive phenotype (results not shown). Such prp17-null cells expressing the tagged Prp17 protein were used in further experiments.
We first examined the snRNP interactions of Prp17 through immunoprecipitation of the tagged Prp17p from splicing extracts. A consistent co-immunoprecipitation of three snRNAs, U2, U6 and U5, was observed (Figures 2A and 2B). Notably, the U2, U6 and U5 interactions were retained up to 500 mM NaCl concentrations (lanes 3–5), and the weak interactions of Prp17 seen with the U1 and U4 snRNP were lost at 150 mM NaCl. To ascertain that these snRNP associations are not merely an artefact of tagged Prp17 which is expressed from a plasmid at higher than endogenous levels, we also examined snRNAs pulled down by anti-Prp17 antibodies. In these experiments the association of Prp17p with snRNPs in a wild-type extract was examined (Figure 2B). The spliceosomal snRNA co-immunoprecipitation profile with these antibodies remained the same as that observed with anti-polyoma antibodies, confirming that the tagged Prp17 retained the normal in vivo interactions of Prp17. We conclude that Prp17 shows a specific and NaCl-stable interaction with the U2, U5 and U6 snRNPs.
We next assessed whether these Prp17 associations were mediated by direct interaction with any one of these snRNPs. Splicing extracts were depleted of U2 snRNA , U5 snRNA  or U6 snRNA  using oligonucleotide-directed RNaseH digestion. We used Northern blot analysis to assess the depletion of these snRNAs in treated extracts. U5 depletion was carried out in the SC261.8 extract that expresses only a modified U5 that is more accessible for ablation . The ablation was nearly complete (Figure 3A, lane 3). U2 depletion from extracts expressing the tagged Prp17 protein produced a shorter cleaved product (lane 2), which, as expected, was inactive due to the loss of a branchpoint recognition sequence. An approx. 90% U6 depletion was obtained in extracts with tagged Prp17 protein (lane 4). We also ascertained that these depleted extracts were inactive for in vitro splicing (Figure 3B), after which they were used in co-immunoprecipitation experiments with anti-Prp17 antibodies. Mock-treated SC261.8 extract and tagged Prp17 extract served as controls to establish that snRNA depletion alone accounted for the loss of splicing activity. The depletion of any one snRNA did not compromise the interaction of Prp17 with the other two snRNPs (Figure 3C). Yet, we noted a reproducible decrease (∼50%) in the levels of U2 snRNA co-immunoprecipitated when U6 was ablated (Figure 3C, lane 4, and Figure 3D). The effect on levels of U5 co-precipitated was less (∼15% decrease). An unstable Prp17–U6 snRNP interaction may therefore weaken, but not completely abolish, the interaction of Prp17 with the U2 snRNP. From these experiments we conclude that Prp17 interacts with the U2/U5/U6 snRNP particle; an association not strictly mediated through any single snRNP, but perhaps with the integral complex.
Prp17 associates with the splicing substrate, intermediates and products
In vitro, in addition to an arrested second step, the first-step reaction is marginally compromised in prp17Δ extracts [7,30]. To determine the in vitro association of the substrate RNA species with Prp17, we performed immunoprecipitations of Prp17 from reactions assembled with actin pre-mRNA (Figure 4). Splicing reactions with tagged Prp17 extract, 20 min after initiation at 22 °C, were the input for immunoprecipitations. As expected, from the second-step Prp17 function, we observed its interaction with products of the first reaction, i.e. lariat intermediate and exon 1 (Figure 4C, lane 1). In addition, the products of the second reaction, i.e. mRNA and the lariat intron, were also precipitated (lane 1). None of these first-step or second-step RNA species bound non-specifically to the beads (Figure 4D, lane 1). A significant amount of the input splicing substrate was also associated with Prp17 (Figure 4C, lane 1, and quantification in Figure 4F, set 1).
Two inferences can be drawn. First, that Prp17 remains associated, at least transiently, with the late spliceosomes and post-splicing complexes as indicated by the pulldown of the spliced mRNA and lariat intron. Release of the spliced mRNA, post-catalysis, is Prp22-helicase-activity-dependent and is facilitated by higher concentrations of ATP [9,31]. We observed a ∼4-fold higher immunoprecipitation of the lariat intron (13.5%) by the anti-Prp17 affinity-tagged antibodies, than the mRNA (3.5%). This suggests an association of Prp17 with the intron-containing post-splicing D complex. Our second inference from these immunoprecipitations is that Prp17 associates with the spliceosome prior to, or during, the first step as suggested by the significant co-immunoprecipitation of pre-mRNA substrate. This feature is not seen with other second-step splicing factors such as Prp16 , Slu7  and Prp22 .
Prp17 associates with the pre-catalytic spliceosome
The progression of spliceosome assembly to catalytic complexes can be stalled at specific stages to enrich for some assembly intermediates. These complexes can be distinguished by their snRNP compositions . Reactions performed with reduced exogenously added ATP proceeded until the formation of the A2-1 complex, which was normally only a transient complex containing all five spliceosomal snRNPs (Figure 4A). In reactions with 2 mM ATP this complex undergoes rapid remodelling/activation whereby U1 and U4 snRNPs are displaced leaving only U2, U5 and U6 snRNPs in association with the substrate. Another stable spliceosome intermediate that can be accumulated in vitro is the pre-catalytic A1 spliceosomes stalled prior to the first catalytic step. A1 complexes can be accumulated by the inclusion of 5 mM EDTA in the reaction  (Figures 4A and 4B, lane 3). We examined whether Prp17 was present in the A1 complex by immunoprecipitation with anti-polyoma antibodies. Substantial co-immunoprecipitation of the pre-mRNA from reactions containing 5 mM EDTA confirmed the association of Prp17 with the A1 complex (Figure 4C, lane 3, and quantified in Figure 4F, set 3). Thus, unlike other second-step factors, Prp17 is present in a pre-catalytic spliceosome.
Association of Prp17 with complexes that precede A1 (Figure 4A), such as the A2-1 complex, was also studied. The stalled A2-1 complex was obtained by limiting the exogenously added ATP to 50 μM  and the immunoprecipitated substrate pre-mRNA compared with that precipitated from the precatalytic A1 and catalytic A2-2 and A2-3 spliceosomes. In each reaction, the immunoprecipitated substrate RNA was compared with the non-specific association, and tested in a matched aliquot of the same reaction to the beads alone (Figure 4D). A consistent and reproducible pulldown of the pre-mRNA was seen, establishing the presence of Prp17 in the A1 complex (Figures 4B–4D, lane 3, and quantified in Figure 4F, set 3, and Supplementary Figures S2B and S2C at http://www.BiochemJ.org/bj/416/bj4160365add.htm). Furthermore, a somewhat weak co-immunoprecipitation of the pre-mRNA, that is marginally higher than non-specific binding of pre-mRNA to the beads alone, was observed in reactions enriched for the A2-1 complex (Figures 4B–4D, lane 2, and quantified in Figure 4F, set 2). The immunoprecipitation of pre-mRNA in A2-1 was however variable in replicate experiments (Supplementary Figures S2A and S2B). Therefore this association was re-examined further as described below. Reactions performed in the absence of any ATP did not have spliceosomes with stably associated U2 snRNP and are thus stalled after CC complex formation. Prp17 association with the pre-mRNA in the reactions was tested and the results showed no enrichment of pre-mRNA in the immunoprecipitate (Supplementary Figure S2D). To ascertain the specificity of the RNA interactions, we carried out mock splicing reactions using an intronless radiolabelled non-specific RNA in reactions containing either 2 mM or 50 μM ATP. The association of Prp17 with this non-specific RNA was examined by immunoprecipitation (Figure 4E). In each case, the amount of non-specific RNA pulled down was reproducibly the same as that associated with the beads alone (Figure 4G). This confirmed that, in the absence of splicing, no specific association of Prp17 occurred with the intronless RNA. Through these experiments, we demonstrated the early association of Prp17 with the pre-catalytic A1 spliceosomes. This association may be initiated with the A2-1 complex, but perhaps the dissociation of U4 snRNP greatly stabilizes Prp17 spliceosome interactions.
Prp17 assembles on spliceosomes after Prp8 and Snu114 (components of the U5 snRNP)
Mutations in Prp17 display synthetic lethal interactions with Prp8, a U5 snRNP factor, and with several components of the Prp19-complex (NTC) . Components of the NTC stably associate with the spliceosome during the A2-1 to A1 transition coincident with the release of U4 snRNP [20,22,33–35]. To substantiate our finding of Prp17 association with pre-catalytic spliceosomes, such complexes were assembled on biotinylated actin substrates and the spliceosomes were affinity purified. The status of the enriched spliceosome, separated from free snRNPs (Figure 5A), was verified in each reaction by determining its snRNP content [21,22].
In reactions with 50 μM ATP, the A2-1 complex accumulated and this was the only complex with all of the five snRNPs (Figure 5A, lane 3). Subsequent to the binding of all snRNPs in A2-1, conformational rearrangements occurred to displace U1 from the 5′ splice-site and U4 from the U4/U6 di-snRNP, leaving only U2, U6 and U5 in the enriched A1 complex (Figure 5A, lane 4). Reaction conditions that arrest at complex A1 possibly slow down the early assembly events and therefore the enriched complex A1 is contaminated by minor amounts of the temporally preceding complexes (note the presence of some U4 snRNA in lane 4). The catalytically active A2-2 and A2-3 complexes formed at ATP concentrations greater than 0.5 mM contained the U2, U5 and U6 snRNPs (lanes 1 and 2). The minor amounts of U4 snRNA even in these reactions perhaps represent the small fraction of early assembly intermediates.
Adopting these procedures, large-scale reactions with tagged Prp17 extract were assembled on biotinylated mutant C303/305 actin pre-mRNA. We examined the presence of Prp17 in A2-1 and A1 spliceosomes. Non-biotinylated unlabelled substrate was used in parallel as the control reactions. Prp17 in the streptavidin-bound spliceosomes was examined by Western blot analysis, using anti-polyoma antibodies (Figure 5B, lower panel). As control spliceosomal proteins, we probed the presence of two U5 snRNP factors, Prp8 and Snu114, using polyclonal antisera (Figure 5B, top panel). The Prp17 protein was clearly discernible and enriched in the affinity-purified A1 spliceosomes (Figure 5B, compare lane 2 with lane 3, and Figure 5C). This observation confirms the earlier conclusions based on the immunoprecipitation of the radiolabelled substrate (Figure 4F, quantified in set 3), but Prp17 association with A2-1 spliceosomes (in reactions with 0.05 mM ATP) was not greater than its non-specific association with beads (Figure 5B, compare lane 4 with lane 5, and Figure 5C). Taken together we conclude that the Prp17 is not present in the A2-1 complex that contains all five snRNPs, but is present in the pre-catalytic A1 complex. The detection of Snu114 and Prp8 proteins in affinity-purified spliceosomes (Figure 5B, lanes 2 and 4) at levels greater than the control (Figure 5B, lanes 3 and 5) established the presence of these U5 proteins and showed that the stability of the complexes was not affected by the experimental procedures.
In parallel reactions, we examined the snRNP content of Prp17-associated spliceosomes by Northern blot analysis of immunoprecipitated in vitro splicing reactions (Figure 6A). We observed interactions of Prp17 with U2, U5 and U6 snRNPs in reactions that accumulated pre-catalytic spliceosomes (Figure 6A, lanes 3 and 4). Also this association was maintained in the subsequent catalytic spliceosomes (Figure 6A, lanes 1 and 2). The levels of Prp17-associated U5 and U6 snRNPs were marginally lower in the conditions that enriched the A2-1 complex as compared with those wherein the A1 or A2-2 and A2-3 complexes were formed (Figure 6A, compare lanes 1, 2 and 4 with lane 3). This suggests that transition from the A2-1 to A1 complex may stabilize the association of Prp17 with the spliceosome, as was also indicated by the increased pulldown of the radiolabelled pre-mRNA in A1 complexes as described above (Figure 4F).
Prp17 is essential for the viability of yeast only at temperatures above 32 °C [13,28], and in vitro splicing with model substrates show a Prp17-dependence for the second-step reaction when reactions are at 37 °C . A predicted reason for the dispensability of Prp17 at lower temperatures, both in vivo and in vitro, is that it plays an auxiliary role in second-step splicing. Certain genetic data support this expected role [13,28]. Although biochemically Prp16 and Prp17 act at the same ATP-dependent stage during the second step, overexpression of Prp16 partially suppresses the lethality of prp17Δ cells at 34 °C, but not at 37 °C  (A. K. Sapra, P. Khandelia and U. Vijayraghavan, unpublished work). The kinetic analysis of the second step in prp17Δ extracts that we report in the present study further support the genetic data and implicate Prp17 in facilitating efficient second-step splicing at high temperatures. A plausible role for Prp17 at higher temperatures is in stabilizing a co-factor or facilitating a fragile interaction. Increased susceptibility of the 3′ splice-site in the lariat intermediate, to RNaseH cleavage, indicates that in prp17Δ spliceosomes the helicase action of Prp16 is compromised. Thus the second-step arrest in prp17Δ probably reflects, at least in part, a Prp16-dependent function for Prp17. In an earlier study , we have shown that, in the absence of Prp17, introns longer than 200 nucleotides are inefficiently spliced at higher temperatures. This is especially so if these introns have a branch point to 3′ splice-site distance >12 nucleotides. It is therefore plausible that the association of Prp17 with the spliceosome enables efficient splicing of such introns. Similarly, a critical role for Prp18 in supporting viability at temperatures greater than 34 °C has been previously reported . An evolutionarily conserved domain in Prp18 conferred this role . The authors suggest that this Prp18 domain stabilizes a protein or RNA conformation that is unstable above 34 °C. Together these results implicate a stabilizing role for these otherwise dispensable yeast second-step factors.
Among the protein factors required for the second step, Prp8, Prp16, Prp17, Prp18, Prp22 and Slu7, only Prp8 is an integral component of the U5 snRNP. In the present study, we report an NaCl-stable Prp17 interaction with the U2, U5 and U6 snRNPs that perhaps occurs with the integral particle. In contrast, another second-step factor, Prp18, has an NaCl-sensitive U5 snRNP association, where U5 ablation eliminates the pulldown of U5, U4 and U6 by anti-Prp18 antibodies . The snRNP associations of Prp17 are, in some respects, similar to those of the affinity-purified Cef1p complex [38,39], of which Prp17 is a component. Furthermore, Prp17 and many components of Cef1 or NTC associate with early spliceosomes. Yet, Prp17 differs in some respects from other factors common to the Cef1-complex and NTC. The Syf1, Syf2 and Isy1 proteins immunoprecipitate U5 and U6 snRNAs, but not U2 snRNA [40,41], whereas Prp19 and Cef1p associate with U2, U5 and U6 snRNPs . However, it is not known whether these interactions arise from contacts with one or more of these snRNPs. Our results suggest that the stable interaction of Prp17 with the U2, U5 and U6 snRNPs is mediated by multiple interactions. The complementation of prp17Δ by yeast–human Prp17 chimaeric proteins both in vivo and in vitro  suggests that human Prp17 may display similar snRNP associations.
The interactions identified in the present study may explain the earlier reports of genetic interactions between prp17 alleles and mutations in the U5 snRNA loop I or mutations in the 5′ end of U2 snRNA that disrupt helix II of the U2–U6 hybrid formed in the spliceosome [7,29,42]. Both of these snRNA elements are important features of catalytic spliceosomes, with the U5 loop I being important for second-step splicing . Although genetic interactions between mutations in U6 snRNA and Prp17 have not emerged, our results from ablation of U6 snRNA hint that the Prp17–U2 snRNP association is influenced by the amount of U6 snRNP. This suggests that U6 snRNP facilitates some of the Prp17–U2 interactions. Thus, as speculated previously , interactions between U2, U5 and factors such as Prp22, Prp17, Prp8 and Slu7 may tether U2, U5 and U6 with the substrate RNA species, thereby assisting in co-ordinating the two steps of splicing.
Among the various factors acting at the second step of splicing, Prp16, Slu7 and Prp18 function exclusively at this step, whereas Prp8 is additionally required for early events in spliceosome assembly and for the first step of splicing. On the other hand, Prp22 performs an additional essential role for the release of spliced mRNA after the second step. Besides the arrested second step, extracts lacking Prp17 are notably slow for kinetics of the first step [7,30]. Thus Prp17 may contribute to improved efficiency of the first and the second steps of splicing. In agreement with this hypothesis, we find an early assembly of Prp17 with pre-catalytic spliceosomes most probably upon the dissociation of U4 snRNP and formation of the A1 complex. In contrast, most other yeast second-step factors associate with the spliceosome after, or concomitant with, the first transesterification reaction. Human Prp17 is detectable in catalytically arrested spliceosomes , but its spliceosomal association prior to catalysis has not been specifically investigated.
The early coalescence of Prp17 with pre-catalytic spliceosomes agrees with the genetic interactions between PRP17 and components of the NTC . The presence of Prp17 in affinity-purified Cef1 complex also supports our biochemical analysis [38,39]. NTC, Cef1-protein complex and most factors common to both of these complexes, i.e. Syf1/Ntc90, Syf2/Ntc31 and Snt309, associate with the spliceosome upon release of U1 and U4 snRNPs during the transition from the A2-1 to the A1 complex. An exception is the Clf1/Syf3/Ntc77 protein, which joins the spliceosome prior to the U4/U6/U5 tri-snRNP and is proposed to aid in the transition from the U1 and U2 snRNP-containing pre-spliceosome to the A2-1 assembly intermediate . Similarly, our results hint at a possible early, but perhaps weak, interaction of Prp17 with A2-1 spliceosomes that may be marginally before the coalescence of the NTC. Since we detect the stable association of Prp17 only after the dissociation of U4 snRNA we speculate that the release of U4 snRNP may stabilize Prp17 interactions.
The relatively efficient immunoprecipitation of the lariat intron by tagged Prp17 leads us to suggest its association with the lariat intron probably when complexed with the U2/U6/U5 snRNPs. Prp17, like some components of the NTC, Syf1/Ntc90, Syf2/Ntc31, Syf3/Ntc77/Clf1 [34,41] and Cef1/Ntc85 , associates with only the lariat intron, in addition to its prior association with pre-mRNA and lariat intermediates. Interestingly, human Prp17 is a component of a 35S U5 snRNP complex, which is part of a larger post-splicing intron-containing complex containing Prp19 [45,46]. The presence of Prp17 in both of these human and yeast complexes corroborates our results indicating Prp17 as a component of the spliced-out intron complex. In summary, the present study maps the spliceosomal snRNP associations of Prp17 that support its role in augmenting essential interactions during the splicing cycle.
This work was supported by an International Senior Research Fellowship from the Wellcome Trust, U.K., to U. V. Partial infrastructure support was provided by ICMR (Indian Council of Medical Research), Government of India. Scholarships to A. K. S. and P. K. were from CSIR (Council for Scientific and Industrial Research), Government of India.
Abbreviations: NTC, nineteen complex; snRNA, small nuclear RNA; snRNP, small nuclear ribonucleoprotein
- © The Authors Journal compilation © 2008 Biochemical Society