Functional domains of the yeast splicing factor Prp22p.

The essential Saccharomyces cerevisiae PRP22 gene encodes a 1145-amino acid DEXH box RNA helicase. Prp22p plays two roles during pre-mRNA splicing as follows: it is required for the second transesterification step and for the release of mature mRNA from the spliceosome. Whereas the step 2 function of Prp22p does not require ATP hydrolysis, spliceosome disassembly is dependent on the ATPase and helicase activities. Here we delineate a minimal functional domain, Prp22(262-1145), that suffices for the activity of Prp22p in vivo when expressed under the natural PRP22 promoter and for pre-mRNA splicing activity in vitro. The biologically active domain lacks an S1 motif (residues 177-256) that had been proposed to play a role in RNA binding by Prp22p. The deletion mutant Prp22(351-1145) can function in vivo when provided at a high gene dosage. We suggest that the segment from residues 262 to 350 enhances Prp22p function in vivo, presumably by targeting Prp22p to the spliceosome. We characterize an even smaller catalytic domain, Prp22(466-1145) that suffices for ATP hydrolysis, RNA binding, and RNA unwinding in vitro and for nuclear localization in vivo but cannot by itself support cell growth. However, the ATPase/helicase domain can function in vivo if the N-terminal region Prp22(1-480) is co-expressed in trans.

The Saccharomyces cerevisiae PRP22 gene encodes an essential 130-kDa protein that plays two roles in pre-mRNA splicing. Prp22p promotes the second transesterification reaction, which entails attack of the 3Ј-OH of the exon on the 3Ј splice site, and it then catalyzes the ATP-dependent release of mature mRNA from the spliceosome (1)(2)(3)(4). Prp22p is a member of the DEXH box family of nucleic acid-dependent NTPases, which is defined by a set of six conserved motifs that are located in the central portion of Prp22p. The central NTPase region is flanked by a 505-amino acid segment N-terminal to motif I (GETGSGKT 513 ) and a 334-amino acid domain C-terminal to motif VI (QRKGRAGR 811 ) (Fig. 1). Three other yeast DEXH box splicing factors, Prp2p, Prp16p, and Prp43p, are organized likewise and share extensive sequence similarity in their central and Cterminal portions (5)(6)(7). Prp22p, Prp2p, and Prp16p are RNAdependent ATPases. Prp22p and Prp16p, but not Prp2p, can utilize the energy of ATP hydrolysis to unwind duplex RNA (2, 3, 8 -10).
ATP hydrolysis by Prp2p, Prp16p, and Prp22p drives sequential steps in the splicing pathway. Prp2p propels the first transesterification step, Prp16p the second transesterification step, and Prp22p the disassembly of the spliceosome (11). Prp43p may also participate in spliceosome disassembly (7).
The segments flanking the NTPase region of DEXH proteins are likely to contribute to their biological specificity, e.g. by facilitating protein-RNA or protein-protein interactions within the spliceosome. Deletion analysis of Prp16p showed that its unique N-terminal segment is important for Prp16p function in vivo and in vitro (12,13). The N terminus of Prp16p appears to play a role in binding Prp16p to the spliceosome (13).
The ATPase and helicase activities are required for the function of Prp22 in pre-mRNA splicing. Alanine substitutions in motifs I-III and VI abolish ATP hydrolysis or uncouple ATP hydrolysis from duplex unwinding. The ATPase-defective Prp22 mutants are lethal in vivo and inactive in spliceosome disassembly in vitro, as are Prp22 mutants that retain ATPase activity but are helicase-defective (2)(3)(4). However, nothing is known about the role of the N-and C-terminal extensions that flank the conserved ATPase/helicase region. Of particular interest is the N-terminal segment, which contains an "S1 motif" spanning amino acids 177-256 (1). The S1 motif, which is encountered in a wide variety of RNA-associated proteins, was originally noted in the ribosomal S1 protein from Escherichia coli (14). It has been suggested that the S1 motif in Prp22p constitutes an RNA binding domain that may be important for the function of Prp22 in pre-mRNA splicing (1).
Here we examine the effects of several N-terminal deletions on Prp22p function in vivo and in vitro. We delineate a minimal functional domain, Prp22(262-1145), that lacks the S1 motif yet suffices for the activity of Prp22p in vivo and for pre-mRNA splicing activity in vitro. We characterize an even smaller catalytic domain, Prp22(466 -1145), that suffices for ATP hydrolysis and RNA unwinding in vitro and for nuclear localization in vivo. We find that the segment 262-350, located upstream of the catalytic domain, enhances Prp22p function in vivo by targeting Prp22p to the spliceosome.

EXPERIMENTAL PROCEDURES
Deletion Mutants of PRP22-N-terminal deletion mutants were generated by PCR 1 amplification, using oligonucleotide primers that either introduced NdeI restriction sites at the codons for Met-202, Met-385, Met-447, and Met-530 or introduced NdeI sites and methionine codons in lieu of the codons for Glu-50, Ser-109, Lys-261, Gln-301, Glu-350, Ile-421, Ser-465, and Asn-499. NdeI-XhoI fragments of the PCR-amplified DNA fragments were inserted into p358-PRP22 (CEN TRP1) in place of the wild-type NdeI-XhoI fragment. In this plasmid the Prp22 deletion mutants ⌬50, ⌬109, ⌬201, ⌬261, ⌬301, ⌬350, ⌬384, ⌬421, ⌬446, ⌬465, and ⌬499 were under the control of the natural PRP22 promoter. The inserts were sequenced completely in order to exclude the acquisition of unwanted mutations during amplification and cloning. The open reading frames encoding wild-type and N-terminal deletion variants * This work was supported by National Institutes of Health Grant GM50288. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
GFP Fluorescence-The coding sequence for green fluorescent protein was amplified by PCR and inserted into pYX132 (TRP1 CEN) so that its expression is driven by the TPI1 promoter. An NdeI site was introduced at the end of the coding sequence to allow insertion of DNA fragments for in-frame fusions to GFP. The resulting plasmid is p132/ GFP (TRP1 CEN). The coding sequences for PRP22 and N-terminal deletion variants were fused to the 3Ј end of the GFP sequence. Wildtype PRP22 cells containing p132/GFP and the various p132/GFP-PRP22 and p132/GFP-PRP22⌬ plasmids were grown in SD(ϪTrp) medium. Aliquots of logarithmically growing cultures were mixed with 50% glycerol and visualized with a fluorescence microscope (Nikon Eclipse E600). Pictures of the same frame were taken with a SPOT camera with Nomarski optics and under fluorescent light.
Dominant-negative Effects of ATPase-defective D603A Mutation-The PRP22⌬ genes were inserted into pYX133 (CEN TRP1), where their expression is under the transcriptional control of the GAL1 promoter (4). A restriction fragment containing the D603A mutation was inserted into these constructs. The p133Prp22⌬ and p133Prp22⌬D603A plasmids were introduced into a wild-type PRP22 strain and transformants were selected on glucose-containing SC(ϪTrp) medium. Trpϩ transformants were streaked onto SC(ϪTrp) agar medium containing glucose or galactose as the carbon source. The plates were incubated at 30, 25, 19, and 15°C.
Expression and Purification of Recombinant Prp22 Protein-Plasmid pET16b-PRP22 expresses an N-terminal His-tagged version of wildtype Prp22p in bacteria under the control of a T7 promoter (2). Here we constructed pET-based plasmids for expression of His-tagged Prp22p N-terminal truncation mutants ⌬261, ⌬301, ⌬350, ⌬384, ⌬421, ⌬446, ⌬465, ⌬499, ⌬529, and Prp22(1-480). The expression plasmids were transformed into E. coli strain BL21-Codon Plus(DE3)RIL (Stratagene). Cultures were inoculated from single colonies of freshly transformed cells and maintained in logarithmic growth at 37°C in LB medium containing 0.1 mg/ml ampicillin to a final volume of 1 liter. When the A 600 reached 0.6 to 0.8, the cultures were chilled on ice for 30 min and then adjusted to 0.4 mM isopropyl ␤-D-thiogalactopyranoside and 2% ethanol. The cultures were incubated for 16 h at 17°C with constant shaking. Cells were harvested by centrifugation, and the pellets were stored at Ϫ80°C.
All subsequent operations were performed at 4°C. The cell pellets were suspended in 100 ml of buffer A (50 mM Tris-HCl, pH 7.4, 250 mM NaCl, 10% sucrose). Lysozyme was added to 0.2 mg/ml, and the suspensions were mixed gently for 30 min and then adjusted to 0.1% Triton X-100. The lysates were sonicated to reduce viscosity, and insoluble material was removed by centrifugation for 30 min at 14,000 rpm in a Sorvall SS34 rotor. The solubility of the N-terminal deletion proteins ⌬261, ⌬301, ⌬350, ⌬384, ⌬421, ⌬446, and ⌬465 was similar to wild-type Prp22p.
The soluble lysates were mixed for 1 h with 10 ml of a 50% slurry of Ni 2ϩ -nitrilotriacetic acid-agarose (Qiagen) that had been equilibrated in buffer A. The resin was recovered by centrifugation, resuspended in 40 ml of buffer A, and collected again by centrifugation. The washed resin was suspended in 40 ml of buffer A and poured into a column. Adsorbed proteins were eluted stepwise with 25, 100, and 500 mM imidazole in buffer E (50 mM Tris-HCl, pH 7.4, 250 mM NaCl, 10% glycerol). The elution profiles of recombinant Prp22p were monitored by SDS-PAGE of the column fractions. Wild-type Prp22p and the truncated polypeptides were recovered predominantly in the 100 mM imidazole eluate (comprising 2-10 mg of protein). Peak fractions were pooled, and aliquots (2-5 mg) of the nickel-agarose Prp22p preparations were diluted 1:5 with buffer D (50 mM Tris, pH 7.4, 2 mM DTT, 1 mM EDTA, 10% glycerol) and then mixed for 1 h with 1 ml of phosphocellulose resin that had been equilibrated with buffer D containing 50 mM NaCl. The resin was recovered by centrifugation, washed twice with 5 ml of the same buffer, and then poured into a column, which was eluted stepwise with buffer D containing 100, 200, 300 and 500 mM NaCl.
Wild-type Prp22p and the Prp22⌬ mutants were recovered predominantly in the 300 mM NaCl eluate. Aliquots (160 -200 g) of the phosphocellulose protein preparations were applied to 4.8 ml of 15-30% glycerol gradients containing 250 mM NaCl, 50 mM Tris-HCl, pH 8.0, 2 mM DTT, 1 mM EDTA, 0.1% Triton X-100. The gradients were centrifuged for 18 h at 47,000 rpm in a Sorvall SW50 rotor. Fractions (0.18 ml) were collected from the tops of the tubes. The elution profiles of the Prp22 variants were gauged by SDS-PAGE. Protein concentrations were determined by using the Bio-Rad dye-binding reagent with bovine serum albumin as the standard.
Pre-mRNA Splicing in Vitro-Yeast whole cell extract from strain BJ2168 was prepared using the liquid nitrogen method (15). The extract was immunodepleted using Prp22p affinity-purified polyclonal antibodies (4). Splicing reaction mixtures (10 l) contained 50% Prp22p-depleted extract, 10 5 cpm (ϳ2 fmol) of 32 P-GMP-labeled actin precursor RNA, 60 mM potassium phosphate, 2.5 mM MgCl 2 , and 2 mM ATP. The reaction mixtures were incubated for 10 min at 23°C, and then 4 -5 ng of wild-type or mutant Prp22p was added and incubation continued for 15 min at 23 or 30°C. The reactions were halted by transfer to ice. For the analyses of spliceosome disassembly, the reaction volumes were increased to 100 l, and the amounts of proteins added were 80 -100 ng. An aliquot (5 l) was removed from each mixture and saved for analysis (Input). The remaining aliquots (95 l) were layered onto 15-40% glycerol gradients containing 20 mM HEPES, pH 6.5, 100 mM KCl, 2 mM MgCl 2 , and then centrifuged at 4°C for 12 h at 35,000 rpm in a Sorvall TH641 rotor. Fractions (400 l) were collected from the tops of the tubes. RNA was recovered by phenol extraction and ethanol precipitation. RNAs from alternate fractions were analyzed by electrophoresis through a 6% polyacrylamide (19:1) gel containing 7 M urea in TBE (89 mM Tris borate, 2 mM EDTA). Radiolabeled RNA was visualized by autoradiographic exposure of the dried gel and quantitated by scanning the gel using a STORM PhosphorImager.
ATPase Assay-Reaction mixtures (20 l) containing 40 mM Tris-HCl, pH 8.0, 2 mM DTT, 2 mM MgCl 2 , 1 mM [␥-32 P]ATP, 0.5 g of poly(A), and Prp22 proteins as specified were incubated for 30 min at 30°C. The reactions were stopped by the addition of 280 l of a 5% (w/v) suspension of activated charcoal (Sigma) in 20 mM phosphoric acid. The samples were incubated on ice for 10 min, and the charcoal was recovered by centrifugation. 32 P radioactivity in the supernatant was quantitated by liquid scintillation counting. The results are average values from duplicate reaction mixes with a deviation of less than 5%.
RNA Binding Assay-Reaction mixtures (20 l) contained 40 mM Tris-HCl, pH 8.0, 2 mM DTT, 25 fmol of RNA substrate, and Prp22 mutant proteins as specified. Mixtures were incubated at 30°C for 30 min and transferred to ice, and glycerol was added to a final concentration of 8%. The reaction products (free substrate and protein-RNA complex) were resolved by electrophoresis through an 8% polyacrylamide (30:0.8) gel containing 50 mM Tris borate, 1 mM EDTA and visualized by autoradiographic exposure of the dried gel.

RESULTS
Mutational Analysis of the S1 Motif-Prp22p contains near its N terminus an S1 motif (aa 177-256), which is found in many RNA-associated proteins. The S1 motif is conserved in the N termini of Prp22p homologues from S. pombe, Caenorhabditis elegans, and human (19 -21). The solution structure of the S1 RNA binding domain from E. coli polynucleotide phosphorylase has been determined (22), and residues that are important for folding and implicated in RNA binding have been identified (Fig. 1).
In order to assess the importance of these conserved residues for the biological function of Prp22p, we substituted Gly-183, Arg-186, Phe-191, Phe-194, His-210, Gly-229, Gln-240, and Lys-244 ( Fig. 1) by alanine. The mutant PRP22 ϪAla alleles were cloned into CEN TRP1 plasmids, and their function was tested in vivo in a prp22⌬ strain using the plasmid shuffle procedure. Growth of prp22⌬ is dependent on a CEN URA3 PRP22 plasmid (4). Trpϩ transformants were selected and then tested for growth on medium containing 5-FOA, a drug that selects against URA3. Wild-type PRP22 cells and the mutant strains G183A, R186A, F191A, F196A, H210A, H210A/G229A, and K244A grew under counterselective conditions. The 5-FOA survivors all formed colonies on YPD medium at 15, 25, 30, and 37°C (data not shown), indicating that individual residues within the S1 motif of Prp22p are not important for growth. This finding prompted us to determine the minimal domain that was essential for Prp22p function.
Prp22 Deletion Mutants Define a Minimum Essential Domain-A series of N-terminal deletion mutants was designed to progressively truncate the 1145 amino acid Prp22 protein. The in vivo function of the truncated alleles, under the transcriptional control of the natural PRP22 promoter, was tested using the plasmid shuffle procedure. Deletion of 50, 109, 201, and 261 amino acid residues from the N terminus did not affect the biological activity of Prp22p as the deletion mutants all formed colonies on 5-FOA medium (Fig. 1), and they grew as well as wild-type cells on YPD medium at 14, 25, 30, 34, and 37°C (not shown). Thus, elimination of the S1 motif had no effect on Prp22p function in vivo.
Deletion alleles prp22⌬301 and prp22⌬350, which were lethal when expressed from the natural PRP22 promoter on a CEN plasmid, could complement the prp22⌬ strain when provided on a 2-m plasmid under the transcriptional control of the strong TPI1 promoter. The more extensively truncated mutants prp22⌬384, prp22⌬421, prp22⌬446, and prp22⌬465 failed to sustain growth of prp22⌬ on 5-FOA even when the mutants were expressed at high gene dosage (not shown).
Determinants of Nuclear Localization of Prp22p-The green fluorescent protein (GFP) was fused to the N terminus of wildtype Prp22p and truncated Prp22 variants. The GFP-PRP22 alleles were placed on CEN TRP1 plasmids under the control of the TPI1 promoter and then tested for complementation of the prp22⌬ strain. The fusion of GFP to full-length Prp22p was functional in vivo as was GFP-Prp22⌬350. In contrast, GFP-Prp22⌬384, GFP-Prp22⌬421, GFP-Prp22⌬446, and GFP-Prp22⌬465 were unable to sustain growth of prp22⌬ cells (data not shown).
Fluorescence microscopy showed that GFP itself was distributed throughout the cell (Fig. 2). However, when GFP was fused to wild-type Prp22p, the fluorescence was concentrated in the nucleus. Truncated variants GFP-Prp22⌬350 and GFP-Prp22⌬446 were also localized to the nucleus (Fig. 2) as was GFP-Prp22⌬465 (not shown). The GFP fluorescence coincided with 4,6-diamidino-2-phenylindole staining of DNA (not shown). These findings show that (i) a nuclear localization (and/or retention) signal resides in Prp22(466 -1145) and (ii) the lethality of the truncated alleles ⌬384, ⌬421, ⌬446, and The S1 motif spans residues 177-256, and a segment of this region in Prp22 is aligned with S1 motifs from polynucleotide phosphorylase (EcoPNP) and ribosomal S1 protein from E. coli (EcoS1). The asterisks above the Prp22 sequence indicate residues that were substituted by alanines. YBST1 was transformed with TRP1 plasmids containing the indicated alleles encoding full-length (WT) and N-terminal truncation mutants (the designations refer to the number of residues that are deleted from the N terminus). The PRP22 alleles were either expressed from the natural PRP22 promoter on CEN plasmids or from the TPI1 promoter on 2-m plasmids. Trpϩ transformants were selected and then streaked to medium containing 5-FOA. Growth was scored as ϩ and failure to form colonies after 10 days is indicated by Ϫ.
⌬465 is not attributable to the failure of nuclear localization of the truncated polypeptides but most likely to a functional defect of these proteins.
Dominant-negative Growth Inhibition by ATPase-defective Mutants-Previous work showed that ATPase-defective Prp22 mutants inhibit growth of wild-type cells when they are overexpressed (4). The dominant-negative effects arise because ATPasedefective mutant proteins bind to spliceosomes and compete with wild-type Prp22p but then fail to release mature RNA. 2 We reasoned that introduction of the ATPase-inactivating D603A mutation (in motif II, the DEAH box) into the ⌬261 and ⌬350 alleles might uncover effects of the N-terminal region on spliceosome binding. The PRP22 genes were cloned into CEN TRP1 plasmids under the transcriptional control of a GAL1 promoter and then introduced into wild-type PRP22 cells. Trpϩ transformants were selected and streaked on agar medium containing either glucose or galactose (Fig. 3). All of the strains grew on glucose-containing medium, when expression of the plasmidencoded Prp22 protein was repressed. PRP22, ⌬261, and ⌬350 grew on galactose. However, galactose-induced expression of the Prp22-D603A protein inhibited cell growth (Fig. 3). A strain carrying the ⌬261-D603A mutant formed pinpoint colonies on galactose-containing medium at 30°C (Fig. 3) and failed to grow at lower temperatures (25,19, and 15°C, not shown). Overexpression of the ⌬350-D603A allele caused a modest inhibition of growth (Fig. 3). The levels of ⌬261 and ⌬350 proteins at 30°C were comparable to the level of wild-type Prp22p in galactosecontaining medium as determined by Western blotting using polyclonal Prp22-specific antibodies (data not shown). We infer from these findings that ⌬261-D603A was capable of competing with wild-type Prp22p. However, ⌬350-D603A was less effective, suggesting a role for residues 262-350 in the interaction of Prp22p with the spliceosome. This finding agrees with the observed requirement for increased expression of ⌬350 for cell growth (Fig. 1).
Effect of N-terminal Deletions on Prp22 Function in Pre-mRNA Splicing-We assayed the deleted Prp22 proteins for their ability to splice actin pre-mRNA in vitro. Prp22p and the truncated variants were expressed in bacteria as Histagged fusions (2,4). The proteins were purified from soluble lysates by Ni 2ϩ -nitrilotriacetic acid-agarose and phosphocellulose chromatography and glycerol gradient sedimentation.
SDS-PAGE gel analysis of the peak glycerol gradient fractions showed that the protein preparations were substantially pure (Fig. 4A).
We tested whether the truncated Prp22 proteins were capable of complementing the step 2 defect in extracts depleted of Prp22p (Fig. 4B). 32 P-Labeled actin pre-mRNA was incubated in depleted extract to allow for spliceosome assembly and step 1 transesterification. Aliquots of the reaction mixture were then supplemented with buffer, wild-type Prp22p, or the mutant proteins ⌬261, ⌬301, ⌬350, ⌬384, ⌬421, or ⌬446. After 15 min at 30°C, the RNAs were extracted, and the products analyzed by denaturing PAGE. In the absence of added protein, very little mature RNA was formed and lariat-exon 2 persisted. Wild-type Prp22p and the ⌬261 protein both promoted step 2 efficiently. The more extensively truncated Prp22 mutants ⌬301, ⌬350, ⌬384, ⌬421, ⌬446, and ⌬465 were inactive for step 2 complementation in vitro, i.e. lariat-exon 2 persisted and little mRNA was produced (Fig. 4B).
Effect of N-terminal Deletion Mutants on Spliceosome Disassembly-Prp22p is also important for the release of mature mRNA from the spliceosome (2). In order to test whether the truncated Prp22 proteins were active for mRNA release, we used a customized precursor RNA (Act 7 ) that does not require Prp22p for step 2 complementation. In the ACT 7 pre-mRNA, the distance between the branchpoint and the 3Ј splice site is reduced, and splicing of this pre-mRNA in vitro is independent of the participation of Prp22p and other splicing factors (2,(23)(24)(25). However, complete release of mature Act 7 mRNA from the spliceosome still does require Prp22p (2). We incubated ACT 7 pre-mRNA substrate in Prp22-depleted extract and supplemented the reactions either with wild-type Prp22p, buffer, or the truncated polypeptides ⌬261, ⌬350, and ⌬465. Mature mRNA was formed in every case (Fig. 5A). Aliquots of the reaction mixtures were sedimented in 15-40% glycerol gradients to assess whether mRNA was released from the spliceosome complex. Fractions were collected, and the RNA species in every odd fraction were analyzed by denaturing PAGE. The distribution of spliced mRNA across each gradient from top to bottom is shown in Fig. 5B. When the reactions were carried out in the presence of wild-type Prp22p or the ⌬261 mutant, released mature mRNA sedimented near the top of the gradient in fractions 7-11. In contrast, when Prp22p was omitted, mRNA sedimented in two peaks, one corresponding to the released mRNA (fractions 7-11) and a second heavier peak (fractions 19 -23) corresponding to the spliceosomes. The same two-peak profile was observed when extracts were supple-2 S. Schneider and B. Schwer, unpublished observations. mented with ⌬350 and ⌬465 (Fig. 5B). We conclude that ⌬261 is active in spliceosome disassembly, whereas ⌬350 and ⌬465 are defective.
ATP Hydrolysis by Truncated Prp22 Polypeptides-Spliceosome disassembly is dependent on ATP hydrolysis by Prp22p.
To determine if loss of disassembly function was attributable to the effects of deletions on ATP hydrolysis, we measured the ATPase activities of the Prp22p deletion variants (Fig. 4). The proteins were incubated with 1 mM ATP in the presence of poly(A) homopolymer as the RNA cofactor for 30 min at 30°C. The extents of ATP hydrolysis increased linearly with the amounts of Prp22 proteins added (Fig. 6A). Titrations of the ⌬261, ⌬350, ⌬44, and ⌬465 polypeptides indicated that the mutants were more active than the wild-type Prp22p (Fig. 6A). The turnover numbers are shown in Fig. 6B. Wild-type Prp22p hydrolyzed 43 and 400 ATP per min in the absence and presence of an RNA cofactor, respectively. The ⌬446 mutant protein hydrolyzed 870 ATP per min in the presence of poly(A) and 128 without RNA cofactor. Further N-terminal deletions of Prp22p to positions 500 or 530 were insoluble when expressed in bacteria. These mutants were thus refractory to purification and biochemical analysis. We conclude that the N-terminal 465 amino acids of Prp22p are not part of the ATPase domain. Indeed, deletion of the N-terminal 465 aa results in increased ATPase activity with and without the poly(A) cofactor.

RNA Binding and Unwinding of Duplex RNA by Prp22deletion Mutants-The
ATPase activity of full-length and truncated Prp22p variants was stimulated by an RNA cofactor. We surmised that the proteins retained their capability to bind to RNA and to unwind duplex RNAs. In order to test this directly, a 3Ј-tailed RNA substrate containing a 25-bp duplex (Fig. 7A) was incubated with the ⌬261 and ⌬446 proteins in the absence of ATP. The mixtures were analyzed by native PAGE (Fig. 7B). In the absence of protein, the labeled RNA molecule migrated as a single species in the native gel. Addition of increasing amounts of ⌬261 and ⌬446 proteins resulted in the appearance of protein-RNA complexes of reduced mobility. We presume that the appearance of two shifted bands at higher protein concentrations reflects the sequential binding of one and two molecules of protein to a single RNA molecule. Note that the ⌬261-RNA complex migrates more slowly than the complexes formed with ⌬446 polypeptide, which we presume reflects the difference in size of the RNA-bound Prp22 proteins. Other truncated polypeptides of intermediate size formed protein-RNA complexes that migrated between the ⌬261-RNA and ⌬465-RNA complexes (data not shown).
The same duplex RNA substrate (Fig. 7A) was used to assess the helicase activity of the truncated Prp22 proteins. Substrate RNA was incubated with the Prp22p variants in the presence of ATP, and the products were analyzed by PAGE after disruption   FIG. 4. N-terminal Prp22p deletion mutants. A, protein gel. Wild-type (WT) and the indicated mutant variants of Prp22p were expressed in bacteria and purified by Ni 2ϩ -nitrilotriacetic acid-agarose and phosphocellulose chromatography, followed by glycerol gradient sedimentation. One of the RNA-protein interactions by 0.1% SDS (Fig. 7C). The unwound single-stranded RNA species migrated faster than the duplex RNA substrate. Full-length Prp22p (wild type) and the truncated mutant ⌬261, ⌬350, ⌬385, ⌬421, ⌬446, and ⌬465 all were active in RNA unwinding. Their helicase activities were dependent on ATP (data not shown). The helicase activities of truncated proteins (adjusted to measure equal molar amounts) were increased compared with wild-type Prp22p. This increase may be due, at least in part, to the higher ATPase activities of the Prp22 deletion mutants. We conclude that the N-terminal 465 amino acids of the Prp22 protein are not involved in ATP hydrolysis, RNA binding, and helicase activities.
Prp22  and Prp22(466 -1145) Can Function in Trans to Support Cell Viability-Neither the ATPase/helicase domain Prp22(466 -1145) nor the N-terminal polypeptide Prp22(1-480) were active in pre-mRNA splicing ( Fig. 4B and not shown). As expected, Prp22(1-480) did not hydrolyze ATP (not shown). Neither Prp22(1-480) nor Prp22(466 -1145) alone could support growth of prp22⌬ cells even when the N-and C-terminal domains were overexpressed at high gene dosage (not shown and Fig. 8). However, the two segments Prp22(1-480) and Prp22(466 -1145) did support growth of the prp22⌬ strain when they were expressed in trans on different plasmid vectors (Fig. 8). The 5-FOA survivors were streaked to YPD medium at temperatures from 14 to 37°C. The strain carrying the prp22  and the prp22(466 -1145) alleles grew at 25, 30, and 34°C, but the cells exhibited cold-sensitive and heat-sensitive growth phenotypes; they failed to grow at 14 and 19°C and formed only pinpoint colonies at 37°C. Wild-type PRP22 cells grew well at all temperatures. We conclude that the Nand C-terminal domains can function in trans to sustain cell growth at temperatures from 25 to 34°C.

Prp22(466 -1145) Constitutes an ATPase/Helicase Domain-
Prior studies have shown that conserved residues in motifs I-III and VI are essential for the function of Prp22p in pre-mRNA splicing. Mutational analyses of DEX(H/D) RNA helicases, including Prp22p, vaccinia virus NPH-II, eIF-4A, and hepatitis C virus NS3 highlight the importance of conserved amino acids in these motifs for NTP hydrolysis and duplex unwinding (2-4, 26 -32). Structural studies of the NS3 helicase show that the catalytic core consists of three globular domains. Motifs I and II are located in domain 1 and motif VI in domain 2. Domains 1 and 2 are connected via a flexible linker segment that includes motif III (33,34). Motif III (T/ SAT) couples ATP hydrolysis to duplex unwinding (4,27,28,30). Kim et al. (34) surmised from their crystal structure of NS3 that (i) ATP bridges domains 1 and 2 and (ii) a conformational change upon ATP hydrolysis leads to opening of the interdomain cleft and translocation of the protein along the polynucleotide.
Whereas the importance of the conserved motifs for ATPase/ helicase function is well studied, the impact of the protein regions flanking motifs I-VI on the biological and enzymatic activities of the DEXH proteins is poorly understood. Prp22p contains an N-terminal segment of 505 amino acids, but only 46 amino acids upstream of Lys-512 in motif I (GETGSGKT) are required for ATPase/helicase activity. The helicase domain of NS3 contains a segment 21 aa N-terminal to the lysine in motif I. In the vaccinia DEXH box protein NPH-I, a DNA-dependent ATPase, the corresponding lysine residue is at position 61 (35). Thus, a functional ATPase domain does not require extensive segments upstream of motif I. In Prp22p, residues 1-465 are not only dispensable for ATP hydrolysis and RNA unwinding, but they appear to repress the enzymatic activities because the deletion mutants are more active than full-length Prp22p.
The C-terminal region of Prp22p is conserved in a subset of DEXH NTPases, including the splicing factors Prp2p, Prp16p, and Prp43p. For example, Prp22p is 40% identical and 62% similar to Prp16p over a 300-aa region downstream of motif VI. The RNA helicases vaccinia NPH-II, Drosophila MLE1, and human RNA helicase A also show similarity to Prp22p beyond motifs I-VI. The structure of NS3 complexed with nucleic acid implicates amino acids in the C-terminal region (domain 3) in RNA binding. The region in Prp22p downstream of motif VI (QRKGRXGR) is 334 amino acids. Preliminary experiments showed that truncating Prp22p by 87 residues, leaving a 247-aa segment downstream of motif VI, abrogated Prp22 function in vivo. 2 Deletion analyses of Prp16p have established that 275 amino acids downstream of motif VI sufficed for cell viability but that 225 did not (12,13). The viral RNA helicases NPH-II and NS3 contain 178 and 159 amino acids, respectively, downstream of motif VI. In the case of the DEXH box protein NPH-I, deletion analysis has established that 65 amino acids downstream of motif VI sufficed for full ATP hydrolysis activity (35). Thus, it appears that a functional ATPase or ATPase/helicase domain requires amino acids beyond motif VI, but the minimum essential length of the C-terminal segments may vary for different DEXH proteins. The C-terminal margin of the ATPase/helicase domain of Prp22(465-1145) remains to be determined.
The Biologically Active Domains of Prp22p-The in vivo and in vitro activities of Prp22(261-1145) are indistinguishable from those of wild-type Prp22p, demonstrating that the Nterminal 260 residues, which include an S1 motif (aa 177-256), are not essential for Prp22p function. However, the minimal ATPase/helicase domain Prp22(466 -1145) does not suffice for the role of Prp22p in pre-mRNA splicing, thereby indicating the importance of residues N-terminal to position 465 in Prp22p. However, Prp22(466 -1145) can function in vivo if Prp22(1-480) is co-expressed, demonstrating that the protein domains can act in trans. We suggest that both domains are required for the productive interaction of Prp22p with the splicing apparatus.
The inference that the N-terminal region in Prp22p from aa 262-465 is involved in spliceosome binding is based on the findings that deletion of 350 and 465 amino acids in the ATPasedeficient Prp22-D603A reduced and abolished, respectively, its effectiveness as a dominant-negative growth inhibitor, whereas ⌬261-D603A was a potent inhibitor. Although the segment from aa 262 to 465 in Prp22p appears to be necessary, it is not sufficient to afford spliceosome binding because Prp22(1-480) did not compete effectively with wild-type Prp22p. Overexpression of Prp22(1-480) in wild-type PRP22 cells did not lead to growth inhibition, and preincubation of spliceosomes with Prp22(1-480) protein did not prevent splicing by subsequently added Prp22p in vitro (not shown). Thus, Prp22p requires both the ATPase/ helicase and a region within the N-terminal domain for spliceosome association. How the two domains cooperate to provide biological activity in trans remains to be investigated. It is possible that they interact directly, or indirectly via another splicing component, to provide a functional Prp22 protein. Alternatively, the N-terminal domain and the ATPase/helicase domain may bind independently to the splicing apparatus and effect splicing.
The DEAH box splicing factor Prp16p also requires its N terminus for spliceosome binding (13). Prp16(1-300)p alone is capable of binding to the spliceosome; however, the binding is stabilized by the C-terminal region (13). As is the case for Prp22p, the two domains of Prp16p do not need to be physically linked, and the N-terminal 300-aa segment functions in trans with Prp16(301-1071) to sustain growth of a prp16⌬ strain (13).
The DEAH box ATPases, Prp2p, Prp16p, and Prp22p act sequentially; each of the proteins associates with the spliceosome at a distinct stage during the splicing pathway and dissociates upon ATP hydrolysis (2,9,36). It will be interesting to determine whether each of these proteins makes similar contacts within the splicing apparatus and to analyze the nature of these contacts at the molecular level.