The Fission Yeast Pre-mRNA-processing Factor 18 (prp18+) Has Intron-specific Splicing Functions with Links to G1-S Cell Cycle Progression*

The fission yeast genome, which contains numerous short introns, is an apt model for studies on fungal splicing mechanisms and splicing by intron definition. Here we perform a domain analysis of the evolutionarily conserved Schizosaccharomyces pombe pre-mRNA-processing factor, SpPrp18. Our mutational and biophysical analyses of the C-terminal α-helical bundle reveal critical roles for the conserved region as well as helix five. We generate a novel conditional missense mutant, spprp18-5. To assess the role of SpPrp18, we performed global splicing analyses on cells depleted of prp18+ and the conditional spprp18-5 mutant, which show widespread but intron-specific defects. In the absence of functional SpPrp18, primer extension analyses on a tfIId+ intron 1-containing minitranscript show accumulated pre-mRNA, whereas the lariat intron-exon 2 splicing intermediate was undetectable. These phenotypes also occurred in cells lacking both SpPrp18 and SpDbr1 (lariat debranching enzyme), a genetic background suitable for detection of lariat RNAs. These data indicate a major precatalytic splicing arrest that is corroborated by the genetic interaction between spprp18-5 and spprp2-1, a mutant in the early acting U2AF59 protein. Interestingly, SpPrp18 depletion caused cell cycle arrest before S phase. The compromised splicing of transcripts coding for G1-S regulators, such as Res2, a transcription factor, and Skp1, a regulated proteolysis factor, are shown. The cumulative effects of SpPrp18-dependent intron splicing partly explain the G1 arrest upon the loss of SpPrp18. Our study using conditional depletion of spprp18+ and the spprp18-5 mutant uncovers an intron-specific splicing function and early spliceosomal interactions and suggests links with cell cycle progression.


spprp18-5 mutant uncovers an intron-specific splicing function and early spliceosomal interactions and suggests links with cell cycle progression.
Pre-mRNA splicing, a fundamental step in the processing of nascent eukaryotic RNA polymerase II transcripts, achieves precise excision of introns coupled with exon ligation to generate functional mRNAs. The spliceosome, which is composed of five U snRNPs 7 and Ͼ100 auxiliary proteins, assembles onto cis splicing signals, namely the 5Ј splice site (5Јss), the branch point nucleotide, the 3Ј splice site (3Јss), and polypyrimidine (Pyn) tracts. The spliceosomal catalytic core consists of the U2, U5, and U6 snRNPs and accessory proteins. This complex mediates the two trans-esterification reactions required for splicing to occur. First, the 5Јss is cleaved to yield the branched lariat intron-exon 2 and exon 1 intermediates, followed by the second reaction, where the 3Јss is cleaved, the exons are joined, and lariat intron is excised (1,2).
Genetic and biochemical analyses in budding yeast and biochemical studies with mammalian cell extracts have established a network of interactions among factors that act at the second step of splicing (i.e. Prp8, Prp16, Prp17, Prp18, Slu7, and Prp22) (3)(4)(5)(6)(7)(8). PRP18, a non-essential budding yeast gene, encodes a U5 snRNP-associated factor. Prp18 has been analyzed extensively in budding yeast and human cell extracts for splicing of a specific model intron containing pre-mRNAs, but the Schizosaccharomyces pombe prp18 ϩ has not been studied in detail. Such studies have the potential to allow mechanistic insights into a genetic model system that better recapitulates splicing of short introns prevalent in several higher eukaryotes. The PRP18 gene is not essential in Saccharomyces cerevisiae, and prp18⌬ cells are temperature-sensitive (ts) and arrested for the second step of splicing both in vivo and in vitro (10,11). Further analysis of S. cerevisiae Prp18 revealed that an N-terminally truncated Prp18 (ScPrp18⌬79) protein, lacking 79 residues, was functional to mediate splicing in vitro. The crystal structure of this truncated protein revealed a five-helix bundle fold with a highly conserved region (CR) loop between helices 4 and 5 (9). Extensive mutational analyses of ScPrp18 identified critical functions for helices 1 and 2 in mediating interactions with the splicing factor ScSlu7. This interaction is important for ScPrp18 spliceosomal association, whereas other regions in its globular domain stabilize U5 snRNA-exonic interactions after the first catalytic reaction (7,9,(12)(13)(14)(15).
In vitro splicing of ␤-globin pre-mRNA in HeLa cell free extracts suggests that the human orthologue of Prp18, hPrp18, functions in the second step of splicing (4) but how widely or even strictly this function is conserved in other short intronrich higher eukaryotic and fungal genomes is not yet clear. When hPrp18 is immunodepleted from HeLa cell extracts, a second step in vitro splicing arrest occurs. However, hPrp18, when expressed in budding yeast scprp18⌬ cells, does not rescue their strong ts phenotype, thus suggesting some distinctions in the spliceosomal associations of these orthologs (4).
Here we exploit the fission yeast system to further define the conserved functions of Prp18 and provide insight into how events in splicing can be coordinated. Salient features of the fission yeast genome include occurrence of multiple short introns per transcript, degenerate cis splicing signals, and unusually located Pyn tracts. These pre-mRNA features, found in many fungal genomes, make it suitable for studies on correlations between cis features and splicing factor requirements (16,17). The splicing of short introns in fission yeast is also a model for other higher eukaryotes like plants, flies, and worms, where an intron definition model for splice site recognition is proposed (18). Interestingly, earlier studies on the role of intronic 3Јss and the Pyn tract sequences for splicing of two fission yeast introns showed that both of these cis elements are required before the first splicing reaction (19). Similar effects are seen for 3Јss mutations in a subset of mammalian introns (20). Genetic studies on predicted second step factor homologs in fission yeast are limited to only SpPrp17 and SpSlu7, yet they reveal certain differences when compared with budding yeast counterparts. Deletion of spprp17 ϩ , a non-essential gene, did not affect the splicing of introns in the tfIId ϩ gene, including the 255-nt intron 1 (21). Genome-wide splicing analyses of transcripts in a missense mutant of spslu7 ϩ revealed its requirement before the first step of the splicing reaction, its widespread but intron-specific functions, and its novel interaction with another precatalytic splicing factor (22). The dependence on fission yeast factors SpPrp2 (23) and SpPrp4 (24) correlates with specific intronic features in each case. Other studies with cwf10-⌬NTE mutant of the SpCwf10 splicing factor revealed that its role in splicing is general and not transcript-specific (25). These studies lend support to the hypothesis of co-evolution of splicing factor functions with changes in gene and intron architecture. These findings warrant investigations on functions for other fission yeast splicing factors. Such studies could uncover mechanisms for splice site selection in the context of short introns.
Here, we investigated the splicing functions of the predicted fission yeast second step factor SpPrp18 through structuredriven mutational and genetic approaches. Our results reveal vital roles for the SpPrp18 conserved domain and flanking helices. Genome-wide splicing studies and genetic interaction analyses using a missense mutant show that widespread SpPrp18 functions are in precatalytic spliceosomes, and its essential functions for early steps in splicing are intron-specific.
Links between splicing and cell cycle progression have been well established by genetic and protein interaction analyses in budding yeast, fission yeast, and mammalian cells (26 -31). In S. pombe, ts mutants in several splicing factors (26 -28, 32) arrest at restrictive temperatures as elongated cells. Many of these mutants (e.g. cdc28-P8 and prp12-1) affect the G 2 -M transition (27,31), whereas the loss of Prp4 kinase activity derails both the G 1 -S and G 2 -M cell cycle transitions (32). But how the splicing functions of these factors and cell cycle progression are related is as yet not entirely clear. Here, using a missense mutant in prp18 ϩ , we uncover an important role in promoting G 1 -S cell cycle transition through intron-specific splicing effects of transcripts encoding some key regulators of this transition.

Comparative Modeling and Mutational Analysis of SpPrp18 -
The essential S. pombe gene SPCC126.14 encodes SpPrp18, a predicted U5 snRNP-associated protein (33). Prp18 proteins from budding yeast, fission yeast, and humans share a high degree of similarity in their C-terminal halves, which adopt the five-helical bundle with a CR loop between helices 4 and 5 (Fig.  1A). S. pombe SpPrp18 has 35% identity and 58% similarity with ScPrp18 and shows a similar degree of relatedness to hPrp18. A comparison of the domain architecture across these three species revealed the splicing factor motif in the N-terminal region of SpPrp18 and hPrp18 that is absent in ScPrp18 (Fig. 1A). We generated a homology model for amino acids 160 -343 of SpPrp18 (Fig. 1B), which shows a five-helix bundle nearly identical to the homologous domain in ScPrp18 (9).
Extensive mutational analysis of budding yeast ScPrp18 (7,13,15) showed that helices 4, 5, and the intervening conserved region have splicing functions separable from those performed by the ScPrp18-ScSlu7 complex (13). For example, a triple mutant G196A/V197A/T198A in the CR region or a double mutant K234A/R235E in helix 5 conferred ts phenotypes in budding yeast (13) (Table 1). For functional studies of fission yeast SpPrp18, we tested whether substitutions in homologous residues create conditional alleles (Fig. 1B, left). An allele with triple alanine replacements in the SpPrp18 CR domain G288A/ V289A/T290A and another with two substitutions in helix 5 (K325A/R326E) were expressed from the nmt42 promoter on a plasmid that also allows the addition of an N-terminal HA epitope. As assessed by a plasmid shuffle assay, both of these mutants are non-functional and cannot support the growth of the spprp18::his3 ϩ null allele (Tables 1 and 2).
To assess the steady-state expression levels of these mutant proteins, the plasmid-expressed wild-type or mutant proteins were analyzed in wild-type cells. The SpPrp18-G288A/V289A/ T290A protein was not detected and thus probably causes aberrant folding and protein turnover (Fig. 1C, lanes 2 and 3, and Table 2). The SpPrp18-K325A/R326E protein, on the other hand, was detected at levels comparable with wild-type SpPrp18 (Fig. 1C, lane 4), yet this protein was unable to complement the spprp18::his3 ϩ null allele (Table 2). Hence, muta-tional analyses confirmed that the helix 5 and CR regions are crucial for spprp18 ϩ cellular functions.
As an alternative approach to obtain conditional mutants in spprp18 ϩ , we examined substitutions in a series of residues buried in the ␣-helical bundle because they could destabilize the protein in a temperature-dependent manner (34,35). Helices 1 and 2 form one face of this helical bundle, with helix 4 and 5 forming another face, and helix 3 bridges the two faces (Fig.  1B, right). As shown in Fig. 2, we obtained a missense partial loss-of-function allele, spprp18-5, by mutating the valine 194 residue in helix 1 to arginine (V194R).
A Missense Mutation in SpPrp18 Helix 1 Alters Its Conformation and Confers Slow Growth-Cells with a plasmid-expressed missense mutant, SpPrp18V194R (nmt81 promoter-driven), were slow growing as compared with a control strain with plasmid-expressed wild-type spprp18 ϩ (data not shown). For stable expression of the WT or prp18-5 alleles, we integrated the expression cassettes at the leu1 chromosomal locus and thus created a set of strains, spprp18⌬ leu1::Pnmt:prp18 ϩ (this strain is referred to as WT; Fig. 2A) and spprp18⌬ leu1::Pnmt:prp18V194R (referred to as prp18-5; Fig. 2A). In these strains, we achieved conditional expression of the wildtype and mutant spprp18 alleles from the weak nmt81 promoter derived from the nmt1 locus whose transcription is repressed on supplementation of thiamine (36). WT cells grew robustly on media lacking thiamine ( Fig. 2A, left, first row) and, as expected, grow poorly upon the addition of thiamine ( Fig. 2A, left, second row). However, the growth of the prp18-5 strain was slow even in medium lacking thiamine at the ambient temperature. Furthermore, this mutant was inviable when expression was repressed by the addition of thiamine ( Fig. 2A, right, top two rows). Thus, wild-type protein expressed from the leu1 locus fully complemented the spprp18⌬ null allele, but the SpPrp18-5 protein only partially supported growth. As expected, the severe growth arrest of WT and prp18-5 cells on thiamine supplementation was rescued when cells were transformed with a plasmid where spprp18 ϩ is driven by endogenous promoter elements ( Fig. 2A, bottom two rows).
Immunoblotting was performed to assess the levels of SpPrp18 wild-type and mutant protein expressed in these strains (WT and prp18-5). These strains were grown in the presence and absence of thiamine. Protein levels were compared with the endogenous level of SpPrp18 generated from its native locus (wild type). We detect approximately equivalent protein levels in these strains (Fig. 2B, lanes 1, 3, and 5). As expected, the growth retardation of WT and arrest of prp18-5 cells upon the addition of thiamine can be attributed to severely depleted SpPrp18 protein levels (Fig. 2B, lanes 2 and 4). Interestingly, immunoblotting analyses of wild-type SpPrp18 protein reveals an additional slower migrating species with ϳ3-4-kDa increased size (Fig. 2B, asterisk). This species is also drastically depleted upon supplementation of thiamine to repress expression of the leu1 locus-integrated wild-type allele. We ruled out the possibility that the slower migrating band represents a nonspecific cross-reacting protein because epitope-tagged wild-type SpPrp18, expressed from a plasmid, also generates two protein species of approximately the same molecular weights (Fig. 2C). This led us to speculate that a posttranslational modification or altered conformational form of wild-type SpPrp18 could exist. Interestingly, this slow migrating species is absent in prp18-5 cells.
To explore the potential structural consequences of substitution of the non-polar (Val) to a charged (Arg) residue for the secondary and tertiary structure of SpPrp18, we examined the biophysical attributes of the bacterially expressed wild-type and mutant proteins. Far UV CD spectra showed negative ellipticity with sharp bands at 208 and 222 nm, indicative of ␣-helical content of the protein. These are discernable for wild-type SpPrp18 over a range of temperatures from 20 to 90°C (Fig. 2D, top left). In contrast, the SpPrp18-5 (V194R) protein showed significant change in secondary structure starting at temperatures Ͼ40°C (Fig. 2D, top right), indicating thermal instability probably due to protein unfolding. The native SpPrp18 protein has three tryptophan residues; Trp-32 (N-terminal), Trp-205 (helix 1), and Trp-285 (CR loop). We exploited these tryptophans to measure the steady-state intrinsic fluorescence of wild-type SpPrp18 and found that the emission maximum was at ϳ350 nm ( Fig. 2E, bottom left). A red shift toward 365 nm was observed for the SpPrp18V194R protein (Fig. 2E, bottom right), which also suggested unfolding. We also observed that increased temperatures had little or no effect on the intrinsic fluorescence intensity for wild-type SpPrp18. In contrast, the SpPrp18V194R protein showed a gradual quenching of fluorescence intensity with increasing temperatures, indicating a loss of tertiary structure (Fig. 2E, bottom right). Thus, both genetic and biophysical data indicate that the SpPrp18V194R is a partially functional protein that is likely to be unstable in vivo.
Global Splicing Profiling of SpPrp18 Reveals Varying Splicing Defects-To probe the cellular functions of spprp18 ϩ , we used RNA from WT and prp18-5 cells, each grown with and without thiamine supplementation (36 h), to analyze the splicing status of a selection of cellular transcripts. Introns studied had diverse features ( Fig. 3 and Table 3). We first analyzed an abundant cellular transcript, tfIId ϩ . Whereas intron 1 and intron 2 in tfIId ϩ are efficiently spliced in wild-type cells, a ϳ3-fold increase in unspliced pre-mRNA was detected when spprp18 ϩ was transcriptionally repressed ( As observed for the introns in tfIId ϩ , the two introns in the ade2 ϩ transcript have differing splicing efficiencies even in wild-type cells. Repression of the WT or the prp18-5 allele caused a ϳ2.5-fold decrease in levels of spliced mRNA across the short ade2 ϩ intron 1 (Fig. 3B, panel M). In contrast, splicing of ade2 ϩ intron 2 in both strains was unaffected (data not shown). These data suggest essential but not ubiquitous splicing functions for spprp18 ϩ and also provide evidence that the mutant Prp18-5 protein is only partially functional.
To assess the genome-wide splicing role of fission yeast SpPrp18, we utilized splicing-sensitive microarrays as used to uncover substrate-specific functions for the S. pombe splicing factor SpSlu7 (22). RNA from cultures of WT and prp18-5 cells, each grown in the presence or absence of 15 M thiamine, was used for these analyses. These experiments are designed to reveal the effects of depleting either wild-type or mutant SpPrp18 and also to assess the function of the SpPrp18-5 protein. A very stringent data set of 253 introns with statistically significant and biologically correlated -fold changes for various probes across several transcripts upon depletion of wild-type SpPrp18 was selected for this analysis. In the absence of wildtype SpPrp18, this set of 253 introns showed a range of splicing defects (Fig. 4, A-E, WTϪT and WTϩT panels). Interestingly, similar splicing phenotypes, although in some instances with differing intensities, were observed in cells that express the

TABLE 2 Complementation profile of SpPrp18 C-and N-terminal mutants
Independent transformants for each of the plasmid in diploids strains heterozygous for null alleles of spprp18ϩ were sporulated. leu ϩ or ura ϩ plasmid-bearing spores were selected and assayed for growth on EMM ade Ϫ and EMM his Ϫ by replica plating at 25°C.

No. of diploids analyzed
No. of spores selected on EMM leu ؊ /25°C No. of spores selected on EMM ura ؊ /25°C No. of "leak-through" diploids growing on EMM ade ؊ /25°C No. of leu ؉ spores growing on EMM his ؊ /25°C mutant SpPrp18-5 protein ( Fig. 4; prp18-5ϪT, probes P and M). Overall, splicing defects could be categorized as introns that accumulate as unspliced pre-mRNA ( Fig. 4B), those that accumulate unspliced pre-mRNA with a corresponding decrease in spliced mRNA levels (Fig. 4D), and those with reduced spliced mRNA levels together with lowered gene expression (Fig. 4E). A small fraction of introns were spliced efficiently even upon metabolic depletion of SpPrp18 (Fig. 4C). The readout from the intron probe (P) for accumulation of unspliced pre-mRNA was further validated by the signals from intron-exon junction probes (IE) (data not shown). These data confirm that the splicing functions for SpPrp18 are substratespecific, and the variant protein is only partly functional for genome-wide splicing, as was suggested by the results presented above on splicing of introns in the tfIId ϩ and ade2 ϩ transcripts.
Semiquantitative RT-PCR Analyses Validate Genome-wide Spectrum of Splicing Defects-The splicing phenotypes inferred from microarrays were validated by semiquantitative RT-PCR analyses of some representative introns. The class with heightened splicing defects manifested as both pre-mRNA accumulation and mRNA reduction was represented by mdm35 ϩ intron 1. In agreement with the microarray data, depletion of the wild-type protein caused a ϳ3-fold increase in the pre-mRNA/mRNA index, a combined effect of accumulated pre-mRNA and reduced spliced mRNA levels. The same phenotype was seen in prp18-5 cells (Fig. 5A, lanes 2-4). To validate the predominant class of splicing defects (i.e. introns with pre-mRNA accumulation alone), we selected sfc9 ϩ intron 2 and spf38 ϩ intron 5 for analysis. Intron 5 in spf38 ϩ was strongly dependent on SpPrp18, as assessed by these semiquantitative RT-PCR analyses (Fig. 5B, lanes 2-4). Similarly, sfc9 ϩ intron 2 splicing was SpPrp18-dependent because increased unspliced pre-mRNA was detected upon metabolic depletion of either wild-type or mutant protein (Fig. 5B, lanes 2-4). Finally, we confirmed that splicing of ubc4 ϩ intron 1 and cwf2 ϩ intron 1 can occur independent of SpPrp18 because depletion of wildtype or SpPrp18V194R mutant protein had no effect on the FIGURE 2. SpPrp18V194R mutant confers reduced growth due to altered thermodynamic stability of the protein. A, diagrammatic representation of WT and prp18-5 strains with integrations at the leu1 ϩ locus of nmt81 promoter-driven spprp18 ϩ or prp18V194R ORFs. 10-fold serial dilutions of these cell cultures grown at 30°C were made and spotted on EMM L-media without (first row) or with (second row) 15 M thiamine. For each strain, a suspension with starting optical density of 0.4 was first spotted, which was followed by subsequent dilutions. The third and fourth rows are serial dilutions of WT and prp18-5 strains that were transformed with a plasmid expressing spprp18 ϩ from native promoter (pDBlet spprp18 ϩ ). All plates were incubated at 30°C for 3-4 days. B, cell lysates from a strain expressing wild-type SpPrp18 protein grown in the absence and presence of thiamine (lanes 1 and 2), lysates from similarly treated prp18-5 cells (lanes 3 and 4), and lysates of wild-type (FY528) (lane 5) were analyzed by Western blotting using SpPrp18 polyclonal sera as described under "Experimental Procedures." *, slow mobility form of SpPrp18. C, immunoblotting analysis on crude whole cell extracts from spprp18:his3 ϩ pREP42HA-spprp18 ϩ cells using monoclonal anti-HA antibodies (top). Two forms of the epitope-tagged wild-type protein are detected. Coomassie-stained gel post-transfer serves as the loading control. D, thermodynamic stability of bacterially expressed and purified wild-type SpPrp18 (left) and mutant SpPrp18 V194R (right) proteins. Far UV circular dichroism spectroscopy of 12 nM protein solutions subjected to temperatures from 20 to 90°C are shown. E, steady-state emission spectra (300 -450 nm) of the intrinsic tryptophan fluorescence of 9.8 M wild-type (left) and mutant (right) protein solutions. Excitation was at 295 nm, and emission spectra were recorded over a range of temperatures.
splicing of these introns (Fig. 5C). Importantly, splicing of these introns was dependent on SpSlu7 (Fig. 5D) because, in the slu7-2 mutant, pre-mRNAs accumulate for both of these introns. These data confirm the substrate-specific roles for SpPrp18 that clearly can differ from the requirement for SpSlu7. Overall, our genome-wide studies indicate an important and general function for SpPrp18 in splicing, although a small proportion of introns can be efficiently spliced independent of SpPrp18.
SpPrp18 Depletion Arrests Splicing before Catalysis-In budding yeast and human cell-free in vitro splicing reactions, Prp18 strongly associates with Slu7 and is required for a second step reaction (4, 10). We investigated roles for fission yeast SpPrp18 before the first catalytic reaction and for second step splicing using tfIId ϩ intron 1 as a substrate because it is a spprp18 ϩ -dependent intron. Primer extension assays were done to detect the cDNAs of distinct lengths that reflect levels of tfIId ϩ intron 1 lariat intermediate (intron-3Ј-exon), the unspliced precursor (E1-I1-E2), and the spliced message (E1-E2) from plasmid-expressed tfIId ϩ E1-I1-E2-GFP minitranscripts (Fig. 6A, schematic). In WT cells, unspliced pre-mRNA accumulated upon metabolic depletion of SpPrp18 (Fig. 6A, left, lanes 1 and 2). Similarly, upon thiamine repression of prp18-5, high levels of unspliced RNA were observed (Fig. 6A, left, lanes 3 and 4). Interestingly, no cDNA species corresponding to lariat intermediates were detected, suggesting precatalytic splicing arrest. To corroborate these findings, experiments were designed to slow the normal rapid turnover of lariat intron-exon 2, which is expected to accumulate in mutants with a slow and/or arrested second step reaction. For this analysis, we employed cells lacking the debranching enzyme, Dbr1. We generated the prp18-5 dbr1⌬ double mutant to inactivate lariat RNA debranching activity in a strain that expresses the mutant SpPrp18-5 protein.
We also made the corresponding WT dbr1⌬ control strain. Strikingly, the prp18-5 dbr1⌬ double mutant showed synthetic sickness at 28°C when compared with the single mutant parents or the control WT dbr1⌬ strain (Fig. 6B, left).
We assessed the splicing status of tfIId ϩ intron 1-containing minitranscript using primer extension assays in these strains (prp18-5 dbr1⌬ and WT dbr1⌬). In WTdbr1⌬ cells, abundant spliced mRNAs and low levels of pre-mRNA were detected. In contrast, elevated unspliced pre-mRNA levels were found upon metabolic depletion of SpPrp18 in the dbr1⌬-sensitized background (Fig. 6A, right, lanes 1 and 2). Similarly, in cells with SpPrp18-5 mutant protein and lacking Dbr1, increased levels of unspliced RNA were observed (Fig. 6A, right, lanes 3 and 4) upon metabolic depletion, indicating an early splicing arrest, yet in cells lacking functional SpPrp18, only a very faint primer extension cDNA product around the expected size of lariat intermediate was detected, indicating a major block before the first catalytic step. A positive control fission yeast strain with a second step splicing defect, either a mutant in a splicing factor or a cis splice site mutant, is as yet lacking in fission yeast; hence, comparative analyses are not possible. Thus, we propose that a  strong and early splicing arrest occurs before catalysis when SpPrp18 is compromised. With these data pointing toward a precatalytic arrest in these cells, we probed for genetic interactions of prp18-5 with the splicing factor SpPrp2/U2AF59, which functions at the very early step of U2 snRNP recruitment to the intronic branch point (37,38). Interestingly, we observed a strong positive genetic interaction, at 28°C, in the double mutant prp18-5 prp2-1, which had better growth as compared with the parent prp2-1. At the permissive temperature for prp2-1, the growth retardation triggered by metabolic depletion of prp18-5 is suppressed (Fig. 6C, right, rows 2 and 3). Further, we assessed whether the near normal growth of prp18-5 prp2-1 double mutant in ϩT conditions relates to a rescue of splicing defects of SpPrp18-dependent introns when prp2-1 is at permissive conditions. Semiquantitative RT-PCR assays showed that the double mutant had only basal levels of unspliced pre-mRNA for spf38 ϩ intron 5 (Fig. 6D, left, lanes 5 and 6), an intron that was strongly dependent on SpPrp18 for splicing (Fig. 6D, left, lanes  1 and 2). Thus, the data indicate a rescue of its splicing defect in the double mutant. This intron was efficiently spliced in prp2-1 mutant grown at the permissive temperature (28°C) regardless of the presence or absence or thiamine (Fig. 6D, left, lanes 3 and  4). However, interestingly, for tfIId ϩ intron 1, we observe persistent accumulation of unspliced pre-mRNA in the double mutant (Fig. 6D, right, lanes 5 and 6). This intron was inefficiently spliced even in parent prp2-1 single mutant at semipermissive 28°C (Fig. 6D, left, lanes 3 and 4). Thus, rescue of splicing defects is observed only for an intron that is spliced in the presence of Prp2-1 protein. These data suggested that splicing suppression in the double mutant is intron-specific and that SpPrp2 acts upstream to SpPrp18.
Efficient Splicing of G 1 -S Regulators Suggests a Role of spprp18 ϩ in Cell Cycle-Several fission yeast pre-mRNA splicing factor mutants, in addition to impaired splicing, also show striking cell cycle defects, indicating a strong link between splicing and cell cycle regulation (28). The cells depleted of the wild-type SpPrp18 arrested as elongated cells, reminiscent of cdc mutants. Microscopy and fluorescence-assisted cell sorting (FACS) of WT and prp18-5 cells grown under conditions when the wild-type or mutant proteins are expressed were compared with cultures treated for 36 h with thiamine. In the absence of thiamine, cells of both genotypes had equivalent cell lengths and a single nucleus (Fig. 7A, left panels). Upon depletion of the wild-type protein, elongated cells, nearly twice the normal length, formed the majority of the population (Fig. 7A, top  right). The prp18-5 cells became ϳ20% longer (p Ͻ 0.001) when the mutant protein was depleted by growth in thiamine (Fig.  7A, bottom right). The FACS analyses of propidium iodidestained cells show, as expected, that WT cultures have largely cells with 2C nuclear DNA content (Fig. 7B, white peak). Strikingly, upon metabolic depletion of SpPrp18, a majority of cells accumulate with 1C nuclear content, indicating an arrest largely before DNA replication (Fig. 7B, light gray peak). In fact, most prp18-5 cells had 1C nuclear content even when the mutant protein was expressed (Fig. 7B, black peak).
We examined the splicing efficiency of transcripts that encode cell division regulators because several of these transcripts are intron-containing. All introns in cdc2 ϩ encoding the central cyclin-dependent kinase Cdc2 were studied in WT and prp18-5 strains. Reduced splicing across the cdc2 ϩ intron 4 was seen when the spprp18 ϩ allele was transcriptionally repressed. The same splicing defect is seen in prp18-5 cells (Fig. 7C, top  left, M). Other cdc2 ϩ introns were unaffected (data not shown). The G 1 -S phase Cdc2 kinase functions are controlled by regulated proteolysis of its inhibitor Rum1 and by the E3-ubiquitin ligase SCF complex-dependent degradation of the S-phase regulator Cdc18 (39 -41). Because rum1 ϩ and cdc18 ϩ are nonintron-containing genes, we investigated the splicing status of intron 1 in skp1 ϩ , encoding an essential adaptor factor of the SCF ubiquitin ligase. Elevated pre-mRNA levels for skp1 ϩ I1 were observed both upon depletion of spprp18 ϩ and in prp18-5 cells (Fig. 7C, top right, P). This defect was enhanced when expression of the mutant protein was reduced (Fig. 7C, top  right, P). In fission yeast, the gene expression peak during G 1 -S phase is regulated by the MluI binding factor (MBF) transcription regulatory complex composed of Res1 and Res2, two zinc finger proteins, and Cdc10 (42,43). We detected elevated pre-mRNA for res2 ϩ intron 1 upon transcriptional repression of spprp18 ϩ (Fig. 7C, bottom left, P), which is recapitulated in prp18-5 cells. The effects on res1 ϩ splicing were moderate (data not shown). To assess whether the splicing defects in transcripts for G 1 -S regulators correlate with changes in the protein levels, we examined endogenous SpCdc2 protein levels by immunoblotting because specific antibodies were available (see "Experimental Procedures"). Lysates from wild-type and . Validation by semiquantitative RT-PCR analysis of the three categories of splicing phenotypes inferred from microarray analyses. A, splicing status of mdm35 ϩ intron 1 measured by semiquantitative RT-PCR as a representative of the subcategory with both pre-mRNA accumulation and mRNA decrease on depletion of SpPrp18. B, splicing status of spf38 ϩ intron 5 and sfc9 ϩ intron 2 as representatives for the intron class showing accumulation of unspliced pre-mRNA without change in mRNA levels in cells depleted of SpPrp18. C, ubc4 ϩ intron 1 and cwf2 ϩ intron 1 represent a subcategory of introns unaffected by the absence of SpPrp18. Semiquantitative RT-PCRs were carried out using tracer label as described under "Experimental Procedures." The -fold change accumulation or reduction of various transcript forms was calculated by densitometric analysis of the amplicon from pre-mRNA and mRNA species after normalization to the intronless act1 ϩ transcript, as described in the legend to Fig. 3. P, pre-mRNA; M, mRNA; A, intronless act1 ϩ mRNA. Bars, average value from three biological replicates; error bars, S.D. *, p Ͻ 0.05; **, p Ͻ 0.01; ***, p Ͻ 0.005, as determined by unpaired Student's t test. ns, non-significant change with p Ͼ 0.05 as determined by unpaired Student's t test. D, in vivo splicing analysis of ubc4 ϩ intron 1 and cwf2 ϩ intron 1 for splicing in slu7 ϩ and slu7-2 mutant cells grown in the absence and presence of 15 M thiamine (ϪT and ϩT) for 28 h using semiquantitative RT-PCR.
prp18-5 mutant cells show a moderate decrease in the steadystate SpCdc2 protein levels in the thiamine-supplemented prp18-5 mutant cells (Fig. 7D, lane 4). These data hint that cumulative splicing defects in multiple transcripts and the ensuing changes to their protein levels may partly contribute to the cell cycle defects caused by loss of SpPrp18.

Discussion
In this study, we deduced critical functions for both the conserved region of SpPrp18 and the flanking helices. Transcriptome profiling of cells lacking functional SpPrp18 revealed intron-specific roles for this protein. Splicing analysis in a dbr1⌬-sensitized background for a SpPrp18-dependent intron and genetic interaction with early acting U2AF59 show that loss of SpPrp18 arrests spliceosomes before catalysis. Reduced splicing efficiency for some transcripts encoding key regulators of the G 1 -S transition indicate a role for SpPrp18 in mediating the splicing of key regulators of cell division cycle progression.
Although second step splicing functions have been demonstrated for budding yeast and human Prp18 using in vitro and in vivo approaches, there have been very limited studies on functionally characterizing Prp18 orthologs from other species (4,9,10). spprp18 ϩ is an essential gene in S. pombe, and in a recent genetic interaction mapping study, a DAmP allele of SpPrp18  DECEMBER 30, 2016 • VOLUME 291 • NUMBER 53

Roles for SpPrp18 in Splicing and Cell Division Cycle
was also found to have extremely poor growth (33,44). Apart from this phenotype, no functional studies on this fission yeast splicing factor have been reported thus far. Here, we undertook detailed analysis of the function of SpPrp18 in fission yeast intron recognition and splicing using strains with depleted levels of wild-type protein and those expressing a recessive missense mutant. Extensive mutational analysis of budding yeast ScPrp18 revealed that the highly conserved CR loop and its flanking helices have splicing functions distinct from regions that interact with ScSlu7 and influence spliceosome recruitment (13). Mutants in the CR of ScPrp18 are dominant negative and ts. On the other hand, whereas the null phenotypes of the G196A/V197A/T198A triple mutant in the CR region and the K325A/R326A mutant in helix 5 in fission yeast SpPrp18 con- and relative cell number are plotted along the x and y axes, respectively. The positions of peaks for 1C and 2C DNA content are indicated. C, splicing status of some intron-containing transcripts encoding factors critical for G 1 -S transition: cdc2 ϩ intron 4, skp1 ϩ intron 1, and res2 ϩ intron 1. Splicing status in WT and prp18-5 cells was determined through limiting cycle semiquantitative RT-PCRs as described in the legend to Fig. 3. For each intron, densitometric values for amplicons representing mRNA or pre-mRNA species were normalized to that from intronless act1 ϩ (y axis). White bar, data from WTϪT cultures; gray bar, data from WTϩT cultures; black bar, data from prp18-5 ϪT cultures; dark gray bar, data from prp18-5ϩT cultures. *, p Ͻ 0.05 calculated using unpaired t test for data from multiple experiments (n ϭ 3 or 4). D, Western blotting analysis of SpCdc2 protein levels in whole cell lysates of WT and prp18-5 cells grown in the absence and presence of thiamine (ϪT and ϩT) at 30°C. Coomassie-stained gel post-transfer served as the loading control. firm a vital functional role for these regions, these alleles are recessive loss-of-function mutants in fission yeast. These data hint at a failure in vital spliceosomal interactions of these nonfunctional variants. In budding yeast ScPrp18, the CR loop is proposed, in a context-dependent manner, to improve interactions between exonic residues and budding yeast U5 snRNA loop 1 (13)(14)(15)45). We speculate that the critical roles for SpPrp18 helix 5 and CR domain may also be to improve U5 loop 1 with exonic sequence interactions.
Interestingly, our missense mutant in helix 1 (V194R), prp18-5, which is also a recessive mutant, supports viability, albeit with compromised growth properties probably due to altered protein stability. We show that compromised functions of SpPrp18V194R may be a consequence of its altered conformation, triggered by the positively charged residue at the protein core. Immunoblotting experiments to probe wild-type SpPrp18 protein levels detected an additional slower migrating species, a form that is intriguingly absent in the SpPrp18 V194R mutant. We speculate that either a modification or altered conformer of wild-type SpPrp18 may cause its altered mobility. A recent study of the proteome and phosphoproteome of S. pombe during various cell cycle stages hints at phosphorylated forms of SpPrp18 (Ser-115 residue) (46).
The splicing defects that we discern from the global splicing analysis for S. pombe introns suggest that its functions are essential but substrate-specific. Cis intronic features, such as intron length, branch point nucleotide to 3Јss distance, 3Јss sequence, and the adenosine and uridine content (AU%) did not differ between SpPrp18-dependent and -independent introns. However, an analysis of the 5Јss hexamer for the occurrence of consensus versus non-consensus nucleotides (log odd score analysis) suggests that a statistically significant proportion of N3 nucleotides are non-consensus in our SpPrp18-dependent intron data set (data not shown). Introns with weak/ abnormal splice sites can have greater dependence on splicing factors, as reported even in budding yeast (47,48). Also, it is suggested that the splicing apparatus is limiting, and thus different cellular transcripts encounter inherent competition for these splicing factors (49). Thus, it is plausible that multiple variables in fission yeast transcripts create dependence on this low abundance (50) but essential factor SpPrp18.
Mutation of the 3Јss, which leads to lariat intermediate accumulation in budding yeast, shows complete splicing arrest before the first step in S. pombe (19). 8 Furthermore, an expanded Pyn tract was unable to compensate for the 3Јss mutation in fission yeast (19). Mutations in the nucleotides of the 5Јss consensus hexamer motif (G 1 -A and U2-A/G) in budding yeast cause lariat intermediate accumulation and second step arrest, whereas surprisingly, mutations in all nucleotides in the 5Јss hexamer sequence cause a block before any catalysis in fission yeast (51). Among the trans-acting factors, studies from our laboratory on SpSlu7, the ortholog of budding yeast second step factor, ScSlu7, also clearly show splicing arrest before catalysis (23). Fission yeast cells deleted for lariat debranching gene (dbr1⌬) have more severe growth defects than the corre-sponding budding yeast mutant. This has been attributed to the higher intronic content of the S. pombe transcriptome that demands rapid turnover of excised lariat introns that are held in stable U2.U5.U6 post-splicing complexes (52). Regardless, because prp18-5 dbr1⌬ cells were viable, we used them to search for accumulation of lariat intron-exon 2 intermediates. Because we detected predominantly accumulated pre-mRNAs in the prp18-5 single mutant and in the prp18-5 dbr1⌬ double mutant, our data point to a major arrest before the first catalytic step in the mutant of SpPrp18. We cannot rule out a continued subsequent role in the second step that again could be substrate context-specific. These studies await additional partial loss of function allelic variants of fission yeast Prp18. Allele-specific splicing functions have been noted for the budding yeast splicing factor, Prp8, wherein mutant alleles inhibiting either first step or second step splicing have been reported (53,54).
In budding yeast, Prp18 acts as a bridging molecule between Slu7 and U5 snRNP to aid 3Јss selection after catalytic step 1 (9,12). The strong association between Prp18 and Slu7 aids their spliceosome assembly (55). Our prior studies suggested weak or no interactions between fission yeast SpPrp18 and SpSlu7 (22). Thus, other domains in SpPrp18 are probably more relevant for its roles in the spliceosome. Remarkably, we found a strong genetic suppression interaction between prp18-5 and spprp2-1, a ts mutant in U2AF59/SpPrp2, because we found that prp2-1 functionally rescued the growth retardation of prp18-5 cells triggered by thiamine supplementation. We also found the suppression is not mutual because the growth arrest of prp2-1 cells at 37°C was not rescued in the double mutant. In the double mutant, the suppression of splicing defects in prp18-5 cells was intron-specific and apparently related to dependence of the intron on prp2-1. These findings indicate that SpPrp2 that assembles in the early precatalytic spliceosomes containing SF1-U2AF59-U2AF23 (38) acts upstream of SpPrp18. In budding yeast, two categories of alleles have been identified in Prp8, Cef1, U6 snRNA, and Prp16, based on their functions in suppressing either the first step or the second step splicing arrest caused by specific intronic cis mutations. Such suppressor alleles have opposing effects on the two catalytic steps (53,54,56). These first step or second step alleles stabilize one catalytic site conformation over the other; thus, combining two mutants in opposing steps can lead to overall improved splicing efficiency as captured in the two-state model (53,56). Fission yeast prp2-1 (U2AF59) is a well studied mutant with an arrest before the first catalytic step. We speculate that the requirement of SpPrp18, possibly for the 3Јss recognition, may be bypassed by a favorable first step conformation when prp2-1 is active; the intermediates thus formed proceed through the second catalytic step.
Several genetic and proteomic studies, in model organisms, establish a relationship between cell cycle regulatory proteins and splicing factors (26 -31, 57). Strikingly, we demonstrate that repression of WT or of prp18-5 generated a G 1 -S transition arrest, yet we do not exclude a role for SpPrp18 at later cell cycle transitions or in other pathways that control cell size and cell cycle progression. Mutations in budding yeast PRP17 or PRP8 that impair splicing cause G 1 -S and G 2 -M cell cycle defects, linking splicing with multiple regulatory checkpoints in the cell cycle (57,58). We correlate the cell cycle arrest observed in the 8 S. Banerjee and U. Vijayraghavan, unpublished data. DECEMBER 30, 2016 • VOLUME 291 • NUMBER 53 absence of functional SpPrp18 with the inefficient splicing of introns in some key G 1 -S transition regulators, including cdc2 ϩ , members of the MBF complex res1 ϩ , res2 ϩ , and skp1 ϩ , an essential subunit of the SCF ubiquitin complex. These data indicate that cumulative defects in multiple transcripts, important for G 1 -S transition, can partly contribute to the cell cycle defects. These could be compounded by effects on other pathways that impinge on the cell cycle (e.g. cell size) and by the indirect effects of SpPrp18 on gene expression that lead to arrest during the G 1 -S transition.

Roles for SpPrp18 in Splicing and Cell Division Cycle
These results uncover a distinct role for Prp18 in S. pombe as compared with budding yeast and human cells. The short, multi-intronic transcripts are a hallmark of S. pombe and many other fungal genomes that sets it apart from S. cerevisiae and humans. Also, this short length allows for the precise recognition of the splice sites at the intronic ends following the intron definition model (59). The precatalytic arrest that we observe in cells depleted of functional SpPrp18, at least for introns examined here, contrast to the second step arrest deduced from studies of few model introns in budding yeast and humans. These data suggest an early role for this fission yeast factor, probably due to early recognition of splice sites before catalysis in the intron definition mode of splicing. As analysis of specific introns becomes more sophisticated, we may continue to refine our understanding of how different splicing factors mediate splicing of specific introns. Further analysis will be required to determine whether the studies performed here point to a general function of Prp18 proteins in early splicing steps for short introns, such as those present in S. pombe.

Experimental Procedures
Yeast Strains and Plasmid Constructions-Procedures for genetic analyses in S. pombe were followed as described (60) and on the PombeNet website. spprp18 ϩ cDNA (1032 bp) was generated by reverse transcription using spprp18 RP on wildtype FY528 RNA. The cDNA was amplified with primers spprp18 FP and RP, restricted with BamHI, and cloned into pBS(KS). The spprp18 ϩ cDNA was excised as a BamHI frag-ment from pBS(KS)-spprp18 ϩ cDNA and ligated with BamHIdigested pREP4X to create pREP4X-spprp18 ϩ . The insert in the pDblet-spprp18 ϩ clone was obtained by PCR on wild-type FY528 genomic DNA of the ORF and 1-kb upstream sequences using a primer pair (spprp18 5Ј-UTR FP and spprp18 RP; supplemental Table S1). The ϳ2-kb PCR product thus obtained was first cloned into the EcoRV site of pBS(KS) and then excised as a KpnI/SacI fragment for cloning into pDblet to generate pDBlet spprp18 ϩ , which was used for complementation studies. S. pombe strains used here are described in Table 4. spprp18 ϩ gene disruption was achieved by transforming a 3.1-kb spprp18::his3 ϩ fragment into S. pombe WT diploid cells to generate the spprp18::his3 ϩ /spprp18 ϩ strain. Subsequently, we created the haploid strain spprp18::his3 ϩ , pREP4x spprp18 ϩ , where the chromosomal null allele was complemented by the pREP4x spprp18 ϩ plasmid. Inverse PCR amplifications, for introducing missense mutations G288A/V289A/ T290A and K325A/R326E, were done using Vent DNA polymerase (New England Biolabs) on the plasmid pBS(KS)-spprp18 ϩ cDNA as the template. For each of these mutants, two complementary mutagenic primers of length ϳ30 -35 nucleotides containing the desired mutation(s), flanked by ϳ15 nucleotides of wild type sequences, were used as primers (supplemental Table S1). The inverse PCR products were DpnItreated and transformed into Escherichia coli competent cells. Plasmids from E. coli transformants were verified for authenticity of the mutation by sequencing the insert in the plasmids from 2-4 independent colonies. For expression of G288A/ V289A/T290A or K325A/R326E mutants or the wild type spprp18 ϩ cDNA in S. pombe, expression clones were made in the pREP41MH-N or pREP42HA-N vectors. In these recombinants, the desired protein would be overexpressed as an N-terminally tagged fusion driven by the nmt41 promoter. In each case, the full-length wild-type or mutant cDNA was excised as a BamHI fragment from pBS(KS)-spprp18mut cDNA (G288A/V289A/T290A or K325A/R326E) and cloned into BamHI-restricted pREP41MH-N or pREP42HA-N vectors. This created the plasmid constructs pREP41MH-spprp18 ϩ or pREP42HA-spprp18 ϩ , pREP42HA-spprp18mut (G288A/ V289A/T290A), and pREP42HA-spprp18mut (K325A/R326E). The buried amino acids Val-194, Leu-239, Ile-259, and Leu-324 in spprp18 ϩ were predicted using the program PREDBUR (35). Random mutagenesis of V194X was done by inverse PCR on bacterial plasmid clone pBS(KS)-spprp18 ϩ , followed by DpnI treatment (New England Biolabs) to recover a pool of potential mutants. We subcloned this pool of mutant cDNAs into the pREP81X vector, and a collection of recombinants was transformed into the spprp18::his3 ϩ /spprp18 ϩ heterozygous diploid. Sporulation was done to obtain viable spprp18⌬ haploids with plasmids carrying various spprp18V194X random mutants. These were screened for growth phenotypes. Plasmid from one of the several ts colonies was sequenced to identify the V194R mutation. Subsequently, the wild-type and mutant (V194R) spprp18 open reading frames were cloned into the pJK148 nmt81 vector for integration at the leu1-32 locus and expression from the nmt81 promoter. The pJK148 nmt81 spprp18 ϩ /spprp18V194R integrants at the leu1-32 locus were confirmed by PCR on genomic DNA.
Comparative Modeling of SpPrp18 -We performed multiple sequence alignment of homologous Prp18 proteins using ClustalW. The alignment obtained was used with ScPrp18⌬79 structure (PDB code 1DVK) as a template for homology modeling using MODELLER 9v7 or Swiss MODEL (61). The default spatial constraints for distances, angles, dihedral angles, pairs of dihedral angles, and other spatial features defined by atoms and pseudo-atoms (62) were used. The models were validated by means of PROCHECK (63) and VERIFY3D (64). The obtained models were saved in PDB format and visualized using PyMOL or chimera version 1.8.1.
Far UV CD Spectrometry and Intrinsic Tryptophan Fluorescence Studies-The E. coli C-41 cells harboring the clones (WT and V194R) in pET15b vector were grown at 37°C to late log phase. This was followed by induction with 0.1 mM isopropyl 1-thio-␤-D-galactopyranoside and further growth for 6 h at 25°C. The proteins obtained in the soluble fraction were purified using nickel-nitrilotriacetic acid affinity columns and used for the biophysical studies. Proteins were exposed to temperatures ranging from 20 to 90°C and were heated at a constant rate of 1º C/min. Circular dichroism measurements were made with a JASCO J-715 spectropolarimeter fitted with a Jasco Peltier-type temperature controller (PTC-348WI). Spectra were recorded with a scan speed of 20 nm/min and with a response time of 1 s. Far UV CD spectra were taken in the wavelength range of 200 -250 nm, at a protein concentration of 12 nM with a 2-mm path length cell.
Fluorescence measurements were carried out on a Hitachi spectrofluorometer (F-7000) equipped with a data recorder DR-3. The fluorescence spectra were measured at a protein concentration of 9.8 M with a 1-cm path length cuvette at a temperature range of 20 -90°C. To determine intrinsic tryptophan fluorescence, the excitation wavelength was set at 295 nm, and the emission spectrum was recorded in the range of 300 -450 nm with 5-and 10-nm slit width for excitation and emission, respectively.
Probe Design, Sample Preparation, Microarray Hybridization, and Data Acquisition-The design of various probes for splicing sensitive microarray analysis was as described (22). RNA was isolated from two biological replicates of WT and prp18-5 cells grown in the presence and absence of 15 M thiamine for 36 h and then harvested. The total RNA was prepared using TRI-reagent (Sigma). The RNA samples were labeled using an Agilent Quick-Amp labeling kit. 500 ng each of the untreated and treated samples were reverse transcribed at 40°C using oligo(dT) primer with a T7 polymerase promoter and random hexamer primer with a T7 polymerase promoter in two individual reactions and converted to double-stranded cDNA. Synthesized double-stranded cDNA was used as a template for cRNA generation. cRNA was generated by in vitro transcription, and the dye Cy3 CTP was incorporated during this step. The cDNA synthesis and in vitro transcription steps were carried out at 40°C. 700 ng of the Cy3-labeled cRNA samples (600 ng of oligo(dT)-labeled ϩ 100 ng of random hexamerlabeled samples) were fragmented at 60°C and hybridized onto customized S. pombe, 8X60K arrays (Agilent Technologies). Fragmentation of labeled cRNA and hybridization were done using a gene expression hybridization kit, and hybridization was carried out at 65°C for 16 h. The hybridized slides were washed using gene expression wash buffers and scanned using the Agilent microarray scanner at 3 m resolution.
Microarray Data Analysis-Feature-extracted data were analyzed using Agilent GeneSpring GX software, and normalization of the data was done using the 75th percentile shift. The log 2 Cy3 fluorescence values for the wild type and mutant were mathematically zero-transformed and analyzed relative to the respective untreated sample (without thiamine, ϪT). Student's t test along with a false discovery rate-adjusted (Benjamini and Hochberg) p value calculated using the R statistical program was used to obtain statistically significant values for various probes (p Ͻ 0.05) in two biological replicates that were taken for hierarchical clustering. Affected introns were considered to be those with a Ͼ0.8-fold (log 2 scale) increase in signal for intronic probes. Also considered were those with a decrease in mRNA levels, minimum 0.8-fold (log 2 scale) splice junction, and intronic probes for both biological replicates of WTϪT and WTϩT samples are provided (supplemental Data Set 1, Sheet 1). The status of these 253 introns in the prp18-5ϪT sample as compared with WTϪT sample is also provided (supplemental Data Set 1, Sheet 2), represented as a heat map in Fig. 4. Also shown is a list of introns with a Ͼ0.8 average -fold change (log 2 ) (equivalent to 1.7-fold) (p Ͻ 0.05) between decrease in splice junction probe signal and that of the untreated sample. List of the stringent set of 253 introns showing -fold change and raw intensity values for gene expression, WTϪT, and WTϩT samples for the intronic probe (supplemental Data Set 1, Sheet 3). Also shown is a list of introns with statistically significant average -fold change less than Ϫ0.8 (log 2 scale) for splice junction probe when WTϪT and WTϩT samples are compared (supplemental Data Set 1, Sheet 4).
Reverse Transcription and Primer Extension Assays-Total RNA from fission yeast was extracted using TRI-reagent (Sigma). 2-5 g of DNase I (New England Biolabs)-treated total RNA was reverse-transcribed with Moloney murine leukemia virus reverse transcriptase (New England Biolabs) using a reverse primer from a downstream exon. [␣-32 P]dATP tracerlabeled PCRs on the cDNAs and gel analyses on 8% native PAGE were performed as described (22). The primers used are listed in supplemental Table S1. Primer extension reactions, at 37°C, were done on 50 g of RNA and a ␥-32 P-5Ј-end-labeled 3Ј exon reverse primer using Moloney murine leukemia virus RT (New England Biolabs). The single-stranded cDNAs were resolved on 4% 8 M urea-polyacrylamide gels.
FACS Analysis-To analyze DNA content, the cells were collected by brief centrifugation, fixed in ice-cold 70% ethanol, and then treated overnight with RNase A (10 mg/ml) in 1ϫ PBS. After staining with propidium iodide (20 g/ml), the fluorescence intensities were measured by flow cytometry using a BD FACScalibur (BD Biosciences). About 10,000 events were scanned for each analysis.
Immunoblotting of SpPrp18 and SpCdc2 in WT or Mutant Cells-Crude whole cell extracts were prepared from cells of the specified strains grown at 30°C in the absence or presence of 15 M thiamine in EMM-selective media as described (65). For Western blotting, 20 -25 l of crude protein lysate corresponding to 20 g of protein was run on a 10% SDS-PAGE. The separated proteins were electroblotted to a Hybond P (GE Healthcare) membrane. The blot was incubated with rabbit polyclonal anti-SpPrp18 antibodies or monoclonal anti HA12CA5 antibodies from Roche Applied Science and subsequently with secondary goat anti-rabbit HRP conjugate/antimouse HRP conjugate from Bio-Rad. Blot was developed with SuperSignal West Pico chemiluminescent substrate (Millipore), and the image was analyzed using ImageQuant LAS 4000 (GE Healthcare). Immunoblotting was done with Cdc2 p34 Y100.4 mouse monoclonal antibody (Santa Cruz Biotechnology, Inc.) on crude whole cell lysates from WT and prp18-5 cells grown in the absence and presence of thiamine (ϪT and ϩT) at 30°C.
Confocal Imaging-Cells were fixed in 70% ethanol, followed by washing with 1ϫ PBS and finally suspended in 1ϫ PBS. DAPIstained cells were imaged at 23-25°C, with a Carl Zeiss LSM 710 confocal microscope. Data were acquired using a ϫ60 oil immersion objective with 5% laser power (wavelength 405 nm) and ϫ1.2 zoom. The images were processed by ImageJ software.
Author Contributions-U. V., R. V., P. K., N. V., and G. M. designed the research; N. V., G. M., P. K., and R. K. performed the experiments; R. K. performed and analyzed the biophysical experiments; P. B. performed the microarray data analysis. Data analysis and manuscript preparation was performed by N. V., G. M., P. K., R. V., and U. V. All authors read and approved the final version of the manuscript.