ψ35 in the Branch Site Recognition Region of U2 Small Nuclear RNA Is Important for Pre-mRNA Splicing in Saccharomyces cerevisiae*

Pseudouridine 35 (ψ35) in the branch site recognition region of yeast U2 small nuclear RNA is absolutely conserved in all eukaryotes examined. Pus7p catalyzes pseudouridylation at position 35 in Saccharomyces cerevisiae U2. The pus7 deletion strain, although viable in rich medium, is growth-disadvantaged under certain conditions. To clarify the function of U2 ψ35 in yeast, we used this pus7 deletion strain to screen a collection of mutant U2 small nuclear RNAs, each containing a point mutation near the branch site recognition sequence, for a synthetic growth defect phenotype. The screen identified two U2 mutants, one containing a U40 → G40 substitution (U40G) and another having a U40 deletion (U40Δ). Yeast strains carrying either of these U2 mutations grew as well as the wild-type strain in the selection medium, but they exhibited a temperature-sensitive growth defect phenotype when coupled with the pus7 deletion (pus7Δ). A subsequent temperature shift assay and a conditional pus7 depletion (via GAL promoter shutoff) in the U2-U40 mutant genetic background caused pre-mRNA accumulation, suggesting that ψ35 is required for pre-mRNA splicing under certain conditions.

The five high abundance spliceosomal small nuclear RNAs (snRNAs), 1 U1, U2, U4, U5, and U6, are integral to pre-mRNA splicing (1,2). During spliceosome assembly, an intricate network of RNA-RNA interactions among the snRNAs and the pre-mRNA facilitates two specific trans-esterification reactions that remove the intron. The well characterized interactions include base pairing between the 5Ј-sequence of U1 and the 5Ј-splice site of the pre-mRNA (3), the base pairing between the U2 branch site recognition sequence and the branch site in the pre-mRNA (4,5), base pairing between U2 and U6 snRNAs (U2-U6 helices I, II, and III) (6 -10), and base pairing between the conserved ACAGAGA sequence in U6 and the 5Ј-splice site of the pre-mRNA (11)(12)(13). In addition, non-Watson-Crick base-pairing interactions between the conserved loop of U5 and the exon sequences at the 5Ј-and 3Ј-splice sites (13)(14)(15)(16)(17), as well as interactions between the 5Ј-and 3Ј-splice sites themselves (18,19), have also been well documented.
A notable feature shared by all five spliceosomal snRNAs is their extensive internal modifications, including pseudouridylation and 2Ј-O-methylation (20,21). Significantly, these modifications appear to be mostly clustered in regions of demonstrated functional importance. For instance, both the 2Ј-O-methylated nucleotides and pseudouridines are concentrated in the sequences involved in the spliceosomal RNA-RNA interactions described above, suggesting a possible functional role in pre-mRNA splicing (21).
Modifications in higher eukaryotic spliceosomal snRNAs are believed to be catalyzed by an RNA-guided mechanism, directed by Box H/ACA sno/scaRNPs for pseudouridylation and Box C/D sno/scaRNPs for 2Ј-O-methylation (23)(24)(25). In some instances, this mechanism has been experimentally verified (26 -28). However, the formation of at least two pseudouridines in S. cerevisiae U2 (⌿35 and ⌿44) is catalyzed by an RNAindependent mechanism. Specifically, ⌿35 formation is catalyzed by Pus7p (29), and ⌿44 is introduced by Pus1p (30). Whether the other modification sites in the yeast spliceosomal snRNAs are introduced by protein-only or RNA-guided mechanisms remains to be determined.
The functional importance of vertebrate U2 modifications has recently been established (31)(32)(33). Using a Xenopus oocyte reconstitution system, we have shown that the 6 pseudouridines in the branch site recognition region as well as the 9 modified nucleotides within the 5Ј-most 27 nucleotides of Xenopus U2, including 3 pseudouridines and 6 2Ј-O-methylated residues, are all required for snRNP assembly and pre-mRNA splicing (31,32). Using NMR, the Greenbaum group (34,35) has recently shown that the base-pairing interaction between ⌿34 (⌿35 in yeast U2) in the branch site recognition region and the nucleotide next to the pre-mRNA branch point adenosine results in the branch point nucleotide being bulged out, a configuration known to be important for the first step of the splicing reaction. Consistent with this result, Valadkhan and Manley (36) have demonstrated that the change of uridine to pseudouridine at the same position (⌿35) greatly activates splicing-related catalysis via protein-free snRNAs (37).
In light of these data, it remains puzzling, however, that the pus7 deletion (pus7⌬) yeast strain, in which U2 pseudouridylation at position 35 is completely blocked, does not show an obvious growth phenotype in rich media. We previously observed only a mild growth defect phenotype when the mutant strain was grown in competition with the wild-type strain or grown under high salt conditions (29). 2 To understand, in more detail, the functional role of ⌿35 in yeast U2 in vivo, we performed a synthetic lethal screen using combinations of the pus7⌬ mutant and unconditionally viable U2 point mutations located adjacent to the branch site recognition region (38, 39). 3 We found that position 40 (U40) in U2, which is immediately downstream of the branch site recognition sequence, is crucial when ⌿35 formation is blocked. Specifically, the inhibition of ⌿35 formation coupled with mutation/deletion of position 40 in U2 gives rise to temperature-sensitive strains in which pre-mRNA splicing is blocked under nonpermissive conditions. Our results indicate that a single pseudouridine, ⌿35, in the yeast U2 snRNA is critical for pre-mRNA splicing and cell growth under certain conditions.
Plasmid pCAU2 (CEN ARS HIS3 SNR20) contains a full-length wild-type U2 gene (40) and contains a single base insertion at the 5Ј-end of the U2 gene to create a unique EcoRI site. All U2 mutant plasmids were derived from pCAU2 by replacing the EcoRI-EcoNI fragment containing the wild-type U2 gene with a PCR fragment containing a mutant U2 gene derived from in vitro T7 expression plasmids (38). All U2 mutant plasmids were then sequenced to verify the presence of the mutation. Plasmid pPUS7 (CEN LEU2 PUS7) was constructed by inserting the wild-type PUS7 gene between the BamHI and PstI sites of Ycplac111 vector (41). Plasmid pPUS7-Gal (CEN URA3 P GAL10 -PUS7) was constructed by inserting the PUS7 gene between the BamHI and PstI sites of pAVA0040 (42).
All U2 mutant plasmids were used for transformation and selection with 5-fluoroorotic acid (5FOA; see below). pCAU2 and pPUS7 were used to rescue the growth phenotype of strains with the pus7 deletion coupled with U2 mutation/deletion at position 40 (Fig. 2C, pus7⌬ϩU2-U40G and pus7⌬ϩU2-U40⌬, and see "Transformation and 5FOA Selec- There are a total of 13 pseudouridines and 10 2Ј-O-methylated residues in mammalian U2 and 3 pseudouridines in yeast U2 (the number of 2Ј-O-methylated residues in yeast U2 has not been established). All three pseudouridines in yeast U2 (⌿35, ⌿42, and ⌿44) have counterparts in mammalian U2 (⌿34, ⌿41, and ⌿43). The base-pairing interaction between the branch site recognition sequence in yeast U2 and the branch site sequence in yeast pre-mRNA (N is the nucleotide immediately upstream of the branch site) is also schematized. The branch site recognition sequences are boxed, and the Sm binding sites are shaded. The sequences involved in U2-U6 interactions, Helices I, II, III (Helix III is not established in yeast) are highlighted by thick lines.
Transformation and 5FOA Selection-Log phase DM2484 or DM2486 cells were collected and adjusted to 50 A 600 /ml in transformation buffer (40% (w/v) polyethylene glycol-3350, 0.1 M lithium acetate). The cells (20 l) were mixed with 1 l of carrier DNA (salmon sperm or calf thymus, 2 mg/ml). The resultant cell mixture then received ϳ1 g of one of the U2 mutant plasmids or pCAU2 as a control and was subsequently incubated at 45°C for 30 min. The cell mixture was then immediately mixed with 150 l of YPD and incubated at 30°C for another hour before plating on solid medium lacking histidine.
A single colony from the above plates was transferred to liquid SDϪHis synthetic medium (0.71% SDϪHis, 2% dextrose) and grown overnight at 30°C. A fixed amount of cells (2 l with 0.006 units of A 600 ) were streaked on 5-FOA plates (SDϪHis, 1 mg/ml 5-FOA) and grown for 4 days at 18, 30, or 37°C. The growth phenotypes were then analyzed ( Fig. 2A).
Temperature-sensitive strains (pus7⌬ϩU2-U40G and pus7⌬ϩU2-U40⌬) were selected, and a single colony of each mutant strain on the 18°C plates was picked for growth in liquid YPD at 18°C overnight. These cells, along with control cells, were either directly plated on solid YPD medium at various temperatures or first transformed with a plasmid containing wild-type U2 or PUS7 genes and then plated on solid YPD medium. We then assessed the ability of wild-type U2 or Pus7p to rescue the growth phenotype of the temperature-sensitive strains (Fig. 2B, pus7⌬ϩU2-U40G and pus7⌬ϩU2-U40⌬).
Temperature-shift Assay-To analyze pre-mRNA splicing at higher temperatures, temperature-sensitive strains (pus7⌬ϩU2-U40G and pus7⌬ϩU2-U40⌬) as well as control strains were first grown in selective synthetic medium at 18°C overnight. The cells were then switched to YPD (1% yeast extract, 2% dextrose, and 2% peptone) and grown at 18°C until the A 600 reached 1.0. At that point, the incubation temperature was changed to 37°C. Cells were collected at various time points both before and after the temperature shift. Total RNA was extracted, and pre-mRNA splicing was analyzed by primer extension (see below).
Conditional Depletion (Promoter Shutoff) Assay-Conditional depletion of Pus7p via promoter shutoff was also used to assess the effect of pus7⌬ϩU2-U40G and pus7⌬ϩU2-U40⌬ on pre-mRNA splicing. Strains pus7⌬ϩU2-U40G and pus7⌬ϩU2-U40⌬, transformed with pPUS7-Gal (CEN URA3 P GAL10 -PUS7), were grown in selective synthetic medium (0.71% SDϪUra, 2% galactose) at 30°C overnight. The yeast cells were then switched to YPGAL (1% yeast extract, 2% galactose, 2% peptone) and grown at 30°C until A 600 reached 2.0 -3.0. At this point, the yeast cells were pelleted and resuspended in YPD. At different time points after switching the medium, cell growth was monitored by A 600 , and a fraction of the cells was collected. Total RNA was extracted from these cells and analyzed for U2 pseudouridylation and pre-mRNA splicing.
RNA Isolation-About 40 A 600 units (ϳ20 ml of 2 OD/ml cells) of yeast cells were collected and resuspended in 0.4 ml of 1ϫ RIB buffer (0.2 M Tris-HCl, pH 7.5, 0.5 M NaCl, 0.01 M EDTA, 1% SDS). PCI-RIB (0.4 ml; phenol:chloroform:isoamyl alcohol (50:49:1) saturated with 1ϫ RIB) and 0.4 cc of sterile acid-washed glass beads were added. The glass beads/cell mixture was vigorously agitated at high speed for 90 s three times, each of which was followed by a 30-s incubation on ice. The cell extracts were pelleted in a microfuge for 2 min, and the supernatants were transferred to fresh tubes and extracted again with 0.4 ml of PCI-RIB. Total RNA was precipitated with ethanol and subsequently treated with RNase-free DNase I (ϳ30 units). After PCI-RIB extraction and ethanol precipitation, the DNA-free total RNA was suitable for U2 pseudouridylation analysis and pre-mRNA measurements (via primer extension or RT-PCR).
RT-PCR-DNA-free total yeast RNA was used as a template for reverse transcription (RT). The reaction was performed for 1 h at 37°C in 20 l containing 5 g of DNA-free total RNA, 1ϫ RT buffer (Promega), 0.5 mM dNTPs, 10 mM dithiothreitol, 2.5 M random DNA hexamer, and 10 units of reverse transcriptase M-MLV (Promega). The cDNA was then amplified by PCR in two steps. In the first step, the reaction was performed in 15 l containing 5% of the RT products, 1ϫ PCR buffer (Invitrogen), 0.25 mM dNTPs, 10 pmol 5Ј-and 3Ј-primers (including 10,000 cpm of the 5Ј 32 P-radiolabeled 5Ј-primer) specific for the gene to be analyzed (see the primers below), and 2.5 units of Taq polymerase (Invitrogen). A total of 5 cycles (94°C for 1 min, 55°C for 1 min, and 72°C for 1 min) was used, with the final cycle being followed by 72°C for 10 min. Subsequently, a 5-l mixture of 1ϫ PCR buffer (Invitrogen), 0.25 mM dNTP, 10 pmol of 5Ј-and 3Ј-primers (including 10,000 cpm of the 5Ј 32 P-radiolabeled 5Ј-primer) for HHF1 (a control intronless gene) was added to the first step PCR reaction, and the second step PCR was performed totaling 20 cycles (94°C for 1 min, 55°C for 1 min, and 72°C for 1 min). PCR products were resolved on 6% polyacrylamide/8 M urea gels and visualized by autoradiography. Under these conditions, the quantities of RT-PCR products were determined to be within the linear range.

RESULTS
pus7 Deletion (pus7⌬) Coupled with Point Mutations in U2 snRNA at Position 40 Yields Temperature-sensitive Yeast Strains-As we have previously demonstrated (29), when incubated in a rich medium (YPD), the yeast pus7⌬ strain, in which U2 pseudouridylation is completely blocked at position 35, shows no obvious growth defect phenotype (Fig. 2B, compare rows 1 and 2). To investigate the functional role of ⌿35, we took advantage of a large collection of yeast U2 snRNAs, each containing a single point mutation near the branch site recognition region. Each of these U2-bearing plasmids was able to substitute for the yeast chromosomal U2 gene without compromising the healthy growth phenotype in rich media. These U2 mutants were combined with pus7⌬ and selected for a synthetic growth defect phenotype. Briefly, we first created a yeast strain in which both the pus7 and U2 (SNR20) genes were deleted, and U2 was supplied by a plasmid containing a wildtype U2 gene and a URA3 selection marker (see "Experimental Procedures"). We then introduced each of the mutant U2 plasmids into the strain and streaked the cells on 5-FOA plates to select against the wild-type U2 plasmid via the URA3 marker. The yeast cells growing on 5-FOA should solely depend on the function of mutant U2 combined with pus7⌬. We screened a total of 35 different U2 mutations (Table I) and identified two temperature-sensitive strains. As shown in Fig. 2A, yeast cells with either pus7⌬ alone or mutation/deletion at position 40 (U40⌬ or U40G) in U2 alone grew normally and showed no obvious growth defect at all temperatures tested. In contrast, yeast cells carrying both pus7⌬ and U2-U40⌬ or U2-U40G mutation (pus7⌬ϩU2-U40G or pus7⌬ϩU2-U40⌬) grew well at 18°C or 23°C, poorly at 30°C, and not at all at 37°C.
To confirm that the growth defect phenotypes were specifically associated with pus7⌬ and the U2 point mutations, we introduced a plasmid containing PUS7 or a plasmid containing a wild-type U2 gene into the temperature-sensitive yeast strains and found that either plasmid rescued the temperature-sensitive growth phenotypes (Fig. 2B, compare rows 7 and 8 with row 6 and rows 11 and 12 with row 10). Taken together, these results suggest that pus7⌬ coupled with a mutation/deletion at U40 in U2 yields temperature-sensitive synthetic lethality.
Yeast Cells with pus7⌬ and U2-U40⌬ Are Defective in U3 Pre-RNA Splicing at 37°C-Because U40 is located immediately downstream of the U2 branch site recognition sequence (Fig. 1), it is possible that mutation at this position, when combined with a blockade of pseudouridylation at position 35 (due to pus7⌬), results in a splicing defect, and thus the observed growth defect, at higher temperatures. To test this possibility, we grew the mutant strains under permissive conditions at 18°C and then shifted them to the nonpermissive temperature (37°C). At various time points after the temperature shift, total RNA was extracted and tested for pre-mRNA splicing by primer extension. Two radiolabeled primers were used. One was complementary to the exon2-intron junction of U3 pre-RNA and should specifically hybridize with two forms of U3 pre-RNA (pre-U3a and pre-U3b), but not with spliced U3 RNA. The other primer was specific for U5 and served as a gel loading control. U3 pre-RNAs did not accumulate in wild-type cells, pus7-deleted cells, or pus7-deleted cells complemented with a PUS7-containing plasmid (Fig. 3, lanes 1-3). When total RNA from U2-U40⌬ cells was used, only a slightly elevated level of U3 pre-RNAs was detected (lane 4). In contrast, cells containing both U2-U40⌬ and pus7⌬ showed a clear accumulation of U3 pre-RNAs (lane 5). Importantly, the levels of U3 pre-RNAs were greatly reduced upon subsequent transformation of a PUS7-containing plasmid into the cells (lane 6). Together, these results suggest that the lack of pseudouridylation at position 35 in combination with the U40 deletion in the U2 snRNA specifically results in a U3 pre-RNA splicing defect.
Conditional Depletion of Pus7p from Mutant U2 Strains (U2-U40⌬ and U2-U40G) Leads to Pre-mRNA Accumulation and Cell Death-Although our temperature shift experiments to measure U3 pre-RNA levels were informative, they were not conclusive because the results could also reflect cellular responses to a sudden temperature change (i.e."heat shock"). For instance, in both the mutant and wild-type strains, the shift from 18 to 37°C elevated the levels of some pre-mRNAs as well as mRNAs (data not shown). Therefore, we took advantage of the widely used conditional promoter shutoff technique to confirm the above splicing results. A plasmid containing the PUS7 gene under the control of the GAL promoter was introduced into the temperature-sensitive yeast strains (pus7⌬ϩU2-U40G and FIG. 2. Selection of the synthetic growth defect. A, pus7⌬ coupled with U2-U40⌬ or U2-U40G mutation (pus7⌬ϩU2-U40⌬ or pus7⌬ϩU2-U40G) results in a temperature-sensitive growth phenotype. A haploid yeast strain (DM2484 or DM2486), in which the chromosomal PUS7 and U2 were deleted and U2 was supplied by a plasmid containing a functional U2 gene and a Ura ϩ selection marker, was transformed with one of the U2 plasmids, each containing a His ϩ marker and a U2 gene with a point mutation/deletion (see Table I). As controls, the parental haploid strain (DM1001), containing the wild-type (WT) PUS7 gene and a plasmid-born wild-type U2 gene, was also transformed with the U2 mutant plasmids. The transformed strains (DM1001*, DM2484*, or DM2486*) were then streaked on selectable solid medium (5FOA, -His). pus7⌬ coupled with a U2-U40⌬ or U2-U40G mutation yielded temperature-sensitive strains: they grew well at 18 or 23°C, poorly at 30°C, and not at all at 37°C (DM2484*(pus7⌬ϩU2-U40⌬), DM2486*(pus7⌬ϩU2-U40⌬), DM2484*(pus7⌬ϩU2-U40G), and DM2486*(pus7⌬ϩU2-U40G)). In contrast, transformed parental strains (DM1001*) that contained a wild-type PUS7 gene, grew normally at all temperatures (DM1001*(WT PUS7ϩU2-U40⌬) and DM1001*(WT PUS7ϩU2-U40G)). Following transformation with a plasmid containing the wild-type U2 gene, the parental strain, DM1001, and strains with pus7⌬ alone (DM2484(pus7⌬ WTU2) and DM2486(pus7⌬ WTU2)) were also streaked on the same solid medium. These control strains grew normally at all temperatures tested (left column). B, a plasmid containing the PUS7 gene or a functional U2 gene rescued the growth phenotype. Temperature-sensitive strains (pus7⌬ϩU2-U40⌬ (row 6) and pus7⌬ϩU2-U40G (row 10)) were grown at 18°C on YPD medium, transformed with a plasmid containing a selection marker and the PUS7 gene (rows 7 and 11) or a functional U2 gene (rows 8 and 12), and then streaked on a solid YPD-drop-out medium (SDϪLeu for the PUS7-containing plasmid and SDϪHis for the U2-containing plasmid) at different temperatures. The growth phenotypes of these strains were rescued at higher temperatures (30 and 37°C). Control strains grew normally, as expected (rows 1, 2, 5, and 9). The healthy growth phenotype of a control strain was not affected by transformation of the PUS7-or functional U2-containing plasmid (compare rows 3 and 4 with row 2). pus7⌬ϩU2-U40⌬). The resulting yeast cells were grown at 30°C, first in galactose medium and then in dextrose medium, thereby shutting off the GAL promoter and hence halting PUS7 transcription. At different time points after switching the medium, total RNA was recovered, treated with DNase, and used to assess U2 pseudouridylation and pre-mRNA splicing.
All yeast strains (wild-type and mutant) grew normally in galactose medium at 30°C (Fig. 4A). However, 18 h after being switched to dextrose medium, mutant yeast strains with chromosomal pus7⌬ and a U2 mutation or deletion at position 40 (closed diamonds or squares, respectively) grew significantly slower compared with the wild-type strain (open circles) or the pus7⌬ strain that had also been transformed with a plasmid containing the PUS7 gene under the control of the P Gal promoter (open triangles). When total RNA recovered from cells either before or 18 h after the medium switch was tested for U2 pseudouridylation, a clear reduction of ⌿35 signal was observed (Fig. 4B 6 -11 and lanes  13-16). Interestingly, for reasons that are yet unclear, the U2-U40G mutation, but not the U2-U40⌬ mutation, also abolished the formation of ⌿42 (compare lanes 4 and 5 with lane 2). Collectively, these results confirm that the depletion of Pus7p in the U2 mutant strains results in a growth defect phenotype as well as a blockade of U2 pseudouridylation at position 35 (and at position 42 for the U2-U40G strain).
We next analyzed pre-mRNA splicing using quantitative RT-PCR. Reverse transcription was performed using random hexamer primers and DNA-free total RNA (as used for U2 pseudouridylation mapping) as a template. The resulting cDNA was then amplified by PCR using 4 pairs of DNA oligonucleotides targeting four randomly chosen genes, including three intron-containing genes, RPL37A, RPL13A, and RPL37B, and the intronless gene, HHF1. Both RPL37A and RPL13A pre-mRNA levels in pus7⌬ϩU2-U40G and pus7⌬ϩU2-U40⌬ strains increased dramatically (Ͼ10-fold for RPL37A and Ͼ3-fold for RPL13A) 18 h after the switch of media (Fig. 4C, top and middle panels, compare lane 5 with  lane 4 and lane 8 with lane 7). The levels of these pre-mRNAs were very low in control strains (lanes 1-3 and 6) and did not change after the medium was switched (data not shown). As expected, the levels of the intronless HHF1 gene remained constant in all strains (lanes 1-8) and did not change after the medium was switched (data not shown). Interestingly, however, the pre-mRNA levels of the other intron-containing gene, RPL37B, apparently were not enhanced (Fig. 4C, bottom  panel). Taken together, our data indicate that the lack of pseudouridylation at position 35 in yeast U2, coupled with a mutation/deletion at position 40 (U40G or U40⌬), significantly impairs the ability of U2 to participate in the splicing of at least a subset of pre-mRNAs. DISCUSSION Using a synthetic growth defect screen, we have determined that Pus7p, which is responsible for ⌿35 formation in yeast U2, is important for pre-mRNA splicing activity. Neither pus7⌬ alone nor mutation/deletion at position 40 in U2 alone had a large impact on pre-mRNA splicing and cell growth. When the two  3. Inhibition of pre-U3 splicing in the temperature-sensitive yeast strains. Yeast strains were first grown at 18°C in YPD to mid-log phase and then shifted to 37°C for another 10 h. Total RNA was isolated and primer extension was performed using two primers specific for U3 pre-RNAs (pre-U3A and pre-U3B) or U5. Lanes 1-4 are controls where template total RNA was isolated from the wild-type (WT) strain (lane 1), from the strain with pus7⌬ only (lane 2), from the strain with chromosomal pus7⌬ but transformed with a PUS7-containing plasmid (lane 3), or from the strain with U2-U40⌬ only (lane 4). In lane 5, total RNA was isolated from the strain with pus7⌬ coupled with U2-U40⌬ (pus7⌬ϩU2-U40⌬). In lane 6, total RNA was isolated from the pus7⌬ϩU2-U40⌬ strain that had been transformed with a plasmid containing the PUS7 gene. The positions of primer extension products of pre-U3A, pre-U3B, and U5 are indicated. The specific hybridization between anti-pre-U3 primer and pre-U3A/B RNAs is also schematized on the right. mutations were combined, however, the cells exhibited a clear temperature-sensitive phenotype. Under nonpermissive conditions (37°C, or even 30°C), the splicing of most pre-mRNAs tested was greatly inhibited, thus providing a direct functional link between ⌿35 (in U2) and pre-mRNA splicing.
We screened 10 nucleotide positions in U2 (a total of 35 different mutations/deletions; Table I). Our data showed that U40 is crucial for pre-mRNA splicing if pseudouridylation at position 35 within the branch site recognition sequence is blocked (Fig. 2). Because U40 is immediately downstream of the branch site recognition sequence (G⌿AGUA) in U2, it may help stabilize the interaction between this sequence and the The closed squares represent the pus7⌬ϩU2-U40⌬ strain transformed with the plasmid containing the PUS7 gene under the control of the P Gal promoter. The closed diamonds represent the pus7⌬ϩU2-U40G strain that had been transformed with the plasmid containing the PUS7 gene under the control of the P Gal promoter. The arrow indicates the time point at which the medium was switched from galactose to dextrose. B, U2 pseudouridylation was analyzed using three different strains, including the wild-type (lanes 1 and 2), pus7⌬ϩU2-U40G transformed with the P Gal -PUS7-containing plasmid (lanes [3][4][5], and pus7⌬ϩU2-U40⌬ transformed with the P Gal -PUS7-containing plasmid (lanes 10 -12). Total RNA was isolated from each of these strains before (lanes 3, 4, 10, and 11) and 18 h after (lanes 1, 2, 5, and 12) the medium was switched. The RNA was assayed for pseudouridylation by CMC modification followed by primer extension using a radiolabeled antisense U2 oligodeoxynucleotide (lanes 2, 4, 5, 11, and 12). Lanes 1, 3, and 10 are controls where CMC modification was omitted. The U2 RNA from pus7⌬ϩU2-U40G cells (lanes 6 -9) or from pus7⌬ϩU2-U40⌬ cells (lanes [13][14][15][16] was also sequenced and subjected to electrophoresis in parallel. The three pseudouridine stops/pauses are indicated by asterisks and arrows. Changes at position 40 in the sequencing lanes are boxed. C, pre-mRNA levels were measured by RT-PCR. Template total RNA used in the study was isolated from various yeast strains, including wild-type (WT) (lane 1), pus7⌬ (lane 2), U2-U40⌬ (lane 3), U2-U40G (lane 6), pus7⌬ϩU2-U40G transformed with the P Gal -PUS7-containing plasmid (pPUS7-Gal; lanes 4 and 5), and pus7⌬ϩU2-U40⌬ transformed with pPUS7-Gal (lanes 7 and 8). For the latter two strains, cells were first grown in galactose medium to the mid-log phase and then switched to YPD (dextrose medium). Total RNA was isolated before (lanes 4 and 7) and 18 h after (lanes 5 and 8) the medium was switched. Reverse transcription was primed with random hexamers, and PCR was performed using a pair of oligodeoxynucleotides (only the forward primer was radiolabeled) designed according to the sequences of the three different pre-mRNAs we analyzed, including RPL37A pre-mRNA (top panel), RPL13A pre-mRNA (middle panel), and RPL37B pre-mRNA (bottom panel). The intronless gene, HHF1, served as a control that was included in each analysis. The positions of RT-PCR products derived from respective pre-mRNAs/mRNA are indicated. branch site in pre-mRNA. A recent NMR study has also shown that conversion of U35 to ⌿35 increases the stability of this RNA-RNA duplex (34,35). It is therefore possible that U40 as well as the branch site recognition sequence, including ⌿35, work in concert to ensure stable binding of the pre-mRNA branch site by U2. Mutations at position 40 or inhibition of pseudouridylation at position 35 alone may not be sufficient to destabilize this base-pairing interaction, whereas changes at both positions may preclude the interaction. Alternatively, mutation at position 40 could also destabilize the U2-U6 Helix III, an interaction that, although not established in the yeast spliceosome (39), appears to be important for mammalian pre-mRNA splicing (10). The hypothesis, linking ⌿35 and U40 to RNA-RNA stability in the spliceosome, is consistent with the fact that the mutant strains (pus7⌬ϩU2-U40G and pus7⌬ϩU2-U40⌬) are heat-sensitive for splicing as well as growth.
Our results also suggest that the role of U2-U40 during splicing is more complex than its possible involvement in a canonical Watson-Crick base-pairing interaction with its partner nucleotide in pre-mRNA. We have tested a total of four pre-mRNAs, two of which (RPL37A and RPL13A) have a purine (G and A, respectively) immediately upstream of their highly conserved branch site (UACUAAC) that could potentially base pair with U40 in U2 snRNA (G-U or A-U pair, respectively). The change of U40 to G40 in U2 would abolish this base-pairing interaction and therefore inhibit splicing. This could explain why these two pre-mRNAs accumulated in the pus7⌬ϩU2-U40G strain. However, U3 pre-RNA, which has a pyrimidine (U) at the equivalent position and should improve pairing with U2-U40G mutant, was also accumulated in the pus7⌬ϩU2-U40G strain (data not shown). Clearly, this result does not agree with the simple Watson-Crick base-pairing interaction involving U40 in U2 and its partner in pre-RNA. Also, we observed that both U2-U40G and U2-U40⌬, when coupled with pus7⌬, had the same inhibitory effect on the splicing of RPL37A and RPL13A pre-mRNAs (Fig. 4C). The basepairing model would not predict this outcome because position 41 of U2 is C, a pyrimidine as well. In other words, deletion of U40 from U2 would place C41 at the original position 40, thus restoring the base pairing with the purine (especially G) in the pre-mRNA, thereby driving the splicing reaction. Furthermore, we also tested the U40A mutant in U2 (Table I), but this mutation did not yield a growth defect phenotype. Like U40G, the U to A change at position 40 would also abolish the base-pairing interaction with RPL37A or RPL13A pre-mRNA, which have a G or A, respectively, at the corresponding position. An alternate hypothesis would be that U40 contributes to protein binding. In this regard, it has been reported that U40 is within the sequence involved in binding with SF3a and SF3b subunits and Prp5p (the DEAD protein) (39). It is therefore possible that mutation/deletion at this position results in a less stable binding with SF3 subunits and/or Prp5p. This defect, when coupled with the lack of ⌿35 in the branch site recognition sequence (and thus a less stable interaction with the pre-mRNA branch site), may result in a splicing defect. Clearly, further work is needed to address how ⌿35 and U40 contribute to pre-mRNA splicing.
It should be noted that because Pus7p also catalyzes tRNA pseudouridylation at certain positions (45), our data cannot exclude the possibility that pseudouridylation on tRNAs contributes to the observed phenotype. However, we think this possibility unlikely for at least three reasons. First, we observed a splicing defect in the mutant strains, and there is no evidence that a deficiency in tRNA pseudouridylation or any alteration in tRNA affects pre-mRNA splicing. Second, a normal phenotype was observed in a yeast strain carrying the pus7⌬ genotype only, and the splicing/growth defect was only evident when an additional U40 mutation in U2 was intro-duced. Because both U40 and ⌿35 are in U2, not tRNA, it is likely that the U40 mutation along with the inhibition of pseudouridylation at position 35 abolishes U2 splicing function at nonpermissive temperatures. Third, we did not observe abnormal tRNA processing in these strains by Northern analysis (data not shown), although some other steps, such as transport, amino-acylation, protein translation, etc., may still have been affected.
It is also interesting that the U40G mutation abolishes U2 pseudouridylation at position 42. It is possible that the mutation completely changes the element necessary for the ⌿42specific enzyme (protein only or sno/scaRNP) to recognize and catalyze the reaction, which would thereby abolish ⌿42 formation. Curiously, however, the deletion of U40 from U2 had no effect on the formation of ⌿42. At present, the mechanism of ⌿42 formation is still an open question. On the other hand, because the U40G mutation inhibits ⌿42 formation, we cannot rule out the possibility that the splicing defect observed in this mutant strain is a direct result of the three alterations in U2: the U40G mutation and the lack of pseudouridylation at positions 42 and 35. However, for at least two reasons, we think that the lack of ⌿42 has a minimal, if any, impact on pre-mRNA splicing in the pus7⌬ϩU2-U40G strain. First, the defects in both splicing and growth of both pus7⌬ϩU2-U40G and pus7⌬ϩU2-U40⌬ strains are virtually identical (Figs. 2 and 4), despite the fact that pseudouridylation at position 42 differs in these two strains. Second, none of the U2-U42 mutants assayed (Table I) shows a growth defect despite the lack of ⌿42, even combined with either pus7⌬ or U40G mutation, suggesting that position 42 is not as important as positions 35 and 40 in contributing to pre-mRNA splicing and cell growth. Further study is necessary to clarify this issue.