Analysis of Synthetic Lethality Reveals Genetic Interactions Between the GTPase Snu114p and snRNAs in the Catalytic Core of the Saccharomyces cerevisiae Spliceosome

Conformational changes of snRNAs in the spliceosome required for pre-mRNA splicing are regulated by eight ATPases and one GTPase Snu114p. The Snu114p guanine state regulates U4/U6 unwinding during spliceosome activation and U2/U6 unwinding during spliceosome disassembly through the ATPase Brr2p. We investigated 618 genetic interactions to identify an extensive genetic interaction network between SNU114 and snRNAs. Snu114p G domain alleles were exacerbated by mutations that stabilize U4/U6 base pairing. G domain alleles were made worse by U2 and U6 mutations that stabilize or destabilize U2/U6 base pairing in helix I. Compensatory mutations that restored U2/U6 base pairing in helix I relieved synthetic lethality. Snu114p G domain alleles were also worsened by mutations in U6 predicted to increase 5′ splice site base pairing. Both N-terminal and G domain alleles were exacerbated by U5 loop 1 mutations at positions involved in aligning exons while C-terminus alleles were synthetically lethal with U5 internal loop 1 mutations. This suggests a spatial orientation for Snu114p with U5. We propose that the RNA base pairing state is directly or indirectly sensed by the Snu114p G domain allowing the Snu114p C-terminal domain to regulate Brr2p or other proteins to bring about RNA/RNA rearrangements required for splicing.

I NTRONS in pre-messenger RNA (pre-mRNA) are removed by the spliceosome to yield mature mRNA, which is then exported into the cytoplasm for translation. The process of pre-mRNA splicing has been reviewed extensively (Moore et al. 1993;Will and Lü hrmann 2006). Splicing is carried out in the nucleus by the assembly of small nuclear ribonucleoprotein particles (or snRNPs) onto the pre-mRNA and occurs by two transesterification steps, which are carried out within the spliceosome by a series of RNA:RNA interactions and RNA:protein rearrangements. Five snRNPs take part in splicing, U1, U2, U4, U5, and U6, of which the last three form a tri-snRNP in which U4 is base paired with U6 (Moore et al. 1993). Although it is thought that the RNAs within snRNPs are responsible for catalyzing the two steps of splicing, it is the many proteins, in particular the NTPases, that bring about the multiple rearrangements that are required for catalysis by the snRNAs (Smith et al. 2008). Key remodelling steps during splicing include unwinding of U4/U6 helices, formation of U2/U6 interactions, unwinding of U2/U6 helices and reestablishment of U4/U6 complexes.
Remodelling is facilitated by eight ATPases and one GTPase, Snu114p, one of the key protein factors required for the regulation of spliceosomal dynamics (reviewed in Frazer et al. 2008).
The single GTPase of the spliceosome, Snu114p (U5-116kDa or hSnu114 in humans), is a protein of the U5 snRNP (Fabrizio et al. 1997). It shares significant sequence homology with translation elongation factor 2 (EF-2) and contains five domains similar to EF-2, with domain I containing the G domain consensus sequence elements G1-G5 that are important for binding and hydrolysis of GTP. hSnu114 has been found to crosslink to GTP specifically and a mutation in Snu114p, expected to abolish GTP binding, is lethal, suggesting that GTP binding and hydrolysis are necessary for Snu114p function in vivo (Fabrizio et al. 1997; Bartels et al. 2003). Snu114p also has a unique acidic amino terminus (N terminus) that is implicated, along with the G domain, in U4/U6 unwinding during spliceosome assembly (Bartels et al. 2002).
The relationship of Snu114p to other spliceosomal proteins has been investigated using conditionally lethal alleles to test for genetic interactions with other splicing factors (Brenner and Guthrie 2005). Genetic interactions of snu114 alleles were found with prp8, brr2, prp28, prp19, sad1, and snu66 alleles. The genetic interactions of SNU114 with PRP8 and BRR2 were expected considering the proteins form a stable complex within the U5 snRNP in human cells . Other genetic interactions were with genes that are all known to act before the first step of splicing, suggesting an important role for Snu114p during spliceosomal activation (Brenner and Guthrie 2005). Although genetic interactions do not necessarily indicate direct interactions, two hybrid studies have confirmed interactions between hSnu114, hPrp8, and hBrr2 (Liu et al. 2006) and other studies have indicated a requirement for Prp8p and Snu114p for correct U5 snRNP assembly (Brenner and Guthrie 2006). Whether Snu114p function requires GTP hydrolysis was addressed by analysis of spliceosome assembly and disassembly, both of which were promoted by Snu114p and Brr2p (Small et al. 2006). The guanine nucleotide state of Snu114p, rather than GTP hydrolysis, regulated Brr2p activity during snRNA unwinding.
Compared to the number of known Snu114p-protein interactions, little is known about how Snu114p interacts with snRNAs of the spliceosome. Evidence of direct Snu114p-RNA interactions include hSnu114 crosslinks to hairpins introduced in pre-mRNA and hSnu114 crosslinks with the pre-mRNA pyrimidine tract after spliceosome assembly but prior to the second step (Chiara et al. 1997;Liu et al. 1997). This suggested a role for hSnu114 in bringing the 39 splice site into the active site for the second step of splicing. Parallels have been drawn between the function of Snu114p in the spliceosome and EF-2 in the ribosome (Staley and Guthrie 1998). The interaction between U5 loop 1 and the 59 exon during splicing was proposed to be analogous to the tRNA anticodon-codon interaction during translation, with a conformational change driving Snu114p to reposition the 59-exon-bound U5 loop 1 for the second step of splicing (Staley and Guthrie 1998). Although Snu114p was found to crosslink to the 59 side of internal loop 1 of U5 snRNA (Dix et al. 1998), a model for Snu114p function based on GTP driven hydrolysis is no longer supported (Small et al. 2006).
As splicing is most likely RNA catalyzed, it is important to determine any interactions between Snu114p and the snRNAs involved in catalysis. We assessed synthetic lethal interactions of SNU114 with the snRNAs that contribute to the catalytic core of the spliceosome, including U4, which dissociates from the spliceosome prior to assembly of the active complex. We found that snu114 G domain mutations were synthetically lethal with mutations in U4 that stabilize the upper portion of U4/U6 stem I and only one mutation within stem I. Similarly, snu114 G domain mutations were synthetically lethal with mutations in U2 and U6 predicted to stabilize or destabilize U2/U6 base pairing in helix I. Synthetic lethality could be suppressed by compensatory mutations that restored U2/U6 base pairing in helix I. In addition, synthetically lethal interactions of G domain mutations of snu114 with U6 snRNA alleles indicated a role for Snu114p during U6 base pairing interactions with the 59 splice site. These results suggest that the Snu114p G domain is important for sensing the integrity of RNA/RNA interactions in the catalytic core of the spliceosome. The role of Snu114p within the catalytic core was confirmed by genetic analysis with U5. We found that snu114 N-terminal and G domain mutations were synthetically lethal with U5 loop 1 alleles in positions important for tethering the exons during splicing and that C-terminal mutations were synthetically lethal with alleles in U5 internal loop 1. These results suggest a spatial orientation for Snu114p with U5 and a possible model for regulation. The Snu114p G domain, and to some extent the N-terminal domain, may sense the state of the snRNAs in the catalytic core and, possibly through a structural rearrangement with U5, transmit this information to the C-terminal domain to regulate Brr2p or other proteins to bring about the rearrangements in the spliceosome required for splicing.

MATERIALS AND METHODS
Yeast strains and manipulation: For analysis of SNU114 mutations, diploid yeast strain Y25023 (Table 1) was obtained from EUROSCARF and transformed with pRS416-Snu114. All transformations were carried out using the lithium acetate procedure (Gietz et al. 1992). G418-resistant and Ura 1 transformants were sporulated and tetrads dissected. Haploid progeny were checked for disruption of SNU114 with kanMX4 by PCR of genomic DNA to identify the strains YSNU114KO1 and YSNU114KO2 (Table 1). SNU114 knockout strain, YSNU114KO1, containing pRS416-Snu114 was transformed with each snu114 allele on a separate plasmid. Transformants were transferred to 5-fluoroorotic acid (5-FOA) to select against the URA3 plasmid by plasmid shuffle (Boeke et al. 1984) and to determine the viability of snu114 alleles as the sole source of Snu114p. Cells viable on 5-FOA were tested for temperature sensitivity by streaking onto yeast peptone dextrose (YPD) agar and incubating at 16°, 25°, 30°, 37°, 39°, and 40°for 3-10 days.
To test genetic interactions between snu114 and snRNA alleles, double deletion strains were constructed by mating strain YSNU114KO2 with strain YU5KO (O'Keefe 2002), YU2KO or YU6KO (Table 1). Mated diploid strains were transferred to 5-FOA to generate diploid strains lacking pRS416 plasmids that were complementing the SNU114 and snRNA knockouts. Diploid strains with single SNU114 and individual snRNA deletions were transformed with a pRS416 plasmid containing both wild-type (WT) genes (pRS416-Snu114-U5, pRS416-Snu114-U2, or pRS416-Snu114-U6). Viable G418-resistant and Ura 1 transformants were sporulated and tetrads dissected. The haploid progeny were analyzed for disruption of SNU114 and the relevant snRNA gene by PCR of genomic DNA and by plasmid shuffle for the requirement of Snu114p and U5, U2 or U6 to identify the strains YSNU114/ U5KO, YSNU114/U2KO, and YSNU114/U6KO (Table 1). YSNU114/U4KO was constructed by replacing the U4 gene with the hphNT1 resistance gene ( Janke et al. 2004) in the SNU114 diploid knockout strain Y25023 with a PCR product amplified from pYM24 using primers U4hphNT1F and U4hphNT1R (supporting information, Table S1). G418resistant and hygromycin B-resistant cells were transformed with pRS416-Snu114-U4, sporulated, and tetrads dissected. Haploid progeny were checked for integration of kanMX4 and hphNT1 by PCR of genomic DNA and in vivo for the requirement of both Snu114p and U4 to identify the strains YSNU114/U4KO1 and YSNU114/U4KO2 (Table 1). Strains YSNU114/U5KO, YSNU114/U2KO, YSNU114/U6KO, and YSNU114/U4KO1 (Table 1) were transformed with snu114 and snRNA alleles on different plasmids and then tested for synthetic lethality by plasmid shuffle.
Plasmid construction and mutagenesis: SNU114 was cloned by PCR amplification using Pfu DNA polymerase (Stratagene) of yeast genomic DNA (Promega) using primers Snu114F and Snu114B (Table S1) complementary to regions upstream and downstream of the open reading frame and containing restriction sites for EcoRI and BamHI. The PCR product was digested with EcoRI and BamHI, then ligated into pRS416 (Sikorski and Hieter 1989) to produce pRS416-Snu114. pRS416 plasmids containing SNU114 and each snRNA gene were made in the following manner: pRS314-U5 (m571) (O'Keefe et al. 1996) was digested with EcoRI/XhoI and the U5 gene fragment was ligated into EcoRI/XhoI digested pRS416-Snu114 to generate pRS416-Snu114-U5. pRS415-U2 (McGrail et al. 2006) was digested with BamHI, filled in with T4 DNA polymerase (Roche), and then digested with XhoI. The U2 gene fragment was ligated into pRS416-Snu114 digested with HindIII, filled in as above, and then digested with XhoI to generate pRS416-Snu114-U2. pRS413-U6 (McGrail et al. 2006) was digested with SacI, blunt ended with T4 DNA polymerase, then digested with XhoI. The U6 gene fragment was ligated into pRS416-Snu114p digested with HindIII, filled in as above, and then digested with XhoI to generate pRS416-Snu114-U6. pRS416-Snu114-U4 was constructed by PCR amplification of the U4 gene using primers U4GF and U4GR containing restriction sites for BamHI and KpnI (Table S1). The PCR product was digested with BamHI and KpnI and then ligated into pRS416-Snu114 to produce pRS416-Snu114-U4. The PCR product was also digested with BamHI and ligated into a BamHI/HincII-digested pRS413 to produce pRS413-U4. The identity of all plasmids was confirmed by sequencing.
Conditionally lethal snu114 alleles were generated by identifying hydrophobic stretches of amino acids in an alignment of SNU114 with the yeast elongation factor, EFT2.
The three-dimensional model of Snu114p was constructed by comparative modeling. Yeast elongation factor 2 (Eft2p, pdb code 1n0v) ( Jorgensen et al. 2003) was selected as the template structure as it was the best sequence match based on BLAST score (Altschul et al. 1990). The sequences of Snu114p and Eft2p were aligned using ClustalW (Larkin et al. 2007). The N-terminal domain (residues 1-121) and a region from the C terminus (residues 982-1008) were omitted from the model as there are no equivalent residues in the crystal structure of Eft2p. The percentage of sequence identity over the aligned region was 25.3%. Regions of buried hydrophobic amino acids within the structure of Eft2p were identified using Modeler (Sali and Blundell 1993) and corresponding homologous regions in Snu114p were then identified and targeted for mutagenesis. This study Ten three-dimensional models were built using Modeler (Sali and Blundell 1993) and the model with the lowest pseudo energy (that when combined best fit to all input data used to generate the model) was chosen.
Snu114 protein levels: To analyze protein levels, overnight YPD cultures of cells containing WT or mutated snu114 as the only source of Snu114p were used to inoculate YPD cultures to an OD 600 of 0.5 and grown for at least three generations at permissive (30°) or restrictive (39°) temperatures. Cells were sedimented by centrifugation and washed with water prior to being lysed in SDS-loading buffer with acid-washed glass beads (Sigma) by vigorous vortexing. Cells were boiled for 5 min prior to centrifugation to remove cell debris and extracts removed and then separated on 12% acrylamide PAGEr gels (Lonza). Levels of Snu114p were detected by Western blots with rabbit anti-Snu114p antibodies generated against a peptide encompassing the first 15 amino acids of Snu114p and standardized by probing with mouse anti-Zwf1p antibodies (Sigma).
Primer extensions: To determine whether snu114 strains exhibited splicing defects, total RNA was extracted with hot phenol (Kohrer and Domdey 1991) from strain YSNU114KO1 in the presence of pBK413-based WT or mutated snu114 after incubation in YPD for 16 hr at the restrictive temperature of 39°. Equal amounts of total RNA were hybridized to a 32 P endlabeled oligonucleotide primer, U3snoRNA-RT (Table S1), and primer extensions were carried out as described previously (Dobbyn and O'Keefe 2004).
Genetic analysis of synthetic lethality: YSNU114/U5KO was transformed with mutated pBK413-Snu114 and pRS414-U5 and transformants selected on synthetic defined (SD) medium (SD À Ura À His À Trp) (BIO 101 Systems). YSNU114/ U2KO was transformed with mutated pRS413-Snu114 and pRS415-U2, YSNU114/U6KO with mutated pRS415-Snu114, and pRS413-U6 and YSNU114/U4KO1 with mutated pRS415-Snu114 and pRS413-U4 and the transformants selected on SD À Ura À His À Leu. Transformants were then tested for synthetic lethality by plasmid shuffle on 5-FOA. Synthetic lethality was scored as the lack of growth after 3 days at 30°. Synthetic sickness was scored as growth of smaller or fewer colonies compared to control sectors after 3 days at 30°.  Compensatory analysis: To determine whether the synthetic lethality observed between snu114 and snRNA mutations could be alleviated, synthetically lethal combinations were transformed with a plasmid containing an snRNA mutation that restored base pairing in U4/U6 stem I or U2/U6 helices. This was achieved by producing a vector with MET15 as the selection marker. A 2491-base pair fragment containing MET15 (Cost and Boeke 1996) was amplified by PCR from yeast genomic DNA with primers Met15F and Met15B (Table  S1). The Met15F primer contained AfeI and the Met15B primer contained BamHI and PstI restriction enzyme recognition sites. The PCR product was digested with BamHI and cloned into pRS414 digested with HincII and BamHI. pRS414 plasmids containing the MET15 fragment were tested for complementation of MET15 in YSNU114/U4KO2. The MET15 was released from a complementing plasmid with AfeI and PstI, and then cloned into pRS413 where most of the HIS3 gene had been removed by digestion with AfeI and NsiI to produce the plasmid pRS413MET15. Mutations in the U2, U4, and U6 snRNAs were transferred into this plasmid.
To assess suppression of Snu114-U2 synthetic lethality, YSNU114/U2KO was transformed with synthetically lethal mutation combinations of pBK413-Snu114 and pRS415-U2, as well as pRS413MET15 containing U6 mutations to restore U2/ U6 base pairing in the presence of wild-type U6. To assess suppression of Snu114-U4 synthetic lethality, YSNU114/U4KO2 was transformed with synthetically lethal combinations of pRS415-Snu114 and pRS413-U4, as well as pRS413MET15 containing a U6 mutation to restore base pairing in U4/U6 in the presence of wild-type U6. To assess suppression of Snu114-U6 synthetic lethality, YSNU114/U6KO was transformed with synthetically lethal mutation combinations of pRS415-Snu114 and pRS413-U6, as well as pRS413MET15 containing U2 mutations to restore U2/U6 base pairing or U4 mutations to restore U4/U6 base pairing in the presence of wild-type U2 or U4. Transformants were selected on SD À Ura À His À Leu À Met agar. Transformants were then tested for growth by plasmid shuffle on 5-FOA alongside the synthetically lethal combinations and controls. Suppression of synthetic lethality was scored after 3 days at 30°.

RESULTS
Identification of additional snu114 mutations that are temperature sensitive and show splicing defects at the first step of splicing: Forty-nine single-and four double-amino-acid-substitution mutations were generated within five predicted hydrophobic regions of Snu114p: L234-I239, C264-I267, and I311-A314 in domain I; I561-I564 in domain II; and P860-I861 in domain IVa. Nineteen mutations were lethal and of the viable alleles five were temperature sensitive (Ts) at 39°(V238D, V266P, F313Q, L563P, and P860K) and one was Ts at 37°( L563F) ( Table 2). The Ts mutations mapped to different locations within Snu114p ( Figure 1A). The V238D mutation was proximal to the putative salt bridge at D233 (with R488) and the V266P mutation was near the critical (N/T)(K/Q)XD G4 motif. This motif is required for GTP binding and hydrolysis (Fabrizio et al. 1997) and stabilizing residues in the G1 motif (Bourne et al. 1991). The F313Q mutation was adjacent to the (G/T)SAL G5 motif, which stabilizes residues in the G4 motif (Bourne et al. 1991), L563F was within domain II, and P860K was in domain IVa ( Figure 1A). Previously published Ts snu114 alleles (K146I, P216N, L381P, G646R, and M842R) (Brenner and Guthrie 2005) are positioned in different locations within the predicted structure from the Ts alleles generated in this study ( Figure 1B). An N-terminal snu114 deletion mutation lacking the first 128 amino acids (DN) (Bartels et al. 2002) was also produced ( Table 2). For clarity, snu114 alleles are annotated with their putative positions within Snu114p, for example, V266P (G4) signifies the mutation is in G domain motif G4 and P860K (IVa) signifies the mutation is in domain IVa.
To determine whether the temperature sensitivity of snu114 alleles was due to defects in protein expression or in pre-mRNA splicing, a SNU114 knockout strain carrying WT or snu114 allele on a plasmid was grown at the permissive and restrictive temperatures. Four strains chosen for analysis were V266P (G4), F313Q (G5), L563F (II), and P860K (IVa) (Figure 2A) as these were the most reliable Ts mutations that spanned the predicted three-dimensional structure of Snu114p (Figure 1B). Protein levels in non-Ts [WT and V266K (G4)] and Ts strains were equivalent at the permissive temper- ature as demonstrated by Western blotting ( Figure 2B, lanes 1-7). When cells were grown at the restrictive temperature for at least three generations, protein levels remained constant in WT ( Figure 2B, lane 8), non-Ts ( Figure 2B, lane 9), and Ts strains ( Figure 2B, lanes 10-11 and 13-14). The exception was the mutation L563P (II) ( Figure 2B, lanes 5 and 12), which displayed a slight decrease in Snu114p levels compared to WT at both permissive and restrictive temperatures and so was excluded from further investigations. Thus the Ts phenotype observed in snu114 mutant strains was not due to aberrant levels of protein expression. To assess whether the temperature sensitivity of snu114 mutant strains was due to a defect in pre-mRNA splicing, total RNA was extracted from strains grown overnight at the restrictive temperature. The levels of spliced and unspliced U3A and U3B snoRNAs were analyzed by primer extension with a primer complementary to their  (Eft2p). Domains are labeled in Roman numerals. The N-terminal domain is unique to Snu114p. Within domain I, the G domain, the GTPase conserved motifs are labeled 1-5. The G domain also contains a G$ element which is unique to Snu114p and Eft2p. Ts snu114 mutations used to analyze genetic interactions with snRNAs are displayed above the linear diagram and mutations from a previous study (Brenner and Guthrie 2005) are shown below the linear diagram. (B) Three-dimensional predicted structure of Snu114p based on sequence homology with Eft2p. The N-terminal domain (residues 1-121) and a region from the C terminus (residues 982-1008) were omitted. Colors correspond to the linear diagram of Snu114p. Yellow spheres indicate location of amino acids mutated in this study and red spheres indicate amino acids mutated by Brenner and Guthrie (2005). strains. Serial dilutions of WT and snu114 strains grown at 16°for 10 days or 25°, 30°, 37°, and 39°for 3 days. snu114 mutant L563F exhibits a growth defect at 37°whereas the remaining snu114 mutants exhibit a growth defect at 39°. (B) Protein expression levels of Snu114p monitored at permissive (30°) and restrictive temperatures (39°). Protein levels in WT and snu114 strains remained constant after a shift to 39°compared to the loading control, yeast glucose 6-phosphate dehydrogenase (Zwf1p). (C) snu114 mutations inhibit pre-mRNA splicing. RNA was extracted from WT or snu114 strains grown at the restrictive temperature and analyzed by primer extension of pre-U3 snoRNA. Pre-U3A, pre-U3B, and mature U3 snoRNA are indicated to the right of the panel. common exon 2. In WT extracts, only a small amount of unspliced U3A or U3B precursor was detectable ( Figure  2C, lane 3). In extracts from snu114 mutant strains, there was a significant accumulation of unspliced U3A and U3B precursor ( Figure 2C, lanes 4-7) compared to levels in WT extracts, indicating that splicing was inhibited and that the block in splicing was prior to the first step. Note that a second step block could not be detected by primer extension in this manner.
snu114 mutations display synthetic lethality with U4 snRNA mutations that alter U4/U6 base pairing: The interactions of U4 with U6 serve to regulate the U6 snRNA during assembly and activation of the spliceosome. Unwinding of U4 from U6 during spliceosome assembly is catalyzed by the DExD/H-box helicase Brr2p (Laggerbauer et al. 1998;Raghunathan and Guthrie 1998), which is regulated by the guanine nucleotide state of Snu114p (Small et al. 2006). It is not clear whether the regulation of Brr2p activity by Snu114p is through direct or indirect interactions with Brr2p, another protein, or the U4 or U6 snRNAs. Genetic interactions were investigated between snu114 alleles and mutations in U4 positions that base pair with U6 snRNA in stems I and II of the U4/U6 complex and U4 mutations adjacent to stem I (Shannon and Guthrie 1991;Hu et al. 1995;Li and Brow 1996) as shown in Figure 3. A mutation adjacent to U4 stem I, U4-cs1 (A66U,A67U,A68G), stabilizes U4/U6 stem I by increasing base pairing with the U6 ACAGAGA region thereby blocking U4/U6 dissociation early in splicing (Li and Brow 1996;Kuhn et al. 1999). Genetic analysis revealed that K146I (G1), V266P (G4), and F313Q (G5) demonstrated synthetic lethality with U4-cs1 ( Figure 3 and Table 4). Interestingly, P216N (G2) and L381P (G$) were not synthetically lethal but noticeably sick with U4-cs1 ( Figure 3 and Table 4). To analyze the effects of other mutations that extend base pairing in the same region of stem I as U4-cs1, mutations U4-U64C, U4-G65A, and U4-U64C,G65A were tested for synthetic lethality with snu114 mutations. Genetic analysis revealed that K146I (G1) was synthetically lethal with the double mutation U4-U64C,G65A and synthetically sick with U4-U64C and U4-G65A (Figure 3 and Table 4). Therefore, it appears that extending U4/U6 base pairing in stem I results in synthetic lethality with snu114 G domain mutations.
Although many positions in U4 stem I have been shown to be tolerant to mutation, only U4-G58 has been shown to have any phenotypic effects (Hu et al. 1995). Analysis of genetic interactions between mutations in snu114 and U4 stem I (U57A, G58A, C59U, U60C, U64G) revealed that only K146I (G1) was synthetically lethal with U4-G58A (Figure 3 and Table 4). No other snu114 alleles were synthetically lethal with U4-G58A (Table 4) although V266P (G4), F313Q (G5), G646R (III), and M842R (IVa) were synthetically sick ( Figure 3 and Table 4). While the U4-U57A, U4-G58A, U4-C59U, U4-U60C, and U4-U64G mutations are all predicted to disrupt base pairing between U4 and U6, U4-G58A is the only mutation that disrupts stem I leading to a Ts phenotype (Hu et al. 1995) and displays synthetic lethality and sickness with snu114 mutations in the G domain and domains III and IVa. These results indicate that only mutation at this nucleotide of U4 in stem I exacerbates snu114 mutations. Figure 3.-Genetic interactions between snu114 and U4 snRNA mutations. Strain YSNU114/U4KO1 containing the WT SNU114 and U4 snRNA alleles on a URA3 plasmid was transformed with WT or snu114 and U4 snRNA alleles on separate plasmids. Transformants were transferred to 5-FOA to evict the URA3 plasmid containing the wild-type genes and determine the synthetic lethality between mutant alleles. The annotated structure of base paired U4/U6 in the conformation within the tri-snRNP is displayed above the panels. Positions of substitutions are underlined. The U4-cs1 mutation is A66U,A67U,A68G.
snu114 mutations display synthetic lethality with U6 snRNA mutations important for base pairing with the 59 splice site and that alter U2/U6 base pairing: The role of U6 snRNA is pivotal in the spliceosome as it is involved in three important base pairing interactions during splicing: with U4 in the U4/U6 di-snRNP (Madhani et al. 1990), with the 59 splice site via a conserved single-stranded ACAGAGA (A47-A53) sequence following tri-snRNP incorporation into the spliceosome (Madhani et al. 1990;Sawa and Shimura 1992;Sontheimer and Steitz 1993) and with U2 snRNA in helix I in the active site before the first step of splicing (Madhani and Guthrie 1992). As Snu114p regulates Brr2p activity and, as shown above, interacts genetically with mutations in U4 that perturb the U4/ U6 complex we determined whether SNU114 interacts genetically with U6.
Interestingly, U6-G50U and U6-G52U both potentially increase the base pairing of U6 with the conserved 59 splice site sequence. Therefore, it appears that stabilization of the interaction of U6 with the 59 splice site exacerbates snu114 G domain mutations.
Genetic interactions were also investigated between snu114 mutations and U6 insertion mutations at positions A53/U54 and G55/A56 to determine whether extending the distance between the ACAGAGA sequence and U4/U6 stem I or U2/U6 helix I had any effects. No synthetic lethality was observed between snu114 alleles and U6 mutations Ins 1U A53/U54 and Ins 1U G55/A56 (Table 5). Substitution mutations at U6-G55 and U6-A56 are expected to weaken stem I by disrupting base pairing between U4 and U6 snRNAs (Li and Brow 1996) and insertion mutations at U6-G55/ A56 have been shown to progressively inhibit the first step of splicing (McGrail et al. 2006). Synthetic lethality was observed between snu114 mutations and substitution mutations at U6-U54 and U6-G55, which would be expected to extend the base pairing at the top of U4/U6 stem I or disrupt the base pairing in U2/U6 helix I. In particular, K146I (G1) was synthetically lethal with U6-U54C, K146I (G1) and V266P (G4) with U6-G55A, and K146I (G1) and V266P (G4) with U6-U54C,G55A (Figure 4). In addition, F313Q (G5) was synthetically sick with all three substitution mutations U6-U54C, U6-G55A, and U6-U54C,G55A (Table 5). This suggests that Further analysis revealed synthetically lethal interactions between snu114 alleles and U6 mutations in the region that mutually exclusively base pairs with U4 in stem I prior to spliceosome activation or with U2 in helix I. Double mutations at U6-A56,U57 are conditionally lethal but can be suppressed by U2 mutations, which restore helix Ia base pairing (Madhani and Guthrie 1992). Our analyses revealed that K146I (G1) and V266P (G4) were synthetically lethal with U6-A56C,U57C (Figure 4 and Table 5). The DN, P216N (G2), F313Q (G5), L381P (G$), and M842R (IVa) mutations were also synthetically sick with U6-A56C,U57C (Figure 4), but no other snu114 alleles were synthetically lethal with U6-A56C,U57C (Table 5). In addition, the only other genetic interaction in this U6 Figure 4.-Genetic interactions between snu114 and U6 snRNA mutations. Strain YSNU114/U6KO containing the WT SNU114 and U6 snRNA alleles on a URA3 plasmid was transformed with WT or snu114 and U6 snRNA alleles on separate plasmids. Transformants were transferred to 5-FOA to evict the URA3 plasmid containing the wild-type genes and determine the synthetic lethality between mutant alleles. The annotated structure of base paired U2/U6 in the helix I and U6/59 splice site conformations of intermolecular interactions required for splicing are displayed above the panels. Positions of substitutions are underlined. region was a synthetically sick interaction between V266P (G4) and U6-C58U (Table 5).
Mutations in snu114 and the U6-AGC triad that base pairs with U4 to form stem I in the U4/U6 complex or with U2 to form helix Ib in the U2/U6 complex were investigated for synthetic lethality. In yeast, substitutions in U6-G60 and U6-C61 result in defects in 59 splice site cleavage and substitutions in U6-A59 cause defects in exon ligation and inhibit the second step of splicing (Fabrizio and Abelson 1990;McGrail et al. 2006). Synthetic lethality was observed between both V266P (G4) and F313Q (G5) and U6-A59C ( Figure 4 and Table  5). In addition, DN and K146I (G1) were synthetically sick with U6-A59C ( Figure 4 and Table 5). Further analysis revealed that all snu114 G domain mutations, as well as DN and M842R (IVa), were synthetically lethal with U6-C61G (Figure 4 and Table 5). No other snu114 alleles were synthetically lethal with U6-C61G (Table 5).
Overall, it appears that snu114 mutations in the N terminus and the G domain are exacerbated by mutations in U6 that disrupt U2/U6 base pairing within helices Ia and Ib.
Finally, genetic interactions between snu114 and U6-A62C were investigated to assess whether disrupting the base pairing with U4-U57 at the base of U4/U6 stem I or with U6-C85 in U6 internal stem loop (ISL) or whether extending the base pairing in helix Ib with U2-G20 caused synthetic lethality. There was no observed synthetic lethality between U6-A62C and any snu114 mutation (Table 5). This supports the lack of synthetic lethality of snu114 G domain mutations with U4-U57A and suggests that Snu114p does not interact genetically with U6-ISL at position U6-A62. These data also suggest that although extending the base pairing of U2/U6 helix Ia with U6 mutations exacerbated snu114 G domain mutations, extending the base pairing of U2/ U6 helix Ib does not have the same effect.
snu114 mutations display synthetic lethality with U2 snRNA mutations that alter U2/U6 base pairing: Two important base pairing interactions of the U2 snRNA are with the branchpoint sequence (BPS) early in spliceosome assembly (Moore et al. 1993) and with U6 to form helix Ia and Ib in the spliceosome active site (Madhani and Guthrie 1992). The U2/U6 helices serve to bring the 59 exon together with the BPS adenosine ready for the first step of splicing (Madhani and Guthrie 1992). We have shown that mutations in the N-terminal and G domains of Snu114p display synthetic lethality with mutations in U6 that increase base pairing with the 59 splice site and disrupt or extend U2/U6 base pairing. Next we determined whether there were genetic interactions between mutations in snu114 and the U2 snRNA. Mutations in snu114 were tested for synthetic lethality with U2 helix Ib mutations at positions U2-C22 and U2-U23 (Madhani and Guthrie 1992). U2-C22 forms intramolecular base pairing interactions with U6-AGC triad nucleotide U6-G60 of helix Ib. U2-U23 is implicated in 59 splice site recognition, base pairs with U6-A59 of helix Ib before the second step of splicing, crosslinks to the 39 exon in the 39 exon-lariat intermediate, and is important for the second step of splicing Guthrie 1992, 1994;Newman et al. 1995;Luukkonen and Séraphin 1998). Our analyses revealed synthetic lethality between V266P (G4) and both mutations at U2-C22 and U2-U23, while F313Q (G5) was synthetically sick with both mutations at -, synthetically lethal interaction; 1/À, synthetically sick interaction; 1, WT growth after 3 days at 30°.
SNU114 and snRNA Genetic Interactions U2-U23 ( Figure 5 and Table 6). No other snu114 alleles were synthetically lethal with these U2 mutations (Table  6). Interestingly, there was no synthetic lethality between any snu114 alleles and several mutations in the bulge, positions U2-U24 and U2-A25, which link U2/U6 helix Ia and Ib (Table 6). These results suggest, similar to that found with U6 helix Ib mutations, that snu114 G domain mutations display synthetic lethality with U2 mutations that disrupt U2/U6 base pairing. Genetic interactions were also investigated between snu114 alleles and U2-A27C,U28C positions involved in base pairing with U6 in helix I. A second U2 mutation in this region, U2-Ins 1U U28/C29, is located in a position where increasing insertions block the second step of splicing (McPheeters and Abelson 1992;McGrail et al. 2006). Genetic analysis revealed that both V266P (G4) and F313Q (G5) were synthetically lethal with U2-A27C,U28C ( Figure 5 and Table 6). In addition, V266P (G4) was also synthetically lethal with U2-Ins 1U U28/ C29 ( Figure 5 and Table 6). No other snu114 alleles were synthetically lethal with these U2 mutations (Table 6).
These results suggest that snu114 G domain mutations are exacerbated by mutations in U2 that disrupt U2/U6 helix Ia. Further analysis revealed that there was no synthetic lethality between any snu114 alleles and insertion mutations U2-Ins 1U A30/A31 and U2-Ins 1U G32/U33, which are situated between helix I and the BPS (Table 6). However, synthetic lethality was observed between V266P (G4) and U2-A31U and L381P (G$) and U2-A31U, which extends the base pairing of U2/U6 helix Ia ( Figure 5). These results support the idea that mutations in U2 that disrupt or extend U2/U6 base pairing display synthetic lethality with snu114 G domain mutations.
snu114 mutations display synthetic lethality with mutations in U5 snRNA loop 1 and internal loop 1: U5 snRNA contributes to the catalytic core of the spliceosome through U5 loop 1, which is required for tethering the 59 and 39 exons for ligation during the second step of splicing (Newman and Norman 1992;O'Keefe and Newman 1998). Protein interactions between Snu114p, Prp8p, and Brr2p are well known (Brenner and Guthrie 2005;Liu et al. 2006) but the only known direct interaction of Snu114p with an snRNA is a crosslink with U5-C79 of U5 internal loop 1 (IL1), found at the base of the stem that carries the important U5 loop 1 (Dix et al. 1998). To determine whether Snu114p displays synthetic lethality with any mutation in U5, genetic analyses were carried out with snu114 alleles and 25 mutations that spanned the entire structure of the U5 snRNA.
As Snu114p was shown to crosslink to U5 IL1 (Dix et al. 1998), genetic interactions between mutations in snu114 and U5 IL1 were investigated. Included were eight substitution and deletion mutations that spanned positions U5-C79 and U5-G80. Increasingly larger deletions on the 59 side of U5 IL1 at or near position U5-C79 were progressively synthetically sick with M842R (IVa) ( Figure 6C and Table 7). Figure 5.-Genetic interactions between snu114 and U2 snRNA mutations. Strain YSNU114/U2KO containing the WT SNU114 and U2 snRNA alleles on a URA3 plasmid was transformed with WT or snu114 and U2 snRNA alleles on separate plasmids. Transformants were transferred to 5-FOA to evict the URA3 plasmid containing the wild-type genes and determine the synthetic lethality between mutant alleles. The annotated structure of base paired U2/U6 in the helix I conformation of intermolecular interactions required for splicing is displayed above the panels. Positions of substitutions/deletions are indicated by underlined nucleotides and insertion points are indicated by arrows.
As mutations on the 59 side of U5 IL1 did not display synthetic lethality with mutations in snu114, mutations on the 39 side of U5 IL1 were investigated. Three deletion mutations on the 39 side of IL1, U5-DC111, U5-DC112,G113, and U5-DC111-G113 were produced. The U5-DC111-G113 mutation was lethal and not studied further (data not shown). The analysis revealed that only M842R (IVa) was synthetically lethal with U5-DC111 but G646R (III), M842R (IVa), P860K (IVa), as well as DN, were synthetically lethal with the U5 IL1 allele, U5-DC112,G113 ( Figure 6D and Table 7). Significantly, no snu114 G domain mutations displayed synthetic lethality with any mutations in U5 IL1 (Table  7). Thus, our results indicate that snu114 mutations in the N-terminal domain and domains III and IV exacerbate mutations in the 59 and 39 side of U5 IL1 whereas snu114 mutations in the N-terminal domain and G domain exacerbate mutations in U5 loop 1.
Compensatory mutations that restore base pairing in U2/U6 helix I suppress synthetic lethality with snu114 mutations: One prediction of the hypothesis that Snu114p senses the state of important RNA/RNA interactions in the spliceosome is that where synthetic lethality was caused by disruption of base pairing, restoration of base pairing with a compensatory mutation would suppress the synthetic lethality. This would support the theory that the synthetic lethality was caused by disruption of base pairing and not an additional effect of the mutations. To test whether compensatory mutations could suppress synthetic le-thality observed between snu114 and snRNA mutations predicted to disrupt base pairing, a plasmid containing a compensatory snRNA mutation designed to restore base pairing was transformed with the synthetic lethal combinations. The compensatory snRNA mutation was introduced in the presence of the wild-type copy of that gene.
In contrast to the examples given above a number of synthetic lethal combinations were suppressed by compensatory mutations. Synthetic lethality between snu114 mutations and U6-A56C,U57C was suppressed when base pairing in U2/U6 helix Ia was restored with U2-A27G,U28G ( Figure 7A). Synthetic lethality of V266P (G4) and F313Q (G5) with U6-A59C was suppressed by restoring base pairing in U2/U6 helix Ib with U2-U23G ( Figure 7B). These results demonstrate that the Snu114p G domain can sense the state of base pairing in U2/U6 helix Ia and Ib.

DISCUSSION
Our aim was to determine the genetic interactions of SNU114 with the snRNAs of the spliceosome. Snu114p is an integral spliceosomal protein that is present throughout the splicing cycle and is known to regulate important snRNA interactions required for activation and disassembly of the spliceosome. By producing new conditional snu114 alleles and combining them with known conditional snu114 alleles, we have identified an extensive genetic interaction network between SNU114 and numerous conditional mutations in the U2, U4, U5, and U6 snRNAs. We propose that the G domain of Snu114p senses the state of important RNA/RNA interactions formed by the spliceosome (Figure 8).
Synthetic lethality of SNU114 with mutations that extend U4/U6 stem I confirms its role in U4/U6 unwinding: Brr2p-dependent unwinding of U4/U6 base pairing during spliceosome activation is regulated by the guanine nucleotide state of Snu114p (Small et al. 2006); therefore, we first investigated genetic interactions between snu114 and various U4 snRNA mutations. Synthetically lethal interactions of snu114 mutations in Figure 6.-Genetic interactions between snu114 and U5 snRNA mutations. Strain YSNU114/U5KO containing the WT SNU114 and U5 snRNA alleles on a URA3 plasmid was transformed with WT or snu114 and U5 snRNA alleles on separate plasmids. Transformants were transferred to 5-FOA to evict the URA3 plasmid containing the wild-type genes and determine the synthetic lethality between mutant alleles. the G domain were found with the U4 mutations, U4-cs1 and U4-U64C,G65A, which both extend the U4/U6 stem I. In addition, synthetic lethality was observed with reciprocal mutations in U6-U54 and U6-G55, which also extend the base pairing with U4 in stem I of the U4/U6 complex. These patterns of synthetic lethality suggest that G domain residues in Snu114p are involved in U4/ U6 unwinding. This supports work that implicates the guanine nucleotide state of Snu114p in Brr2p regulation (Small et al. 2006) and work with a snu114 GTP binding mutation within G domain motif G4 that blocks U4/U6 unwinding (Bartels et al. 2003). That the snu114 N-terminal deletion mutation did not interact genetically with U4-cs1 or U4-U64C,G65A was surprising since this snu114 allele has been shown to block U4/U6 unwinding at elevated temperature (Bartels et al. 2002). Intriguingly, in another study no genetic interactions were identified between snu114 G domain alleles and the cold-sensitive BRR2 allele brr2-1, which decreases U4/U6 unwinding activity of Brr2p and is synthetically lethal with U4-cs1 (Raghunathan and Guthrie 1998;Kuhn and Brow 2000;Brenner and Guthrie 2005). Rather, the brr2-1 allele is synthetically lethal with a C-terminal deletion mutation of snu114 (Brenner and Guthrie 2005), an interaction that has been substantiated by two hybrid assays (Liu et al. 2006). Therefore, as known genetic and physical interactions of Snu114p with Brr2p appear to be outside the G domain of Snu114p, our genetic interactions of Snu114p with U4 suggest that the G domain may sense the state of the U4/U6 helix and signal to Brr2p when to unwind it. This signal may be transmitted through the domains of Snu114p and Brr2p, which have been shown to interact (Brenner and Guthrie 2005;Liu et al. 2006) or through another protein mediator, such as Prp8p. Many mutations throughout Prp8p suppress the U4-cs1 mutation (Kuhn et al. 1999;Kuhn and Brow 2000) and prp8-1 and prp8-brr mutations impair formation of U5 snRNP and tri-snRNPs (Brown and Beggs 1992;Collins and Guthrie 1999). Snu114p has been shown to interact genetically with these prp8 mutations through its C-terminal domain (Brenner and Guthrie 2005). This suggests that Snu114p may directly or indirectly sense U4/U6 base pairing with its G domain and signal the state of the U4/U6 helix through its C terminus to Brr2p. This signaling to Brr2p may require a (II) G646R (III)  structural rearrangement between the G domain and C terminus of Snu114p and occur either directly or through Prp8p. Recent in vitro evidence suggests that the C terminus of Prp8p is required for the ATPdependent unwinding of U4/U6 by Brr2p (Maeder et al. 2009) and electron microscopy localization of Snu114p to a hinge region of the yeast tri-snRNP suggests Snu114p may promote rearrangements in the tri-snRNP (Hacker et al. 2008).
snu114 mutations are exacerbated by only one U4/ U6 stem I mutation: Whereas mutations that extend the U4/U6 complex are synthetically lethal with snu114 G domain mutations, the U4-G58A mutation, which is Ts and shows a block in splicing before or at the first step of splicing (Hu et al. 1995), is synthetically lethal or sick with snu114 mutations in the G domain, domain III, and domain IVa. The compensatory mutation U6-C61U that could restore base pairing in U4/U6 stem I with U4-G58A did not rescue synthetic lethality. This suggests that in addition to base pairing with U6-C61 in stem I, U4-G58 has another role during splicing. This role may be in steps prior to tri-snRNP assembly or in steps following tri-snRNP assembly but prior to spliceosome activation, since the U4-G58C mutation leads to accumulation of the B complex (Hu et al. 1995). Other mutations in U4 stem I predicted to disrupt base pairing with U6 do not display synthetic lethality with snu114 mutations. Further work is required to understand the specific sensitivity of only position U4-G58A in U4 stem I and the reason for its genetic interactions with SNU114.
Snu114p senses the state of U6-59 splice site and U2/ U6 base pairing: In addition to sensing the state of U4/ U6, we show that Snu114p interacts genetically with key positions in the U6-ACAGAGA that base pairs with the 59 splice site. This suggests that successive interactions of Snu114p with U4 and U6 may facilitate the transitions that occur during spliceosome activation. Notably, Prp8p also interacts with the U6-ACAGAGA sequence: both Prp8p and U6-A51 crosslink to the 59 splice site before both first and second steps of splicing (Sontheimer and Steitz 1993;Kim and Abelson 1996;Reyes et al. 1999). However, since prp8 mutations only partially suppress U6-A51 mutations that cause 59 splice site misalignment and result in defective exon ligation (Collins and Guthrie 1999), Snu114p, together with Prp8p, may coordinate U6 at the 59 splice site. That snu114 alleles interact genetically with mutations of U6-G52 but not with U2-A25 suggests that the Snu114p genetic interactions with U6-G52 occur during U6-59 splice site base pairing rather than during formation of a tertiary interaction with U2 for the second step of splicing (Madhani and Guthrie 1994). Additional mutations in U6 or the pre-mRNA designed to stabilize the U6-59 splice site interactions combined with snu114 G domain mutations would provide further support for the hypothesis that the G domain senses U6-59 splice site interactions.
Brr2p-dependent unwinding of U2/U6 base pairing during spliceosome disassembly is also regulated by the guanine nucleotide state of Snu114p (Small et al. 2006); therefore, we investigated genetic interactions between snu114 and various U2 and U6 snRNA mutations that influence U2/U6 base pairing. Formation of U2/U6 helix I is important for promoting the first and second chemical steps of splicing (Madhani and Guthrie 1992;Ryan and Abelson 2002;Mefford and Staley 2009) and unwinding of the U2/U6 helix I is required for disassembly of the spliceosome (Small et al. 2006). Synthetic lethality is only observed between N terminus and G domain snu114 mutations in combination with mutations in U2 and U6 that are predicted to disrupt or stabilize U2/U6 helix I base pairing. The synthetic lethality patterns observed between Snu114p with U6 and U2 helix I mutations suggest that Snu114p senses the state of U2/U6 base pairing interactions. This hypothesis is supported by the suppression of synthetic lethality by compensatory mutations that restore base pairing in U2/U6 helix I. Therefore, at least for the U2/ U6 helix I it is clear that synthetic lethality was solely due to the disruption of base pairing between U2 and U6 in this region. The lack of suppression of synthetic lethality by compensatory mutations at other positions predicted to restore base pairing does not preclude a role for Snu114p in sensing base pairing at these positions. Rather it suggests a possible additional role, besides base pairing, for the positions where synthetic lethality cannot be suppressed by compensatory mutations.
The specificity of genetic effects of U6 mutations to the U2/U6 conformation is confirmed by the lack of genetic interactions between Snu114p and U4 in U4 stem I. Although positions in U4 stem I that participate in the U4/U6 complex opposite U6-A56,U57 were not assayed, three other U4 stem I alleles displayed no synthetic lethality with snu114 alleles. This suggests that SNU114-U6 genetic interactions are related to U6 in the U2/U6 conformation rather than the U4/U6 conformation. That synthetically lethal effects are specific to disruption of U2/U6 helix I is supported by the lack of synthetic lethality with mutations and deletions at U2-U24 and U2-A25, which form a nonbase paired bulge in U2/U6 helix I.
Three differences are observed between U2 and U6 synthetically lethal and compensatory interactions. First, the snu114 N-terminus deletion mutation is synthetically lethal with mutations in U6 that disrupt base pairing whereas there is no synthetic lethality of the N-terminus deletion mutation with U2 mutations that disrupt base pairing. Second, insertion mutations in opposite sides of helix Ia display different genetic effects with SNU114. An insertion mutation of U6 in helix Ia (U6-Ins 1U A53/U54) is viable with all snu114 mutations tested. On the other hand, an insertion mutation of U2 in helix Ia (U2-Ins 1U U28/A29) displays synthetic lethality with snu114 G domain mutation V266P (G4). It is known that insertions at this location in U6 inhibit the first step of splicing whereas insertions at this location in U2 inhibit the second step of splicing (McPheeters and Abelson 1992;McGrail et al. 2006). Third, suppression of synthetic lethality by compensatory mutations was successful with mutations in U2 that restored base pairing with U6 mutations in helix Ia and Ib; however, mutations in U6 that restored base pairing with U2 mutations in helix Ia and Ib were not found to suppress synthetic lethality with snu114 mutations. Overall, these differences suggest additional roles for nucleotides in U2 that form helix Ia and Ib during splicing. These roles could be manifested when U2 and U6 are not base paired either before formation of U2/ U6 helix I, after formation of U2/U6 helix I, or when U2/U6 helix I is destabilized by Prp16p between the catalytic steps of splicing (Mefford and Staley 2009).
Snu114p interactions and orientation with U5 snRNA: As Snu114p is an integral protein of the U5 snRNP, we investigated synthetic lethality between snu114 and 25 U5 mutations. Mutations in U5 displayed synthetic lethality with the snu114 N-terminal deletion mutation and G domain mutations. U5 was the only snRNA to show synthetic lethality with the snu114 Cterminal P860K (IVa) mutation. The snu114 G domain mutation V266P (G4) was synthetically lethal with U5 loop 1 mutations while the snu114 C-terminal domain mutations G646R (III) and P860K (IVa) were synthetically lethal with a U5 IL1 mutation. This suggests an orientation of Snu114p with U5 within the spliceosome such that U5 links the G domain to the other domains of Snu114p. We observed spatially distinct genetic interactions between the C terminus of Snu114p, which displayed synthetic lethality with U5 IL1, and the G domain which did not. This orientation suggests that U5 and Snu114p may work together to transmit information from the G domain to the C-terminal domain which has been shown to interact with other splicing factors (Brenner and Guthrie 2005;Liu et al. 2006). It is possible that the state of the snRNAs sensed through the G domain of Snu114p is transduced through the Cterminal end of Snu114p, perhaps through a structural change/domain shift, which is the mechanism of function of other regulatory GTPases (Jorgensen et al. 2003).
The SNU114 N-terminal domain also interacts genetically with U5 IL1 which suggests that it may also be involved in contacting U5 in this region along with U5 loop 1. This is the first evidence that suggests a conformation for the N-terminal domain of Snu114p ( Figure  8), which cannot be modeled on EF-2. Whether the Nterminal domain interacts with U5 loop 1 and U5 IL1 at the same or at different times is unknown. For both the N-and C-terminal domains of Snu114p to interact with U5 IL1 suggests a long range intramolecular interaction such that the N terminus and domain IV come into close proximity with U5 IL1 if these interactions are simultaneous. These interactions may also involve Prp8p since it has been found to crosslink to position C112 in U5 IL1 (Dix et al. 1998). The only other genetic interactions found with the snu114 DN mutation were with mutations in the U6 snRNA. The fact that no U2 or U4 synthetically lethal interactions were found with the snu114 DN mutation suggests that the Snu114p N terminus may be involved in U5 and U6 interactions at the 59 splice site.
We have identified an extensive genetic interaction network between the spliceosomal GTPase Snu114p and the snRNAs of the spliceosome (Figure 8). By combining four new snu114 mutations with five previously known mutations we have distinguished different roles for distinct regions of Snu114p and propose an orientation for Snu114p with U5 snRNA. It appears that Snu114p recognizes the state of important RNA/RNA interactions within the spliceosome (Figure 8). Snu114 G domain mutations exacerbate mutations in U4 that change U4/U6 and U2/U6 base pairing. Snu114 G domain mutations also exacerbate mutations in U6 that increase base pairing with the 59 splice site. Mutations in the snu114 G domain and N-terminal domain exacerbate mutations in U5 at positions that align the exons in the catalytic core. Thus the Snu114p G domain may have a significant role in signaling the state of the snRNAs throughout splicing (Figure 8). Additionally, snu114 mutations in the N-terminal domain and domains III and IV exacerbate mutations in U5 internal loop 1, which could serve to link the N-and C-terminal domains of Snu114p functionally. These genetic interactions with U5 also provide a structural link between the G domain and C terminal domain, and therefore a link between the proposed sensing and transducing regions of Snu114p. Thus by analysis of the genetic interactions between distinct domains in Snu114p and the snRNAs of the spliceosome we have defined a functional and spatial network of interactions that we propose is essential for spliceosome function. Further work is required to test the new snu114 mutations with alleles in proteins that have been found to interact with Snu114p previously (Brenner and Guthrie 2005). This work would establish how the genetic interactions we found between Snu114p and snRNAs fit into the network of Snu114p-protein interactions in the spliceosome and would provide an insight into whether these genetic interactions represent direct or indirect physical interactions with Snu114p. Conversely, genetic interactions between snRNA mutations and other splicing protein alleles should be investigated to discover how the Snu114p-snRNA interactions fit in with other potential snRNA-protein networks.