Nam8, a component of yeast U1 snRNP, is optional for mitotic growth but required during meiosis, because Nam8 collaborates with Mer1 to promote splicing of essential meiotic mRNAs AMA1, MER2 and MER3 . Here, we identify SPO22 and PCH2 as novel targets of Nam8-dependent meiotic splicing. Whereas SPO22 splicing is co-dependent on Mer1, PCH2 is not. The SPO22 intron has a non-consensus 5′ splice site (5′SS) that dictates its Nam8/Mer1-dependence. SPO22 splicing relies on Mer1 recognition, via its KH domain, of an intronic enhancer 5′-AYACCCUY. Mutagenesis of KH and the enhancer highlights Arg214 and Gln243 and the CCC triplet as essential for Mer1 activity. The Nam8-dependent PCH2 pre-mRNA has a consensus 5′SS and lacks a Mer1 enhancer. For PCH2 , a long 5′ exon and a non-consensus intron branchpoint dictate Nam8-dependence. Our results implicate Nam8 in two distinct meiotic splicing regulons. Nam8 is composed of three RRM domains, flanked by N-terminal leader and C-terminal tail segments. The leader, tail and RRM1 are dispensable for splicing meiotic targets and unnecessary for vegetative Nam8 function in multiple synthetic lethal genetic backgrounds. Nam8 activity is enfeebled by alanine mutations in the putative RNA binding sites of the RRM2 and RRM3 domains.
Regulated pre-mRNA splicing figures prominently in the control of eukaryal gene expression. Alternative splice site utilization amplifies the information content of the human genome ( 1 , 2 ), allowing a single transcription unit to generate multiple protein products that may differ in their domain composition, cellular localization, macromolecular interactions or biochemical activities. Splicing control receives numerous inputs in metazoan organisms: (i) from cis -acting RNA sequences or secondary structures that modulate interactions of the splicing machinery with pre-mRNA ( 3 ) and (ii) from a fleet of RNA-binding proteins ( 4 ) that enhance or silence splicing in response to these RNA signals, to changes in the cellular milieu, and to developmental cues ( 2 , 5 , 6 ). Defects or imbalances in pre-mRNA splicing underlie a broad spectrum of human diseases, either via mutations of core splicing machinery and splicing regulatory factors or, more commonly, mutations of RNA signals that direct or regulate splice site utilization ( 7 ).
Yeast has been a powerful model system to elucidate the composition and function of the core splicing machinery ( 8 ) and the dynamic changes in the spliceosome that propel splicing chemistry and fidelity ( 9 ). Although budding yeast has relatively few natural examples of splicing regulation via alternative 5′ splice site choice, alternative 3′ splice site choice or exon skipping ( 10–12 ), it relies heavily in meiosis on controlling the mode of alternative splicing known as intron retention ( 2 ). To wit, the meiotic developmental program entails a shift in the processing patterns of target pre-mRNAs from a vegetative ‘off’ state, in which single introns are included, to a meiotic ‘on’ state in which the target introns are removed. The discovery of Mer1-regulated meiotic splicing and the appreciation that Mer1 functions together with Nam8 to compensate for inherently weak introns are landmarks in the field ( 13–15 ).
Nam8 is an RNA-binding component of the yeast U1 snRNP and is present in the commitment complex of U1 snRNP at the 5′ splice site, where it can be crosslinked to the pre-mRNA ( 16–18 ). The Nam8 protein contains three tandem RRM domains and is a putative yeast homolog of the mammalian RNA-binding protein and pre-mRNA splicing factor TIA-1 ( 19 , 20 ). Nam8 is inessential for yeast mitotic growth, but is essential for yeast sporulation because it promotes splicing of a small set of specific mRNAs that encode proteins required for meiotic recombination and cell division ( 15 , 21–23 ). Nam8-dependent splicing of three meiotic mRNAs – AMA1 , MER2 and MER3 – is activated by the meiotic splicing regulator Mer1 ( 13 , 24 ). Mer1 is produced only in meiotic cells under the control of the meiotic transcription factor Ime1 ( 25 ). Mer1 activates splicing by binding to an intronic splicing enhancer sequence (5′-AYACCCUY-3′) present in the AMA1 , MER2 and MER3 pre-mRNAs ( 15 ). Mer1 bound to the intronic enhancer is thought to promote assembly of the U1 and U2 snRNPs on the pre-mRNA ( 22 , 26 ). Transcripts subject to Mer1/Nam8 splicing regulation have suboptimal 5′ splice sites and, in some cases, suboptimal branchpoints or a large 5′ exon that might, alone or together, dictate their reliance on otherwise inessential splicing factors ( 14 , 22 , 27 ).
Recent studies have illuminated additional meiotic intron-containing genes as potential targets of regulated splicing, several of which do not contain a Mer1 enhancer ( 10 , 28–31 ). Here, by evaluating the splicing of the known meiotic intron-containing RNAs in wild-type and nam8Δ yeast diploids during attempted sporulation, we discovered that Nam8 has two additional meiotic mRNAs targets: SPO22 and PCH2 . The SPO22 intron contains a Mer1 binding site and requires both Nam8 and Mer1 for splicing. By contrast, PCH2 lacks a Mer1 site and is indifferent to the presence of Mer1 for its splicing. We interrogated the prevailing hypothesis that meiotic splicing controllers such as Mer1 and Nam8 compensate for weak intronic splicing signals and/or override negative influences exerted by other RNA elements. We elucidated these RNA features for SPO22 , MER3 and PCH2 . A structure-guided mutational analysis of the putative RNA binding site of Mer1 identified individual amino acid functional groups required for its splicing enhancement activity in vivo . Mutational analysis of Nam8 demarcated domain requirements for meiotic splicing and for Nam8 function in vegetative cells that require Nam8 for growth because they lack other inessential splicing factors.
Yeast strains for sporulation studies
The meiosis/sporulation experiments were carried out with isogenic diploids in the SK1 background. To generate the nam8 Δ diploids, haploid derivatives of the SK1 strain, SKY163 ( MAT a ho::LYS2 lys2 ura3 leu3::hisG ) and SKY164 ( MATα ho::LYS2 lys2 ura3 leu3::hisG ) were used. In brief, a DNA segment encompassing the nam8::kanMX disruption cassette was generated by polymerase chain reaction (PCR) amplification of genomic DNA from a nam8::kanMX yeast strain, the PCR product was introduced into SKY163, and geneticin-resistant kanMX integrants were selected. SKY164 was transformed with a nam8::natMX deletion cassette ( 32 , 33 ) and nourseothricin-resistant natMX integrants were selected. The targeted insertions were confirmed by diagnostic Southern blotting. The SKY haploids were then mated and homozygous nam8 Δ diploids were selected on YPD agar containing 100 µg/ml nourseothricin (clonNat, Werner BioAgents, Jena) and 150 µg/ml geneticin (Invitrogen).
Sporulation and analysis of meiotic RNA splicing
Single colonies of wild-type and nam8 Δ diploid yeast strains were patched on agar plates with glycerol as the carbon source for at least 6 h to select for cells with healthy mitochondria. Cells were streaked on YPD agar plates and incubated for 3 days at 30°C. Single colonies were then inoculated into YPD liquid medium and grown at 30°C to stationary phase ( A600 of 6–8). Aliquots were inoculated into 12.5 ml of presporulation medium (0.5% yeast extract, 1% peptone, 0.67% yeast nitrogen base (without amino acids), 1% potassium acetate, 0.05 M potassium biphthalate (pH 5.5), 0.002% antifoam 204) to attain an A600 of 0.8. The cultures were incubated for 7 h at 30°C and added to 100 ml of fresh presporulation medium to attain an A600 of 0.025. After 16 h incubation at 30°C, the cells were washed twice with 100 ml of sporulation medium (2% potassium acetate, 0.001% polypropylene glycol) and resuspended in 15 ml of sporulation medium to attain an A600 of 6. For analysis of meiotic RNA, 2-ml aliquots of the cultures were withdrawn immediately prior to transfer to sporulation medium and at 4 h and 8 h post-transfer to sporulation medium. Cells were harvested by centrifugation and RNA was extracted from the cells by using the MasterPure Yeast RNA Purification Kit (Epicentre BioTechnologies). The DNase digestion step in the RNA purification protocol was modified such that the samples were incubated for 1 h with 20 U of RNase-free DNase I (New England Biolabs) to eliminate genomic DNA. First-strand cDNA was synthesized in reaction mixtures containing 10 ng/μl total RNA, 25 ng/μl oligo(dT) 12–18 primer, 10 U/μl SuperScript II (Invitrogen), 1 X First Strand Buffer, 10 mM DTT and 0.5 mM dNTPs. RNA, dNTPs and primers were preincubated for 5 min at 65°C before quick chilling on ice. DTT and First Strand Buffer were added, and the mixture was incubated at 42°C for 2 min. Finally, reverse transcriptase was added and the reaction mixture was incubated at 42°C for 50 min and then at 70°C for 15 min. The cDNAs for meiotic transcripts were then PCR-amplified in 25 μl reaction mixtures containing 1X native Pfu buffer, 0.2 mM dNTPs, 0.8 μM gene-specific sense strand primer and 0.8 μM 5′- 32 P-labeled gene-specific antisense strand primer ( Supplementary Table S1 ), 0.05 U/μl Pfu DNA Polymerase (Agilent Technologies) and 2 μl of each cDNA sample. The PCR cycles ( n = 31) entailed incubations at 94°C for 30 s, 55°C for 90 s, and 72°C for 2 min. The RT–PCR products were analyzed by electrophoresis through 2% native agarose gels. After electrophoresis, the gels were stained with ethidium bromide and then dried under vacuum on DEAE paper. The 32 P-labeled PCR products were visualized by autoradiography and quantified by scanning with a Fuji BAS-2500 imager.
Assay of splicing of meiotic RNAs in vegetative cells
A haploid ‘wild-type’ NAM8 yeast strain ( MAT a NAM8 leu2Δ ura3Δ ), derived from S288c, and a nam8::kanMX and variants thereof ( 32 ) were transformed with plasmid pYX212-MER1 (2 μ URA3 MER1 ) expressing the MER1 open reading frame under the transcriptional control of the yeast TPI1 promoter (see below), or with the empty pYX212 vector. Cells were cotransformed with 2 µ plasmids bearing SPO22 , MER3 or PCH2 genes as specified in the figure legends. pRS425TPI-SPO22 (2 μ LEU2 SPO22 ) carries the intron-containing SPO22 gene driven by the TPI1 promoter. pRS425TPI-MER3 (2 μ LEU2 MER3 ) carries the intron-containing MER3 gene driven by the TPI1 promoter. pRS425-PCH2 (2 μ LEU2 PCH2 ) bears the intron-containing PCH2 gene (from 540 bp upstream of the translation start codon to 265 bp downstream of the stop codon) under the control of its native promoter. Intron mutants were generated by two-stage overlap extension PCR with mutagenic primers; the mutated DNAs were inserted into pRS425. The inserts of all plasmid clones were sequenced to exclude the acquisition of unwanted mutations during amplification and cloning.
The haploid plasmid-bearing yeast strains were grown in SD-(Ura − Leu − ) liquid medium at 30°C until A600 reached 2–4. Cells were harvested by centrifugation from 2-ml aliquots of the cultures. RNA extractions and reverse transcription–polymerase chain reaction (RT–PCR) were performed as described above, with exceptions as follows: (i) cDNA synthesis was primed with 0.1 µM gene-specific antisense primers ( Supplementary Table S1 ) instead of oligo(dT) 12–18 ; (ii) the numbers of PCR cycles for amplification of MER2, PCH2, SPO22 and MER3 cDNAs were 28, 27, 25 and 23, respectively; (iii) the PCR reactions were quenched by adding EDTA and SDS to a final concentrations of 12.5 mM and 3.3%, respectively; and (iv) the 32 P-labeled PCR products were analyzed by electrophoresis through native 5% polyacrylamide gels containing 90 mM Tris-borate, 1.2 mM EDTA.
The MER1 open reading frame was amplified by PCR from yeast genomic DNA using primers designed to introduce an EcoRI site immediately upstream of the translation initiation codon and an XhoI site immediately downstream of the stop codon. The PCR product was inserted into yeast expression vector pYX212 ( 2 µ URA3 ) to yield pYX212-MER1. Truncated MER1-NΔ alleles were constructed by PCR amplification of pYX212-MER1 template with sense-strand primers that introduced a new start codon at the positions specified plus a flanking EcoRI site. Missense mutations were introduced into the MER1 ORF via a two-stage overlap extension PCR. The PCR products of the truncated and mutated MER1 ORFs were digested with EcoRI and XhoI and inserted into EcoRI/XhoI-cut pYX212-MER1 in lieu of the wild-type gene.
Yeast Nam8 expression vectors
To construct the NAM8 vectors for plasmid shuffle complementation assays, a 1.6-bp DNA fragment comprising the NAM8 ORF plus 427 bp of upstream (5′) and 277 bp of downstream (3′) chromosomal DNA was amplified by PCR from Saccharomyces cerevisiae genomic DNA using primers that introduced a HindIII site at the 5′ end and a SacI site at the 3′ end. We then used overlap extension PCR to introduce a PstI site immediately upstream of the NAM8 translation initiation codon and an XbaI site immediately downstream of the stop codon. The NAM8 cassette (ORF plus flanking DNA) was then inserted into yeast expression vector pRS415 ( CEN LEU2 ) to yield pRS415-NAM8. The truncated NAM8 alleles were constructed by PCR amplification with: (i) sense strand primers that introduced a new start codon at the positions specified plus a flanking PstI site and (ii) antisense strand primers that introduced a new stop codon after the positions specified plus a flanking XbaI site. Missense mutations were introduced into the NAM8 ORF via two-stage overlap extension PCR. The PCR products containing the truncated and mutated NAM8 ORFs were digested with PstI and XbaI and inserted into PstI/XbaI-cut pRS415-NAM8 in lieu of the wild-type gene.
To construct vectors for NAM8 overexpression, the NAM8 ORFs encoding wild-type, truncated or mutated Nam8 were amplified by PCR from their respective pRS415 plasmids using sense primers that introduced a KpnI site immediately flanking the start codon. The PCR products were digested with KpnI/XbaI and inserted into KpnI/XbaI-cut pYES2 ( 2 µ URA3 GAL ) so as to place NAM8 under the transcriptional control of a glucose-repressible/galactose-inducible GAL1 promoter.
All of the NAM8 plasmid inserts were sequenced completely to exclude the acquisition of unwanted mutations during amplification and cloning.
Yeast strains for assays of Nam8 function in vegetative cells
Saccharomyces cerevisiae strain yTN ( Mata tgs1Δ nam8Δ p360-TGS1), described in Hausmann et al . ( 32 ) was used for plasmid shuffle complementation assays of Nam8 function. We constructed other nam8Δ double mutant strains by crossing and dissecting haploid strains that were described previously ( 32 , 33 ). In brief, haploid nam8::natR cells were mixed with lea1::kanMX , mud1::kanMX and mud2::kanMX cells of the opposite mating type, and diploids were selected on YPD agar containing nourseothricin and geneticin. The heterozygous diploids were transformed with a URA3 CEN plasmid carrying the NAM8 gene. Sporulation was induced and tetrads were dissected. Individual spores were tested to select for nam8Δ::natR mud1Δ::kanMX p360-NAM8, nam8Δ::natR mud2Δ::kanMX p360-NAM8 or nam8Δ::natR lea1Δ::kanMX p360-NAM8 cells. Using the same strategy, we also generated nam8Δ::natR mud2Δ::kanMX p360-MUD2, in which the complementing URA3 plasmid harbors the MUD2 gene ( 33 ).
HIS3 reporter assay for PCH2 intron splicing
To generate integration cassettes for the HIS3 reporter genes, a 1-bp genomic DNA fragment spanning the HIS3 ORF plus 297 and 20 bp of upstream and downstream sequences was amplified by PCR using primers His3-F and His3-R ( Supplementary Table S2 ) and inserted in between the XmaI and PstI sites in pUC19. To insert the PCH2 intron (or the mutated PCH2-BP intron) at position 430 in the HIS3 ORF, we individually amplified the HIS3-5 ′ gene fragment (727 bp) with primers His3-F and exon1R ( Supplementary Table S2 ) and the HIS3-3 ′ (253 bp) gene fragment with His3-R and exon2F (Table S2). The PCH2 and PCH2-BP introns (123 bp) were amplified by PCR with primers IntronF and IntronR. The exon and intron DNA fragments with overlapping terminal sequences were then assembled into intron-punctuated HIS3 cassettes by overlap-extension PCR and the cassettes were inserted into pUC19 plasmids. The HIS3 , HIS3-[PCH2] and HIS3-[PCH2-BP] cassettes were excised with SmaI and PstI and transformed into nam8Δ::natR p360-NAM8 ( URA3 NAM8 ) cells and isogenic wild-type cells (harboring a URA3 plasmid) that were histidine-auxotrophs. His + Ura + transformants were selected and analyzed for integration of the respective cassettes at the HIS3 locus by diagnostic Southern blotting and by sequencing of PCR-generated DNA fragments using primers designed to amplify the HIS3 gene. The cells were then streaked to medium containing 5-FOA (and histidine) to select for cells that had lost the URA3 plasmids. Individual colonies were patched to YPD agar. Cells were grown in liquid YPD medium until the cultures attained A600 between 0.7 and 0.9. The cultures were diluted in water to A600 of 0.01. Serial 10-fold dilutions were prepared and aliquots (3 µl) of each were spotted in parallel on YPD agar and SD agar medium lacking histidine.
Meiotic splicing in nam8 Δ cells
RNA was isolated from wild-type and nam8 Δ diploids immediately prior to (time 0) and 4 h and 8 h after transfer from pre-sporulation medium to sporulation medium. cDNA was prepared from each RNA sample by RT and then used for gene-specific PCR amplification of meiotic spliced transcripts ( Figure 1 and Table 1 ). The sense and antisense primers corresponded to sequences flanking the introns so that the longer products of amplification of cDNA derived from unspliced pre-mRNAs could be easily resolved by native gel electrophoresis from the shorter products of amplification of cDNAs copied from spliced mRNA ( Figure 1 ). One of the primers was 5′ 32 P-labeled in each PCR reaction so that we could quantify the distributions of unspliced and spliced cDNAs for each gene of interest ( Figure 1 , lanes 3–5). An aliquot of a PCR amplification reaction using genomic DNA as template provided a marker for the unspliced species (lane 2). No labeled products were generated from PCR reactions programmed by RNA that had not been subjected to prior treatment with reverse transcriptase (lane 1), indicating that the RNA samples were effectively free of contaminating genomic DNA.
|RNA|| Percent spliced at 4 h |
|AMA1||84 ± 5||6 ± 3|
|MER2 (REC107)||76 ± 2||6 ± 3|
|MER3 (HFM1)||80 ± 2||2 ± 1|
|HOP2||88 ± 6||84 ± 4|
|REC114||87 ± 1||76 ± 2|
|MEI4||80 ± 1||71 ± 2|
|REC102||84 ± 4||74 ± 4|
|DMC1||95 ± 1||95 ± 3|
|PCH2||73 ± 2||9 ± 2|
|SAE3||66 ± 1||37 ± 2|
|SPO1||84 ± 3||80 ± 4|
|SPO22||86 ± 1||4 ± 2|
|MND1||90 ± 4||82 ± 3|
|SRC1||94 ± 2||87 ± 2|
|RNA|| Percent spliced at 4 h |
|AMA1||84 ± 5||6 ± 3|
|MER2 (REC107)||76 ± 2||6 ± 3|
|MER3 (HFM1)||80 ± 2||2 ± 1|
|HOP2||88 ± 6||84 ± 4|
|REC114||87 ± 1||76 ± 2|
|MEI4||80 ± 1||71 ± 2|
|REC102||84 ± 4||74 ± 4|
|DMC1||95 ± 1||95 ± 3|
|PCH2||73 ± 2||9 ± 2|
|SAE3||66 ± 1||37 ± 2|
|SPO1||84 ± 3||80 ± 4|
|SPO22||86 ± 1||4 ± 2|
|MND1||90 ± 4||82 ± 3|
|SRC1||94 ± 2||87 ± 2|
Figure 1 shows exemplary data demonstrating induction of meiotic transcription and regulated meiotic splicing in wild-type cells, while focusing on some of the meiotic transcripts for which splicing efficiency was most acutely affected by the absence of Nam8. Transcriptional induction of the MER3 gene in wild-type cells was evinced by the increase in total RT–PCR products at 4 and 8 h post-sporulation (lanes 4 and 5) compared to the level at time 0 (lane 3), which was accompanied by a sharp increase in the percentage of the RT–PCR product derived from spliced versus unspliced MER3 RNA. In agreement with previous studies ( 34 , 35 ), we also observed transcriptional induction of SPO22 and PCH2 ( Figure 1 ) and of several other intron-containing meiotic genes analyzed (data not shown). Transcriptional induction of MER3 , SPO22 , PCH2 and other meiotic genes was also evident in nam8 Δ cells after 4 and 8 h in sporulation medium ( Figure 1 and data not shown).
The role of Nam8 in meiotic splicing was verified for MER3 (one of three known Mer1/Nam8 targets), splicing of which was barely detectable in nam8 Δ cells at either 4 or 8 h ( Figure 1 ). Nam8 was also required for meiotic splicing of SPO22 and PCH2 pre-mRNAs ( Figure 1 ), which are hereby identified as new Nam8 targets. By contrast, the control GLC7 RNA was constitutively expressed and very efficiently spliced in wild-type cells (≥98% of the RT–PCR product was from spliced GLC7 RNA) and this was unaffected by nam8 Δ ( Figure 1 ).
The results of our survey of splicing of 14 meiotic transcripts at 4 h post-induction of sporulation are compiled in Table 1 , wherein each datum for splicing efficiency – [spliced/(spliced + unspliced)] × 100 – is the average of three independent sporulation experiments and RT–PCR analyses. The absence of Nam8 severely affected meiotic splicing of AMA1 , MER2 and MER3 (the three Nam8 targets described previously) plus novel targets SPO22 and PCH2 . Nam8 ablation had little or no impact on the efficiency of splicing of the HOP2 , REC114 , MEI4 , REC102 , DMC1 , SPO1 , MND1 and SRC1 transcripts ( Table 1 ), though nam8 Δ did reduce SAE3 splicing modestly.
Inspection of the features of the 14 meiotically spliced transcripts reveals that four of them— AMA1 , MER2 , MER3 and SPO22 —have a Mer1 intronic enhancer, thereby extending the correlation of coregulation by Mer1 and Nam8 to the SPO22 pre-mRNA identified presently as Nam8 dependent. The magnitudes of the nam8 Δ splicing defects for the four Mer1 enhancer-containing transcripts were between 12- and 40-fold ( Table 1 ). To our knowledge, PCH2 is the first instance of a Nam8-dependent meiotic pre-mRNA that does not have a Mer1 intronic consensus sequence. The decrement in PCH2 splicing in the nam8 Δ strain was 8-fold.
Differential requirements for Mer1 in splicing of SPO22 and PCH2 mRNAs
Regulated splicing of Mer1-dependent meiotic mRNAs can be recapitulated in vegetative yeast cells by forced expression of the Mer1 splicing enhancer protein (13,15.23). The efficacy of this maneuver is readily demonstrated for splicing of the MER2 pre-mRNA, which is constitutively transcribed in vegetative cells, but is spliced inefficiently (20%) because Mer1 is absent ( Figure 2 A). Expression of Mer1 during vegetative growth (by transformation with a plasmid bearing MER1 under the control of a constitutive promoter) increased MER2 splicing efficiency to 87% ( Figure 2 A). MER2 splicing was inhibited severely (5% efficiency) in nam8 Δ cells expressing Mer1 ( Figure 2 A).
To test whether SPO22 or PCH2 pre-mRNA splicing relies on Mer1, we transformed vegetative cells with 2 µ plasmids bearing SPO22 or PCH2 genes plus a second plasmid for constitutive MER1 expression (or an empty vector control). Splicing of SPO22 mRNA was acutely dependent on Mer1 (86% in Mer1-expressing cells versus 14% in control cells) and Nam8 (4% splicing in Mer1-expressing nam8 Δ cells) ( Figure 2 B). By contrast, splicing of PCH2 (at 67% efficiency in wild-type vegetative cells, comparable to that seen in meiotic cells; Table 1 ) was unaffected by Mer1 expression ( Figure 2 C). These results consolidate the following points: (i) SPO22 , which contains an intronic Mer1 enhancer element, joins three other meiotic genes— AMA1 , MER2 and MER3 —to comprise a splicing regulon controlled by both Mer1 and Nam8; (ii) PCH2 is the first instance of a Mer1-independent Nam8-regulated meiotic transcript.
Why does SPO22 splicing depend on Mer1 and Nam8?
SPO22 has a consensus Mer1 enhancer element (5′-AUACCCUU) situated between the 5′ splice site and the branchpoint ( Figure 3 A). SPO22 is especially interesting, vis à vis other Mer1 targets, in that its intron has multiple features that could confer feeble splicing and hence reliance on upregulation. These include: (i) a rare non-consensus 5′ splice site (GUAUAU) distinct from those in other Mer1-dependent transcripts; (ii) a rare non-consensus branchpoint (AACUAAC); and (iii) a non-consensus 3′ splice site (AAG) ( Figure 3 A). Analysis of the SPO22 intron in Mfold ( 36 ) revealed no stable secondary structure that would mask any of these signals.
We installed consensus 5′SS , BP and 3′SS elements in the SPO22 gene and tested these SPO22 mutants on 2 µ plasmids for their splicing efficiency in vegetative wild-type and nam8 Δ cells that do or do not express Mer1. Our aim here was to pinpoint which of the intron manipulations might render SPO22 splicing independent of Mer1 and/or Nam8. In particular, we wished to see if the degrees of Nam8 and Mer1 dependence tracked together, or if we could find an intronic manipulation that dissociated Mer1 from Nam8 (e.g. relieving the requirement for one factor, but not the other). The SPO22 splicing efficiencies are plotted in Figure 3 A, with each datum representing the average of three separate experiments ± SEM.
We found that installing a perfect GUAUGU 5′ splice site (via a single A-to-G change in mutant SPO22-5′SS ) eliminated the dependence of SPO22 splicing on Mer1, i.e. basal splicing efficiency was increased to 92% in wild-type yeast lacking Mer1 and this value was unaffected by Mer1 expression ( Figure 3 A). Similarly, the consensus 5′SS change effaced the dependence of SPO22 splicing on Nam8, elevating splicing efficiency to 86% in nam8 Δ cells lacking Mer1 ( Figure 3 A). We surmise that the nonconsensus 5′ splice site is the critical determinant of the Mer1/Nam8 requirement for efficient SPO22 splicing in vivo .
By contrast, whereas the installation of a consensus branchpoint (entailing a A-to-U mutation in SPO22-BP ) or a consensus 3′ splice site (UAG in SPO22-3′SS ) raised basal splicing efficiency in wild-type cells to 51% and 28%, respectively (versus 14% for native SPO22 ), the SPO22-BP and SPO22-3′SS transcripts remained dependent on Mer1 to attain their optimal splicing efficiencies of 86% and 90%, respectively ( Figure 3 A). The increases in basal splicing in nam8 Δ cells for the SPO22-BP and SPO22-3′SS mutants (to 28% and 16%, respectively, versus 6% for the native SPO22 transcript) were of lesser magnitude than that seen in NAM8 cells. Moreover, there was no further increase in splicing of the SPO22-BP and SPO22-3′SS mutants in nam8 Δ cells attendant on Mer1 expression ( Figure 3 A). These results indicate that: (i) all of the salutary effects of Mer1 on SPO22 splicing are mediated via Nam8, but not vice versa ; (ii) even in the absence of Mer1, Nam8 has a modest positive effect (∼2-fold stimulation) on basal SPO22 splicing, be it in the context of the feebly spliced native SPO22 intron or the better-spliced consensus BP and 3′SS variants; and (iii) the non-consensus branchpoint and 3′ splice site of SPO22 each exerts a suppressive effect on intron removal that is ‘independent’ of Nam8, with the non-consensus branchpoint effect being of greater magnitude.
Determinants of MER3 pre-mRNA splicing
Inspection of the MER3 intron, which requires Mer1 and Nam8 for splicing in meiosis ( 35 ), reveals several features not found in other Mer1/Nam8 targets. These include: (i) a deviant 5′SS (GUAGUA) found in no other yeast introns ( 37 , 38 ); (ii) a rare non-consensus branchpoint ( G ACUAAC); (iii) an exceptionally long distance (83-nt) from the BP to the 3′SS ( 38 ); and (iv) a 12-nt polypyrimidine tract (5′-CCUCUUUCUUUU) between the BP and 3′SS, placed close to the branchpoint ( Figure 3 B). Unlike SPO22 , MER3 has a typical UAG 3′ splice site that is utilized to generate the correctly spliced MER3 mRNA that encodes the 1187-aa Mer3 polypeptide ( Figure 3 B). MER3 also has three other YAG triplets (potential 3′ splice sites) located between the branchpoint/polypyrimidine tract and the correct 3′SS ( Figure 3 B); utilization of any of these alternative 3′SS would generate out-of-frame transcripts with premature stop codons.
We initiated an analysis of MER3 splicing requirements by installing consensus 5′SS and BP signals in the MER3 gene, singly and in combination. Native and mutated MER3 alleles on 2 µ plasmids were then assayed by RT–PCR for their splicing efficiency in vegetative wild-type and nam8 Δ cells that do or do not express Mer1. Splicing of native MER3 mRNA was dependent on Mer1 (53% in Mer1-expressing cells versus 3% in control cells) and Nam8 (2% splicing in Mer1-expressing nam8 Δ cells) ( Figure 3 B). Introducing a perfect GUAUGU 5′ splice site (via a single U insertion in mutant MER3-5′SS ) increased basal splicing efficiency to 75% in wild-type yeast lacking Mer1, and enhanced the extent of splicing in the presence of Mer1 to 84% ( Figure 3 B). By contrast, the consensus BP change per se elicited a relatively modest elevation in basal MER3 splicing (to 12%) while the MER3-BP intron remained Mer1-dependent, with splicing efficiency rising to 71% in wild-type cells expressing Mer1 ( Figure 3 B). The effects of combining the MER3 5′SS and BP changes (resulting in 78% and 86% splicing in the absence and presence of Mer1, respectively) were virtually identical to those seen for the single 5′SS change ( Figure 3 B). These results signify that the aberrant MER3 5′ splice site is the principal determinant of the Mer1 requirement for splicing in vivo .
The 5′SS change eliminated the dependence of MER3 splicing on Nam8, elevating splicing efficiency to 83% in nam8 Δ cells lacking Mer1 ( Figure 3 B). The consensus BP change conferred no such benefit per se (6% splicing with or without Mer1) and had no additive effects in nam8Δ cells when combined with the consensus 5′SS mutation ( Figure 3 B). We surmise that the weak 5′ splice site is the critical determinant of the Nam8 requirement for efficient MER3 splicing.
Requirements for Mer1 enhancer elements in the MER3 and SPO22 introns
Spingola and Ares ( 15 ) identified a cis -acting intronic splicing enhancer, of consensus sequence 5′-AYACCCUY, located within 25-nt of the 5′ splice site in three of the Mer1-responsive meiotic transcripts: AMA1 , MER2 and MER3 . Focusing on the AMA1 intron, they showed that multi-base substitutions within the enhancer element abolished Mer1 splicing activation without affecting the basal level of splicing in vegetative cells not expressing Mer1 ( 15 ). Here, we conducted a mutational analysis of the corresponding Mer1 enhancer elements in the MER3 and SPO22 introns. First, the invariant central CCC triplets in the putative enhancers of the MER3 and SPO22 transcripts were changed to GGG. In both cases, this maneuver abolished Mer1-activated splicing ( Figure 4 A and B). We then proceeded to conduct a finer analysis of the SPO22 intronic enhancer, by replacing the conserved cytidines singly with guanosine. The GCC and CGC variants of SPO22 were unresponsive to Mer1; their splicing efficiencies in the presence of Mer1 (24% and 26%, respectively) were little better than that of wild-type SPO22 in the absence of Mer1 (18%) ( Figure 4 C). The SPO22-CCG variant was partially responsive to Mer1, attaining 41% splicing efficiency in Mer1-expressing cells, versus 88% for the wild-type SPO22 control ( Figure 4 C).
To better define structure-activity relations in the SPO22 enhancer, the central cytidines were replaced with uridines, singly and en masse ( Figure 4 D). The SPO22-UUU mutant was spliced with 45% splicing efficiency in Mer1-expressing cells, versus 86% for the wild-type SPO22 control ( Figure 4 D). The triple C-to-U change was clearly less deleterious than the triple C-to-G mutation ( Figure 4 B). Virtually all of the impact of the UUU mutation could be attributed to the change in the first pyrimidine position, insofar as the UCC variant (52% splicing) resembled UUU , whereas the single changes as the second and third positions had no apparent effect on the extents of SPO22 splicing in cells expressing Mer1 (85% and 89% for the CUC and CCU variants, respectively) ( Figure 4 D).
Effects of Mer1 mutations
Saccharomyces cerevisiae Mer1 is a 270-aa polypeptide composed of two functional modules: a C-terminal KH domain implicated in RNA binding and an N-terminal domain that interacts with the spliceosome and promotes splicing of target mRNAs ( 22 ). The KH domain is an ancient protein fold, widely distributed among taxa, that binds single-stranded RNA or DNA ( 39 ). KH domains of the type found in Mer1 consist of a three-stranded antiparallel β sheet plus three α-helices packed against one face of the sheet ( Figure 5 A, top panel). Fenn et al . ( 40 ) solved the 1.6 Å crystal structure of a KH domain from the human poly(C)-binding protein PCBP2 in complex with a 7-nt single-stranded DNA oligonucleotide that included the internal pentamer sequence 5′-ACCCT, which is the DNA counterpart of the Mer1 RNA enhancer sequence 5′-ACCCU. Reference to the PCBP2–ssDNA complex shows that recognition of the pyrimidine-rich sequence is achieved, in large part, via direct and water-mediated hydrogen bonding interactions of KH amino acid side chains with the polar edge atoms of the T and C nucleobases ( Figure 5 A, top panel). The four side chains that engage the bases in PCBP2 emanate from the β2 and β3 strands and the α2 helix. Three of these pyrimidine-binding residues are conserved in yeast Mer1 (as Arg209, Arg214 and Lys222); the fourth residue, an arginine in PCBP2, is replaced by glutamine in the yeast Mer1 and homologous KH domains found in several other fungal proteomes ( Figure 5 A, bottom panel).
Guided by the PCBP2 structure, we introduced alanine mutations in lieu of Mer1 residues Arg209, Arg214, Lys222 and Gln243 and tested the mutant alleles for their ability to activate splicing of SPO22 pre-mRNAs in vegetative yeast cells ( Figure 5 B). Whereas R209A and K222A were as active as wild-type MER1 in promoting SPO22 intron removal (82–84% spliced), the R214A and Q243A mutants were nonfunctional. The 16% SPO22 splicing seen in MER1-R214A and MER1-Q234A cells was little better than the 13% SPO22 splicing in yeast cells that did not express Mer1 ( Figure 5 B, top panel). Structure-activity relations in the Mer1 KH domain were garnered by introducing conservative substitutions for the essential Arg214 and Gln243 side chains, whereby Arg214 was changed to lysine and glutamine and Gln243 was changed to asparagine, glutamate and arginine (arginine being the native residue at the equivalent position in PCBP2). We found that whereas the R214K change revived SPO22 splicing activity to the level of wild-type MER1 , the R214Q mutant was no better than R214A ( Figure 5 B, bottom panel). We surmise that a positively charged side chain at position 214 of the KH domain is necessary for Mer1 activity. The Q243E change elicited a significant, though incomplete recovery of Mer1 function (to 69% SPO22 splicing versus 87% for wild-type MER1 ) that contrasted with the Q243N and Q243R mutations (22% and 16% SPO22 splicing respectively, which were similar to the 18% splicing efficiency in control cells that did not express Mer1) ( Figure 5 B, bottom panel). We conclude that: (i) a critical distance is required from the protein main chain to the Gln243 amide (i.e. the shorter Asn residue does not suffice); (ii) the putative RNA contacts made by Gln243 differ from those made by the corresponding arginine in the human PCBP2–DNA complex, insofar as a basic arginine side chain was inactive in Mer1, though an acidic glutamate residue functioned almost as well as the native glutamine.
The MER1 mutants were also surveyed for their activity in promoting the splicing of a different meiotic mRNA target— MER2 ( Figure 5 C). The results agreed with the findings for SPO22 , to wit: (i) Arg209 and Lys222 were not important for MER2 splicing, but Arg214 and Gln243 were essential; and (ii) R214K and Q243E restored activity in MER2 splicing, though the other conservative changes did not.
Determinants of the Nam8-dependence of PCH2 splicing
Unlike the Mer1-regulated Nam8-dependent transcripts, the PCH2 pre-mRNA has a perfect consensus 5′ splice site and lacks a Mer1 intronic enhancer sequence. Rather, the PCH2 transcript is distinguished by its exceptionally long 5′ exonic open reading frame (1551-nt) and a non-consensus intron branchpoint sequence, 5′-CACUAAC ( Figure 6 ). We probed the PCH2 splicing determinants by: (i) installing a consensus branchpoint sequence in an otherwise native PCH2 gene; (ii) deleting most of the long upstream exon (while installing a new in-frame AUG codon) to create a PCH2-Δ5 ′ variant with a 51-nt 5′ exon; and (iii) combining the BP and Δ5′exon changes. We tested the PCH2 mutants on 2 µ plasmids for their splicing efficiency in NAM8 and nam8 Δ yeast cells. The results implicate the non-consensus branchpoint and long 5′ exon as separate negative influences on PCH2 splicing in wild-type cells. For example, shortening the 5′ exon per se increased splicing efficiency to 89% for PCH2-Δexon versus 70% for the PCH2 transcript ( Figure 6 ). The consensus branchpoint change per se increased splicing efficiency to 91% for PCH2-BP ( Figure 6 ). There was no apparent effect of combining the BP and Δexon changes in wild-type cells. The PCH2 transcript was spliced with 22% efficiency in nam8Δ cells, a 3-fold decrement compared to wild-type cells. The PCH2-Δexon transcript was spliced at 83% efficiency in nam8 Δ cells, signifying that shortening the 5′ exon sufficed to override the Nam8 requirement. The PCH2-BP transcript was spliced with 63% efficiency in nam8Δ cells, implying that the nonconsensus branchpoint is an independent determinant of Nam8-dependence of PCH2 splicing, albeit perhaps not as strong a factor as 5′ exon length. Combining the BP and exon changes elicited a further gain of splicing in nam8Δ cells to 92% ( Figure 6 ).
Nam8 is a stable component of the U1 snRNP and can be crosslinked to intronic RNA flanking the 5′ splice site ( 16–18 ), though the sites of RNA contact within Nam8 have not been mapped. Nam8 contains three tandem RRM (RNA recognition motif) domains that are candidates to mediate the imputed intron RNA-binding function of Nam8. The mammalian TIA-1 protein—also a triple RRM-containing polypeptide and a putative homolog of yeast Nam8—has been shown to regulate pre-mRNA splicing by binding to U-rich sequences flanking the 5′ splice site ( 19 ). Thus, it is noteworthy that the PCH2 intron contains two U-rich polypyrimidine tracts: an 11-mer element (5′-UUUUUUCUUU) immediately adjacent to the 5′ splice site, and a 24-mer element (5′-UUCUUUUUCUUUCUUUCCUUCUUC) located between the 5′ splice site and the branchpoint ( Figures 6 and S1). We considered that one or both of these PCH2 polypyrimidine tracts might act as Nam8 enhancers (akin to the Mer1 enhancer discussed above) and thereby be important for the observed Nam8-dependence of PCH2 splicing. To test this idea, we deleted the proximal or distal U-rich tracts en bloc from an otherwise wild-type PCH2 transcript (retaining the long 5′ exon and deviant branchpoint) and assayed splicing efficiency of these mutants pre-mRNAs (named UΔ1 and UΔ2 respectively) in NAM8 and nam8Δ yeast ( Supplementary Figure S1 ). Splicing of the UΔ1 transcript in NAM8 cells (72%) was similar to that of the wild-type PCH2 RNA (69%). Deleting the distal polypyrimidine tract in UΔ2 had a mild effect of splicing efficiency (56% in NAM8 cells) but did not alleviate Nam8 dependence, i.e. 18% of the UΔ2 transcript was spliced in nam8 Δ cells ( Supplementary Figure S1 ). Splicing of mutant transcript UΔ1 + UΔ2 , in which both U-tracts were deleted, was similar to the UΔ2 single-deletant ( Supplementary Figure S1 ). These results exclude the idea that Nam8 promotes PCH2 splicing via obligate specific interactions with one or both of the U-rich polypyrimidine element of the PCH2 intron.
The PCH2 intron is a portable determinant of Nam8 dependence
We inserted the 113-nt PCH2 intron within the otherwise intron-less chromosomal yeast HIS3 gene of NAM8 and nam8 Δ haploid yeast cells. The PCH2 intron was initially placed near the 3′ end of the HIS3 ORF (after nucleotide +430) to reflect its distal position in the native PCH2 pre-mRNA ( Figure 7 A). HIS3 provides a convenient reporter for gene expression, manifest as histidine prototrophy. The native HIS3 gene is functional in NAM8 and nam8 Δ strains, both of which grow on agar medium lacking histidine ( Figure 7 B). By contrast, the HIS3-[PCH2] gene containing the distally inserted PCH2 intron was functional in NAM8 cells, but not in the nam8Δ background ( Figure 7 B). The salient finding was that a single C-to-U mutation in the intron of the HIS3-[PCH2] reporter that restored a consensus branchpoint ( Figure 7 A) sufficed to restore HIS3 function in nam8 Δ cells ( Figure 7 B). These results show that the PCH2 intron, by virtue of its deviant branchpoint, is an autonomous determinant of Nam8-dependence in the absence of other PCH2 gene elements (i.e. promoter, 5′ and 3′ exons, 5′ and 3′ untranslated regions). To evaluate the contributions of 5′ exon length to the Nam8-dependence of PCH2 intron removal, we repositioned the PCH2 intron proximally within the HIS3 reporter (inserting it at nucleotide +48). This maneuver sufficed to allow nam8 Δ cells to grow in the absence of added histidine (data not shown). These results implicate Nam8 in recognition and/or utilization of the nonconsensus PCH2 branchpoint in the context of a pre-mRNA substrate with a relatively long 5′ exon.
Domain requirements for Nam8 function
The present findings that Nam8 governs two distinct meiotic splicing pathways prompted us to demarcate the functional domains of Nam8, about which virtually nothing is known, and to gauge whether different Nam8-dependent events rely on different structural features of the Nam8 protein. Nam8 is a 523-aa polypeptide ( Figure 8 A) that includes three RRM domains ( Figure 8 B). The closely spaced RRM1-RRM2 tandem unit is preceded by a 40-aa N-terminal leader peptide and separated from the downstream RRM3 domain by a hydrophilic 63-aa inter-RRM linker peptide rich in Ser ( n = 12), Asn ( n = 11) and Gln ( n = 8) ( Figure 8 A). Distal to RRM3 is a hydrophilic 130-aa C-terminal extension that is also rich in Ser ( n = 17), Asn ( n = 18) and Gln ( n = 15) ( Figure 8 A). Proteins with primary structure similarity across the entire Nam8 polypeptide can be found in the proteomes of other fungi, including Neurospora crassa (491-aa), Kluyveromyces lactis (589-aa), Candida glabrata (555-aa) and Ashbya gossypii (566-aa). Many other fungi encode shorter proteins (∼400-aa) that retain the three RRMs, but are missing the Nam8 C-terminal domain. An alignment of the primary structures of S. cerevisiae and A. gossypii Nam8 ( Figure 8 A) helped guide the design of serial N and C terminal truncations of S. cerevisiae Nam8 at the putative domain (or subdomain) boundaries indicated by the forward and reverse arrowheads in Figure 8 A. The truncated open reading frames were cloned into CEN plasmids under the control of the native NAM8 promoter and tested for function in vivo .
Because deletion of NAM8 has no apparent impact on yeast growth at 18°C to 37°C, we took advantage of synthetic genetic interactions between nam8Δ and deletions of other nonessential yeast genes implicated in pre-mRNA splicing to assay for complementation of growth by the Nam8 truncation mutants. For example, a nam8Δ tgs1Δ strain that lacks Nam8 and Tgs1 (a nonessential enzyme that synthesizes the trimethylguanosine cap found on the U1, U2, U4 and U5 snRNAs) is slow growing at 37°C, but inviable at ≤30°C ( 32 ). Growth of nam8Δ tgs1Δ at 30°C was restored by a CEN plasmid encoding full-length Nam8 (aa 1–523), whereas the empty CEN vector was ineffective ( Figure 8 B). Deletion of the N-terminal 40-aa leader peptide preceding RRM1 had no apparent impact on complementation of nam8Δ tgs1Δ , nor did truncations at the C-terminus to positions 489 or 454 ( Figure 8 B). Moreover, simultaneous deletion of aa 1–40 and 455–523 in NAM8-(41–454) also did not affect complementation ( Figure 8 B), signifying that the N and C terminal segments are dispensable. By contrast, the next incremental C-terminal deletions, to positions 400 or 280, ablated nam8Δ tgs1Δ complementation ( Figure 8 B). We conclude that: (i) Nam8-(1–400), which contains all three RRMs, is not sufficient for Nam8 activity in this genetic context and (ii) the Nam8 segment from aa 401–454 flanking RRM3 is essential for Nam8 activity. [Prior information on the C-terminus of Nam8 was limited to the observation that a premature stop codon at amino acid 140 eliminated Nam8 function in a different synthetic lethal complementation assay ( 16 ).]
The effects of incremental N-terminal deletions that serially removed the RRMs were also instructive. Deletion of the entire RRM1 domain (in the 151–523 and 151–454 alleles) had little impact on nam8Δ tgs1Δ complementation, signifying that RRM2 plus RRM3 sufficed for activity. However, deletions of both RRM1 and RRM2 (in the 251–523 and 251–454 alleles) resulted in diminished complementation activity, as gauged by the smaller size of NAM8-(251–523) tgs1Δ colonies vis à vis the preceding longer versions of Nam8 ( Figure 8 B). Whereas further deletions into the inter-RRM linker did not worsen Nam8 function (see 281–523 and 291–523 ), complementation was abolished by truncating the N terminus to aa 304 ( Figure 8 B), which is the predicted proximal margin of the RRM3 domain ( Figure 8 A). This result implicates the 291 MSQFIYPVQQQPS 303 peptide flanking RRM3 as important for Nam8 function in this genetic assay; it is noteworthy that this peptide sequence at the distal end of the inter-RRM linker is well conserved in A. gossypii Nam8 (11/13 positions of amino acid identity), whereas the rest of the linker is not conserved ( Figure 8 A).
To gauge whether the domain requirements for Nam8 function varied with the genetic background in which Nam8 becomes necessary for yeast growth, we performed a parallel series of complementation assays in a nam8 Δ lea1 Δ strain. A synthetic interaction of Nam8 and Lea1 (an inessential stable constituent of the U2 snRNP) was detected in a recent genome-wide genetic array analysis ( 41 ). A lea1Δ single mutant is temperature-sensitive; it grows well at 20° to 34°C, but very slowly at 37°C ( 32 , 42 ). Here, we constructed a nam8Δ lea1Δ double mutant, which was synthetic lethal and required plasmid p360-NAM8 ( CEN URA3 NAM8 ) for viability at 30°C ( Supplementary Figure S2A ). The nam8Δ lea1Δ p360-NAM8 strain enabled us to test complementation by plasmid shuffle. Cells transformed with a CEN LEU2 plasmid bearing full-length NAM8 plasmid grew on medium containing FOA (a drug that selects against the URA3 plasmid), while cells transformed with the empty CEN LEU2 vector (or a vector encoding an inactive Nam8 mutant) did not (Figure S2A). We found that NAM8 truncation mutants 41–523 , 151–523 and 1–454 permitted growth of ‘wild-type’ sized colonies on FOA, while the 251–523 , 281–523 and 291–523 alleles yielded small colonies (data not shown). However, cells transformed with mutants 304–523 or 1–400 did not grow on FOA (data not shown). The viable strains were grown in liquid medium and then spot-tested for growth on YPD agar (Figure S2B). Deletion of the N leader peptide plus RRM1 or the C-terminal domain (aa 455–523) had little impact on cell growth per se , or in combination ( Supplementary Figure S3B ). Further N-terminal deletions of RRM2 and parts of the inter-RRM linker resulted in slowed growth, e.g. reflected in smaller colony size of the 291–523 and 251–454 strains ( Supplementary Figure S2B ). We surmise that the domain requirements for Nam8 function trend similarly in the lea1 Δ and tgs1 Δ genetic backgrounds.
The synthetic genetic interactions of Nam8 also embrace Mud1 ( 41 ), an inessential component of the U1 snRNP ( 43 ). We constructed a nam8Δ mud1Δ strain bearing a CEN URA3 NAM8 plasmid. Although this strain did yield small FOA-resistant colonies under selection at 30°C when transformed with an empty CEN LEU2 plasmid (data not shown), the vector control cells grew poorly on YPD agar ( Supplementary Figure S3 ). By contrast, transformation with a CEN LEU2 plasmid expressing full-length NAM8 resulted in normal-sized FOA-resistant colonies and the cells derived from these colonies grew well on YPD agar ( Supplementary Figure S4 , 1–523 ). The salient findings were that the Nam8 RRM domain requirements for complementation of nam8Δ mud1Δ were less stringent than what we observed for nam8Δ tgs1Δ and nam8Δ lea1Δ . To wit, Nam8-(291–523) and Nam8-(291–454), which lack the leader, RRM1, RRM2 and most of the inter-RRM linker, were as effective as full-length Nam8 in supporting growth in the mud1Δ background ( Supplementary Figure S3 ). Moreover, the 304–523 mutant, which was inactive in the tgs1Δ and lea1Δ strains ( Figures 8 B and S2), displayed weak complementation activity in mud1Δ ( Supplementary Figure S3 ).
A different trend was observed when we tested Nam8 activity in complementation in the synthetic lethal nam8Δ mud2Δ background ( 44 ). Mud2 is an inessential RRM-containing protein that forms a stable complex with the yeast branchpoint binding protein (BBP) that recruits the U2 snRNP during spliceosome assembly ( 45 , 46 ). We constructed a nam8Δ mud2Δ double-mutant, which required either plasmid p360-MUD2 ( CEN URA3 MUD2 ) or p360-NAM8 ( CEN URA3 NAM8 ) for viability (Figure S4A). Growth of these strains on FOA agar was allowed after transformation with CEN LEU2 plasmids bearing either MUD2 or NAM8 , but not after transformation with the empty CEN LEU2 vector ( Supplementary Figure S4A ). We found that NAM8 N-terminal truncation mutants 41–523 , 151–523 , 251–523 and 281–523 permitted growth of nam8Δ mud2Δ cells on FOA, whereas mutants 291–523 and 304–523 did not (data not shown). The viable NAM8 N-deletants were grown in liquid medium and then spot-tested for growth on YPD agar ( Supplementary Figure S4B ). These results indicate that the leader, RRM1, and RRM2 are dispensable for Nam8 function in the mud2 Δ background, though the distal part of the inter-RRM linker is essential. The C-terminal deletant 1–454 supported normal growth in the mud2 Δ background ( Supplementary Figure S4B ), but the 1–400 mutant was lethal (no growth on FOA; not shown). Simultaneous deletion of aa 1–40 and 455–523 in NAM8-(41–454) did not affect complementation of nam8Δ mud2Δ ( Supplementary Figure S4B ). However, combining the RRM1 deletion with the C-terminal truncation clearly weakened NAM8-(151–454) function in the mud2 Δ background ( Supplementary Figure S4B ).
Finally, we tested Nam8 function in a different genetic assay predicated on the growth-inhibitory effects of Nam8 overexpression ( 47 ). NAM8 and truncated versions thereof were installed on 2 µ URA3 plasmids under the control of a galactose-inducible promoter. The plasmids were introduced into wild-type cells, which were then tested for growth on –Ura medium containing either glucose ( NAM8 expression repressed) or galactose ( NAM8 expression induced). Whereas all of the strains grew normally on glucose, the galactose-induced expression of full-length Nam8 from the 2 µ plasmids inhibited colony formation, an effect not seen with the empty 2 µ vector ( Supplementary Figure S5 ). Galactose-dependent growth inhibition was evident for the single N- and C-terminal truncations that scored as active in the synthetic lethal complementation assays: 41–523 , 151–523 , 1–489 and 1–454 , again attesting the dispensability of the leader, RRM1, and C-terminal segment ( Supplementary Figure S5 ). The more extensively truncated versions of NAM8 that were inactive in the complementation assays ( 1–400 , 1–280 and 304–523 ) had little or no effect on growth when induced by galactose at high gene dosage. Between these extremes, we saw that N-terminal deletions embracing RRM2 ( 251–523 and 281–523 ) alleviated the inhibitory effects of galactose induction, albeit not completely ( Supplementary Figure S5 ). The enfeebled dominant negative activity of the RRM2 deletants was consistent with their weakened complementation activity in the tgs1Δ and lea1Δ backgrounds.
Nam8 domain requirements for splicing of SPO22 and PCH2 pre-mRNAs
The full-length and truncated NAM8 alleles were tested in nam8 Δ haploids for their activities in Mer1-dependent splicing of the MER2 pre-mRNA ( Figure 9 A) and Mer1-independent splicing of PCH2 pre-mRNA ( Figure 9 B). MER2 splicing efficiency was unaffected by loss of the N-terminal leader and C-terminal tail segments, singly or in combination ( Figure 9 A). Deletion of RRM1 had little impact −66% MER2 splicing efficiency in NAM8-(151–523) cells versus 77% in NAM8-(41–523) cells. However, deletion of RRM1 and RRM2 sharply reduced MER2 splicing efficiency—to 13% in NAM8-(251–523) cells, a level barely above the 7% splicing seen in the nam8Δ strain ( Figure 9 A). A C-terminal deletion of the segment from aa 401–454 abolished MER2 splicing ( Figure 9 A). Thus, the domain requirements for MER2 splicing in the presence of Mer1 reflect the trends seen for complementation of nam8Δ tgs1Δ (and are apparently more demanding than those for Nam8-dependent vegetative growth in mud1 Δ and mud2 Δ backgrounds).
The effects of Nam8 domain deletions on PCH2 pre-mRNA splicing ( Figure 9 B) were qualitatively similar to those for MER2 splicing, although the incremental removal of RRM2 caused a larger reduction in PCH2 splicing—42% in NAM8-(151–523) cells versus 62% in NAM8-(41–523) cells and 21% in the nam8Δ controls.
Targeted mutations in Nam8 RRM2 and RRM3 guided by the structure of TIA-1
The RRM fold comprises a four-stranded β-sheet flanked on one side by two α-helices, as exemplified in the crystal structure of the RRM2 domain of human TIA-1 ( 48 ) ( Figure 10 A), the putative homolog of the Nam8 RRM2 domain, to which it aligns with 43/80 positions of amino acid identity/similarity ( Figure 10 B). RRM domains typically bind RNA along the exposed surface of the β-sheet opposite the α-helices ( 49 ). The central β1 and β3 strands of the sheet (referred to as the RNP2 and RNP1 motifs, respectively, and denoted by brackets under the Nam8 primary structure in Figure 8 A) make most of the atomic contacts with bound RNA.
The present findings that interval deletions of RRM2 plus RRM3 abolished all Nam8 activities prompted us to test the effects of alanine mutations in the putative RNA-binding surfaces of RRM2 and RRM3. The alanine changes were introduced into Nam8-(41–454), a biologically active allele that includes all three tandem RRMs, and the mutants were assayed for complementation of the nam8Δ tgs1Δ strain. With respect to RRM2, we tested a single-alanine substitution at Phe166 in the β1 strand RNP2 motif and a triple alanine substitution for Lys205, Tyr207 and Phe209 in the β3 strand RNP1 motif; these side chains project outward on the surface of the β-sheet opposite the two α-helices ( Figure 10 A) and would correspond to residues in canonical RRMs that interact with RNA. We also combined the RNP2 and RNP1 mutations to generate a quadruple-alanine cluster allele, F166A-K205A-Y207A-F209A . The F166A change per se had little effect on Nam8 function in this assay, as gauged by colony size; however, the triple RNP1 mutant and the quadruple mutant resulted in slowed growth and tiny colonies ( Figure 10 C), similar to the effects of deleting the RRM2 domain on Nam8 function in the tgs1Δ background ( Figure 8 B, 251–454 ).
In RRM3, we introduced a single-alanine substitution at Phe316 in the β1 strand and a double-alanine substitution for Lys348, and Phe352 in the β3 strand; also we combined these to make a triple mutant, F316A-K348A-F352A . Here, the double RNP1 mutant and the combined RNP2/RNP1 triple mutant were extremely sick ( Figure 10 C), attesting to a critical contribution of RRM3, and its putative RNA-binding surface, to Nam8 activity when RRMs 1 and 2 are present.
The effects of the RRM mutations on Nam8-dependent splicing of MER2 pre-mRNA were gauged by transforming nam8Δ haploids constitutively expressing Mer1 with CEN plasmids bearing the ‘wild-type’ or alanine-substituted NAM8-(41–454) alleles. Subtraction of the side chains at the putative RNA binding surface of the RRM2 and RRM3 domains reduced MER2 splicing to background levels ( Figure 10 C).
Regulated meiotic splicing in budding yeast is an elegant example of a post-transcriptional ‘off–on’ switch linked to a developmental program. Whereas only ∼5% of yeast genes contain introns, there is an evident enrichment for introns among the genes encoding meiosis-specific proteins, and a tendency for meiotic introns to bear splicing signals that deviate from the norm ( 31 , 38 ). It has been suggested that upregulation of splicing during meiosis—via programmed transition from the splicing ‘off’ state during vegetative growth—provides added protection against the untimely production of meiotic proteins that could be deleterious to vegetative cells ( 11 , 31 ). Prior studies had characterized a single meiotic splicing regulon controlled by Mer1 and Nam8 and embracing three pre-mRNA targets. The regulatory inputs (if any) to the splicing of the other known meiosis-specific pre-mRNAs are a virtual tabula rasa ( 31 ). Here, by surveying the full catalogue of meiosis-specific spliced pre-mRNAs for their dependence on Nam8, we’ve expanded the scope of the Mer1/Nam8 regulon to embrace SPO22 and illuminated a novel Mer1-independent role for Nam8 in splicing of PCH2 pre-mRNA. The two flavors of Nam8-dependent splicing differ fundamentally. The Mer1/Nam8-regulated meiotic transcripts share two properties: (i) they have an intronic enhancer sequence, to which Mer1 is thought to bind via its KH domain and (ii) they have non-consensus 5′ splice sites. By contrast, the PCH2 transcript lacks a Mer1 intronic enhancer and contains a perfect 5′ splice site.
We focused here on dissecting the contributions of the Mer1 enhancer sequence, Mer1 KH domain, and other SPO22 intronic signals to splicing of SPO22 pre-mRNA, the newly identified Nam8 target. Whereas SPO22 has multiple potentially enfeebling non-consensus intronic signals (5′SS, BP and 3′SS), we find that the atypical 5′SS is the decisive factor in Nam8/Mer1 dependence, insofar as a single nucleotide change in the SPO22 intron that restores a consensus 5′SS overrides the requirements for Mer1 and Nam8. The same scenario applies to the MER3 intron.
Our studies of PCH2 splicing implicate the non-consensus 5′-CACUAAC branchpoint and the exceptionally long 5′ exon as separable negative influences on PCH2 splicing in wild-type cells and as concerted determinants of Nam8-dependence. The finding that Nam8-dependence is portable with the PCH2 intron, when inserted into the HIS3 reporter, certifies the importance of the intron and its deviant branchpoint as decisive elements per se in the Nam8 requirement. The action of Nam8 in countering a suboptimal PCH2 branchpoint in the context of a perfect 5′ splice site differs from the case of Mer1/Nam8-dependent splicing, where the intron-bound Mer1 facilitates recruitment of the Nam8-containing U1 snRNP to a suboptimal 5′ splice site ( 14 , 15 , 26 ) ( Figure 11 A). Current models invoke bridging contacts between U1 snRNP components Nam8, Snu56 and Snu71 with the enhancer-bound Mer1 in mediating U1 recruitment to the deviant 5′ splice site ( 22 , 26 ) ( Figure 11 A). We speculate that Nam8 facilitates PCH2 splicing via macromolecular interactions that recruit the U2 snRNP to the deviant branchpoint ( Figure 11 B). The participants in this putative interaction network remain to be identified. Toward that end, the Nam8-responsive HIS3-[PCH2] reporter that we developed here may prove useful in forward genetic screening for yeast mutants that either lose their His prototrophy in NAM8 cells (as candidate cofactors for Nam8-dependent splicing) or mutants that acquire His prototrophy in nam8 Δ cells (as candidate suppressors of Nam8-dependency).
The present study illuminates a first set of instructive structure–function relations for the yeast Nam8 protein. As depicted in Figures 8 and 11 , Nam8 consists of three RRM domains flanked by N-terminal ‘leader’ and C-terminal ‘trailer’ segments. Whereas the distal segment of the trailer domain (aa 455–523) is unnecessary for any Nam8 activity surveyed, our results demonstrate an essential contribution by the proximal part of the Nam8 C-terminal domain (aa 401–454) in all aspects of Nam8 function. One or more of the RRMs are plausible candidates for the imputed intron RNA-binding properties of Nam8. Here, we showed that the N leader and RRM1 are dispensable for Nam8-dependent splicing of meiotic pre-mRNA targets and for Nam8 function in vegetative yeast growth in a variety of synthetic genetic backgrounds. Thus, RRM1 is not implicated in Nam8′s principal contacts to RNA or other components of the splicing apparatus.
Deleting RRM2 along with RRM1 resulted in: (i) disabled Nam8-dependent splicing of MER2 and PCH2 pre-mRNAs; (ii) variable effects, ranging from loss of activity to no impact, on Nam8-dependent growth, according to the genetic background being assayed. Because it is possible that RRM1 and RRM2 are functionally redundant, we made an independent test of RRM2 function by targeting alanine substitutions to the putative RNA-binding surface of RRM2 in the context of an otherwise active Nam8 protein that contains all three RRMs. The decrements in vegetative growth in the tgs1 Δ background and in splicing of MER2 pre-mRNA implicated RRM2 as a direct participant in Nam8 activity. Nakagawa and Ogawa ( 21 ) reported previously that a single mutation, Leu170Pro, in the RNP2 motif (IFVGDL 170 ) of Nam8′s RRM2 domain resulted in a severe defect in meiotic recombination, comparable to that of a nam8 Δ null. However, we deem this mutational effect as poorly instructive with regard to RRM function, because reference to the equivalent VFVGDL motif in the TIA-1 RRM2 crystal structure ( Figure 10 B) shows us that the Leu170 side chain is not projecting outward onto the putative RNA-binding surface, but rather points into the hydrophobic core of the RRM (data not shown). Thus, a proline change at this position would be more likely to impact the folding of the RRM2 module than to specifically affect an RNA-binding site.
Extending the N-terminal deletions to the proximal margin of RRM3 (aa 304) abolished or severely weakened all of the Nam8 functions tested. As mentioned above, we attempted to isolate RRM3′s contributions by making alanine changes in the predicted RRM3 RNA-binding surface. The severe defects in vegetative growth in the tgs1 Δ background and in splicing of MER2 pre-mRNA caused by the RRM3 alanine clusters implicated RRM3, too, as important for Nam8 activity. Although the results are consistent with an RNA-binding role for either RRM2 or RRM3, it is plausible that one (or both) of the Nam8 RRMs mediates important protein–protein interactions, rather than RNA binding, as has been described for other RRM domains ( 50 ).
It is interesting to note that the least demanding genetic function of Nam8 is seen in the mud1 Δ background, where Nam8-(291–454)—which lacks RRMs 1 and 2—suffices for normal growth. Mud1, like Nam8, is an intrinsic protein component of the yeast U1 snRNP. We presume that the synthetic growth phenotype of the nam8Δ mud1Δ double-mutant reflects a gross defect in U1 snRNP structure and function, in which case its full complementation by Nam8-(291–454) implies that this minimized version of Nam8 is assimilated into the U1 snRNP lacking Mud1 and renders it active. That Nam8-(291–454) is not adept at (i) splicing Nam8 meiotic targets with suboptimal introns; (ii) complementing yeast synthetic lethal interactions with splicing factors that are not U1 snRNP components (Lea1, Mud2, Tgs1); or (iii) inhibiting vegetative growth when overexpressed suggests that upstream domains (i.e. RRM2 and the interdomain linker) mediate the invoked interactions of Nam8 with non-U1 splicing factors or the pre-mRNA.
Supplementary Data are available at NAR Online.
U.S. National Institutes of Health grants GM52470 (to S.S.); GM50288 (to B.S.). Funding for open access charges: National Institutes of Health (grants GM52470 and GM50288).
Conflict of interest statement . None declared.
S.S. is an American Cancer Society Research Professor.