https://orcid.org/0000-0002-6299-8641
Figures & data
Figure 1. prp45(1–169) impaired the inducibility of PHO5 synthetically with rad6. (A) primary structure and interacting partners of Prp45. Predicted helices are drawn as dark boxes over the primary structure. The 3D distances were taken from the spliceosomal P complex [Citation48]. The diagram of disorder content is based on the values of the AlphaFold prediction [Citation49]. Values below 50 were found to predict intrinsically disordered regions with high reliability [Citation50]. The vertical lines over the primary structure indicate the extents of truncation of the variants used in this study. Interacting partners above the primary structure are drawn based on the proximity between Prp45 and the indicated components in the yeast spliceosomal structures [Citation51]. Contacts with Prp46 (NTR) and Prp8 (tri-snRNP) are the most extensive. Contacts with Prp22 (marked with *), which are not resolved in the cryo-EM structures, are drawn according to the published two-hybrid data [Citation52]. (B) induction of the PHO5 gene in response to phosphate depletion was delayed in cells with truncated Prp45 as compared to WT. WT and prp45(1–169) cells were incubated in non-inducing conditions, washed, and transferred to a medium without phosphate. Total RNA was isolated at the indicated time points and reverse transcribed. RNA levels were quantified by qPCR, normalized to TOM22 reference, and related to the signal obtained from WT cells before the shift. p values were calculated for the comparisons between WT and prp45(1–169) strains and a mutant strain using the t-test (see Methods). Stars indicate differences with p < 0.05. (C, D) the effect of prp45(1–169) on PHO5 induction was tested with rad6Δ (C), paf1Δ, and set1Δ (D) using the same setup as in (B). Error bars represent standard deviations calculated from independent biological replicates. p values, which were obtained for comparisons between strains using the t-test with Holm correction for multiple testing, are listed in STab. 11 (see Methods).
![Figure 1. prp45(1–169) impaired the inducibility of PHO5 synthetically with rad6. (A) primary structure and interacting partners of Prp45. Predicted helices are drawn as dark boxes over the primary structure. The 3D distances were taken from the spliceosomal P complex [Citation48]. The diagram of disorder content is based on the values of the AlphaFold prediction [Citation49]. Values below 50 were found to predict intrinsically disordered regions with high reliability [Citation50]. The vertical lines over the primary structure indicate the extents of truncation of the variants used in this study. Interacting partners above the primary structure are drawn based on the proximity between Prp45 and the indicated components in the yeast spliceosomal structures [Citation51]. Contacts with Prp46 (NTR) and Prp8 (tri-snRNP) are the most extensive. Contacts with Prp22 (marked with *), which are not resolved in the cryo-EM structures, are drawn according to the published two-hybrid data [Citation52]. (B) induction of the PHO5 gene in response to phosphate depletion was delayed in cells with truncated Prp45 as compared to WT. WT and prp45(1–169) cells were incubated in non-inducing conditions, washed, and transferred to a medium without phosphate. Total RNA was isolated at the indicated time points and reverse transcribed. RNA levels were quantified by qPCR, normalized to TOM22 reference, and related to the signal obtained from WT cells before the shift. p values were calculated for the comparisons between WT and prp45(1–169) strains and a mutant strain using the t-test (see Methods). Stars indicate differences with p < 0.05. (C, D) the effect of prp45(1–169) on PHO5 induction was tested with rad6Δ (C), paf1Δ, and set1Δ (D) using the same setup as in (B). Error bars represent standard deviations calculated from independent biological replicates. p values, which were obtained for comparisons between strains using the t-test with Holm correction for multiple testing, are listed in STab. 11 (see Methods).](/cms/asset/b3b6eac7-0849-4ea4-9c9a-753a1e1f5aa8/krnb_a_2348896_f0001_b.gif)
Figure 2. Synthetic genetic array analysis revealed strong negative genetic interactions of prp45(1–169) with genes involved in transcription and chromatin regulations. (A) pie-chart of manually curated categories illustrating the proportions of negative genetic interactions (NGIs) of prp45(1–169). The categories ‘transcription and chromatin’ and ‘RNA processing’ represent 33% and 14% of the NGIs found. Genes were grouped into categories using their GO terms as listed in STab. 6. (B) the prp45(1–169) allele had strong negative interactions with deletions of HTZ1 and SWR1 complex members. prp45(1–169) mutant cells were crossed with htz1Δ, swr1Δ, vps71Δ, vps72Δ, swc3Δ, and swc5Δ strains from the yeast deletion collection. Haploids with indicated combinations of mutations obtained from tetrad dissections of diploid strains harbouring an URA3 plasmid with full length PRP45 (p416ADH-His6-PRP45; [Citation22]) were cultivated to the mid-log phase, serially 5× diluted and spotted on SD plates and SD plates with 5-FOA to get rid of the complementing plasmid. (C) prp45(1–169) shares a high proportion of its genetic interactions with other components of transcription and chromatin regulatory complexes. We overlaid the NGIs of prp45(1–169) onto the network of 165 chromatin regulators constructed in a previously published perturbation analysis [Citation23]. The parts of the network which were enriched for the NGIs of prp45(1–169) (see STab. 7) were re-clustered, using the relative proportions of shared NGIs/all NGIs as a measure (increasing Jaccard indexes were used to cluster the heat-map). As a data source of NGIs genomewide, BioGRID (S. cerevisiae; thebiogrid.Org; downloaded 210127) categories ‘negative genetic’, ‘synthetic growth defect’, ‘synthetic lethality’ and ‘phenotypic enhancement’ were used. Multiple gene list comparator tool (https://www.molbiotools.com/listcompare.Php) was employed for the comparisons between genetically interacting genes and the calculation of Jaccard indexes. Interaction profiles of individual genes were hierarchically clustered and visualized as a heatmap.
![Figure 2. Synthetic genetic array analysis revealed strong negative genetic interactions of prp45(1–169) with genes involved in transcription and chromatin regulations. (A) pie-chart of manually curated categories illustrating the proportions of negative genetic interactions (NGIs) of prp45(1–169). The categories ‘transcription and chromatin’ and ‘RNA processing’ represent 33% and 14% of the NGIs found. Genes were grouped into categories using their GO terms as listed in STab. 6. (B) the prp45(1–169) allele had strong negative interactions with deletions of HTZ1 and SWR1 complex members. prp45(1–169) mutant cells were crossed with htz1Δ, swr1Δ, vps71Δ, vps72Δ, swc3Δ, and swc5Δ strains from the yeast deletion collection. Haploids with indicated combinations of mutations obtained from tetrad dissections of diploid strains harbouring an URA3 plasmid with full length PRP45 (p416ADH-His6-PRP45; [Citation22]) were cultivated to the mid-log phase, serially 5× diluted and spotted on SD plates and SD plates with 5-FOA to get rid of the complementing plasmid. (C) prp45(1–169) shares a high proportion of its genetic interactions with other components of transcription and chromatin regulatory complexes. We overlaid the NGIs of prp45(1–169) onto the network of 165 chromatin regulators constructed in a previously published perturbation analysis [Citation23]. The parts of the network which were enriched for the NGIs of prp45(1–169) (see STab. 7) were re-clustered, using the relative proportions of shared NGIs/all NGIs as a measure (increasing Jaccard indexes were used to cluster the heat-map). As a data source of NGIs genomewide, BioGRID (S. cerevisiae; thebiogrid.Org; downloaded 210127) categories ‘negative genetic’, ‘synthetic growth defect’, ‘synthetic lethality’ and ‘phenotypic enhancement’ were used. Multiple gene list comparator tool (https://www.molbiotools.com/listcompare.Php) was employed for the comparisons between genetically interacting genes and the calculation of Jaccard indexes. Interaction profiles of individual genes were hierarchically clustered and visualized as a heatmap.](/cms/asset/d1b7c5c7-cb26-4949-a9b8-c552d0f579ce/krnb_a_2348896_f0002_oc.jpg)
Figure 3. Negative genetic interactions of prp45(1–169) do not overlap with genes impacted on mRNA levels. To compare the genetic interactions of prp45(1–169) with the effect of the prp45(1–169) mutant allele on transcription, we plotted mRNA levels of intron containing genes as transread ratios mutant/WT, or splicing efficiency, using previously published RNA-seq data and our workflow [Citation67,Citation75]. (A) scatter plot shows the effects of prp45(1–169) on mRNA levels of intron containing genes, expressed as transread ratios mutant/WT, plotted against mean gene expression levels (total RNA count across all samples). Differentially scored transreads are shown in red. (B) for comparison, the effects of prp45(1–169) on splicing efficiency (S.E.) is plotted as in (A). Splicing efficiency was calculated by dividing transread counts by 5’ intron end first base coverage and expressed as mutant/WT ratios. Black data points are genes with sufficient coverage in both WT and mutant strain, i.e. ≥5 transreads and ≥5 reads covering intron end base. Gray data points represent genes which did not fully meet our criteria of sequence read data coverage (see S Fig.3A for details). (C) scatter plot showing relative transread ratios of intron containing genes (same data set as in (A)) as a function of intron length. (D) Genetic interaction enrichment landscapes generated using SAFE annotated genetic interaction similarity network of S. cerevisiae [Citation4,Citation76]. The maps were obtained from the web interface of TheCellMap database (https://thecellmap.org/). Downregulated differentially scored transreads of prp45(1–169) (Subset_45_genes) (STab. 12) (1), complete set of intron containing genes of S. cerevisiae (STab. 13) (2), and all NGIs we found (SGA hit list including manually curated alleles; STab. 14) (3) were visualized in green overlay using the interface. NGIs of prp45(1–169) but not the subset_45 genes showed overlap with the subnetworks of transcription and chromatin regulators. (E)–(G) relative mRNA levels of intron containing genes (same data set as in (A)) were plotted as a function of splicing and chromatin related parameters measured previously. Relative co-transcriptional efficiency (taken from [Citation5]) (E), relative Htz1 density at transcription start sites (TSS) of genes [Citation83] (F), and relative nucleosome turnover at transcription start sites [Citation87] (G) were plotted on the x axis. Only the genes present in both datasets used for the comparison are shown. In (F), the graph does not show the gene YJR145C (coordinates 9.06112; 0.017245623). The parameter ‘observed delta’ increases with increasing co-transcriptional splicing efficiency (‘1’ = 100% co-transcriptional splicing). The parameter ‘Z-score’ increases with decreasing nucleosome turnover. Highly expressed genes, which are mostly ribosomal protein coding genes, cluster at longer intron length (in C), high co-transcriptional splicing efficiency (in E), low levels of Htz1 at transcription start sites (in F), and low relative nucleosome turnover at transcription start sites (in G). See text for discussion.
![Figure 3. Negative genetic interactions of prp45(1–169) do not overlap with genes impacted on mRNA levels. To compare the genetic interactions of prp45(1–169) with the effect of the prp45(1–169) mutant allele on transcription, we plotted mRNA levels of intron containing genes as transread ratios mutant/WT, or splicing efficiency, using previously published RNA-seq data and our workflow [Citation67,Citation75]. (A) scatter plot shows the effects of prp45(1–169) on mRNA levels of intron containing genes, expressed as transread ratios mutant/WT, plotted against mean gene expression levels (total RNA count across all samples). Differentially scored transreads are shown in red. (B) for comparison, the effects of prp45(1–169) on splicing efficiency (S.E.) is plotted as in (A). Splicing efficiency was calculated by dividing transread counts by 5’ intron end first base coverage and expressed as mutant/WT ratios. Black data points are genes with sufficient coverage in both WT and mutant strain, i.e. ≥5 transreads and ≥5 reads covering intron end base. Gray data points represent genes which did not fully meet our criteria of sequence read data coverage (see S Fig.3A for details). (C) scatter plot showing relative transread ratios of intron containing genes (same data set as in (A)) as a function of intron length. (D) Genetic interaction enrichment landscapes generated using SAFE annotated genetic interaction similarity network of S. cerevisiae [Citation4,Citation76]. The maps were obtained from the web interface of TheCellMap database (https://thecellmap.org/). Downregulated differentially scored transreads of prp45(1–169) (Subset_45_genes) (STab. 12) (1), complete set of intron containing genes of S. cerevisiae (STab. 13) (2), and all NGIs we found (SGA hit list including manually curated alleles; STab. 14) (3) were visualized in green overlay using the interface. NGIs of prp45(1–169) but not the subset_45 genes showed overlap with the subnetworks of transcription and chromatin regulators. (E)–(G) relative mRNA levels of intron containing genes (same data set as in (A)) were plotted as a function of splicing and chromatin related parameters measured previously. Relative co-transcriptional efficiency (taken from [Citation5]) (E), relative Htz1 density at transcription start sites (TSS) of genes [Citation83] (F), and relative nucleosome turnover at transcription start sites [Citation87] (G) were plotted on the x axis. Only the genes present in both datasets used for the comparison are shown. In (F), the graph does not show the gene YJR145C (coordinates 9.06112; 0.017245623). The parameter ‘observed delta’ increases with increasing co-transcriptional splicing efficiency (‘1’ = 100% co-transcriptional splicing). The parameter ‘Z-score’ increases with decreasing nucleosome turnover. Highly expressed genes, which are mostly ribosomal protein coding genes, cluster at longer intron length (in C), high co-transcriptional splicing efficiency (in E), low levels of Htz1 at transcription start sites (in F), and low relative nucleosome turnover at transcription start sites (in G). See text for discussion.](/cms/asset/7856384f-5cec-452f-a22e-d82e92f91ef5/krnb_a_2348896_f0003_oc.jpg)
Figure 4. Intron deletion of SRB2 repaired part of the growth defect of prp45(1–169). (A) intron deletion in SRB2 partially repaired the growth defects of prp45(1–169) double mutants carrying deletions of RAD6, PAF1, and GCN5. Cells were cultivated to mid-log phase, serially diluted (ratio 1:4), spotted onto YPAD plates, and incubated at 30°C or 37°C for 6 days. (B) intron deletion in SRB2, but not in HRB1 or VPS75, partially repaired the growth phenotype of prp45(1–169) rad6Δ double mutant. The growth of liquid culture in the YPD medium at 37°C was measured in VarioSkan. (C) SRB2 intron removal had only negligible capacity to rescue the strong negative genetic interaction between htz1Δ and prp45(1–169). Cells were cultivated to mid-log phase, serially diluted (ratio 1:4), spotted onto SD plates containing 5-FOA to dispose of complementing plasmid (p416ADH-His6-PRP45; [Citation22]), and incubated at 30°C, 37°C, or 16°C for 7 days.
![Figure 4. Intron deletion of SRB2 repaired part of the growth defect of prp45(1–169). (A) intron deletion in SRB2 partially repaired the growth defects of prp45(1–169) double mutants carrying deletions of RAD6, PAF1, and GCN5. Cells were cultivated to mid-log phase, serially diluted (ratio 1:4), spotted onto YPAD plates, and incubated at 30°C or 37°C for 6 days. (B) intron deletion in SRB2, but not in HRB1 or VPS75, partially repaired the growth phenotype of prp45(1–169) rad6Δ double mutant. The growth of liquid culture in the YPD medium at 37°C was measured in VarioSkan. (C) SRB2 intron removal had only negligible capacity to rescue the strong negative genetic interaction between htz1Δ and prp45(1–169). Cells were cultivated to mid-log phase, serially diluted (ratio 1:4), spotted onto SD plates containing 5-FOA to dispose of complementing plasmid (p416ADH-His6-PRP45; [Citation22]), and incubated at 30°C, 37°C, or 16°C for 7 days.](/cms/asset/098bffe3-59b9-4a79-a661-2eeddd34f1b2/krnb_a_2348896_f0004_oc.jpg)
Figure 5. Less extensive truncation of Prp45 is sufficient to cause htz1-dependent pre-mRNA accumulation. (A) extended chain of Prp45 connects distinct parts of the spliceosomal architecture. In yeast Bact spliceosome (structure 5gm6; [Citation54]), Prp45 contacts extensively Prp8 (grey) and more than 10 other proteins. We highlighted Prp46 (dark blue), Cef1 (purple), Bud13 (green), Snu17 (blue grey), Pml1 (light blue), and Hsh155 (ocre). The Prp45 fragments are coloured in dark red (1–169), orange (170–247), light orange (248–330), and yellow (331–350). The complex was visualized using ChimeraX [Citation86]. Prp45(1–169) lacks parts of the chain that lines the interface between Prp8 (RT-like domain) and Cef1 (left). Prp45(1–247) is devoid of the contacts to the RES complex components (Pml1, Snu17 and Bud13) and the RNAse H-like domain of Prp8 (right). Truncation of the C-terminal 49 amino acids in Prp45(1–330) should impinge on the interactions with Prp8, Ist3, and Hsh155. Amino acids 1–30 and 350–379 (C-term) were not resolved in the structure and may contain additional interacting partners. (B) growth rate comparison of prp45(1–169), prp45(1–247), and prp45(1–330) cells. Cultivations in YPAD at 30°C were monitored using VarioSkan. In contrast to prp45(1–169) cells, the growth of prp45(1–247) and prp45(1–330) mutants was indistinguishable from the WT strain. (C) Intron deletion of SRB2 partially repaired the cold sensitive growth phenotype of htz1Δ mutant but not of the double mutant htz1Δ prp45(1–330). Cells were cultivated to mid-log phase, serially diluted (ratio 1:4), spotted onto YPAD plates, and incubated at 16°C or 30°C for the indicated number of days. (D) prp45(1–247) and prp45(1–330) mutants accumulated increased levels of pre-mRNA. The mRNA (left) and pre-mRNA levels (right) of ECM33 and ACT1 genes were measured by qPCR in WT, prp45(1–169), prp45(1–247), and prp45(1–330) cells. While the mRNA levels were approximately the same in all four strains, the pre-mRNAs were accumulated to the highest extent in prp45(1–169), followed by prp45(1–330). qPCR values were normalized to TOM22 mRNA and expressed relative to WT strain. Error bars represent the standard deviation of four biological replicates for WT and prp45(1–169) cells and six biological replicates for prp45(1–247) and prp45(1–330) cells. (E) prp45(1–330) and htz1Δ negatively interacted on the level of pre-mRNA accumulation. The mRNA (left) and pre-mRNA levels (right) of ECM33 and COF1 genes were measured by qPCR in WT, htz1Δ, prp45(1–330) and htz1Δ prp45(1–330) cells. The pre-mRNAs showed highest accumulation in the double mutant. qPCR values were normalized to TOM22 mRNA and expressed relative to WT strain. Error bars represent the standard deviation of 8 biological replicates. Statistical significance of the differences between strains in (D) and (E) is indicated as (*) for p ≤ 0.05, (**) for p ≤ 0.01, and (***) for p ≤ 0.001 based on the t-test with Holm correction for multiple testing (see methods and STab. 11).
![Figure 5. Less extensive truncation of Prp45 is sufficient to cause htz1-dependent pre-mRNA accumulation. (A) extended chain of Prp45 connects distinct parts of the spliceosomal architecture. In yeast Bact spliceosome (structure 5gm6; [Citation54]), Prp45 contacts extensively Prp8 (grey) and more than 10 other proteins. We highlighted Prp46 (dark blue), Cef1 (purple), Bud13 (green), Snu17 (blue grey), Pml1 (light blue), and Hsh155 (ocre). The Prp45 fragments are coloured in dark red (1–169), orange (170–247), light orange (248–330), and yellow (331–350). The complex was visualized using ChimeraX [Citation86]. Prp45(1–169) lacks parts of the chain that lines the interface between Prp8 (RT-like domain) and Cef1 (left). Prp45(1–247) is devoid of the contacts to the RES complex components (Pml1, Snu17 and Bud13) and the RNAse H-like domain of Prp8 (right). Truncation of the C-terminal 49 amino acids in Prp45(1–330) should impinge on the interactions with Prp8, Ist3, and Hsh155. Amino acids 1–30 and 350–379 (C-term) were not resolved in the structure and may contain additional interacting partners. (B) growth rate comparison of prp45(1–169), prp45(1–247), and prp45(1–330) cells. Cultivations in YPAD at 30°C were monitored using VarioSkan. In contrast to prp45(1–169) cells, the growth of prp45(1–247) and prp45(1–330) mutants was indistinguishable from the WT strain. (C) Intron deletion of SRB2 partially repaired the cold sensitive growth phenotype of htz1Δ mutant but not of the double mutant htz1Δ prp45(1–330). Cells were cultivated to mid-log phase, serially diluted (ratio 1:4), spotted onto YPAD plates, and incubated at 16°C or 30°C for the indicated number of days. (D) prp45(1–247) and prp45(1–330) mutants accumulated increased levels of pre-mRNA. The mRNA (left) and pre-mRNA levels (right) of ECM33 and ACT1 genes were measured by qPCR in WT, prp45(1–169), prp45(1–247), and prp45(1–330) cells. While the mRNA levels were approximately the same in all four strains, the pre-mRNAs were accumulated to the highest extent in prp45(1–169), followed by prp45(1–330). qPCR values were normalized to TOM22 mRNA and expressed relative to WT strain. Error bars represent the standard deviation of four biological replicates for WT and prp45(1–169) cells and six biological replicates for prp45(1–247) and prp45(1–330) cells. (E) prp45(1–330) and htz1Δ negatively interacted on the level of pre-mRNA accumulation. The mRNA (left) and pre-mRNA levels (right) of ECM33 and COF1 genes were measured by qPCR in WT, htz1Δ, prp45(1–330) and htz1Δ prp45(1–330) cells. The pre-mRNAs showed highest accumulation in the double mutant. qPCR values were normalized to TOM22 mRNA and expressed relative to WT strain. Error bars represent the standard deviation of 8 biological replicates. Statistical significance of the differences between strains in (D) and (E) is indicated as (*) for p ≤ 0.05, (**) for p ≤ 0.01, and (***) for p ≤ 0.001 based on the t-test with Holm correction for multiple testing (see methods and STab. 11).](/cms/asset/a1b98646-3c6e-4eed-a745-468140acbaa8/krnb_a_2348896_f0005_oc.jpg)
Supplemental Material
Download Zip (16.3 MB)Data availability statement
The authors confirm that the data supporting the findings of this study are available within the article and its supplementary materials.