Broad variation in response of individual introns to splicing inhibitors in a humanized yeast strain

  1. Manuel Ares Jr1
  1. 1Center for Molecular Biology of RNA, University of California, Santa Cruz, California 95064, USA
  2. 2Illumina, Inc., Madison, Wisconsin 53719, USA
  1. Corresponding author: ares{at}ucsc.edu

  1. Handling editor: John Woolford

 
Next Section

Abstract

Intron branchpoint (BP) recognition by the U2 snRNP is a critical step of splicing, vulnerable to recurrent cancer mutations and bacterial natural product inhibitors. The BP binds a conserved pocket in the SF3B1 (human) or Hsh155 (yeast) U2 snRNP protein. Amino acids that line this pocket affect the binding of splicing inhibitors like Pladienolide-B (Plad-B), such that organisms differ in their sensitivity. To study the mechanism of splicing inhibitor action in a simplified system, we modified the naturally Plad-B resistant yeast Saccharomyces cerevisiae by changing 14 amino acids in the Hsh155 BP pocket to those from human. This humanized yeast grows normally, and splicing is largely unaffected by the mutation. Splicing is inhibited within minutes after the addition of Plad-B, and different introns appear inhibited to different extents. Intron-specific inhibition differences are also observed during cotranscriptional splicing in Plad-B using single-molecule intron tracking to minimize gene-specific transcription and decay rates that cloud estimates of inhibition by standard RNA-seq. Comparison of Plad-B intron sensitivities to those of the structurally distinct inhibitor Thailanstatin-A reveals intron-specific differences in sensitivity to different compounds. This work exposes a complex relationship between the binding of different members of this class of inhibitors to the spliceosome and intron-specific rates of BP recognition and catalysis. Introns with variant BP sequences seem particularly sensitive, echoing observations from mammalian cells, where monitoring individual introns is complicated by multi-intron gene architecture and alternative splicing. The compact yeast system may hasten the characterization of splicing inhibitors, accelerating improvements in selectivity and therapeutic efficacy.

Keywords

Previous SectionNext Section

INTRODUCTION

Eukaryotic gene expression relies on accurate and timely splicing of pre-mRNA transcripts by the spliceosome. The spliceosome is an RNA–protein complex that assembles anew on each intron and proceeds through a litany of compositional and conformational changes as it executes the reactions necessary to remove the intron and ligate the flanking exons (for review, see Wilkinson et al. 2020). For the most part, the early assembly steps presage the splicing outcome for the mRNA. Recognition of the 5′ splice site (5′ss) by the U1 snRNP and the branchpoint (BP) by the U2 snRNP capture the intron reactive groups that will be gathered at the catalytic site by subsequent rearrangements. Selection of the 3′ splice site (3′ss) is dictated in large part by BP choice, and is very often found at the first YAG that is >7 nt from the BP (Smith et al. 1993; Chua and Reed 2001). Because the early steps in spliceosome assembly have an outsized influence on the spliced product, much effort has been focused on understanding their mechanism and regulation.

The molecular events leading to the stable association of the U2 snRNP with the intron BP are only partially understood. The E-complex, containing pre-mRNA and U1 snRNP (Michaud and Reed 1991, 1993; Li et al. 2019), must merge with a form of the U2 snRNP carrying the SF3a and SF3b proteins (called 17S U2 snRNP in human) (Brosi et al. 1993a,b) to recognize the BP and form the prespliceosome or A-complex, within which the 5′ss and BP are base paired to snRNAs U1 and U2, respectively (Plaschka et al. 2018; Zhang et al. 2020b). As the prespliceosome forms, proteins present in E-complex and the U2 snRNP are released to allow new RNA–RNA interactions to be established. For example, in the yeast E-complex, the intron BP region is bound by a heterodimer of Msl5/Bbp and Mud2 (Berglund et al. 1998; Li et al. 2019), and is inaccessible to its future RNA pairing partner, the BP interacting stem–loop (BSL) of U2 snRNA (Perriman and Ares 2010). In yeast, genetic interactions between the U2AF65 homolog Mud2 and the RNA helicase family member Sub2 (UAP56, BAT1, or DDX39B in humans) suggest that the intron BP might be cleared of proteins by Sub2 (Kistler and Guthrie 2001).

In the U2 snRNP, a U2 snRNA stem–loop structure called the BSL (Perriman and Ares 2010) is bound by Cus2 (yeast), or TATSF1 (humans) (Zhang et al. 2020b; Tholen et al. 2022), which must be removed to allow the BSL to open and pair with the intron to establish the U2-BP helix (Perriman and Ares 2010). In yeast, Cus2 is released by the ATP-dependent activity of the RNA helicase family member Prp5 (DDX46 in humans) (Perriman et al. 2003; Talkish et al. 2019a). Once formed, the U2-BP helix is bound through conformational changes of Hsh155 (in humans, SF3B1), and the BP-adenosine (BP-A) residue enters the BP-A binding pocket formed by HEAT repeats (HRs) 15 and 16 of this conserved U2 protein (Rauhut et al. 2016; Yan et al. 2016; Plaschka et al. 2018). In yeast, Prp5 releases from correctly formed U2-BP helices allowing splicing to proceed, but remains associated with malformed U2-BP helices, serving as a fidelity check (Xu and Query 2007) that prevents the tri-snRNP from binding (Liang and Cheng 2015; Kao et al. 2021). Once bound in the BP-A binding pocket of SF3B1/Hsh155, the BP is held awaiting the arrival of the tri-snRNP to form the pre-B-complex. Upon activation of the B-complex by the Prp2 RNA helicase (DHX16 in humans), the SF3B proteins are destabilized, allowing the BP to dock in the active site for 5′ss cleavage and branch formation (Wan et al. 2019; Bai et al. 2021).

Correct BP recognition is critical for correct 3′ss selection, which is in turn essential for producing mRNA with the correct reading frame. The importance of correct BP selection is underscored by two additional findings. One is that BP selection is altered in cancerous cells carrying recurrent mutations in SF3B1 that influence tumor progression (Darman et al. 2015; Alsafadi et al. 2016), and the other is that bacteria produce small molecule splicing inhibitors (Spliceostatin-A, Pladienolide-B, etc.) that block loading of the BP into SF3B1 (Yokoi et al. 2011; Cretu et al. 2021). In vitro, formation of the prespliceosome or A-complex is blocked by this group of inhibitors (Roybal and Jurica 2010; Corrionero et al. 2011; Folco et al. 2011) as a consequence of their binding the BP-A binding pocket between HRs 15–16 of SF3B1 and PFH5A (Teng et al. 2017; Cretu et al. 2018, 2021; Finci et al. 2018). The relationships between the antitumor activity of splicing inhibitors and the contribution of recurrent mutations along a pathway of tumor evolution are difficult to attribute to changes in the splicing of specific introns; however, for many cancers, a general inhibition of splicing is selectively lethal, opening therapeutic opportunities (Hsu et al. 2015; Huang et al. 2020; Zhang et al. 2020a; Alors-Perez et al. 2021). The potential for developing intron-specific inhibitors is supported by the observation that different mammalian splicing events show different levels of sensitivity to this class of inhibitors (Corrionero et al. 2011; Vigevani et al. 2017; Cretu et al. 2018; Finci et al. 2018; Seiler et al. 2018). The sequence neighborhood surrounding and including the BP-A and polypyrimidine tract influence inhibitor sensitivity in vitro in the few cases tested (Vigevani et al. 2017; Finci et al. 2018). In vivo estimates of splicing changes using steady-state RNA level measurements are generally partly confounded by gene-specific transcription and decay rates; however, short GC-rich introns seem most affected (Vigevani et al. 2017; Seiler et al. 2018). Much about the specific intron features that might predict sensitivity remains to be learned.

We set out to study intron-specific features of splicing inhibition by splicing inhibitors using Saccharomyces cerevisiae. Previous studies have shown that yeast and other organisms are naturally more resistant to splicing inhibitors than mammals, and have patterns of amino acid substitutions in the BP-A binding pocket of their SF3B1 homolog that account for this resistance (Carrocci et al. 2018; Hansen et al. 2019; Serrat et al. 2019). Moreover, replacing the existing amino acids with those from human SF3B1 increases sensitivity to splicing inhibitors, at least for S. cerevisiae and Caenorhabditis elegans (Carrocci et al. 2018; Hansen et al. 2019; Serrat et al. 2019). We used CRISPR/Cas9 to introduce 14 specific amino acid changes in the BP-A binding pocket of the yeast SF3B1 homolog HSH155, making it identical to human SF3B1 through the region of the protein that binds the BP-A and splicing inhibitors. Addition of Pladienolide-B (Plad-B) to this yeast strain results in rapid splicing inhibition. RNA-seq analysis of the transcriptome after treatment with Plad-B or Thailanstatin-A (Thail-A) reveals patterns of response to inhibitors that we suggest are due to intron-specific differences in the rates of mechanistic steps that lead to BP selection. Using simple single-intron transcripts such as are common in yeast, may enable more direct and parallel determination of structure–activity relationships for splicing inhibitors, possibly leading to inhibitors with improved therapeutic efficacy.

Previous SectionNext Section

RESULTS

Design and creation of a humanized HSH155 inhibitor-sensitive protein

To create an allele of yeast HSH155 sensitive to splicing inhibitors, we aligned HRs 15 and 16 from human SF3B1 and Hsh155 and identified residues that differ (Fig. 1A, see also Hansen et al. 2019). Previous humanized alleles (Carrocci et al. 2018; Hansen et al. 2019) replaced HRs 5–16 from human SF3B1 in their entirety, or just one to two amino acids (N747 and L777) in the region of the pocket. Our goal was to replace the surface surrounding the yeast BP-A binding pocket with the human side chains, in the hope of producing a replica of the human drug binding surfaces for multiple different inhibitors, without substituting other surfaces of Hsh155 in a way that might compromise its interaction with its endogenous cofactors like Cus1, Cus2, Prp5, or Rse1. Using CRISPR/Cas9, we cleaved HSH155 in a strain deleted for three “drug exporter” genes (PDR5, SNQ2, and YOR1) (Jeong et al. 2007; Hansen et al. 2019), in the presence of a homologous repair fragment designed to direct the substitution of the 22 amino acids that differ between human and yeast to “humanize” HRs 15 and 16. In analyzing multiple independent colonies of edited yeast, it became clear that strains with complete humanization of HRs 15–16 were invariably also heat sensitive (not shown). One strain, confirmed by DNA sequencing to be derived from partial incorporation of the rescue sequence, replaced the 14 most carboxy-terminal yeast-specific residues with their human counterparts, and did not have a temperature-sensitive growth phenotype (Fig. 1A,B). Since the replacement of yeast with human HRs 5–16 is not temperature sensitive (Carrocci et al. 2018), we suspect that interactions between HR15 and HR14 are compromised when the amino-terminal segment of HR15 is humanized. In more carefully evaluating the utility of this serendipitous allele of HSH155, we found that amino acids that had not been humanized did not contribute to the binding surfaces for several different inhibitors of this class based on X-ray and cryo-EM analysis of human splicing complexes (Supplemental Fig. S1; Cretu et al. 2018, 2021). This allele, named hsh155-ds (ds, for drug-sensitive, see below), is expressed equivalently to wild-type when fused to a carboxy-terminal GFP-tag and evaluated by an anti-GFP western blot (Fig. 1C). The surface of the binding pocket formed by PHF5A would be contributed in yeast by its homolog Rds3. There are no amino acid differences between human and yeast on the PHF5A/Rds3 surface of the pocket, thus editing of Rds3 was not necessary. We conclude that humanizing yeast in this manner does not create significant differences in growth or Hsh155 protein expression.

FIGURE 1.
View larger version:
FIGURE 1.

Creation and characterization of a yeast HSH155 allele humanized for sensitivity to splicing inhibitors. (A) Sequence alignment of the BP-A binding pockets of HSH155 (top line) and SF3B1 (middle line). HRs 15 and 16 are shown with their paired α helices in blue above the alignment. The bottom line shows the sequence of hsh155-ds with blue boxes indicating the yeast-specific amino acids, and gold boxes indicating the human-specific amino acids. Amino acids identical in yeast and humans are unshaded. The amino acid numbers are shown above (yeast) and below (human). Below the sequence is a model of SF3B1 (left) or just the HRs 15 and 16 (right) bound to Plad-B (black, Cretu et al. 2018, PDB: 6EN4). Human-specific amino acids replaced in the yeast protein are shown in gold. (B) Serial dilution of WT HSH155 and hsh155-ds yeast strains on YPD plates to test for temperature sensitivity. (C) Western blot of strains expressing wild-type and hsh155-ds tagged with GFP. Lane 1, markers; lane 2, WT Hsh155 carboxy-terminally tagged with GFP; lane 3, hsh155-ds tagged with GFP; lane 4, untagged wild-type Hsh155; lane 5, positive control SIR1 tagged with GFP. The blot was probed with mouse anti-GFP antibodies, visualized using secondary goat anti-mouse antibody, then scanned on a Li-Cor infrared scanner. (D) Measurement of Plad-B sensitivity by RT-PCR of the MATa1 first intron. Samples were taken at times after the addition of DMSO (top) or Plad-B to 5 µM (bottom) at 0, 1, 4, and 8 min. Positions of the PCR product derived from unspliced RNA (US) and spliced RNA (S) are indicated by the labels.

Plad-B inhibits splicing in hsh155-ds yeast within minutes in vivo

To test for Plad-B inhibition of splicing in the humanized yeast strain, we treated cells with Plad-B (or DMSO carrier), isolated RNA, and analyzed splicing of the MATa1 first intron by RT-PCR. MATa1 mRNA (Miller 1984) has high synthesis and turnover rates, allowing for the rapid and sensitive detection of splicing inhibition. Hansen and colleagues studied similarly modified yeast and found that growth is greatly inhibited at 5–10 µM Plad-B, but cells remain viable (Hansen et al. 2019). Inhibition of splicing of the MATa1 first intron is evident after 4 min of treatment with 5 µM Plad-B in YEPD (rich media) at 30°C, and is nearly complete by 8 min (Fig. 1D). This experiment shows that splicing is strongly inhibited almost immediately upon addition of Plad-B to the culture, suggesting that this strain will be useful for experiments that require a strong and immediate block to splicing. This amount of Plad-B does not inhibit wild-type cell growth (Hansen et al. 2019) and has no effect on splicing in wild-type yeast even after an hour of exposure (see below).

Splicing in humanized yeast extracts is sensitive to Plad-B at nM concentrations

The BP-A pocket binding class of splicing inhibitors (Plad-B, Spliceostatin-A, Herboxidiene) block splicing at or before ATP-dependent prespliceosome formation (Roybal and Jurica 2010; Corrionero et al. 2011; Cretu et al. 2018). To estimate the concentration of Plad-B needed to inhibit in vitro splicing to 50% of untreated extracts (IC50), we prepared cell-free splicing extracts from the humanized and wild-type control strains and tested them using a radiolabeled ACT1 pre-mRNA substrate (Fig. 2A,B). We observe splicing inhibition (as judged by the decreased production of splicing products: first step lariat-3′ exon, second step lariat and final ligated product) at Plad-B concentrations above ∼10 nM (Fig. 2A,C) relative to untreated hsh155-ds and wild-type strains. We also observe inhibition of prespliceosome (A-complex) formation at similar concentrations of Plad-B in vitro using native gels (Fig. 2B,D). Quantification of these gels indicates that the IC50 for splicing and for ATP-dependent splicing-complex formation are both in the range of ∼26–28 nM (Fig. 2C,D), similar in magnitude to an adenovirus late leader-derived substrate in HeLa splicing extracts (∼100 nM) (Effenberger et al. 2014). This suggests that the hsh155-ds strain provides a robust platform to study details of splicing inhibition by this class of compounds (see also Carrocci et al. 2018; Hansen et al. 2019).

FIGURE 2.
View larger version:
FIGURE 2.

Plad-B blocks splicing during ATP-dependent prespliceosome formation in extracts from humanized yeast. (A) In vitro splicing assays of a 32P-radiolabeled actin pre-mRNA. Indicated extracts were incubated at 23°C for 20 min with either H2O or 2 mM ATP, and in the presence of DMSO or the indicated concentration of Plad-B. RNA was extracted from the reactions, resolved on a 6% acrylamide/8 M urea gel and visualized by autoradiography. Lariat-3′ exon, lariat, pre-mRNA, and mature mRNA are shown. Lane “m” is a 50 bp marker, and “no n.e.” is a no extract control lane. (B) Half of the in vitro splicing reactions in panel A were also directly loaded on nondenaturing gels to visualize splicing-complex formation. Prespliceosomes/spliceosomes (psp/sp) and commitment complexes (cc) are indicated. (C) Quantification of splicing reactions in panel A measuring relative splicing efficiency in the presence of increasing concentration of Plad-B. (D) Quantification of splicing reaction assay in panel B measuring relative ATP-dependent complex formation in the presence of an increasing concentration of Plad-B.

We also tested wild-type and hsh155-ds splicing extracts for sensitivity to the structurally distinct but mechanistically similar BP-A pocket binding inhibitor herboxidiene (HB) (Hasegawa et al. 2011). We observe similar inhibition of splicing and splicing-complex formation in hsh155-ds extracts with IC50 ∼47 nM and ∼74 nM, respectively (Supplemental Fig. S2A–D). Unlike Plad-B, HB weakly but detectably inhibits in vitro splicing in wild-type yeast extracts at the ∼2 µM level (Supplemental Fig. S2A–D). To test whether HB inhibits splicing in vivo, we constructed a reporter gene with a MATa1 first intron modified to have an open reading frame so that Neon fluorescent protein would be expressed when splicing is inhibited, and integrated it into the wild-type and humanized hsh155-ds strains. Growth of wild-type yeast for 1 h in 5 µM HB in rich medium results in significant fluorescence above background and controls (Supplemental Fig. S2E), indicating splicing inhibition in vivo by HB. As expected, the humanized strain is more sensitive, producing about 3× more fluorescence than the wild-type over the hour incubation (Supplemental Fig. S2E), consistent with the in vitro results here and elsewhere (Supplemental Fig. S2A–D, Hansen et al. 2019). These experiments show that the substitution of 14 amino acids in Hsh155 protein with the corresponding human SF3B1 residues creates a binding pocket for Plad-B (or HB), substantially increasing the ability of these compounds to inhibit splicing in yeast. The amount of inhibitor needed to block splicing in vitro is similar to amounts needed to block human splicing in vitro, with the block occurring at or near the step of prespliceosome or A-complex formation in both systems.

Transcriptome-wide effect on splicing in the presence and absence of Plad-B and Thail-A

To examine the global impact of treatment with Plad-B and Thail-A on splicing and gene expression in yeast, we performed total RNA sequencing (RNA-seq) on wild-type HSH155 and hsh155-ds cells treated with either DMSO carrier, 0.5 µM Plad-B, 5 µM Plad-B, or 5 µM Thail-A, a Spliceostatin A-like compound expected to make a covalent complex with PHF5A/Rds3 (Liu et al. 2013; Cretu et al. 2021). We chose to treat for 1 h at 30°C to allow the effects of the drug on splicing to manifest for genes with a variety of expression rates. Duplicate cultures were treated, RNA was isolated, and rRNA depleted using DNA oligonucleotides complementary to rRNA and RNase H (sequences in Supplemental Table S3.1, see Materials and Methods), and libraries were prepared and sequenced. We obtained about 1.7 billion reads across all 12 samples, mapped them, and performed splicing and differential gene expression analysis as described in Materials and Methods.

We estimated a weighted splicing efficiency (SE) for each intron (Fig. 3A; Xia 2020; see Materials and Methods) resulting in measurements for 267 splicing events. Comparisons of each replicate pair (Supplemental Fig. S3A–F) were highly correlated, allowing us to average replicates from each treatment and use them to compare treatments to each other with confidence (Fig. 3B–E). Comparison of wild-type cells treated for 1 h with 5 µM Plad-B to those treated with DMSO show no evidence that splicing is inhibited (r2 = 0.99, Fig. 3B). Additionally, comparison of SEs in the wild-type and humanized strain in the absence of inhibitors suggests that splicing of most introns is unchanged in the mutant (r2 = 0.98, Fig. 3C). There are a small number of introns that appear to be spliced slightly more efficiently in the humanized strain compared to wild-type. A histogram of the change in SE (mutant − wt) for each intron shows that most introns change <3%, and that the few that improve seem enriched for noncanonical BP sequences (Supplemental Fig. S3G). We conclude that wild-type yeast is completely insensitive to Plad-B, and that although the humanizing mutation may advantage introns with variant BPs (see Carrocci et al. 2018; Hansen et al. 2019), it has little effect on the overall landscape of splicing in yeast in the absence of inhibitors.

FIGURE 3.
View larger version:
FIGURE 3.

Comparisons of SE under different genetic conditions and drug treatments. (A) Splicing efficiency is calculated as percent spliced using junctionCounts (see Materials and Methods). Reads spanning the splice junction (NEE, blue) and the unprocessed splice sites (NIE, orange) were used to calculate a percent spliced value as shown using the equations on the right. (B) Scatter plot comparing SE for each intron in wild-type cells incubated with either DMSO (x-axis) or 5 µM Plad-B (y-axis). (C) Scatter plot comparing SE for each intron in wild-type cells (x-axis) or in hsh155-ds (y-axis) cells without the drug (DMSO only). (D,E) Scatter plots comparing SE for each intron in hsh155-ds cells treated only with DMSO (x-axis), 0.5 µM (D), or 5 µM (E) Plad-B (x-axis). In panels B–E, pink dots represent ribosomal protein genes. (F) Normalized read coverage tracks over part of the yeast genome encoding divergently transcribed genes RPS16B (bottom strand right to left, orange) and RPL13A (top strand left to right, blue). Top track shows ORF locations, middle track shows humanized strain expression in DMSO (no drug), bottom track shows humanized strain expression after 1 h in 5 µM (Plad-B). Note accumulation of intron reads in the Plad-B-treated sample. Browsable coverage data may be found at https://genome.ucsc.edu/s/mannyares/Hunter_etal.

Apparent splicing inhibition is Plad-B dose-dependent and different for different introns

To estimate the change in SE for each intron in the humanized yeast treated with Plad-B, we compared intron SEs from DMSO-treated cells against those treated with 0.5 µM (Fig. 3D) or 5 µM (Fig. 3E) Plad-B. We observe a strong, dose-dependent reduction in SE across the transcriptome of the humanized yeast strain (Fig. 3D,E) with 5 µM showing the highest amount of retained intron for most genes (Fig. 3F). None of the introns escape inhibition, but the ribosomal protein gene (RPG) introns (pink dots) appear generally less inhibited than the non-RPG introns (black dots). Many non-RPGs appear almost completely blocked after 1 h at 5 µM (Fig. 3E). Many introns spliced at 85% or more in the absence of drug show a wide range of apparent SEs (10%–80%) in 0.5 µM Plad-B (see below). Because these numbers are derived from steady-state measurements of spliced and unspliced RNA, it is likely that some fraction of the observed difference in intron SEs between introns are contributed by gene-specific differences in rates of transcription and decay, a concern for any analysis of splicing changes captured by measuring steady-state RNA levels.

To explore the extent that gene-specific transcription and decay rates obscure potential intron-specific differences in Plad-B sensitivity, we plotted SE as a function of Plad-B dose (Fig. 4A). As suggested by the scatter plot, introns appear inhibited to different extents by Plad-B; however, introns also have different apparent intrinsic SEs in the absence of drug. To avoid complications from mRNAs with two introns or multiple alternatively spliced isoforms, we focused on single-intron genes. Some, but not all, introns with high intrinsic (unperturbed) apparent SEs have SEs of ∼50% or more in the highest concentration of Plad-B (5 µM), while other introns that are less efficiently spliced in the absence of drug are completely inhibited at 0.5 µM (Fig. 4A).

FIGURE 4.
View larger version:
FIGURE 4.

Introns can be classified by their sensitivity to Plad-B. (A) Splicing efficiency (SE) dose-response curves for all 247 introns analyzed at 0, 0.5, and 5 µM Plad-B for the humanized strain. The data have been separated into quintiles and color-coded: Blue represents quintile 1 (most efficient), gold represents quintile 2, pink represents quintile 3, green represents quintile 4, and teal represents quintile 5. (B) Scatter plot comparing Plad-B sensitivity (the difference in SE between 0 and 0.5 µM Plad-B) with the intrinsic SE of the uninhibited introns (0 Plad-B, DMSO). Pink dots represent RPG introns. For introns >75% intrinsic SE versus ΔSE in 0.5 µM Plad-B, r2 = 0.08. (C) Relative SE dose-response for six representative genes. The data shown in (A) have been normalized to hsh155-ds DMSO sample as described in Materials and Methods. (D) Hierarchical clustering of relative SE dose-response curves for single-intron genes. Cluster nodes were used to call intron groups as “most sensitive,” “intermediate sensitivity,” and “most resistant.” Color indicates sensitivity to Plad-B, with yellow being the most resistant, black being intermediately sensitive, and blue being the most sensitive. The top panel shows the complete set with the lower panels showing the groups. Introns with nonconsensus BP sequences are marked with asterisks.

Before bringing these introns into alignment for comparison, we asked whether there was a strict relationship between SE in the absence of inhibitor (“intrinsic” SE or strength) and intron-specific Plad-B resistance by plotting the change in SE between 0 and 0.5 µM Plad-B (ΔSE0.5) against the intrinsic SE of the uninhibited introns (Fig. 4B; Supplemental Table S4.1; note: small changes in ΔSE0.5 indicate higher Plad-B resistance, points on the diagonal are introns with no apparent resistance to 0.5 µM Plad-B). As can be seen from the graph, the group of introns with intrinsic SEs above 50% or so have a wide range of responses (10%–70% inhibition) to Plad-B, whereas most of the ∼15 introns with intrinsic SE below 50% are completely inhibited by 0.5 µM Plad-B. The intrinsic SE of a well-spliced intron (>75%) is poorly predictive of its sensitivity to Plad-B (r2 = 0.08, Supplemental Table S4.2). In order to begin comparing introns with each other, we normalized their responses by adjusting their intrinsic SE in the absence of inhibitor to 100% and increasing the SEs under inhibition by the same ratio, to produce relative SE at a given inhibitor concentration for each intron (Fig. 4C; for simplicity, five introns with distinct plots are shown; Supplemental Table S4.3). Going forward, we will use the relative SE for each intron measured after 1 h in 0.5 µM Plad-B as a measure of response to Plad-B, the dose that best spreads out the intron-specific responses observed.

Effects of gene expression dynamics on apparent splicing inhibition

To assess the extent to which the apparent intron-specific resistance values derived from steady-state RNA measurements are influenced by gene-specific expression dynamics, we compared them to published gene-specific transcription, splicing, and expression measurements (Supplemental Fig. S4; Supplemental Tables S4.4–4.6). Transcript synthesis rates measured by Cramer and colleagues (Miller et al. 2011) are under ∼20 molecules/cell/150 min for most genes; however, the RPGs produce between 50 and 400 mol/cell/150 min. If splicing of all introns were completely blocked, unspliced RNA from highly transcribed genes would accumulate more rapidly and to a greater extent in 1 h than would less frequently transcribed genes, making them appear more sensitive to Plad-B, producing a negative correlation between apparent resistance and transcript synthesis rate. However, the overall correlation is positive with an r2 = 0.18, with only a few of the very most highly transcribed RPGs appearing somewhat less resistant than more slowly transcribed RPGs (Supplemental Fig. S4A). Thus, for most genes, differences in transcript synthesis rate contribute little to apparent differences in resistance.

If all introns were completely inhibited by Plad-B, spliced mRNAs with long half-lives would disappear more slowly than would mRNAs with short half-lives, making the former appear more resistant and creating a positive correlation between apparent resistance and mRNA half-life. A plot of the relative SEs versus mRNA half-life (Miller et al. 2011) shows a modest positive correlation (Supplemental Fig. S4B, r2 = 0.42), suggesting that mRNA decay rates may partially influence estimates of resistance, making introns from more stable mRNAs appear more resistant to Plad-B. If the mRNA decay rate played a large role in the apparent splicing inhibition measurement, then correcting for decay would make introns appear more similar in resistance, but it does not, possibly because the intron population is dominated by RPGs that appear both more resistant as a group and have very similar half-lives. To assess the relationship between splicing rate and apparent resistance to Plad-B, we used the “splicing speed” measurements of Granneman and colleagues (2015) for about 100 yeast introns (Barrass et al. 2015). Splicing speed is poorly predictive of Plad-B resistance (Supplemental Fig. S4C, r2 = 0.06), suggesting that features governing the rate-limiting step(s) of splicing for most introns in the absence of drug are different from those governing sensitivity to Plad-B. We suggest that except for the possible contribution by mRNA decay rates, the observed differences in response to Plad-B as measured by steady-state RNA levels are largely due to intron-specific differences in inhibition of splicing by Plad-B.

To evaluate broad relationships in intron-specific resistance patterns, we clustered the introns by Plad-B dose-response using their relative SEs in low (0.5 µM) and high (5 µM) Plad-B. Clustering revealed three loosely defined groups of introns separated by their sensitivity to Plad-B (Fig. 4D; Supplemental Tables S4.7–4.9). The node with the highest apparent relative resistance (Fig. 4D, “most resistant”) is greatly enriched for RPGs but includes other introns such as those from ACT1 and UBC4. The “intermediate sensitivity” class is more heterogeneous, with some RPGs at the more resistant end of the cluster, and considerable numbers of genes with diverse functions. Finally, the “most sensitive” node is devoid of RPG introns but contains other essential genes. Analysis of intron features that might explain or predict sensitivity to Plad-B or other inhibitors with statistical power is limited by the small number of introns in yeast. Introns are enriched in the functional class of genes associated with ribosome function (e.g., RPGs), and intron size is bimodal, with RPG introns occupying the long (∼400 nt) class (Spingola et al. 1999). We ranked the 241 introns analyzed here from most to least resistant, as well as from longest to shortest. By their ranks, intron length and resistance are positively correlated (Spearman's rho = 0.65, Supplemental Table S4.3); however, this could be confounded by unknown factors that correlate with RPG function or expression. We also evaluated the ranks of the 39 introns whose BPs deviate from UACUAAC (Supplemental Table S4.3) and found that 33 of 39 have ranks below 154 of this set of 241 introns, placing the vast majority of BP variant introns in the bottom third of resistance. The highest ranking BP variant intron at rank 15 is PMI40, but it has two overlapping potential BPs (CACUAACUAAC) within 11 nt, possibly representing a strong BP. Ranked third is the well-studied ACT1 intron that has a consensus BP plus a strong cryptic BP upstream that plays a mysterious positive role in SE (Kao et al. 2021). In contrast, the introns with nonstandard 5′ss (standard is GTATGT) are not consistently more resistant or sensitive (Supplemental Table S4.3). It did not matter whether we considered the very common variant GTACGT as the standard or nonstandard site; the effect of the 5′ss sequence on resistance appears weaker than that of the BP. Pinpointing the physical features of introns that potentiate or depress the effect of the splicing inhibitors must await more detailed experimentation, but to a first approximation BP strength appears to be associated with Plad-B resistance.

Intron-specific sensitivity to Plad-B manifests during cotranscriptional splicing

To measure the effect of Plad-B during cotranscriptional splicing, we used single-molecule intron tracking (SMIT), which evaluates splicing on chromatin-associated nascent transcripts (Oesterreich et al. 2016; Alpert et al. 2020). We treated the wild-type and the humanized strains with 5 µM Plad-B for 15 min and processed cells for SMIT, sequencing chromatin-associated nascent RNA from 63 intron-containing genes, mapping their 3′ ends to obtain the position of RNA polymerase, and determining the status of the intron on the nascent transcript (spliced or unspliced, Fig. 5A; Oesterreich et al. 2016; Alpert et al. 2020). Since SMIT requires the use of gene-specific 5′ primers to capture nascent transcript reads, the number of genes analyzed was limited. However, we obtained sufficient data to graph the relationship between RNA polymerase progression and nascent transcript splicing for 37 intron-containing genes (six are shown in Fig. 5B, for complete data see Supplemental Table S5). To estimate cotranscriptional splicing efficiency (CTX-SE), we calculated the area under the curve (AUC) within a defined window starting from the polymerase position at the onset of splicing to the position 200 bp downstream from the onset point, a narrow snapshot of the gene expression pathway which should allow fair gene to gene comparisons (Fig. 5A). The area under the SMIT curve (fraction spliced vs. pol II position) for introns in control cells was set to 1, and the AUC from Plad-B-treated humanized cells was subtracted to calculate ΔAUC, thus an estimate of cotranscriptional Plad-B resistance (CTX-ΔSE) is given by (1 − ΔAUC) × 100% (see Materials and Methods, Fig. 5B,C). Because this assay measures only nascent transcripts at steady state within this narrowly defined window, it is less likely to be confounded by gene-specific transcription or decay rates. The set of 37 introns for which we could obtain cotranscriptional Plad-B sensitivity data shows a wide range of responses (Fig. 5C), some with <20% residual splicing of the nascent transcripts within the window of analysis, whereas others achieve 80% of the splicing observed in the absence of drug.

FIGURE 5.
View larger version:
FIGURE 5.

Intron-specific difference in Plad-B sensitivity during cotranscriptional splicing. (A) Model depicting cotranscriptional splicing in Plad-B for insensitive wild-type cells (top) and sensitive humanized cells (bottom). An example gene is represented by the horizontal boxes, with RNAP II shown as green circles moving left to right during transcription. Nascent transcripts extend from RNAP II and their splicing status is shown by the presence or absence of the intron (thin line). Exon 1 is purple, exon 2 is blue. The bracket below the gene shows the range of RNAP II positions from which nascent transcripts are counted to evaluate SE. Inhibition is detected by counting the fraction of nascent transcripts within this window that have had their introns removed. (B) Plots of fraction spliced versus RNAP II distance from the 3′SS for six genes. Pink line is WT HSH155 with 5 µM Plad-B; blue line is hsh155-ds with 5 µM Plad-B. The difference of the normalized area under the curve values (ΔAUC) is shown. (C) Plot of SE in 0.5 µM Plad-B measured using RNA-seq of total RNA with cotranscriptional SE in Plad-B as measured by SMIT for 37 genes. Pink points are RPGs; genes plotted in B are shown in blue.

We compared the cotranscriptional estimates of Plad-B splicing resistance to RNA-seq derived estimates for the genes we could evaluate by SMIT (Fig. 5C). These are reasonably similar over this subset of introns (r2 = 0.55). We conclude that introns differ from each other in some fundamental way that affects (i) how well Plad-B binds the spliceosome, or (ii) how well bound Plad-B can interfere with their splicing, or both. Furthermore, this behavior is not explained by gene-specific differences in rates of transcription, splicing, and decay. It is possible that unspliced nascent transcripts in the nucleus are subject to decay by the nuclear exosome or Rat1 nuclease at intron-specific rates (Bousquet-Antonelli et al. 2000; Egecioglu et al. 2012) after Plad-B inhibition; however, such measurements have yet to be made. Some of the outliers that show higher cotranscriptional sensitivity than expected based on their RNA-seq estimated resistance, such as ERD2 or ARF9, might be spliced after release from chromatin more readily than other pre-mRNAs. We conclude that a major underlying explanation for intron-specific differences in apparent sensitivity to Plad-B is that introns possess intrinsic differences that affect their inhibition by Plad-B at the spliceosome.

Intron-specific differences in relative sensitivity to Plad-B versus Thail-A

To test whether intron-specific differences in sensitivity to Plad-B are similar to other BP-pocket binding splicing inhibitors, we compared the SE of yeast introns in 5 µM Plad-B to their SE in 5 µM Thail-A. Thail-A is in the same class as Spliceostatin-A, whose scaffold is distinct from the scaffolds of either Plad-B or HB, and which can covalently bind to a cysteine in PHF5A (Cretu et al. 2021). At 5 µM, Thail-A is overall a more potent inhibitor of splicing in the humanized yeast strains than Plad-B; however, its effects are not systematically proportional across introns to that of Plad-B (Fig. 6A). This is more evident in a plot of the relative SE in 5 µM Plad-B (x-axis) against the difference in SE between Thail-A and Plad-B (y-axis, Fig. 6B). A swath of introns whose relative SE in Plad-B is between 30% and 50% vary with respect to the difference in relative SE between the two inhibitors over a range from 5% (slightly more sensitive to Thail-A than Plad-B, e.g., RPS9B, COF1, and RPS13) to 25% (much more sensitive to Thail-A than to Plad-B, e.g., RPL17B and GLC7). Although a large number of the well-separated introns in the plot are RPGs (pink dots) that are more resistant to Plad-B than other introns, non-RPGs (black dots) also differ in their sensitivity to Thail-A relative to Plad-B. Considering that each intron should have the same gene-specific transcription and decay rates in these two experiments, the differences observed are likely to be due to intron-specific, drug-specific differences in splicing inhibition. Based on intron-specific differences in inhibition by Thail-A as compared to Plad-B, we conclude that introns are differently susceptible to inhibition by chemically distinct members of the general class of splicing inhibitors that bind to the BP pocket of humanized Hsh155.

FIGURE 6.
View larger version:
FIGURE 6.

Introns respond differently to Plad-B and Thail-A. (A) Scatter plot comparing the relative SE of introns in 5 µM Plad-B (x-axis) compared to their relative SE in Thail-A (y-axis). As Thail-A is overall a more potent inhibitor than Plad-B at 5 µM, all introns are more sensitive to Thail-A. The regression line is derived from a nonlinear least squares fit. Pink dots are RPGs. Introns above this line are relatively less sensitive to Thail-A than to Plad-B as compared to the average intron, whereas those below the line are relatively more sensitive to Thial-A than Plad-B as compared to the average intron. (B) Relationship between sensitivity to Plad-B and the difference between Thail-A and Plad-B sensitivity. The y-axis represents the difference between the resistance to Thail-A and Plad-B; values are negative because Thail-A is a more potent inhibitor. A high position in the Y dimension indicates an intron whose sensitivity to Thail-A is much greater than to Plad-B, whereas those lower in the Y dimension have more similar sensitivities to both inhibitors.

Splicing inhibitors alter the expression of intronless genes

A striking feature of the S. cerevisiae genome is the large number of intronless genes. The existence of introns in a small subset of genes fundamental to cell function offers an opportunity to investigate the complex integration of splicing into the gene expression landscape of a eukaryotic cell by observing the indirect effects of splicing inhibition on intronless genes. As observed for splicing of intron-containing genes (Fig. 3B), comparison of mRNA levels for all genes in Plad-B-treated versus untreated wild-type HSH155 cells showed no significant changes using DESeq2 (Love et al. 2014), indicating that Plad-B has no effect on gene expression in wild-type yeast. To explore how a strong coherent block to splicing reverberates across the transcriptome, we similarly analyzed changes in mRNA levels for all genes in Plad-B-treated versus untreated humanized hsh155-ds cells and plotted the log2 fold change in expression versus the negative log10 of their adjusted P-values using volcano plots (Fig. 7A–C). As expected, we observe down-regulation of most intron-containing genes (Fig. 7A, green points; Fig. 7B), as well as numerous, presumably secondary changes in the expression of intronless genes (Fig. 7A, black points; Fig. 7C). A few intron-containing genes appear up-regulated (Fig. 7B, e.g., PCH2 and NBL1). Some meiotic genes like PCH2 appear transcriptionally induced (see below), and the unspliced transcripts from NBL1 accumulate to a level about twofold above normal NBL1 mRNA levels for unknown reasons, despite the block to splicing.

FIGURE 7.
View larger version:
FIGURE 7.

Impact of splicing inhibition on intronless genes. (A) Volcano plot showing differential gene expression analysis of the humanized hsh155-ds strain in 5 µM Plad-B compared to DMSO. Y-axis shows −log10 adjusted P-value and the x-axis shows log2 fold change in expression. Points not meeting P-value < 0.05 (−log Padj > 1.3) or log2 Fold Change = |0.26| (Fold Change = ±1.2) are gray. Both intron-containing (green) and intronless genes (black) are shown. (B) Same plot as in A but only shows intron-containing genes. RPGs are pink and non-RPGs are green. (C) Same plot as in A but only showing intronless genes. ADE/HIS/URA genes are in orange, RPGs are in pink, and meiotic genes are in blue. (D) GO analysis identifies the enrichment of the ribonucleotide metabolic process and the ribosome (component) in the set of down-regulated intronless genes. At left is a diagram of four key protein complexes required for chromatin state and transcription along with the names of intron-containing genes encoding at least one of their subunits.

The prevalence of introns in RPGs (102 of 137 RPGs have introns) results in reduced expression of intron-containing RPGs after a 1 h Plad-B treatment in the humanized strain (Fig. 7B). This likely results in insufficient translation of ribosomal proteins to sustain ribosome biogenesis, which then feeds back on RPG transcription (de la Cruz et al. 2018; Albert et al. 2019), resulting secondarily in reduced expression of intronless RPGs (Fig. 7C, pink dots). GO analysis (Boyle et al. 2004) using only intronless genes changing >1.5-fold as input confirmed that intronless RPGs are enriched in the down-regulated class of intronless genes (Fig. 7D). Another prominent class of down-regulated intronless genes concerns nucleobase synthesis, for example, numerous genes in the connected adenine (ADE1, ADE5,7, ADE8,…, ADE17) and histidine (HIS4, HIS5) pathways, and some for uracil synthesis (URA1, URA4) and import (FUI1, orange points, Fig. 7C). Although there is one intron-containing gene IMD4 (encoding inosine monophosphate dehydrogenase, also carrying an intronic snoRNA SNR54) in the de novo purine biosynthetic pathway, expression of intronless paralogs IMD2 and IMD3 is unchanged, and it is unclear how the loss of this enzymatic activity would repress ADE genes (Ljungdahl and Daignan-Fornier 2012). Mutations originally identified through their strong block to rRNA synthesis (Hartwell et al. 1970), were ultimately revealed to be in splicing factor genes (Lustig et al. 1986), likely blocking expression of intron-containing RPGs. Perhaps a sudden block to rRNA transcription produces an excess of free nucleotides that could feed back on those pathways.

At a lower threshold of 1.2-fold change, the GO analysis revealed subtle up-regulation of Ty retrotransposons. Together with the apparent slight up-regulation of meiotic genes (blue points, Fig. 7C), the loss of expression of ADE genes and the up-regulation of Ty elements and meiotic genes is suggestive of a general transcriptional and chromatin state defect. Considering the presence of introns in genes for at least one subunit of four main protein complexes essential for transcription (RNAP II core, Mediator, SAGA, and SWI/SNF, see Fig. 7D), we suggest that general transcription is impacted within 1 h of robust splicing inhibition. Depletion of chromatin remodeling and RNA polymerase II activities could lead to reduced expression of genes already induced for growth in YEPD (a poor source of adenine and uracil), and chromatin changes could reflect the loss of repression of meiotic genes and Ty elements. Reduced growth efficiency might also explain the up-regulation of other genes in Figure 7C that are part of stress responses and glucose exhaustion in the Plad-B-treated culture.

Numerous splicing factor genes have introns, including MUD1, LSM2, LSM7, SMD2, YSF3, and PRP5; however, we considered the splicing block to be so strong at the inhibitor concentrations used here that the secondary effect of loss (or gain in the case of PRP5) of expression of these genes would make little additional contribution to the loss of splicing activity. We did not detect a transcriptional response for splicing factor genes that would indicate the cell could sense and up-regulate its splicing capacity in response to the inhibitors. We conclude that the indirect effects of splicing inhibition on the expression of the intronless class of yeast genes are complex and widespread, likely due to the importance of splicing to correct expression of the translation and transcription machinery.

Previous SectionNext Section

DISCUSSION

In this study, we created a yeast strain sensitized to splicing inhibitors to explore the mechanism and consequences of splicing inhibition in a well-characterized splicing system. Previous work captured the effects of four different combinations of amino acid substitutions in plasmid-borne HSH155 (yeast SF3B1) on Plad-B and HB sensitivity (Carrocci et al. 2018; Hansen et al. 2019). Here, we constructed a chromosomal allele of HSH155 (hsh155-ds) with just 14 humanizing amino acid substitutions in the pocket where Plad-B and the intron BP-A compete for binding (Fig 1; Teng et al. 2017; Cretu et al. 2018, 2021; Finci et al. 2018). The effect of the mutations in this allele (hsh155-ds) agrees with the previous studies of Plad-B and HB sensitivity in humanized yeast (Carrocci et al. 2018; Hansen et al. 2019). Using a reporter that detects translation of unspliced RNA, we confirmed the sensitivity of wild-type yeast to a high concentration of HB in vivo (Supplemental Fig. S2), that is not apparent at lower concentrations (Hansen et al. 2019). This weak natural sensitivity to HB is greatly augmented by humanizing mutations (Supplemental Fig. S2; Hansen et al. 2019). Sensitivity of wild-type yeast to HB but not Plad-B is consistent with structure modeling that suggests these two inhibitors bind to overlapping but not identical surfaces in the BP-A binding pocket (Cretu et al. 2018). Plad-B rapidly inhibits splicing in vivo (Fig. 1D), and IC50 values in hsh155-ds yeast splicing extracts are similar to those observed in human extracts (Fig. 2; Supplemental Fig. S2, see also Carrocci et al. 2018).

By examining the effect of splicing inhibitors across the transcriptome, we find that individual introns have distinct Plad-B inhibition profiles (Figs. 3 and 4). Intron-specific differences in sensitivity to Plad-B extend to cotranscriptional splicing as well (Fig. 5), under conditions that should be agnostic to gene-specific differences in expression dynamics, with the exception of unknown features of nascent transcript decay. Furthermore, individual introns have different relative sensitivities to the related but not identical inhibitors Thail-A and Plad-B (Fig. 6). Finally, we assessed the effects of splicing inhibition on the expression of intronless genes (Fig. 7) due secondarily to loss of expression of intron-containing genes encoding core transcription factors and ribosomal proteins.

Transcriptome-wide patterns of splicing inhibition in engineered S. cerevisiae

Plad-B has no effect on the splicing of any intron in wild-type yeast; however, every intron in the humanized hsh155-ds strain is susceptible to inhibition (Fig. 3). Direct comparison of splicing and gene expression between two genetically different yeast cultures is noisier than comparing different treatments on aliquots of the same yeast culture due to the lack of synchrony in the diauxic shift in different cultures (DeRisi et al. 1997); however, most introns appear unaffected by the humanizing mutation in the absence of inhibitor (Fig. 3C). Mutation of certain residues in this part of HSH155 subtly improves splicing of mutant BP ACT1 reporter introns (Carrocci et al. 2018), although similar effects were observed with human-yeast HSH155 domain swaps that did not include HRs 15 and 16. The humanizing allele used here is mostly neutral in the absence of inhibitor, and there are hints that it subtly improves the splicing of introns with noncanonical BPs (Supplemental Fig. S3G).

By capturing splicing inhibition at two Plad-B concentrations, we obtained simple dose-response curves for many introns that we could cluster and rank by relative resistance to Plad-B. For introns spliced at >50% efficiency in the absence of inhibitors, there is a wide range of sensitivity to Plad-B (Fig. 4B), suggesting that each intron relies to a different extent on the rate(s) of the specific step(s) affected by Plad-B that together determine its overall rate of splicing. In addition, it is not possible to use a given Plad-B sensitivity to determine the sensitivity of an intron to Thail-A (Fig. 6). RPG introns as a class appear to be more resistant to Plad-B and Thail-A than introns in other genes (Figs. 4 and 6). The RPG introns are also characteristically longer (∼400 nt) and often contain secondary structure elements that influence splicing (Charpentier and Rosbash 1996; Howe and Ares 1997; Spingola et al. 1999; Rogic et al. 2008; Meyer et al. 2011; Rangan et al. 2023). Even within the more resistant RPG class, a range of sensitivities is evident (Figs. 4 and 6). The 39 yeast introns whose BP sequences deviate from the consensus UACUAAC are among the most Plad-B sensitive (Fig. 4D; Supplemental Table S4.3), suggesting that yeast sequences including the BP consensus contribute to the differences in sensitivity.

These observations echo findings in vertebrate cells where strong BPs are implicated as resistance features in mammalian introns (Corrionero et al. 2011; Vigevani et al. 2017; Finci et al. 2018), and are hypothesized to mediate escape and resistance (Cretu et al. 2021). A recent high-resolution structure of a human A-like complex with a pre-mRNA fragment and a spliceostatin-A molecule covalently attached to PFH5A provides evidence that drug binding prevents the complete formation of the U2-BP helix (Cretu et al. 2021), leaving it unloaded, with SF3B1 in the open conformation. Although Thail-A, like Spliceostatin-A should be able to form the same irreversible covalent bonds with yeast Rds3 (homolog of PHF5A), we do not know how efficient this is in vivo and whether this feature contributes to intron-specific or drug-specific differences in resistance. Inhibitors may differ in their affinity for the binding pocket (or in their off-rate from specific splicing complexes) to produce overall stronger or weaker inhibition of the spliceosome; however, specific features of the intron must interact with the characteristics of a particular drug-spliceosome complex to produce idiosyncratic splicing outcomes for each intron with each compound. In an investigation of Plad-B and its close derivatives, clear inhibition of specific splicing events was observed for compounds that scored weakly if at all in cell growth assays, or in an in vitro splicing test using a robust substrate, consistent with intron-specific activity (Effenberger et al. 2014). Combined with the surprising selectivity of a small molecule modulator of SMN2 exon 7 5′ss use (Naryshkin et al. 2014), we suggest that BP-pocket inhibitors with high selectivity for specific introns and low general toxicity may exist. How intron features contribute to the activity of splicing inhibitors on individual introns will require further experimentation, but the possibility that a spliceosome operating on a given splicing substrate will be selectively druggable does not seem that remote.

How does the block to U2-BP helix loading result in intron-specific sensitivity to inhibitors?

Yeast introns with noncanonical BP sequences appear more sensitive to Plad-B (Fig. 4D). Variant U2-BP helices may naturally be slower to load into and trigger closing of Hsh155 in the absence of inhibitor, increasing the chance that they may be recognized as “incorrect” by the Prp5-dependent discard pathway (Xu and Query 2007; Liang and Cheng 2015; Kao et al. 2021; Zhang et al. 2021). In the presence of a bound inhibitor molecule, this slow loading would be exacerbated, enhancing discard, resulting in greater sensitivity by magnifying the tendency toward discard. Conversely, introns with consensus BPs in a “fast loading” context would either outrace inhibitor binding to the pocket (or respond quickly to inhibitor dissociation), or actively displace inhibitor, or otherwise delay or pass the fidelity check before the discard pathway recruits them, resulting in resistance. In this model, drug sensitivity is mediated at the molecular level by the discard pathway, such that different introns with different intrinsic rates of Prp5 release (also influenced by BP sequence context) differ in sensitivity to an inhibitor. The Prp5 fidelity check has also been linked to an escape pathway involving the use of alternative BP-As by “scanning” upstream (Kao et al. 2021); however, molecular details concerning the components of the discard pathway that disassemble Prp5-stalled complexes are scant.

The role of the human Prp5 homolog DDX46 in enforcing BP fidelity in human cells is unclear. Rather than stringent adherence to UACUAAC as observed for yeast, mammalian BPs conform to the more degenerate sequence URAY and are thus not able to form the uniform 6 bp + bulged A version of the U2-BP helix found commonly in yeast. Differences from yeast in the process of loading the more variable mammalian U2-BP helix into SF3B1 include the protein p14 (SF3B6, not present in yeast) that directly binds the BP-A and the amino-terminal region of SF3B1 (Schellenberg et al. 2006; Tholen et al. 2022). This protein is hypothesized to stabilize the BP-A duplex and promote SF3B1 closing during loading into SF3B1 (Yazhini et al. 2021; Tholen et al. 2022), and thus could affect DDX46 recognition of authentic U2-BP helices. Instead, mammalian BP use is affected by recurrent cancer mutations in SF3B1 that reduce BP fidelity to the detriment of accurate 3′ss selection (Darman et al. 2015; Alsafadi et al. 2016). These mutations lie outside the BP-A binding pocket and disrupt the binding of the G-patch protein SUG1P, which promotes DHX15/PRP43 disassembly of early splicing complexes (Maul-Newby et al. 2022; Zhang et al. 2022; Feng et al. 2023). Failure to disassemble incorrectly recognized BPs results in the loss of fidelity. Thus, mammalian introns that may be intrinsically more susceptible to disassembly by SUG1P and DHX15 in the absence of inhibitors or mutations could also be more sensitive to inhibitors that disrupt U2-BP helix loading. Relevant to this model is the finding that certain mutations in SUG1P produce a subtle Plad-B resistance in human cells, possibly by delaying the discard of a Plad-B inhibited complex to provide time for escape of the block (Beusch et al. 2023). Although we currently have only partial understanding of the factors involved in the recognition and discard of complexes poised to use an incorrect BP in either system, this class of inhibitors is positioned to take advantage of BP fidelity mechanisms to create the observed intron-specific blocks to splicing.

Are splicing inhibitors of bacterial origin driving the evolution of eukaryotic microbial spliceosomes and gene architecture in the wild?

The class of splicing inhibitors that includes spliceostatins and pladienolides were originally identified as natural products with antitumor activity (Kaida et al. 2007). Further exploration of this class of molecules has revealed new derivatives, in particular from the ubiquitous soil bacteria Pseudomonas aeruginosa, Streptomyces platensis, and Burkholderia sp. (Zhao et al. 2019), which can be induced to make gram amounts of these compounds per liter of culture (Eustáquio et al. 2016; Adaikpoh et al. 2023). At the same time, evolutionary studies have noted that ancestors of present-day fungi including Saccharomyces sp., once had intron-rich genomes (Irimia and Roy 2008) with more relaxed BP and 5′ss sequences, as well as factors like SF3B6 (Yazhini et al. 2021) and U2AF1 (Schirman et al. 2021), and other splicing proteins (Sales-Lee et al. 2021; Black et al. 2023) now lost in present-day S. cerevisiae. So far, no ecological conditions under which intron loss, splice site sequence constraint, and loss of spliceosome proteins might be adaptive have been identified. It seems clear how producing a splicing inhibitor would advantage a bacterium in competition with eukaryotic microbes for limiting resources, but what evolutionary pathways might enable eukaryotes to succeed where bacteria are producing splicing inhibitors? In addition to resistance mechanisms that reduce the accumulation of foreign molecules in cells, or alter the inhibitor binding site, we imagine that events such as losing introns, simplification of the spliceosome by loss of protein subunits or auxiliary factors, and accumulation of intron BP mutations that lead to stronger U2-BP helices might minimize the impact of splicing inhibitors on cell reproduction over time. As we learn more about which organisms are naturally resistant to splicing inhibitors and whether they or their ancestors naturally resided with inhibitor-producing bacteria, the possible role of bacterial splicing inhibitors in shaping the genomes of eukaryotic microbes can be assessed.

Previous SectionNext Section

MATERIALS AND METHODS

Yeast strains

All strains used throughout this work are based on JRY8012 (BY4741 MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0 with pdr5Δ::KanMX6, snq2Δ::KanMX6 yor1Δ::KanMX6) (Jeong et al. 2007), which we refer to as “wild-type” for this study. The deletion of PDR5, SNQ2, and YOR1 is essential for allowing the splicing inhibitors to accumulate in yeast (Jeong et al. 2007; Hansen et al. 2019).

OHY001 was constructed from JRY8012 using the CRISPR/Cas9 yeast system. Construction of the CRISPR plasmid was done essentially as described (Talkish et al. 2019b), using a single plasmid p416-TEF1p-Cas9-NLS-crRNA-BaeI. This plasmid was cut with BaeI, and a guide sequence DNA targeting the genomic region encoding Hsh155 HRs 15–16 region (Supplemental Table S1) was inserted using HiFi DNA Assembly (New England Biolabs) to make p416-TEF1p-Cas9-NLS-crRNA-HSH155.

Cotransformation of p416-TEF1p-Cas9-NLS-crRNA-HSH155 with a 1021 bp rescue DNA (Supplemental Table S1) built by SGI-DNA carrying humanizing mutations was performed. Transformant colonies were selected on SCD–ura, and then restreaked on 5-fluoroorotic acid (5-FOA) plates to select for colonies that lost the URA-CRISPR plasmid. These colonies were subjected to colony PCR to amplify the genomic DNA coding region of HRs 15 and 16. Humanizing mutations in this region were confirmed by Sanger sequencing of the PCR product. Although the rescue synthetic construct contained humanizing mutations from Hsh155 amino acids 719–793, rescued clones varied with respect to which mutations were incorporated. After testing several individual clones, we selected one for which sequencing data confirmed that only amino acids 745–793 had changed. This specific strain OHY001 (humanizing mutations 745–793, Fig. 1, here called hsh155-ds) was used in all experiments unless otherwise indicated.

GFP-tagging

The OHY001 (hsh155-ds) and JRY8012 (HSH155) strains were transformed with a 2493 bp PCR product bearing GFP and Saccharomyces kluyveri HIS3 genes with homologous arms to the carboxyl terminus of the DNA coding region of the HSH155 gene. The PCR product was generated from a pYM28-kHIS3MX6 plasmid (Janke et al. 2004) using S2/S3 primers (Janke et al. 2004) adapted for carboxy-terminal HSH155 gene-tagging. The exact primer sequence can be found in Supplemental Table S1. Transformants were selected on SCD–hist. Colony PCR was performed to confirm the integration of GFP, and the PCR product was further confirmed by Sanger sequencing. This produced strains JRY8012-HSH155:GFP and OHY001-hsh155-ds:GFP. The Sir2-GFP yeast strain shown in the western blot analysis was a kind gift from Rohinton Kamakaka.

Amino acid sequence alignment

Amino acid sequences for human SF3B1 and yeast Hsh155 were taken from the UCSC Genome Browser and aligned using Clustal Omega to generate the alignment shown in Figure 1A.

Temperature-dependent growth assay

JRY8012 and OHY001 cells were cultured overnight in 5 mL YPD media at 30°C shaking at 220 rpm. Cells were spun down for 5 min and washed with 5 mL sterile water twice. Cells were then diluted down and allowed to grow up to OD600 = 0.5. The cultures were loaded in one column of a 96-well plate and diluted by 10-fold four times. An amount of 5 µL of cells from each dilution was dropped onto YPD plates and incubated at 18°C, 30°C, and 37°C for 2 d. The experiment was done in triplicate.

Western blots

OHY001, JRY8012-HSH155:GFP, OHY001-hsh155-ds:GFP, and Sir2-GFP were grown to log phase in 30°C overnight with 220 rpm shaking. The next day, cells were harvested at OD600 = 0.5. Whole-cell extracts were prepared by trichloroacetic acid (TCA) method exactly as described in Gallina et al. (2015). Total protein was then mixed with 2× protein gel loading buffer and heated at 95°C for 1 min to denature proteins. Samples were loaded onto a 4%–12% SDS–PAGE gel (Bio-Rad), transferred to a PVDF membrane, and blocked with 5% low-fat dry milk in 1× phosphate-buffered saline (PBS) (blocking buffer). The membrane was then washed with 1× PBS with 0.1% Tween-20 (washing buffer). Next, the membrane was probed with anti-GFP antibodies in blocking buffer (primary anti-GFP mouse monoclonal, Santa Cruz Biotechnology 1:200), visualized using goat anti-mouse IRDye 680 secondary antibody in blocking buffer at 1:15,000 (Li-Cor), then scanned on a Li-Cor infrared scanner.

Time course of inhibition

OHY001 cells were cultured overnight and diluted down to OD600 = 0.5 the next day. Plad-B was added to 5 µM and the control was supplemented with an equivalent amount of carrier DMSO. Cultures were incubated for either 0, 1, 4, or 8 min. At each time point, 8–9 mL cells were rapidly placed in 10 mL aliquots of 100% ethanol prechilled in a dry ice-ethanol bath to stop metabolism instantaneously (Koš and Tollervey 2010). Cells were pelleted and total RNA was extracted from all samples by hot phenol/chloroform extraction essentially as described (Ares 2012). Reverse transcription polymerase chain reaction (RT-PCR) was performed using primers annealing to the first intron of the MATA1 gene (Supplemental Table S1). Bands were quantified using ImageJ (NIH) and data were visualized using GraphPad Prism (v9.4.0 for macOS).

Splicing extract preparation and in vitro splicing assays

32P-radiolabeled actin pre-mRNA was transcribed in vitro using the MEGAscript T7 Transcription kit (Invitrogen). Saccharomyces cerevisiae splicing extracts were prepared from the hsh155-ds strain OHY001 using the liquid nitrogen method as described (Stevens and Abelson 2002), except frozen cells were disrupted using a Retsch MM301 ball mill for 5 × 3 min cycles at 10 Hz according to Jon Staley (protocol at https://voices.uchicago.edu/staleylab/retsch-ball-mill-protocol-by-the-staley-lab/, or https://www.retsch.com/knowledge-base/extract-of-saccharomyces-cerevisiae-cells-using-retsch-ball-mill/). ATP was depleted from extracts for 20 min at 23°C, using 1 U of S. cerevisiae hexokinase (Sigma-Aldrich) and 16 mM d-glucose in the presence of DMSO, or the indicated concentration of Pladienolide-B (Santa Cruz Biotechnology), or herboxidiene (Focus Biomolecules). After depletion of ATP, 32P-radiolabeled actin pre-mRNA was added to a final concentration of 0.4 nM, and standard splicing reactions were carried out in the presence or absence of 2 mM ATP at 23°C for 20 min as described (Ares 2013). To visualize the precursors and products of the splicing reaction, reactions were quenched in 200 µL of RNA extraction buffer (0.3 M NaOAc, 0.2% SDS, 1 mM EDTA, 10 µg/mL proteinase K) and incubated at 65°C for 10 min. RNA was extracted from the reactions using 200 µL of acid phenol (VWR), ethanol precipitated, resolved by electrophoresis on 6% acrylamide/8 M urea gels, and detected by phosphorimaging. Splicing complexes were visualized by mixing splicing reactions with 2× native loading dye (20 mM Tris/glycine, 25% glycerol, 0.1% bromophenol blue, and 1 mg/mL heparin), loaded directly on 2.1% agarose gels as described (Effenberger et al. 2013) and visualized by phosphorimaging. Splicing efficiency and ATP-dependent complex formation were quantified using ImageJ (NIH), and data (Supplemental Table S2) were visualized using GraphPad Prism (v9.4.0 for macOS).

Reverse transcription polymerase chain reaction (RT-PCR) and quantification

After extraction, total RNA was subjected to DNase treatment (TURBO DNase, Invitrogen). First-strand (FS) synthesis was performed using a 5× FS master mix composed of 0.5 µL 1× FS buffer (Invitrogen), 0.5 µL random primer mix, 5 µg RNA, and up to 7 µL of water. This mixture was heated to 95°C for 1 min, followed by 65°C for 1 min, then at room temperature for 1–2 min. All 7 µL were added to 5 µL of 1× reverse transcriptase (RTase) master mix. 1× RTase master mix included 1.5 µL 5× FS buffer, 1 µL 0.1 M dithiothreitol (DTT, Invitrogen), 1 µL 10 mM dNTPs (Thermo Scientific), 0.5 µL 40 U/µL RNase inhibitor (RNasin, Promega), and 0.5 µL 200 U/µL SuperScript III RTase (Invitrogen) or water for no RTase control. The final 12 µL mixture was incubated at room temperature for 5 min followed by 48°C for 25 min. The Zymo Research DNA Clean & Concentrator-5 kit was used to purify single-stranded DNA (ssDNA). PCR was performed on ssDNA, and products were run on a 2% agarose gel. Bands were quantified from unsaturated tif files using ImageJ (NIH), and data were visualized using GraphPad Prism (v9.4.0 for macOS).

Sequencing of rRNA-depleted total yeast RNA

Libraries were created from the following six treatments of two biological replicates (12 libraries total): (i) WT Hsh155 treated with 0.05% DMSO (0 µM inhibitor), (ii) WT Hsh155 treated with 5 µM Plad-B, (iii) hsh155-ds cells treated with 0.05% DMSO (0 µM inhibitor), (iv) hsh155-ds treated with 0.5 µM Plad-B, (v) hsh155-ds treated with 5 µM Plad-B, and (vi) hsh155-ds treated with 5 µM Thail-A. Inhibitor and DMSO treatments were made by growing WT Hsh155 and hsh155-ds in 50 mL YPD to OD600 = 0.5 and splitting them into 10 mL cultures for each treatment (DMSO, 0.5 µM Plad-B, 5 µM Plad-B, or 5 µM Plad-B Thail-A). Replicates were done at separate times with different starting cultures. Total RNA was then extracted from cells as for the time course. The quality of extracted RNA was assessed using High Sensitivity RNA TapeStation (Agilent). Yeast rRNA depletion was done using the Illumina Stranded Total RNA Prep, Ligation Ribo-Zero Plus kit (cat. no. 20040525) using RNase H treatment, with modification. The kit provides human rRNA oligos; however, we ordered and used a substitute DNA oligonucleotide set specific for yeast rRNA (see Supplemental Table S1 for sequences). The fragmentation step was skipped due to the smaller average size of yeast mRNAs compared to human mRNAs. After barcoding and amplification, libraries were sequenced using the NovaSeq 6000 system (Illumina, Inc.). Reads were trimmed to 120 × 120 bp using Cutadapt. Before mapping with STAR, the trimmed reads were aligned to a set of repeat elements from the S. cerevisiae genome using Bowtie2. Approximately 6% of the reads from each sample matched repeat elements, consisting mainly of residual ribosomal RNA (including mitochondrial 15S rRNA, 21S rRNA) and yeast Ty elements (Ty1, Ty2), and were excluded from further analysis. The remaining repeat-filtered reads were mapped to the sacCer3 genome assembly with STAR version 2.5.3 (Dobin et al. 2013).

Splicing efficiency calculations using junctionCounts

junctionCounts 0.1.0 (https://github.com/ajw2329/junctionCounts) was first used to extract “retained intron” (RI) alternative splicing events using an annotated genome (gtf) file and the “infer_pairwise_events” command. It was then run against the dup-removed STAR alignments with the “junctionCounts” command against the 344 RI events derived above, using default parameters. The SE (percentage spliced) was calculated by dividing the exon–exon junction count by the sum of the exon–exon and intron–exon count using the equations given in Figure 3A. RI splicing events with less than 100 average read counts between replicates were removed from the total normalized read count, resulting in 287 RI splicing events total (247 single-intron RI splicing events and 20 double intron RI splicing events; two intron genes were not further analyzed). The relative SE was calculated by normalizing to SE values for the appropriate DMSO control. Regression and two-tailed Pearson correlations were calculated using GraphPad Prism v9.4.0. Data were visualized using GraphPad Prism v9.4.0 for macOS.

Differential gene expression analysis

The dup-removed STAR alignments were converted to position-by-position coverage of the sacCer3 genome assembly using the bedtools (Quinlan 2014, v2.22.1) genomeCoverageBed command. Then for each gene, the total coverage of all its exonic positions was calculated. This was divided by 240, because each aligned paired end read in the STAR mapping covers 240 positions in the genome. The result is an estimate of the counts of reads aligned to that gene. These counts were used as inputs to DESeq2 1.30.1 (Love et al. 2014). DESeq2 uses a negative binomial distribution-based model to estimate the variance-mean relationship and perform statistical tests for differential gene expression. Differential gene expression was determined based on a log2 fold change threshold of >0.6 or <−0.6, and an adjusted P-value (Benjamini–Hochberg correction) significance threshold of P < 0.001. The DESeq2 analysis generated a results table containing the log2 fold change, P-value, and adjusted P-value for each gene (Supplemental Table S7.1).

Ranking and clustering intron resistance estimates

To compare intron resistance profiles to each other on a common scale, we normalized the raw (measured) SEs in 0.5 and 5 µM Plad-B by dividing each by the SE in the absence of the drug and multiplying by 100. This sets the resistance to 100% in the absence of the drug and proportionately adjusts the two drug-treated measurements. We then sorted and assigned ranks to introns based on their resistance (adjusted % SE, 1 = most resistant) for each of the two drug concentrations and then averaged the two ranks for each intron to obtain the “average rank” (Supplemental Table S4.7). Resistance profiles were clustered using Cluster3 (de Hoon et al. 2004). Clusters (.cdt and .gtr files) were visualized with Java Treeview (Saldanha 2004). The input and output files for Figure 4 are given in Supplemental Tables S4.7–4.9.

Single-molecule intron tracking (SMIT)

Wild-type JRY8012 and humanized OHY001 cells were each grown overnight in 50 mL YPD (yeast extract, peptone, dextrose) media at 30°C with 220 rpm overnight shaking. The following day, cells were diluted to OD600 = 0.1 in prewarmed media and grown to OD600 = 0.5. The 50 mL cultures were then split into two 25 mL cultures and treated with either DMSO or 5 µM Pladienolide-B (Plad-B) for 15 min to induce splicing inhibition. After treatment, chromatin was prepared essentially as described (Oesterreich et al. 2016) with additional details as outlined previously (Vo et al. 2021). Gene targeting, sequencing, mapping, and data processing were performed following the methods described in Oesterreich et al. (2016).

All scripts for analysis and processing of SMIT data can be found on GitHub at https://github.com/donoyoyo/SMITstuff. The raw sequencing reads for this SMIT experiment are available in the Sequence Read Archive with accession number PRJNA975902. Reads were preprocessed using bbmap clumpify (version 37.90) to remove duplicate and clump reads, which could arise from PCR artifacts or sequencing errors. Next, Cutadapt (version 1.11) was used to remove adapters and filter reads based on the presence of adapters, error rate, minimum length, and minimum overlap between the read and adapter. The following parameters were used with Cutadapt: -O 8 (minimum overlap of 8 bases), -n 2 (allowing up to two mismatches in the overlap region), -m 23 (keeping only reads with a minimum length of 23 bases), -e 0.11 (allowing up to 11% error rate), and ‐‐discard-untrimmed (discarding reads where no adapter was found). After adapter trimming and quality filtering, the processed reads were aligned to the sacCer3 genome using hisat2-2.1.0 with the following parameters: ‐‐no-mixed (disallowing reads that align to multiple reference sequences), and ‐‐max-intronlen 10000 (setting the maximum intron length to 10,000 bases to filter out potential false positive alignments).

The SMIT curves were generated using the following workflow. Only reads associated with the 63 primed genes and intronless control genes were considered. Reads are categorized by whether they are spliced or unspliced, as well as by the position of their 3′ end relative to the 3′ss. Splicing is calculated as the ratio of spliced count to the sum of unspliced count and spliced count (“raw fraction spliced”). The distribution and probability of insert length are determined from the data, and the fraction spliced is normalized using the position, insert length, and lengths of the gene products (spliced and unspliced), as well as the insert length probability function as originally described in Oesterreich et al. (2016) using the SMIT script provided in the GitHub link above.

To quantify the SE for each gene in each sample before comparing splicing with and without Plad-B, we defined a specific area under the splicing curve (AUC) over comparable gene positions where cotranscriptional splicing is robust in untreated cells for most genes. This specific area was bounded on the 5′ side by the gene-specific point of onset of splicing (the polymerase position at which spliced nascent transcripts first begin to accumulate) and on the 3′ side by the position 200 nt downstream from the point of onset of splicing. This definition measures cotranscriptional splicing over equivalent regions of each gene and is insensitive to differences in gene structure and activity such as the absolute distance between the 3′ss and the point of onset, or the length of the second exon (all of which extend beyond the 200 bp window). Files showing the fraction spliced versus RNA polymerase II position for all genes in all conditions are given in the Supplemental Material and in Supplemental Table S5.

To quantify the difference in SE between conditions for a given gene, the fractional difference in AUC values (ΔAUC) within the 200 nt beyond the onset of splicing was determined by subtracting the AUC of one sample (for example hsh155-ds in Plad-B) from the AUC of another sample (for example hsh155-ds in DMSO), and then taking that value as a percentage to estimate the percentage of SE relative to control for a given gene under a given treatment. To assess the significance of the ΔAUC, Wilcoxon paired P-values were calculated. The results of these calculations are shown in Supplemental Table S5.

To obtain the best estimate of the average SE under Plad-B treatment derived from multiple samples, we considered six pairs of sample comparisons, three of which compare treatments for which the standard RNA-seq analysis showed no significant splicing changes (HSH155 + Plad-B vs. HSH155 + DMSO, HSH155 + Plad-B vs. hsh155-ds + DMSO, and HSH155 + DMSO vs. hsh155-ds + DMSO), and three of which were expected based on standard RNA-seq to reveal splicing inhibition by Plad-B (hsh155-ds + Plad-B vs. hsh155-ds + DMSO, hsh155-ds + Plad-B vs. HSH155 + DMSO, and hsh155-ds + Plad-B vs. HSH155 + Plad-B). The last three comparisons represent the hsh155-ds strain treated with Plad-B (in which splicing is inhibited) in the numerator compared with three different samples in which Plad-B is either absent (vs. either strain in DMSO) or inert (vs. HSH155 in Plad-B) in the denominator. Thus, the best estimate for percent cotranscriptional splicing under Plad-B treatment would be represented by the average of these three comparisons (Supplemental Table S4.3).

Gene Ontology analysis

To identify Gene Ontology (GO) terms associated with differentially expressed genes, we utilized the Saccharomyces Genome Database (SGD) Gene Ontology Term Finder tool (version 2023-05-10). First, we obtained the differentially expressed genes 1.2× fold change P < 0.05 (low stringency) and 1.5× fold change P < 0.01 (high stringency) from the DESEQ2 hsh155-ds DMSO versus hsh155-ds 0.5 µM Plad-B analysis. For all three SGD ontologies (process, function, component), we obtained GO terms for both up- and down-regulated genes using a background set of intronless genes. These are shown in Supplemental Table S7.

Previous SectionNext Section

DATA DEPOSITION

The raw reads for the rRNA-depleted yeast total RNA-sequencing experiment are available in the Sequence Read Archive (SRA) with accession number PRJNA972180. Sequencing data for the SMIT experiment are at the Sequence Read Archive, accession number PRJNA975902. Publicly available UCSC Genome Browser coverage tracks for the samples can be accessed at https://genome.ucsc.edu/s/mannyares/Hunter_etal.

Previous SectionNext Section

SUPPLEMENTAL MATERIAL

Supplemental material is available for this article.

Previous SectionNext Section

ACKNOWLEDGMENTS

This work was supported by National Institutes of Health (NIH) grants R01 GM040478 (including a Diversity Supplement in support of O.H.) and R35 GM145266 to M.A. J.Q.-C. was supported by a Jane Coffin-Childs postdoctoral fellowship and a UC President's Postdoctoral Fellowship. Thanks to Jessie Lopez Suzuki for initiating this study during her rotation. We are very grateful to Karla Neugebauer for generously hosting us in her laboratory to learn the SMIT protocol, and in particular to Tara Alpert for detailed assistance with the analysis. Many thanks to Sarah Hansen, Aaron Hoskins, Suzanne Mays, and Juan Valcarcel for communicating unpublished results, and to Irene Beusch, Aaron Hoskins, Suzanne Mays, and Juan Valcarcel, and anonymous reviewers, for helpful comments on early versions.

  • Received October 5, 2023.
  • Accepted November 14, 2023.

This article is distributed exclusively by the RNA Society for the first 12 months after the full-issue publication date (see http://rnajournal.cshlp.org/site/misc/terms.xhtml). After 12 months, it is available under a Creative Commons License (Attribution-NonCommercial 4.0 International), as described at http://creativecommons.org/licenses/by-nc/4.0/.

Previous SectionNext Section

REFERENCES

  1. ↵
    Adaikpoh BI, Romanowski SB, Eustáquio AS. 2023. Understanding autologous spliceostatin transcriptional regulation to derive parts for heterologous expression in a bacterial host. ACS Synth Biol 12: 1952–1960. doi:10.1021/acssynbio.3c00228
    CrossRefGoogle Scholar
  2. ↵
    Albert B, Kos-Braun IC, Henras AK, Dez C, Rueda MP, Zhang X, Gadal O, Kos M, Shore D. 2019. A ribosome assembly stress response regulates transcription to maintain proteome homeostasis. Elife 8: e45002. doi:10.7554/eLife.45002
    CrossRefGoogle Scholar
  3. ↵
    Alors-Perez E, Blázquez-Encinas R, Alcalá S, Viyuela-García C, Pedraza-Arevalo S, Herrero-Aguayo V, Jiménez-Vacas JM, Mafficini A, Sánchez-Frías ME, Cano MT, et al. 2021. Dysregulated splicing factor SF3B1 unveils a dual therapeutic vulnerability to target pancreatic cancer cells and cancer stem cells with an anti-splicing drug. J Exp Clin Cancer Res 40: 382. doi:10.1186/s13046-021-02153-9
    CrossRefGoogle Scholar
  4. ↵
    Alpert T, Straube K, Oesterreich FC, Herzel L, Neugebauer KM. 2020. Widespread transcriptional readthrough caused by Nab2 depletion leads to chimeric transcripts with retained introns. Cell Rep 33: 108496. doi:10.1016/j.celrep.2020.108496
    CrossRefGoogle Scholar
  5. ↵
    Alsafadi S, Houy A, Battistella A, Popova T, Wassef M, Henry E, Tirode F, Constantinou A, Piperno-Neumann S, Roman-Roman S, et al. 2016. Cancer-associated SF3B1 mutations affect alternative splicing by promoting alternative branchpoint usage. Nat Commun 7: 10615. doi:10.1038/ncomms10615
    CrossRefMedlineGoogle Scholar
  6. ↵
    Ares M. 2012. Isolation of total RNA from yeast cell cultures. Cold Spring Harb Protoc 2012: 1082–1086. doi:10.1101/pdb.prot071456
    CrossRefMedlineGoogle Scholar
  7. ↵
    Ares M Jr.. 2013. Analysis of splicing in vitro using extracts of Saccharomyces cerevisiae. Cold Spring Harb Protoc 2013: 982–985. doi:10.1101/pdb.err089888
    CrossRefGoogle Scholar
  8. ↵
    Bai R, Wan R, Yan C, Jia Q, Lei J, Shi Y. 2021. Mechanism of spliceosome remodeling by the ATPase/helicase Prp2 and its coactivator Spp2. Science 371: eabe8863. doi:10.1126/science.abe8863
    Abstract/FREE Full Text
  9. ↵
    Barrass JD, Reid JEA, Huang Y, Hector RD, Sanguinetti G, Beggs JD, Granneman S. 2015. Transcriptome-wide RNA processing kinetics revealed using extremely short 4tU labeling. Genome Biol 16: 282. doi:10.1186/s13059-015-0848-1
    CrossRefGoogle Scholar
  10. ↵
    Berglund JA, Abovich N, Rosbash M. 1998. A cooperative interaction between U2AF65 and mBBP/SF1 facilitates branchpoint region recognition. Genes Dev 12: 858–867. doi:10.1101/gad.12.6.858
    Abstract/FREE Full Text
  11. ↵
    Beusch I, Rao B, Studer MK, Luhovska T, Šukytė V, Lei S, Oses-Prieto J, SeGraves E, Burlingame A, Jonas S, et al. 2023. Targeted high-throughput mutagenesis of the human spliceosome reveals its in vivo operating principles. Mol Cell 83: 2578–2594.e9. doi:10.1016/j.molcel.2023.06.003
    CrossRefGoogle Scholar
  12. ↵
    Black CS, Whelan TA, Garside EL, MacMillan AM, Fast NM, Rader SD. 2023. Spliceosome assembly and regulation: insights from analysis of highly reduced spliceosomes. RNA 29: 531–550. doi:10.1261/rna.079273.122
    Abstract/FREE Full Text
  13. ↵
    Bousquet-Antonelli C, Presutti C, Tollervey D. 2000. Identification of a regulated pathway for nuclear pre-mRNA turnover. Cell 102: 765–775. doi:10.1016/S0092-8674(00)00065-9
    CrossRefMedlineWeb of ScienceGoogle Scholar
  14. ↵
    Boyle EI, Weng S, Gollub J, Jin H, Botstein D, Cherry JM, Sherlock G. 2004. GO::TermFinder—open source software for accessing Gene Ontology information and finding significantly enriched Gene Ontology terms associated with a list of genes. Bioinformatics 20: 3710–3715. doi:10.1093/bioinformatics/bth456
    CrossRefMedlineWeb of ScienceGoogle Scholar
  15. ↵
    Brosi R, Gröning K, Behrens SE, Lührmann R, Krämer A. 1993a. Interaction of mammalian splicing factor SF3a with U2 snRNP and relation of its 60-kD subunit to yeast PRP9. Science 262: 102–105. doi:10.1126/science.8211112
    Abstract/FREE Full Text
  16. ↵
    Brosi R, Hauri HP, Krämer A. 1993b. Separation of splicing factor SF3 into two components and purification of SF3a activity. J Biol Chem 268: 17640–17646. doi:10.1016/S0021-9258(19)85380-2
    Abstract/FREE Full Text
  17. ↵
    Carrocci TJ, Paulson JC, Hoskins AA. 2018. Functional analysis of Hsh155/SF3b1 interactions with the U2 snRNA/branch site duplex. RNA 24: 1028–1040. doi:10.1261/rna.065664.118
    Abstract/FREE Full Text
  18. ↵
    Charpentier B, Rosbash M. 1996. Intramolecular structure in yeast introns aids the early steps of in vitro spliceosome assembly. RNA 2: 509–522.
    Abstract
  19. ↵
    Chua K, Reed R. 2001. An upstream AG determines whether a downstream AG is selected during catalytic step II of splicing. Mol Cell Biol 21: 1509–1514. doi:10.1128/MCB.21.5.1509-1514.2001
    Abstract/FREE Full Text
  20. ↵
    Corrionero A, Miñana B, Valcárcel J. 2011. Reduced fidelity of branch point recognition and alternative splicing induced by the anti-tumor drug spliceostatin A. Genes Dev 25: 445–459. doi:10.1101/gad.2014311
    Abstract/FREE Full Text
  21. ↵
    Cretu C, Agrawal AA, Cook A, Will CL, Fekkes P, Smith PG, Lührmann R, Larsen N, Buonamici S, Pena V. 2018. Structural basis of splicing modulation by antitumor macrolide compounds. Mol Cell 70: 265–273.e8. doi:10.1016/j.molcel.2018.03.011
    CrossRefGoogle Scholar
  22. ↵
    Cretu C, Gee P, Liu X, Agrawal A, Nguyen T-V, Ghosh AK, Cook A, Jurica M, Larsen NA, Pena V. 2021. Structural basis of intron selection by U2 snRNP in the presence of covalent inhibitors. Nat Commun 12: 4491. doi:10.1038/s41467-021-24741-1
    CrossRefGoogle Scholar
  23. ↵
    Darman RB, Seiler M, Agrawal AA, Lim KH, Peng S, Aird D, Bailey SL, Bhavsar EB, Chan B, Colla S, et al. 2015. Cancer-associated SF3B1 hotspot mutations induce cryptic 3′ splice site selection through use of a different branch point. Cell Rep 13: 1033–1045. doi:10.1016/j.celrep.2015.09.053
    CrossRefMedlineGoogle Scholar
  24. ↵
    de Hoon MJL, Imoto S, Nolan J, Miyano S. 2004. Open source clustering software. Bioinformatics 20: 1453–1454. doi:10.1093/bioinformatics/bth078
    CrossRefMedlineWeb of ScienceGoogle Scholar
  25. ↵
    de la Cruz J, Gómez-Herreros F, Rodríguez-Galán O, Begley V, de la Cruz Muñoz-Centeno M, Chávez S. 2018. Feedback regulation of ribosome assembly. Curr Genet 64: 393–404. doi:10.1007/s00294-017-0764-x
    CrossRefGoogle Scholar
  26. ↵
    DeRisi JL, Iyer VR, Brown PO. 1997. Exploring the metabolic and genetic control of gene expression on a genomic scale. Science 278: 680–686. doi:10.1126/science.278.5338.680
    Abstract/FREE Full Text
  27. ↵
    Dobin A, Davis CA, Schlesinger F, Drenkow J, Zaleski C, Jha S, Batut P, Chaisson M, Gingeras TR. 2013. STAR: ultrafast universal RNA-seq aligner. Bioinformatics 29: 15–21. doi:10.1093/bioinformatics/bts635
    CrossRefMedlineWeb of ScienceGoogle Scholar
  28. ↵
    Effenberger KA, Perriman RJ, Bray WM, Lokey RS, Ares M Jr., Jurica MS. 2013. A high-throughput splicing assay identifies new classes of inhibitors of human and yeast spliceosomes. J Biomol Screen 18: 1110–1120. doi:10.1177/1087057113493117
    CrossRefMedlineGoogle Scholar
  29. ↵
    Effenberger KA, Anderson DD, Bray WM, Prichard BE, Ma N, Adams MS, Ghosh AK, Jurica MS. 2014. Coherence between cellular responses and in vitro splicing inhibition for the anti-tumor drug pladienolide B and its analogs. J Biol Chem 289: 1938–1947. doi:10.1074/jbc.M113.515536
    Abstract/FREE Full Text
  30. ↵
    Egecioglu DE, Kawashima TR, Chanfreau GF. 2012. Quality control of MATa1 splicing and exon skipping by nuclear RNA degradation. Nucleic Acids Res 40: 1787–1796. doi:10.1093/nar/gkr864
    CrossRefMedlineWeb of ScienceGoogle Scholar
  31. ↵
    Eustáquio AS, Chang L-P, Steele GL, Donnell CJ O, Koehn FE. 2016. Biosynthetic engineering and fermentation media development leads to gram-scale production of spliceostatin natural products in Burkholderia sp. Metab Eng 33: 67–75. doi:10.1016/j.ymben.2015.11.003
    CrossRefGoogle Scholar
  32. ↵
    Feng Q, Krick K, Chu J, Burge CB. 2023. Splicing quality control mediated by DHX15 and its G-patch activator SUGP1. Cell Rep 42: 113223. doi:10.1016/j.celrep.2023.113223
    CrossRefGoogle Scholar
  33. ↵
    Finci LI, Zhang X, Huang X, Zhou Q, Tsai J, Teng T, Agrawal A, Chan B, Irwin S, Karr C, et al. 2018. The cryo-EM structure of the SF3b spliceosome complex bound to a splicing modulator reveals a pre-mRNA substrate competitive mechanism of action. Genes Dev 32: 309–320. doi:10.1101/gad.311043.117
    Abstract/FREE Full Text
  34. ↵
    Folco EG, Coil KE, Reed R. 2011. The anti-tumor drug E7107 reveals an essential role for SF3b in remodeling U2 snRNP to expose the branch point-binding region. Genes Dev 25: 440–444. doi:10.1101/gad.2009411
    Abstract/FREE Full Text
  35. ↵
    Gallina I, Colding C, Henriksen P, Beli P, Nakamura K, Offman J, Mathiasen DP, Silva S, Hoffmann E, Groth A, et al. 2015. Cmr1/WDR76 defines a nuclear genotoxic stress body linking genome integrity and protein quality control. Nat Commun 6: 6533. doi:10.1038/ncomms7533
    CrossRefMedlineGoogle Scholar
  36. ↵
    Hansen SR, Nikolai BJ, Spreacker PJ, Carrocci TJ, Hoskins AA. 2019. Chemical inhibition of pre-mRNA splicing in living Saccharomyces cerevisiae. Cell Chem Biol 26: 443–448.e3. doi:10.1016/j.chembiol.2018.11.008
    CrossRefGoogle Scholar
  37. ↵
    Hartwell LH, McLaughlin CS, Warner JR. 1970. Identification of ten genes that control ribosome formation in yeast. Mol Gen Genet 109: 42–56. doi:10.1007/BF00334045
    CrossRefMedlineWeb of ScienceGoogle Scholar
  38. ↵
    Hasegawa M, Miura T, Kuzuya K, Inoue A, Won Ki S, Horinouchi S, Yoshida T, Kunoh T, Koseki K, Mino K, et al. 2011. Identification of SAP155 as the target of GEX1A (Herboxidiene), an antitumor natural product. ACS Chem Biol 6: 229–233. doi:10.1021/cb100248e
    CrossRefMedlineWeb of ScienceGoogle Scholar
  39. ↵
    Howe KJ, Ares M Jr.. 1997. Intron self-complementarity enforces exon inclusion in a yeast pre-mRNA. Proc Natl Acad Sci 94: 12467–12472. doi:10.1073/pnas.94.23.12467
    Abstract/FREE Full Text
  40. ↵
    Hsu TY-T, Simon LM, Neill NJ, Marcotte R, Sayad A, Bland CS, Echeverria GV, Sun T, Kurley SJ, Tyagi S, et al. 2015. The spliceosome is a therapeutic vulnerability in MYC-driven cancer. Nature 525: 384–388. doi:10.1038/nature14985
    CrossRefMedlineGoogle Scholar
  41. ↵
    Huang HH, Ferguson ID, Thornton AM, Bastola P, Lam C, Lin Y-HT, Choudhry P, Mariano MC, Marcoulis MD, Teo CF, et al. 2020. Proteasome inhibitor-induced modulation reveals the spliceosome as a specific therapeutic vulnerability in multiple myeloma. Nat Commun 11: 1931. doi:10.1038/s41467-020-15521-4
    CrossRefGoogle Scholar
  42. ↵
    Irimia M, Roy SW. 2008. Evolutionary convergence on highly-conserved 3′ intron structures in intron-poor eukaryotes and insights into the ancestral eukaryotic genome. PLoS Genet 4: e1000148. doi:10.1371/journal.pgen.1000148
    CrossRefMedlineGoogle Scholar
  43. ↵
    Janke C, Magiera MM, Rathfelder N, Taxis C, Reber S, Maekawa H, Moreno-Borchart A, Doenges G, Schwob E, Schiebel E, et al. 2004. A versatile toolbox for PCR-based tagging of yeast genes: new fluorescent proteins, more markers and promoter substitution cassettes. Yeast 21: 947–962. doi:10.1002/yea.1142
    CrossRefMedlineWeb of ScienceGoogle Scholar
  44. ↵
    Jeong H, Herskowitz I, Kroetz DL, Rine J. 2007. Function-altering SNPs in the human multidrug transporter gene ABCB1 identified using a Saccharomyces-based assay. PLoS Genet 3: e39. doi:10.1371/journal.pgen.0030039
    CrossRefMedlineGoogle Scholar
  45. ↵
    Kaida D, Motoyoshi H, Tashiro E, Nojima T, Hagiwara M, Ishigami K, Watanabe H, Kitahara T, Yoshida T, Nakajima H, et al. 2007. Spliceostatin A targets SF3b and inhibits both splicing and nuclear retention of pre-mRNA. Nat Chem Biol 3: 576–583. doi:10.1038/nchembio.2007.18
    CrossRefMedlineWeb of ScienceGoogle Scholar
  46. ↵
    Kao C-Y, Cao E-C, Wai HL, Cheng S-C. 2021. Evidence for complex dynamics during U2 snRNP selection of the intron branchpoint. Nucleic Acids Res 49: 9965–9977. doi:10.1093/nar/gkab695
    CrossRefGoogle Scholar
  47. ↵
    Kistler AL, Guthrie C. 2001. Deletion of MUD2, the yeast homolog of U2AF65, can bypass the requirement for sub2, an essential spliceosomal ATPase. Genes Dev 15: 42–49. doi:10.1101/gad.851601
    Abstract/FREE Full Text
  48. ↵
    Koš M, Tollervey D. 2010. Yeast pre-rRNA processing and modification occur cotranscriptionally. Mol Cell 37: 809–820. doi:10.1016/j.molcel.2010.02.024
    CrossRefMedlineWeb of ScienceGoogle Scholar
  49. ↵
    Li X, Liu S, Zhang L, Issaian A, Hill RC, Espinosa S, Shi S, Cui Y, Kappel K, Das R, et al. 2019. A unified mechanism for intron and exon definition and back-splicing. Nature 573: 375–380. doi:10.1038/s41586-019-1523-6
    CrossRefGoogle Scholar
  50. ↵
    Liang W-W, Cheng S-C. 2015. A novel mechanism for Prp5 function in prespliceosome formation and proofreading the branch site sequence. Genes Dev 29: 81–93. doi:10.1101/gad.253708.114
    Abstract/FREE Full Text
  51. ↵
    Liu X, Biswas S, Berg MG, Antapli CM, Xie F, Wang Q, Tang M-C, Tang G-L, Zhang L, Dreyfuss G, et al. 2013. Genomics-guided discovery of thailanstatins A, B, and C As pre-mRNA splicing inhibitors and antiproliferative agents from Burkholderia thailandensis MSMB43. J Nat Prod 76: 685–693. doi:10.1021/np300913h
    CrossRefGoogle Scholar
  52. ↵
    Ljungdahl PO, Daignan-Fornier B. 2012. Regulation of amino acid, nucleotide, and phosphate metabolism in Saccharomyces cerevisiae. Genetics 190: 885–929. doi:10.1534/genetics.111.133306
    Abstract/FREE Full Text
  53. ↵
    Love MI, Huber W, Anders S. 2014. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol 15: 550. doi:10.1186/s13059-014-0550-8
    CrossRefMedlineGoogle Scholar
  54. ↵
    Lustig AJ, Lin RJ, Abelson J. 1986. The yeast RNA gene products are essential for mRNA splicing in vitro. Cell 47: 953–963. doi:10.1016/0092-8674(86)90810-X
    CrossRefMedlineWeb of ScienceGoogle Scholar
  55. ↵
    Maul-Newby HM, Amorello AN, Sharma T, Kim JH, Modena MS, Prichard BE, Jurica MS. 2022. A model for DHX15 mediated disassembly of A-complex spliceosomes. RNA 28: 583–595. doi:10.1261/rna.078977.121
    Abstract/FREE Full Text
  56. ↵
    Meyer M, Plass M, Pérez-Valle J, Eyras E, Vilardell J. 2011. Deciphering 3′ss selection in the yeast genome reveals an RNA thermosensor that mediates alternative splicing. Mol Cell 43: 1033–1039. doi:10.1016/j.molcel.2011.07.030
    CrossRefMedlineWeb of ScienceGoogle Scholar
  57. ↵
    Michaud S, Reed R. 1991. An ATP-independent complex commits pre-mRNA to the mammalian spliceosome assembly pathway. Genes Dev 5: 2534–2546. doi:10.1101/gad.5.12b.2534
    Abstract/FREE Full Text
  58. ↵
    Michaud S, Reed R. 1993. A functional association between the 5′ and 3′ splice site is established in the earliest prespliceosome complex (E) in mammals. Genes Dev 7: 1008–1020. doi:10.1101/gad.7.6.1008
    Abstract/FREE Full Text
  59. ↵
    Miller AM. 1984. The yeast MATa1 gene contains two introns. EMBO J 3: 1061–1065. doi:10.1002/j.1460-2075.1984.tb01927.x
    CrossRefMedlineWeb of ScienceGoogle Scholar
  60. ↵
    Miller C, Schwalb B, Maier K, Schulz D, Dümcke S, Zacher B, Mayer A, Sydow J, Marcinowski L, Dölken L, et al. 2011. Dynamic transcriptome analysis measures rates of mRNA synthesis and decay in yeast. Mol Syst Biol 7: 458. doi:10.1038/msb.2010.112
    Abstract/FREE Full Text
  61. ↵
    Naryshkin NA, Weetall M, Dakka A, Narasimhan J, Zhao X, Feng Z, Ling KKY, Karp GM, Qi H, Woll MG, et al. 2014. Motor neuron disease. SMN2 splicing modifiers improve motor function and longevity in mice with spinal muscular atrophy. Science 345: 688–693. doi:10.1126/science.1250127
    Abstract/FREE Full Text
  62. ↵
    Oesterreich FC, Herzel L, Straube K, Hujer K, Howard J, Neugebauer KM. 2016. Splicing of nascent RNA coincides with intron exit from RNA polymerase II. Cell 165: 372–381. doi:10.1016/j.cell.2016.02.045
    CrossRefGoogle Scholar
  63. ↵
    Perriman R, Ares M Jr.. 2010. Invariant U2 snRNA nucleotides form a stem loop to recognize the intron early in splicing. Mol Cell 38: 416–427. doi:10.1016/j.molcel.2010.02.036
    CrossRefMedlineWeb of ScienceGoogle Scholar
  64. ↵
    Perriman R, Barta I, Voeltz GK, Abelson J, Ares M Jr.. 2003. ATP requirement for Prp5p function is determined by Cus2p and the structure of U2 small nuclear RNA. Proc Natl Acad Sci 100: 13857–13862. doi:10.1073/pnas.2036312100
    Abstract/FREE Full Text
  65. ↵
    Plaschka C, Lin P-C, Charenton C, Nagai K. 2018. Prespliceosome structure provides insights into spliceosome assembly and regulation. Nature 559: 419–422. doi:10.1038/s41586-018-0323-8
    CrossRefMedlineGoogle Scholar
  66. ↵
    Quinlan AR. 2014. BEDTools: the Swiss-army tool for genome feature analysis. Curr Protoc Bioinformatics 47: 11.12.1–11.12.34. doi:10.1002/0471250953.bi1112s47
    CrossRefMedlineGoogle Scholar
  67. ↵
    Rangan R, Huang R, Hunter O, Pham P, Ares M, Das R. 2023. RNA structure landscape of S. cerevisiae introns. bioRxiv doi:10.1101/2022.07.22.501175
    Abstract/FREE Full Text
  68. ↵
    Rauhut R, Fabrizio P, Dybkov O, Hartmuth K, Pena V, Chari A, Kumar V, Lee C-T, Urlaub H, Kastner B, et al. 2016. Molecular architecture of the Saccharomyces cerevisiae activated spliceosome. Science 353: 1399–1405. doi:10.1126/science.aag1906
    Abstract/FREE Full Text
  69. ↵
    Rogic S, Montpetit B, Hoos HH, Mackworth AK, Ouellette BF, Hieter P. 2008. Correlation between the secondary structure of pre-mRNA introns and the efficiency of splicing in Saccharomyces cerevisiae. BMC Genomics 9: 355. doi:10.1186/1471-2164-9-355
    CrossRefMedlineGoogle Scholar
  70. ↵
    Roybal GA, Jurica MS. 2010. Spliceostatin A inhibits spliceosome assembly subsequent to prespliceosome formation. Nucleic Acids Res 38: 6664–6672. doi:10.1093/nar/gkq494
    CrossRefMedlineWeb of ScienceGoogle Scholar
  71. ↵
    Saldanha AJ. 2004. Java Treeview—extensible visualization of microarray data. Bioinformatics 20: 3246–3248. doi:10.1093/bioinformatics/bth349
    CrossRefMedlineWeb of ScienceGoogle Scholar
  72. ↵
    Sales-Lee J, Perry DS, Bowser BA, Diedrich JK, Rao B, Beusch I, Yates JR III., Roy SW, Madhani HD. 2021. Coupling of spliceosome complexity to intron diversity. Curr Biol 31: 4898–4910.e4. doi:10.1016/j.cub.2021.09.004
    CrossRefGoogle Scholar
  73. ↵
    Schellenberg MJ, Edwards RA, Ritchie DB, Kent OA, Golas MM, Stark H, Lührmann R, Glover JNM, MacMillan AM. 2006. Crystal structure of a core spliceosomal protein interface. Proc Natl Acad Sci 103: 1266–1271. doi:10.1073/pnas.0508048103
    Abstract/FREE Full Text
  74. ↵
    Schirman D, Yakhini Z, Pilpel Y, Dahan O. 2021. A broad analysis of splicing regulation in yeast using a large library of synthetic introns. PLoS Genet 17: e1009805. doi:10.1371/journal.pgen.1009805
    CrossRefGoogle Scholar
  75. ↵
    Seiler M, Yoshimi A, Darman R, Chan B, Keaney G, Thomas M, Agrawal AA, Caleb B, Csibi A, Sean E, et al. 2018. H3B-8800, an orally available small-molecule splicing modulator, induces lethality in spliceosome-mutant cancers. Nat Med 24: 497–504. doi:10.1038/nm.4493
    CrossRefMedlineGoogle Scholar
  76. ↵
    Serrat X, Kukhtar D, Cornes E, Esteve-Codina A, Benlloch H, Cecere G, Cerón J. 2019. CRISPR editing of sftb-1/SF3B1 in Caenorhabditis elegans allows the identification of synthetic interactions with cancer-related mutations and the chemical inhibition of splicing. PLoS Genet 15: e1008464. doi:10.1371/journal.pgen.1008464
    CrossRefGoogle Scholar
  77. ↵
    Smith CW, Chu TT, Nadal-Ginard B. 1993. Scanning and competition between AGs are involved in 3′ splice site selection in mammalian introns. Mol Cell Biol 13: 4939–4952. doi:10.1128/mcb.13.8.4939-4952.1993
    Abstract/FREE Full Text
  78. ↵
    Spingola M, Grate L, Haussler D, Ares M Jr.. 1999. Genome-wide bioinformatic and molecular analysis of introns in Saccharomyces cerevisiae. RNA 5: 221–234. doi:10.1017/S1355838299981682
    Abstract
  79. ↵
    Stevens SW, Abelson J. 2002. Yeast pre-mRNA splicing: methods, mechanisms, and machinery. Methods Enzymol 351: 200–220. doi:10.1016/S0076-6879(02)51849-8
    CrossRefMedlineWeb of ScienceGoogle Scholar
  80. ↵
    Talkish J, Igel H, Hunter O, Horner SW, Jeffery NN, Leach JR, Jenkins JL, Kielkopf CL, Ares M. 2019a. Cus2 enforces the first ATP-dependent step of splicing by binding to yeast SF3b1 through a UHM-ULM interaction. RNA 25: 1020–1037. doi:10.1261/rna.070649.119
    Abstract/FREE Full Text
  81. ↵
    Talkish J, Igel H, Perriman RJ, Shiue L, Katzman S, Munding EM, Shelansky R, Donohue JP, Ares M. 2019b. Rapidly evolving protointrons in Saccharomyces genomes revealed by a hungry spliceosome. PLoS Genet 15: e1008249. doi:10.1371/journal.pgen.1008249
    CrossRefGoogle Scholar
  82. ↵
    Teng T, Tsai JH, Puyang X, Seiler M, Peng S, Prajapati S, Aird D, Buonamici S, Caleb B, Chan B, et al. 2017. Splicing modulators act at the branch point adenosine binding pocket defined by the PHF5A-SF3b complex. Nat Commun 8: 15522. doi:10.1038/ncomms15522
    CrossRefGoogle Scholar
  83. ↵
    Tholen J, Razew M, Weis F, Galej WP. 2022. Structural basis of branch site recognition by the human spliceosome. Science 375: 50–57. doi:10.1126/science.abm4245
    CrossRefGoogle Scholar
  84. ↵
    Vigevani L, Gohr A, Webb T, Irimia M, Valcárcel J. 2017. Molecular basis of differential 3′ splice site sensitivity to anti-tumor drugs targeting U2 snRNP. Nat Commun 8: 2100. doi:10.1038/s41467-017-02007-z
    CrossRefGoogle Scholar
  85. ↵
    Vo JM, Mulroney L, Quick-Cleveland J, Jain M, Akeson M, Ares M Jr.. 2021. Synthesis of modified nucleotide polymers by the poly(U) polymerase Cid1: application to direct RNA sequencing on nanopores. RNA 27: 1497–1511. doi:10.1261/rna.078898.121
    Abstract/FREE Full Text
  86. ↵
    Wan R, Bai R, Yan C, Lei J, Shi Y. 2019. Structures of the catalytically activated yeast spliceosome reveal the mechanism of branching. Cell 177: 339–351.e13. doi:10.1016/j.cell.2019.02.006
    CrossRefMedlineGoogle Scholar
  87. ↵
    Wilkinson ME, Charenton C, Nagai K. 2020. RNA splicing by the spliceosome. Annu Rev Biochem 89: 359–388. doi:10.1146/annurev-biochem-091719-064225
    CrossRefGoogle Scholar
  88. ↵
    Xia X. 2020. RNA-Seq approach for accurate characterization of splicing efficiency of yeast introns. Methods 176: 25–33. doi:10.1016/j.ymeth.2019.03.019
    CrossRefGoogle Scholar
  89. ↵
    Xu Y-Z, Query CC. 2007. Competition between the ATPase Prp5 and branch region-U2 snRNA pairing modulates the fidelity of spliceosome assembly. Mol Cell 28: 838–849. doi:10.1016/j.molcel.2007.09.022
    CrossRefMedlineWeb of ScienceGoogle Scholar
  90. ↵
    Yan C, Wan R, Bai R, Huang G, Shi Y. 2016. Structure of a yeast activated spliceosome at 3.5 Å resolution. Science 353: 904–911. doi:10.1126/science.aag0291
    Abstract/FREE Full Text
  91. ↵
    Yazhini A, Sandhya S, Srinivasan N. 2021. Rewards of divergence in sequences, 3-D structures and dynamics of yeast and human spliceosome SF3b complexes. Curr Res Struct Biol 3: 133–145. doi:10.1016/j.crstbi.2021.05.003
    CrossRefGoogle Scholar
  92. ↵
    Yokoi A, Kotake Y, Takahashi K, Kadowaki T, Matsumoto Y, Minoshima Y, Sugi NH, Sagane K, Hamaguchi M, Iwata M, et al. 2011. Biological validation that SF3b is a target of the antitumor macrolide pladienolide. FEBS J 278: 4870–4880. doi:10.1111/j.1742-4658.2011.08387.x
    CrossRefMedlineGoogle Scholar
  93. ↵
    Zhang D, Hu Q, Liu X, Ji Y, Chao H-P, Liu Y, Tracz A, Kirk J, Buonamici S, Zhu P, et al. 2020a. Intron retention is a hallmark and spliceosome represents a therapeutic vulnerability in aggressive prostate cancer. Nat Commun 11: 2089. doi:10.1038/s41467-020-15815-7
    CrossRefGoogle Scholar
  94. ↵
    Zhang Z, Will CL, Bertram K, Dybkov O, Hartmuth K, Agafonov DE, Hofele R, Urlaub H, Kastner B, Lührmann R, et al. 2020b. Molecular architecture of the human 17S U2 snRNP. Nature 583: 310–313. doi:10.1038/s41586-020-2344-3
    CrossRefMedlineGoogle Scholar
  95. ↵
    Zhang Z, Rigo N, Dybkov O, Fourmann J-B, Will CL, Kumar V, Urlaub H, Stark H, Lührmann R. 2021. Structural insights into how Prp5 proofreads the pre-mRNA branch site. Nature 596: 296–300. doi:10.1038/s41586-021-03789-5
    CrossRefGoogle Scholar
  96. ↵
    Zhang J, Huang J, Xu K, Xing P, Huang Y, Liu Z, Tong L, Manley JL. 2022. DHX15 is involved in SUGP1-mediated RNA missplicing by mutant SF3B1 in cancer. Proc Natl Acad Sci 119: e2216712119. doi:10.1073/pnas.2216712119
    CrossRefGoogle Scholar
  97. ↵
    Zhao H, Zhang B, Ma L-F, Shi L-M, Zhan Z-J. 2019. Cytotoxic spliceostatin analogs from Pseudomonas sp. Chem Biodivers 16: e1900266. doi:10.1002/cbdv.201900266
    CrossRefGoogle Scholar
Previous Section
 

MEET THE FIRST AUTHOR

Oarteze Hunter
View larger version:
Oarteze Hunter

Meet the First Author(s) is an editorial feature within RNA, in which the first author(s) of research-based papers in each issue have the opportunity to introduce themselves and their work to readers of RNA and the RNA research community. Oarteze Hunter is the first author of this paper, “Broad variation in response of individual introns to splicing inhibitors in a humanized yeast strain.” Oarteze did this work as a graduate student in the laboratory of Dr. Manuel Ares Jr. at the University of California, Santa Cruz Center for the Molecular Biology of RNA. His research focused on understanding the mechanism of splicing inhibition using a new humanized yeast strain. Oarteze is currently a postdoctoral fellow in the laboratory of Dr. Hilary Coller at UCLA.

What are the major results described in your paper and how do they impact this branch of the field?

In this study, we wanted to understand the mechanism of small molecule splicing inhibitors such as Pladienolide-B (Plad-B) and Thailanstatin-A (Thail-A). Human SF3B1 and its yeast homolog Hsh155 both have a pocket that binds the branchpoint early in splicing, but only the human pocket binds splicing inhibitors Plad-B and Thail-A, so yeast is resistant to these compounds. We wanted to use the relatively simple yeast system with its few introns and genetic tractability to see what the consequences of blocking splicing were on a per-gene basis. We used CRISPR/Cas9 in yeast to humanize 14 amino acids in the pocket to make yeast sensitive to inhibitors. We measured total and cotranscriptional splicing and found that some introns are more resistant to inhibitors than others, with introns that deviate from the standard UACUAAC branchpoint sequence among the most sensitive. We also compared the sensitivity profiles of Plad-B and Thail-A and found that resistance to one inhibitor does not predict resistance to a different inhibitor. Finally, we observed an indirect response of intronless genes when humanized yeast were treated with a drug, suggesting a complex and widespread gene expression program between intron-containing and intronless genes. Our main results echo what has been previously seen in mammalian studies, such as the importance of the branchpoint sequence on splicing inhibition. This strain can be used for any future experimentation that requires splicing to be inhibited.

What led you to study RNA or this aspect of RNA science?

I always credit the environment of UCSC's Center for the Molecular Biology of RNA for introducing me to the world of RNA. Manny was super influential to me in this space when he offered to have me as an undergrad in his lab to help study splicing. I was able to truly understand how complex RNA is, from a molecular, biochemical, and evolutionary perspective. I also met a lot of fun and outgoing people in this field including a few at RNA Society meetings. Overall, I had a great time and learned about all the awesome things RNA can do.

What are some of the landmark moments that provoked your interest in science or your development as a scientist?

I was very interested in nature as a kid. I spent most of my childhood outside so there was plenty to look at and wonder about. I also moved around Los Angeles a lot as a child, so I was exposed to different people and places. I used to catch bugs and was fascinated by bug behavior. I looked up at night and saw stars and the moon and wondered where they came from. I wanted to know about what things were made of down to their simplest parts. I think it wasn't until graduate school that I found my place in the study of biology. I was pushed to my intellectual and emotional limits. Grad school changed me in ways I didn't anticipate. It was the hardest thing I've done in my life so far. I found my identity as a scientist during that time of my life.

If you were able to give one piece of advice to your younger self, what would that be?

If I were to give a piece of advice to my younger self, it would be to take your ideas seriously. I took a lot of walks around the beaches of Santa Cruz coming up with molecular mechanisms that could explain the results I was seeing in the lab. I had all these big ideas about how biology worked but would also let them pass without putting them into the world. Maybe it was fear of having my ideas rejected, or maybe it was more related to self-doubt and not having enough background knowledge on a certain topic to even feel like I could have my own ideas about it. It wasn't really until grad school that I learned that although these could be valid fears, I shouldn't let them stop me from contributing to the field. I was lucky enough to have a good community of people who wanted to listen. And yes it was uncomfortable at first, but I got the hang of it. And I'm glad I did because a few of those big ideas made their way into this paper.

Articles citing this article

  • A Yeast-Based High-Throughput Screening Platform for the Discovery of Novel pre-mRNA Splicing Modulators bioRxiv November 2, 2025 0: 2025.10.20.683303v2-2025.10.20.683303
  • The pre-mRNA Splicing Modulator Pladienolide B Inhibits Cryptococcus neoformans Germination and Growth bioRxiv March 12, 2025 0: 2025.03.06.641977v1-2025.03.06.641977
  • Selected humanization of yeast U1 snRNP leads to global suppression of pre-mRNA splicing and mitochondrial dysfunction in the budding yeast RNA August 1, 2024 30: 1070-1088
  • Intron-lariat spliceosomes convert lariats to true circles: implications for intron transposition bioRxiv April 4, 2024 0: 2024.03.26.586863v1-2024.03.26.586863
  • Intron lariat spliceosomes convert lariats to true circles: implications for intron transposition Genes Dev. April 1, 2024 38: 322-335
  • The evolutionary dynamics of alternative splicing during primate neuronal differentiation bioRxiv February 22, 2024 0: 2024.02.20.581203v1-2024.02.20.581203
| Table of Contents

This Article

  1. Published in Advance November 28, 2023, doi: 10.1261/rna.079866.123 RNA 30: 149-170 © 2024 Hunter et al.; Published by Cold Spring Harbor Laboratory Press for the RNA Society

Article Category

ORCID

  1. Profile for Hunter, O.http://orcid.org/0000-0003-0458-4229
  2. Profile for Talkish, J.http://orcid.org/0000-0002-0260-3345
  3. Profile for Quick-Cleveland, J.http://orcid.org/0000-0002-4070-8842
  4. Profile for Igel, H.http://orcid.org/0000-0001-6468-3128
  5. Profile for Katzman, S.http://orcid.org/0000-0002-1787-1246
  6. Profile for Donohue, J. P.http://orcid.org/0000-0002-5497-1940
  7. Profile for S. Jurica, M.http://orcid.org/0000-0002-4488-1386
  8. Profile for Ares, M.http://orcid.org/0000-0002-2552-9168

Share

Navigate This Article

Current Issue

  1. July 2026, 32 (7)
  1. Current Issue
  1. Alert me to new issues of RNA