Identification of translation events that drive nonsense-mediated mRNA decay reveals functional roles for alternative RNA isoforms

Wait 5 sec.

IntroductionThe nonsense-mediated mRNA decay (NMD) pathway degrades mRNAs undergoing premature translation termination in eukaryotes1,2,3. Genes encoding the core components of the NMD pathway were initially identified in Saccharomyces cerevisiae – NAM7 (UPF1), NMD2 (UPF2), and UPF34,5,6,7,8. Despite extensive research, the exact biological function of NMD remains poorly understood. A compelling model for recognition of premature termination (stop) codons (PTCs) is the “faux-3’UTR model” that links the efficiency of termination to NMD3,9. Efficient translation termination depends on interactions between the translation termination release factors (eRF1 and eRF3) and proteins bound to the 3’UTR and polyA tail; these interactions are weakened when the ribosome terminates upstream at a PTC, and this allows the NMD machinery to associate with the ribosome and initiate NMD9,10,11,12,13,14,15,16,17,18. In some eukaryotes, detection of PTCs can be enhanced by downstream exon-junction complexes19,20,21 or inhibited by other proteins22,23. Nevertheless, the core NMD pathway is conserved in S. cerevisiae, making it a tractable organism for addressing basic questions of NMD function and mechanism24.Several structural classes of NMD targets have been identified in yeast based on their stabilization upon inactivation of the NMD pathway, including: (1) genes with mutations that introduce nonsense codons, including pseudogenes1,25,26, (2) mRNAs that have retained their introns either due to splicing errors or developmental programming e.g., RPL28 (CYH2) and REC107 (MER2)27, (3) mRNAs containing upstream open reading frames (uORFs) in their 5’UTR e.g., CPA126,28,29,30,31,32, (4) mRNAs with long 3’UTRs e.g., PGA1 and COX1915,16,33, (5) mRNAs that utilize programmed + 1 frameshifting in their translation e.g., EST326, and (6) a subfraction of mRNAs with short polyA tails34 that likely have reduced binding of Pab1 resulting in inefficient termination and NMD.NMD is likely modulated by many factors. Only some mRNAs with uORFs are targeted by the NMD pathway26,29,30. At least some that are resistant to NMD appear to exhibit efficient termination30, potentially due to sequences in proximity to the uORF stop codon that recruit factors that facilitate termination, e.g., Pub129. Similarly, only some mRNAs with long 3’UTRs are targeted to NMD. The basis of this resistance is unclear, though it may depend on the length of the main ORF17. In addition, the level of ribosome queuing upstream of a stop codon or the identity of the codon immediately prior to the stop codon (“penultimate codon”) have been proposed to affect termination efficiency and could therefore be important11,12,13,14. Studies have also shown that the level of readthrough at stop codons can affect NMD35,36. Apparent NMD sensitivity is also affected by the efficiency of competing mRNA turnover pathways37. Given the above, it remains a challenge to predict how effective any particular PTC is at promoting degradation.Aside from the annotated genome, many unannotated transcripts have been identified as NMD substrates in yeast. These include stable unannotated transcripts (SUTs)38 and RNAs stabilized in the absence of genes encoding specific RNA decay factors, including the cytoplasmic 5’−3’ RNA exonuclease XRN1 (XUTs)39,40 and the catalytic subunit of the decapping complex DCP241. It has been shown that 70% of XUT transcripts are NMD substrates40,42, and another study reported that of ~ 800 unannotated transcripts identified, 16% of were NMD-sensitive43.A recent exhaustive effort defined NMD targets in yeast by performing RNA-seq on WT, upf1∆, upf2∆, and upf3∆ yeast strains44. Among protein-coding genes and functional RNAs, 88% of NMD-targeted transcripts appeared to lack an obvious PTC that would fit into one of the five structural classes described above. These are likely to be direct targets of NMD because prior studies showed that RNAs stabilized in upf∆ strains directly associate with Upf1 protein45. The genes without a recognizable PTC were generally translated inefficiently, suggesting that alternative translation events could be responsible for NMD44. Alternative NMD-targeting translation events have been reported, and many relate to an often-underappreciated observation that multiple transcript isoforms can be associated with a given yeast gene46,47,48. Whereas early studies using RNA-seq and tiling microarrays limited each gene to only one transcription start site (TSS)38,49, more recent studies using transcript isoform sequencing (TIF-Seq), which sequences the 5’ and 3’ ends of each transcript, and TL-seq, a high-throughput version of 5’ Rapid Amplification of cDNA Ends (RACE), confirmed multiple TSSs for a majority of S. cerevisiae genes31,42,50. These later studies identified several additional classes of transcript isoforms that are targeted to NMD, including one where two or more distinct TSSs, which were at least 50 nucleotides apart, exist for at least 6.3% of genes31.Large numbers of two-transcript gene regulatory systems where transcription generates a 5’-extended RNA that represses the promoter of a downstream canonical transcript have recently been described51,52,53,54,55. Transcription of a 5’-extended transcript, known as a long undecoded transcript isoform (LUTI), results in repressive chromatin marks being deposited on the downstream (proximal) promoter52,55,56. There is no expression of the canonical protein from the LUTI (hence it is “undecoded”) due to the presence of multiple uORFs at the 5’ end of the transcript. Transcript isoform switching between LUTIs and canonical transcripts occurs frequently in meiosis51,53. LUTIs also have key roles in the unfolded protein response (UPR)54, and similar two-transcript systems have functions in other areas of biology57,58,59. Many SUTs and XUTs also map 5’ of coding genes38,39,40, and most DCP2-sensitive transcripts map proximal to inducible genes involved in metabolism41. This suggests they may also play a role in regulating expression of canonical transcripts.Here, we use an RNA-seq and 40S ribosome profiling strategy to identify alternative translation events (mainly uORFs and internal open reading frames [iORFs]) on transcripts associated with NMD-sensitive genes. We report a plausible rationale to account for most NMD-sensitive transcripts. Intriguingly, we find many NMD-sensitive, 5’-extended transcripts that are associated with inducible genes and function as LUTIs. We show that the NMD-sensitive DAL5 LUTI has a functional role in repressing expression of the canonical DAL5 transcript.ResultsDetection of premature termination events that cause NMDTo identify genes that increased in abundance in the upf1∆ strain (i.e., NMD targets), we performed RNA-seq on BY4741 (WT) and upf1∆ S. cerevisiae strains followed by differential expression analysis on reads from main open reading frames (ORFs) with DESeq260. Because NMD has been reported to target genes that, to varying degrees, retain an intron27, intron read counts were also analyzed (Supplementary Fig. S1a). Combining both analyses after duplicate removal (4 genes were found in both datasets, see Methods) yielded a high confidence set of 552 genes that were significantly changed and upregulated > 2-fold in upf1∆ cells (Fig. 1a, b and Supplementary Data 1 Sheet1).Fig. 1: Strategy for identification of PTCs and 40S peak analysis. See also Supplementary Figures S1, S4.Full size imagea Pipeline for identification of NMD-sensitive genes, including genes with retained introns. DESeq2 was used to determine significantly changing RNA levels ( > 2-fold) within ORFs and introns (552 total). Retained introns reported as a distinct category, and other genes subjected to further analysis in (d). b Volcano plot shows cutoffs for significance and fold-change, with each dot representing 1 gene. Differential gene expression analysis was performed using DESeq2, which applies a two-sided Wald test to assess the statistical significance of log2 fold changes. To account for multiple comparisons, raw p-values were adjusted using the Benjamini-Hochberg false discovery rate (FDR) procedure. c The three possible locations for PTCs: uORFs (top), iORFs (middle), main stop codon rendered a PTC by some other factor, such as a long 3’UTR, as shown (bottom). d Pipeline for identification of 40S peak location to facilitate categorization of transcripts. e Categorization of where 40S peaks were identified in NMD-sensitive genes.We previously used 40S ribosome profiling to show that 40S subunits accumulate at start codons and that these peaks can be enhanced by reducing the level of 60S subunits via deletion of RPL11B, one of two genes that encode Rpl11p (uL5). We also showed that deletion of genes encoding 40S subunit recycling factors, TMA64 and TMA20, results in a dramatic accumulation of 40S subunits at stop codons61. We reasoned that the peaks at start and stop codons could be utilized as markers to precisely identify translation events that would otherwise be missed by 80S ribosome profiling. Supporting this association, we found that the 40S start or stop codon peak height of a given gene showed a strong correlation with the number of 80S ribosome footprints that mapped to the respective gene’s coding sequence (Supplementary Fig. S1b, c). Given the better correlation for start codon peaks, we focused on them for much of our analysis (see Methods). We therefore performed 80S ribosome profiling on WT and upf1∆ strains, and 40S ribosome profiling on rpl11b∆ and tma64∆ tma20∆ (abbreviated 64∆ 20∆) in an upf1∆ strain background. Confirming this approach, we could detect a 40S stop codon peak at the previously identified PTC for the NMD target RPL28 (CYH2) (Supplementary Fig. S1d).Classically, the term premature termination codon (PTC) has been used to denote termination events at premature stop codons introduced into the coding sequence upstream of the canonical stop codon by mutation that result in NMD. Here we use a broader definition to include any stop codon, generally 5’ of the canonical stop codon, that results in a termination event and NMD. Theoretically, there are three possible locations for premature termination events: after upstream open reading frames (uORFs) in the 5’ untranslated region (5’UTR) (uORFs; Fig. 1c top), after internal out-of-frame ORFs within the main ORF (iORFs; Fig. 1c middle), and “canonical” termination events that are rendered premature by some other feature of the transcript, e.g., a long 3’UTR (Fig. 1c bottom). We therefore developed a customized bioinformatic pipeline (Fig. 1d) to identify putative premature termination events for the first two categories (Fig. 1e) and offer examples and potential candidates for the third case where the canonical termination event is recognized as premature. As not every premature termination event triggers NMD, our approach focuses on the set of 552 NMD-sensitive genes we identified by RNA-seq and then offers a rationalization for the NMD sensitivity of each transcript.Classification of premature termination events in 5’UTRsWe detected 320 NMD-sensitive genes with 5’UTR translation events (Fig. 1e) using 40S profiling peak score analysis (see Methods) (Supplementary Data 1 Sheet2). Start codons were identified in all frames by extending the annotated 5’UTR upstream to capture any cases of alternative upstream TSSs (see Methods). As noted above, many yeast genes utilize multiple TSSs (i.e., LUTIs), and this results in the production of multiple transcripts per gene. Preliminary observations revealed two different classes of NMD-sensitive transcripts with uORFs: traditional annotated transcripts and a newly defined class of 5’-extended transcripts, often with multiple uORFs, similar to LUTIs (Fig. 2a). Analysis of the distribution of yeast 5’UTR lengths based on a more extensive annotation of 5’UTR TSSs62 revealed that the average 5’UTR length is approximately 100 nt (Fig. 2b). We therefore used 100 nt as a cut-off to define uORFs in short 5’UTR transcripts (Supplementary Data 1 Sheet3) vs extended 5’UTR transcripts (Fig. 2c and Supplementary Data 1 Sheet4).Fig. 2: Categorization of 5’UTR 40S peak containing transcripts. See also Supplementary Figure S1.Full size imagea Potential transcription architecture of yeast genes with uORFs: extended 5’UTR and canonical transcripts (top) or a short 5’UTR transcript (bottom). Note the similarity of the architecture of the extended 5’UTR and canonical transcripts to the architecture of LUTI 2-transcript systems. PAS = polyadenylation signal. b Distribution of annotated 5’UTR lengths in yeast grown in glucose. c Categorization of NMD-sensitive genes with 5’UTR 40S peaks into extended 5’UTR transcripts (184 genes) and short 5’UTR transcripts (136 genes) based on the average 5’UTR length (100-nt cutoff). d Sequencing reads for the DAL5 gene in rich media (arrow corresponds to main ORF) show an example of a 5’-extended isoform. 80S ribosome profiling (top) reveals translation on uORFs. 40S profiling (middle tracks) shows locations of start and stop codons on uORFs. mRNA-Seq (bottom) shows expression of the extended 5’UTR transcript and its sensitivity to NMD. e Close-up of d to reveal peaks on uORFs.We identified the DAL5 gene as an NMD-sensitive extended 5’UTR transcript (Fig. 2d). The gene’s main ORF encodes allantoate permease and is repressed by nitrogen catabolite repression (NCR) under rich nitrogen growth conditions63,64. The extended 5’UTR transcript includes a 5’UTR that measures approximately 1500 nt in length and includes 33 AUG codons that could serve as uORF start codons. We observed 40S start and stop codon peaks at the first four uORFs of the extended 5’UTR transcript (Fig. 2e). Interestingly, these uORFs appear to inhibit expression of the main DAL5 ORF based on the lack of 80S profiling reads (Fig. 2d). In contrast, an example of an NMD-sensitive gene with a uORF that was classified to be part of a short 5’UTR transcript is QDR3 (Supplementary Fig. S1e).Classification of premature termination events in coding sequencesWe detected 194 NMD-sensitive genes with internal translation events within the main ORF in a frame different from the coding sequence (Fig. 1e) using 40S profiling peak score analysis (see Methods) (Supplementary Data 1 Sheet5).There are several possible explanations for the observed translation events within main ORFs: (1) The 48S pre-initiation complex (PIC) occasionally does not recognize the annotated start codon. This could be due either to poor start codon context (leaky scanning) or a short transcript leader that causes the PIC to sometimes attach downstream of the annotated start codon. Both alternatives should result in initiation at the next available start codon within the main ORF. We call these collectively inefficient initiation transcripts (IITs; Fig. 3a top). (2) The TSS for the transcript is within the main ORF, resulting in a 5’ truncated transcript. This ORF-internal transcript isoform (OTI) lacks the canonical ORF start codon, so translation must initiate at an internal AUG (Fig. 3a bottom).Fig. 3: Categorization of ORF 40S peak containing transcripts. See also Supplementary Figure S2.Full size imagea Potential transcription architecture of yeast genes with iORFs: inefficient initiation transcript (IIT) of a canonical isoform due to leaky scanning or short transcript leader (top) or ORF-internal transcript isoform (OTI) where the transcript starts within the main ORF (bottom). PAS = polyadenylation signal. b Categorization of NMD-sensitive genes with main ORF 40S peaks into IITs or OTIs based on the presence or absence of a 40S main ORF start codon peak and the split-gene test. c Sequencing reads for the TCA17 gene (arrow corresponds to main ORF) show an example of an IIT. 80S ribosome profiling (top) fails to show iORF translation. 40S profiling (middle tracks) shows locations of the start and stop codons for the iORF. mRNA-Seq (bottom) shows the expression of the canonical transcript and its sensitivity to NMD. d Close-up view of c to show individual peaks on the translated iORF. e Sequencing reads for the HFM1 gene (arrow corresponds to main ORF) show an example of an OTI. 80S ribosome profiling (top) shows translation of iORFs. 40S profiling (middle tracks) shows the locations of start and stop codons on iORFs. mRNA-Seq (bottom) shows expression of the OTI and its sensitivity to NMD. f Close-up view of (e) to show individual peaks on translated iORFs.We divided NMD-sensitive genes with 40S ORF peaks into IITs and OTIs using several bioinformatic criteria. A simple way to separate IITs and OTIs is based on the presence or absence of a 40S peak at the start codon of the main ORF (Supplementary Fig. S2a). NMD-sensitive genes with 40S start peaks were categorized as IITs (101 genes) (Fig. 3b and Supplementary Data 1 Sheet6), whereas genes without were categorized as OTIs (93 genes) (Supplementary Data 1 Sheet7). Another bioinformatic method to identify OTIs based on upf1∆ RNA-seq data was developed (see Methods). This split-count test (Supplementary Fig. S2b) compared RNA-Seq read density in the first and last 100 nt of the annotated main ORF. Those with a more than 2-fold higher read count in the final 100 nt of the annotated main ORF relative to the first 100 nt were identified as OTIs (49 genes) (Supplementary Data 1 Sheet8). Comparison of the OTIs identified by the two tests showed strong overlap (Supplementary Fig. S2c), so the lists were combined to yield a total of 101 OTIs (Fig. 3b and Supplementary Data 1 Sheet9). The TCA17 gene expresses an IIT that appears to inefficiently initiate translation at the main start codon, possibly due to a short (20 nt) 5’UTR (Fig. 3c), and often initiates translation on a 7-aa iORF (Fig. 3d). The TCA17 gene reveals the power of the 40S profiling technique, as the internal translation event cannot be detected in the 80S ribosome profiling track. The HFM1 gene expresses an OTI (Fig. 3e) that exhibits 40S start and stop codon peaks at three out-of-frame iORFs within the main ORF (Fig. 3f).Overlap and additional causes for peaksWe noted that many of the 90 genes (Fig. 1e) that have both 5’UTR and ORF peaks appear to express two NMD-sensitive transcript isoforms, such as BDS1 (Supplementary Fig. S2d). In addition, our 40S profiling method identified programmed ribosomal frameshift sites. Of 3 known65,66 sites, OAZ1, EST3, and ABP140, only EST3 shows significant NMD-sensitivity (Supplementary Fig. S2e). In total, we found at least one translation event that likely results in premature termination for 77% of the 552 NMD-sensitive genes we identified by 40S peak analysis (Fig. 1e).Confirming the role of 5’ uORFs in targeting DAL5 to NMDTo confirm the ability of uORFs to trigger NMD on extended 5’UTR transcripts, we developed a system (based on the YCplac33 sc URA3 plasmid) for studying the extended 5’UTR transcript of DAL5 (Fig. 2d, e and Methods). Northern blots of RNA samples from YCplac33-DAL5 cells recapitulated what we observed in our RNA-seq data with the extended 5’UTR DAL5 transcript being stabilized in the upf1∆ background strain (Fig. 4a). Having successfully established a system for studying the extended 5’UTR transcript of DAL5, we then mutated the 33 AUG start codons in the extended 5’UTR transcript upstream of the YCplac33-DAL5 plasmid to AAA to test if these uORFs target the transcript for NMD (Supplementary Fig. S3a, b). Mutation of the 5’ uORF AUGs resulted in stabilization of the DAL5 extended 5’UTR transcript in UPF1+ (WT) cells (Fig. 4b), proving that the uORFs are responsible for the transcript’s NMD sensitivity. The level of stabilization is equivalent to that observed in the upf1∆ strain, suggesting that these uORF termination events account for all the NMD sensitivity. To test the generality of this result, we repeated the experiment with another NMD-sensitive gene, DAL7 (Fig. 4c). While the DAL7 5’UTR does meet the 100-nt cutoff to be classified as an extended 5’UTR transcript, experiments below suggest it is part of a two-transcript system. A similar plasmid-based system was established to mutate its 3 5’UTR AUG start codons (Supplementary Fig. S3c, d). As with DAL5, these mutations resulted in the stabilization of DAL7 (Fig. 4d). These results show that PTCs that we identified with 40S ribosome profiling target these transcripts to NMD.Fig. 4: uORFs in the extended 5’UTR transcripts of DAL5 and DAL7 are responsible for NMD sensitivity. See also Supplementary Figure S3.Full size imagea Northern analysis of the DAL5 extended 5’UTR transcript expressed from the YCplac33-DAL5 plasmid confirms NMD sensitivity. b Northern analysis of the DAL5 extended 5’UTR transcript expressed from YCplac33-DAL5 and YCplac33-DAL5_noAUG, where all uORF AUGs are mutated to AAA, shows that the DAL5 uORFs confer NMD sensitivity. The diagram shows mutations made to two representative uORFs. c Sequencing reads for the DAL7 gene in rich media (arrow corresponds to main ORF). 80S ribosome profiling (top) reveals translation on uORFs. 40S profiling (middle tracks) shows locations of the start and stop codons on uORFs. mRNA-Seq (bottom) shows expression of the long transcript and its sensitivity to NMD. d Northern analysis of the DAL7 long transcript expressed from YCplac33-DAL7 and YCplac33-DAL7_noAUG, where all uORF AUGs are mutated to AAA shows uORFs fully confer NMD sensitivity. e Western analysis of the YCplac33-DAL5-GFP and YCplac33-DAL5-GFP_noAUG reporters shows that loss of all uORFs in the DAL5 extended 5’UTR transcript leader leads to translation of the main ORF. The second lane for each experiment is 2X loaded. In blots above, SCR1 or β-actin serve as loading controls. Two replicates are shown for each condition. Blots were quantified using Fiji (ImageJ); individual data points are shown as open circles. For 4e both the 1X and 2X loadings for each replicate were quantified.Given the length of the DAL5 extended transcript’s 5’UTR ( ~ 1.5 kb), we wondered whether, in the absence of 5’ uORFs, the 48S PIC would be able to scan the entire 5’UTR and initiate translation at the main ORF start codon. We therefore constructed versions of YCplac33-DAL5 and YCplac33-DAL5_noAUG where the DAL5 main ORF was replaced with the coding sequence for GFP. We found that, in the absence of 5’ uORFs, GFP was expressed (Fig. 4e), demonstrating that the 5’ uORFs in the extended 5’UTR transcript are critical for preventing expression of the Dal5 protein.Potential role of long 3’UTRs in stimulating NMDWe found that 128 genes (including those with retained introns) that increased in abundance in the upf1∆ strain lack 40S peaks in their 5’UTR or ORF (Fig. 1e). In many cases, therefore, we are left to conclude that it is the main stop codon that is recognized as a PTC (Fig. 1c bottom). While multiple factors could be responsible for this effect, previous studies have shown that inefficient termination at a canonical stop codon can be driven by a long 3’UTR15,16,33. For example, the COX19 gene has been reported to be targeted to NMD and 3’UTR swapping experiments showed that the long 3’UTR is responsible33. We therefore examined the distribution of 3’UTR lengths (Supplementary Fig. S4a left) and chose the average plus two standard deviations ( > 396 nt) to generate a list of 143 genes with “long” 3’UTRs. While we found this 3’UTR length was, on average, more sensitive to NMD (Supplementary Fig. S4a right), it is not robust as a predictor. We therefore utilized a process of elimination to find the best cases for where a transcript’s 3’UTR length is a likely trigger NMD and offer a resource for future study (Supplementary Fig. S4b).First, we noted that in pioneering work, Kebaara and Atkin (2009) bioinformatically identified 56 genes with 3’UTRs > 350 nt as strong candidates where 3’UTR length would drive NMD16 [Supplemental Table 1 therein]. We found that 19 of these genes were significantly NMD-sensitive and exhibited 3’UTRs with a predicted length similar to that observed in our RNA-seq data (Supplementary Data 1 Sheet10). Second, based on our analysis (Supplementary Fig. S4a and Methods), we used the 396 nt cutoff to add 27 NMD-sensitive genes where a long 3’UTR could be the driver of NMD (Supplementary Data 1 Sheet11). Consistent with the main stop codon being recognized as a PTC, the main ORF exhibited an increase in 80S ribosome profiling footprints in the upf1∆ strain in all these cases.Next, we considered the observation that “bicistronic” mRNAs had been previously suggested as a type of NMD substrate26. The use of the term bicistronic can be misleading; it is unlikely that these transcripts produce two protein products from a single mRNA. We therefore suggest that a more accurate term is pseudo-bicistronic transcripts. Pseudo-bicistronic transcripts are potential NMD targets since translation termination at the stop codon of the first ORF could be recognized as a PTC, with the second ORF effectively acting as a long 3’UTR. In many cases, a second transcript covering the second ORF in the pair is also expressed, thus enabling production of both protein products (Supplementary Fig. S4c).A study using TIF-Seq identified 279 bicistronic transcripts in yeast50 [Supplementary data 10 therein]. After manual editing for false positives (see Methods), we found 50 bicistronic transcripts where either one or both genes (68 total genes) were NMD targets in our RNA-Seq data (Supplementary Data 1 Sheet12). Anecdotally, we noticed an additional 13 genes involved in pseudo-bicistronic pairs that did not meet the threshold defined for bicistronic transcripts in the TIF-seq study50. These genes were part of 9 pseudo-bicistronic transcripts (Supplementary Data 1 Sheet13). SLO1-ISC10 is an example of a gene pair expressing an NMD-sensitive pseudo-bicistronic transcript originally identified by TIF-seq (Supplementary Fig. S4d). Our 80S data show that only SLO1 is translated, making it likely that its canonical stop codon acts as a PTC. In another example, Kebaara and Atkin (2009)16 determined that the MAK31 gene is part of two transcript isoforms, the longer of which is NMD-sensitive and constitutes a pseudo-bicistronic transcript extending over the PET18 ORF (Supplementary Fig. S4e). Our data again show that only MAK31 is translated, with its canonical stop codon acting as a likely PTC. Our data therefore help explain how NMD controls the expression of genes encoded within pseudo-bicistronic transcripts.Lastly, we considered pseudogenes as another case that effectively acts as a long 3’UTR. We define a pseudogene here as a tandem gene pair separated by a (premature) stop codon or other mutation that is not present in closely related Saccharomyces species (see Methods). We identified 22 NMD-sensitive genes that are paired into 11 likely pseudogene transcripts (Supplementary Data 1 Sheet14). As an example, the NIT1-YIL165C pseudogene encodes a nitrilase with a single premature stop codon between NIT1 and YIL165C (Supplementary Fig. S4f). While uORFs likely drive NMD under the conditions shown, the stop codon after NIT1 is expected to be important for any shorter induced transcript isoforms that lack uORFs (see below).Taken together, the above analysis and observations argue that long 3’UTRs could be a driver of NMD for many genes, including many that may be targeted for additional reasons due to 40S peaks we identified above. Of the 128 genes with no 40S peaks, we suggest that long 3’UTRs, where the canonical stop codon acts as a premature stop codon, and pseudogenes may explain NMD-sensitivity for an additional 57 genes.With the 424 genes with 40S peaks and 10 genes with retained introns, our full analysis offers a rationalization of NMD sensitivity for 491 of 552 genes or 89% (Supplementary Data 2). For the remaining 11% of transcripts that lack rationalization, increased sequencing depth in our 40S profiling datasets could potentially reveal the PTCs that drive NMD since this cohort is low in abundance. Given that we found transcription start site (TSS) to be a major determinant of NMD sensitivity, we next explored functional roles for genes with multiple TSSs.NMD-sensitive LUTIs regulate metabolismThe NMD-sensitive extended 5’UTR transcripts identified here resemble LUTIs that have been previously identified to play a role in two-transcript gene regulatory systems (Fig. 2a top)51,52,53,55. Genes utilizing this system have two promoters: a proximal promoter close to the main ORF from which a canonical (or induced) transcript isoform is expressed and a 5’ distal promoter from which the LUTI is expressed. Transcription of the canonical transcript from the proximal promoter results in translation of the main ORF protein (uORFs block main ORF translation on LUTI transcripts). Transcription factors have been proposed to bind to both promoters and respond to changing environmental conditions or developmental programs, though the magnitude of change (particularly for the LUTI transcript, which does not always change) varies considerably between genes55. The key function of the LUTI is that its transcription represses the downstream proximal promoter via repressive chromatin marks52,55,56. We compared the list of genes with putative LUTIs associated with meiosis from (Cheng et al. 2018)53 [Supplementary Table S2 therein] with our list of 320 NMD-sensitive genes with 5’UTR 40S peaks and found 84 in common, including DAL5 and DAL7. This overlap suggests that NMD-sensitive extended 5’UTR transcripts have functional roles as LUTIs. To better assess whether the NMD-sensitive extended 5’UTR transcripts we observed in rich media are LUTIs, we asked whether a second, canonical transcript could be induced under alternate growth conditions. Similar to previous studies of NMD-targets26,45, many genes in our list are involved in nitrogen catabolism and thiamine biosynthesis. We therefore grew WT and upf1∆ cells under either poor nitrogen (proline) or no thiamine growth conditions and performed RNA-seq.We identified 48 genes that were > 5-fold upregulated in poor nitrogen (Supplementary Data 1 Sheet15). A 5-fold cut-off was utilized to examine genes that were more likely to be directly upregulated by poor nitrogen rather than genes upregulated due to secondary effects. We then compared changes in expression due to poor nitrogen with changes in expression due to loss of UPF1 (Fig. 5a). We identified a group of genes with NMD-sensitive long transcripts that showed increased expression in poor nitrogen (upper right quadrant), suggestive of a LUTI architecture (Fig. 5a). To further investigate if these transcripts were LUTIs, we focused on the DAL5 and DAL7 genes. Both expressed shorter, canonical transcripts that were not sensitive to NMD under poor nitrogen conditions in addition to NMD-sensitive long transcripts that were constitutively expressed under rich and poor conditions (Fig. 5b and Supplementary Fig. S5a). To confirm these observations, we performed northern blots under poor nitrogen conditions (Fig. 5c and Supplementary Fig. S5b). We observed similar stability patterns for the long (candidate LUTI) transcripts as we did under rich nitrogen growth conditions (Fig. 4b, d, i.e., loss of uORFs stabilized the extended transcripts). We also observed distinct shorter, induced DAL5 and DAL7 transcripts that were not subject to NMD, confirming that at least two transcript isoforms are expressed from these genes.Fig. 5: Identification of regulated transcript pairs that include an NMD-sensitive 5’-extended (LUTI) transcript and downstream (canonical) transcript. See also Figure S5.Full size imagea Comparison of differential expression in mRNA-Seq data for poor vs rich nitrogen (y-axis) vs loss of UPF1 (x-axis). Genes most affected by both conditions (upper right quadrant of zoom-up) are enriched for putative LUTIs. b mRNA-Seq reads mapped to the DAL5 gene in order from top to bottom: WT in nitrogen-rich media, upf1∆ in nitrogen-rich media, WT in nitrogen-poor media, upf1∆ in nitrogen-poor media. Black arrows on top show approximate TSSs. Colored arrows at the bottom show transcripts. Data reveal that the NMD-sensitive, putative LUTI is present under both media conditions and the existence of a shorter canonical NMD-insensitive transcript that is induced under poor nitrogen. c Northern analysis confirms expression of both a DAL5 putative LUTI and a canonical induced transcript under poor nitrogen conditions. The 5’ uORFs confer NMD sensitivity to only the putative LUTI. SCR1 serves as a loading control. Two biological replicates are shown for each condition. d Comparison of differential expression in mRNA-Seq data for loss of thiamine (y-axis) vs loss of UPF1 (x-axis). The genes most affected by both conditions are enriched for putative LUTIs (upper right quadrant). e mRNA-Seq reads mapped to the MAK31-PET18 pseudo-bicistronic gene pair in order from top to bottom: WT in media with thiamine, upf1∆ in media with thiamine, WT in media without thiamine, upf1∆ in media without thiamine. Black arrows on top show approximate TSSs. Arrows at the bottom show transcripts. Data reveal an NMD-sensitive, pseudo-bicistronic transcript that is present under both media conditions and an NMD-insensitive induced PET18 transcript that is only present in the absence of thiamine. f Comparison of differential expression in mRNA-Seq data for +DTT vs −DTT (y-axis) vs loss of UPF1 (x-axis). The genes most affected by both conditions are enriched in putative LUTIs, such as ULI1. g Comparison of differential expression in mRNA-Seq data over the region 200 nt upstream of the start codon for +DTT vs −DTT in a upf1∆ background (y-axis) vs loss of UPF1 in +DTT conditions (x-axis). This analysis reveals cases where NMD-sensitive LUTIs (both putative and previously reported) may be turned on by DTT. For (a, d, and f), included differentially expressed genes are those that have FDR  10 fold) and peak height ( > 137 rpkm). A more stringent filtering was used for the ORF 40S peak analysis than for the 5’UTR 40S peak analysis due to the higher 40S background within ORFs caused by artifactual 80S ribosome separation during sample processing61,93. However, based on visual inspection of strong peaks that did not make the pause score cutoff, we noted 3 genes, YOL159C-A, MAK31, KTI11, where the low pause score appeared to be explained by other 40S peaks that artificially elevated the pause score denominator (baseline average). The 40S peaks for these genes were manually added back to the 40S ORF peak list.An ORF 40S peak representing the true start codon was also detected for LTO1/YNL260C. The misannotated main start codon has been corrected in more recent yeast gffs so this peak was removed from the ORF 40S peak list. In addition, twenty-two 0-frame (in frame with main AUG) 40S peaks were removed from the ORF 40S peak list since they do not result in a premature termination event. They are likely to be main start codons (annotated start codon is not used and, therefore, likely misannotated) or internal start codons that would lead to N-terminally truncated protein (annotated start codon is used) (Supplementary Data 4). For example, the structure of Atg1694 is consistent with the annotated start codon being misannotated. Four of the genes where we found an ORF-internal AUG codon is likely to be the true start codon (WSS1, YHL042W, AEP1, and PEX4) had additional (out-of-frame) “ORF” 40S peaks upstream of the newly identified start codon. These peaks were transferred to the 5’UTR 40S peak list as these peaks lie within the newly established 5’UTR of these genes. One ORF AUG 40S peak for the gene NQM1 was removed as it was likely caused by reads that mapped to the neighboring main ORF stop codon (AUGUGA). Combining all the ORF AUG 40S peaks resulted in the identification of 194 NMD-sensitive genes with at least one ORF AUG 40S peak (Supplementary Data 1 Sheet2).To determine the number of NMD-sensitive genes with 5’UTR 40S peaks within the last 100 nt of their 5’UTRs, start codons were identified in all frames within the first 100 nt upstream of the canonical start codon using findseqinstance2_run_wf.py from a custom genome annotation (GFF) file (Supplementary Data 5.zip yeast_5UTRc_toCDS_max2000_longest.gff). Peak heights and pause scores for these start codons were calculated from the rpl11b∆ upf1∆ 40S data using makeposstats_multidensity_wf_run.py. Similar to the previous 5’UTR AUG peak analysis described above, raw peak height was used to detect 40S peaks (92 rpkm). As done with the longer 5’UTR analysis above, three false positive genes that were assigned 5’UTR 40S peaks on upstream unannotated transcripts and tRNAs were removed from the peak list. The three genes whose detected 5’UTR peaks are likely misannotated start codons (ASA1, LST7 and PPT2) were also removed for the list. The four genes (WSS1, YHL042W, AEP1, and PEX4) identified in the ORF 40S peak with misannotated start codons had their upstream 40S peaks added to the list. Combining all the 5’UTR AUG 40S peaks within the first 100 nt of the 5’UTR resulted in the identification of 136 NMD-sensitive genes.Main start codon peak analysis of NMD-sensitive genes with ORF AUG 40S peakslist_wf_run.py was used to calculate main start codon peak height on NMD-sensitive genes from the rpl11b∆ upf1∆ 40S Ribo-seq data. Two previously identified genes (Supplementary Data 4) with misannotated start codons (AYT1 and ATG16) were removed from the merged list. Genes with main start codon peak height of less than 0.137 rpkm were scored as lacking a peak at the main ORF start codon, resulting in the identification of 93 NMD-sensitive genes with ORF 40S peaks expressing ORF-internal transcript isoforms (OTIs). Genes with main start codon peak heights of greater than 0.137 rpkm were scored as having a peak at the main ORF start codon. The two genes with misannotated start codons, AYT1 and ATG16, were added to this list, resulting in the identification of 101 NMD-sensitive genes with ORF 40S peaks expressing inefficient initiation transcripts (IITs).Split-count test for identification of ORF-internal transcript isoforms (OTIs)A modified version of listavg.py – listavgintAUG_splitcounts_2.py – was used to calculate raw read counts for the first 100nt and the last 100 nt of each main ORF from upf1∆ (DY231M, DY232M, and DY233M) RNA-seq data for genes that were sensitive to NMD. Differential expression analysis was performed on reads for the first 100 nt and the last 100 nt of each ORF with DESeq260. The split counts list was then sorted by FoldChange_split and filtered using an FDR (padj_split) cutoff of 5%. This resulted in the identification of 34 genes with a greater than 2-fold increase in reads in the last 100 nt of their ORF relative to the first 100 nt, consistent with them expressing OTIs. We also generated a hand curated list of fifteen additional OTI expressing genes with a greater than 2-fold increase that failed to make the FDR cut-off of 5% (0.05). We combined these two lists to give a final total of 49 OTIs identified using the ORF split gene test. Comparison of the list of OTIs identified with the split-count test with OTIs identified through the absence of a 40S main start codon peak (above) showed strong overlap (Supplementary Fig. S2c). The two lists were therefore combined, generating a final list of 101 OTIs.Identification of NMD-sensitive genes with long 3’UTRs3’UTR lengths were calculated from the 3’UTR annotation published in most_abundant_full-ORF_transcripts_ypd.gff92. The list of 3’UTR lengths was merged with the list of NMD-sensitive genes using R. The combined list was sorted by 3’UTR length to yield 42 NMD-sensitive genes with a 3’UTR length greater than 396 nt. If NMD-sensitivity is caused by inefficient termination due to a long 3’UTR, one expectation is that stabilization of the transcript in an upf1∆ strain should result in a concomitant increase in 80S ribosome footprints. (Translation efficiency should remain unchanged.) list_wf_run.py was used to calculate raw read counts for the coding sequence of genes from WT (DY258F and DY259F) and upf1∆ (DY260F and DY261F) 80S Ribo-seq data. Differential expression analysis was then performed on reads from main ORFs with DESeq260. The results were filtered using an FDR (padj) cutoff of 1% and fold increase >1.5. This list was merged with the list of 42 NMD-sensitive long 3’UTR genes to yield a combined list with 27 NMD-sensitive genes with a 3’UTR length greater than 396 nt and a significant increase in 80S ribosome footprints in the upf1∆ strain (Supplementary Data 1 Sheet11).A previous TIF-Seq study [Supplementary Data 10 therein]50 identified 279 bicistronic transcripts in yeast in the form of two gene lists (referred to as gene1 and gene2 in Supplementary Data 1 Sheet12). We combined the two gene lists and removed duplicate genes, generating a list of 536 genes. Comparison of this list to our list of NMD-sensitive genes resulted in the identification of 89 NMD-sensitive genes that were potentially involved in a pseudo-bicistronic transcript. The TIF-Seq study determined transcript isoforms in WT yeast grown with two different carbon sources: glucose and galactose. The list therefore likely contains bicistronic transcripts from both growth conditions. As all our studies were performed in glucose media, we confirmed the presence of bicistronic isoforms in the TIF-Seq glucose data and checked that reads on the corresponding NMD-sensitive transcript in our upf1∆ RNA-seq data extend across both genes involved in the pseudo-bicistronic transcript. This resulted in the removal of 21 false positives, leaving a list of 68 NMD-sensitive genes involved in 50 pseudo-bicistronic transcripts (Supplementary Data 1 Sheet12).Identification of candidate split pseudogenesCandidates for consideration as split pseudogenes were derived by searching the Saccharomyces Genome Database (SGD) for genes where the description provided information that the gene and adjacent gene constituted a single ORF in closely related Saccharomyces species. Comparison with our list of NMD-sensitive genes identified 22 genes that formed 11 candidate pseudogenes expressing an NMD-sensitive transcript that covers both ORFs (Supplementary Data 1 Sheet14).Analysis of related yeast sequences and sequencing dataFor Fig. 7c, DAL5 orthologs were identified by Blastn (v2.15.0 + )95 using S. cerevisiae DAL5 as a query against seven genomes with parameters -word_size 7 -evalue 1e-2 -outfmt 6, and the top hit in each genome was considered the candidate ortholog. The protein sequences of Dal5 orthologs were then aligned using ClustalW in MEGA1296 with default settings, and a phylogenetic tree was subsequently constructed using the neighbor-joining method97.The mRNA-Seq from related yeast species71 was downloaded from Sequence Read Archive (SRA) with Project ID SRP187756. GFF annotations and genome references were found here: https://figshare.com/articles/dataset/Combination_of_novel_transcripts_from_de_novo_assembly_and_genes_from_reference_annotations_for_11_species/7851521/2For L. kluyveri, the genome was obtained from the Saccharomyces Genome Database using the strain name NRRL Y-12651.Raw bulk RNA-seq data were processed with fastp (v0.24.0)98 to remove adapters, and the clean reads were aligned to the corresponding genomes using STAR (v2.7.11b)99 using default parameters except for outFilterMismatchNmax = 2 and alignEndsType = EndToEnd. BigWig100 files were generated with bamCoverage from deepTools (v3.5.6)101 using the parameters --binSize 10 --normalizeUsing BPM, and the tracks were visualized with IGV102. The codes for processing RNA-seq data can be accessed on GitHub: https://github.com/TriLab-bioinf/GUYDOSH_LAB_TK_207.For Fig. 7d (S. pombe data), RNA-Seq and ribosome profiling data72 were processed the same. Reads were processed with: cutadapt -u 9 -a AAAAAAAA -m 15. Then, reads were aligned to ncRNA to subtract reads that were not ribosomal footprints and then aligned to a spliced transcriptome as described103. Reads up to 50 nt were retained for mRNA-Seq and 25-34 nt for ribosome profiling. We used data files from ArrayExpress: E_pat1_Oh_mRNA and E_pat1_Oh_ribosome-protected RNA from E-MTAB-2179, which is starved for nitrogen. We used S_mRNA_wt_1A and S_ribo_wt_1A from E-MTAB-2176, which is grown in rich media.Northern blottingNorthern blotting was conducted using the DIG Northern Starter Kit (Roche; Cat. No. 12039672910).Growth of cultures and RNA extraction for Northern blotTo determine if the 5’ uORFs identified by 40S ribosome profiling in the extended 5’UTR transcript of DAL5 are responsible for triggering NMD, dal5_ext∆ UPF1 and dal5_ext∆ upf1∆ cells transformed with the YCplac33-DAL5 and YCplac33-DAL5_noAUG plasmids were grown in SC-U media to an OD600 of 1.0. The cells were harvested by being spun down at 4,000 rpm (approximately 3739 max g) for 5 min at 4 °C, the supernatant was discarded, and the cells frozen in liquid nitrogen. Similarly, for DAL7, dal7_ext∆ UPF1 and dal7_ext∆ upf1∆ cells were transformed with the YCplac33-DAL7 and YCplac33-DAL7_noAUG plasmids and were grown and harvested as described above for DAL5. To confirm that the DAL5 and DAL7 genes express two transcript isoforms, a longer and a shorter induced canonical transcript (similar to LUTI 2-transcript systems), the above cells were also grown in poor nitrogen (proline) media to an OD600 of 1.0 and harvested as described above.For the proline induction experiments, cells were grown in rich nitrogen (glutamine) media to an OD600 of 0.5 before the media was changed to poor nitrogen (proline). Samples were taken prior to induction, and at 0 m, 15 m, 30 m, 45 m, and 1 h after induction. Samples were processed as described above. For the glutamine repression experiments, cells were grown in poor nitrogen (proline) media to an OD600 of 0.5 before the media was changed to rich nitrogen (glutamine). Samples were taken prior to repression, and at 0 m, 5 m, 10 m, 20 m, and 1 h after induction. Samples were processed as described above.Cells were resuspended in lysis buffer (8.4 mM EDTA, 60 mM Sodium acetate) and total RNA was purified using hot phenol chloroform.Labeling of Northern probesSingle stranded antisense digoxigenin-labeled RNA probes were generated for DAL5, DAL7, and SCR1 (loading control). To generate DNA templates for probe labeling, the last 500 nt of the DAL5 and DAL7 genes was PCR amplified using primer pairs DAL5_500nt_f/DAL5_probe_r and DAL7_500nt_f/DAL7_probe_r from the YCplac33-DAL5 and YCplac33-DAL7 plasmids. To generate a probe template for SCR1 the last 500 nt of the SCR1 gene was amplified by RT-PCR (SuperScript IV One-Step RT-PCR System with ezDNase; Invitrogen; Cat. No. 12595025) using primer pair SCR1_500nt_f and SCR1_probe_r from 1 μg of yeast total RNA. The reverse primers contain the T7 RNA Polymerase promoter sequence to allow transcription of antisense RNA probes. 200 ng of the purified probe templates were labeled in an in vitro transcription reaction with Digoxigenin-11-UTP according to the manufacturer’s instructions.Formaldehyde gelsFor DAL5 northern blots 1 μg of denatured RNA/sample was loaded onto a 1.2 % agarose formaldehyde gel and run in 1x MOPS buffer at 80 V for 2 h 30 min. For DAL7 northern blots grown under poor nitrogen conditions a 2.5 % agarose was run for 3 h 30 m to ensure separation of the similarly sized 5’-extended and induced transcripts.RNA transfer and fixationRNA was transferred overnight to positively charged nylon membranes (Roche; Cat. No. 11209299001) using a Whatman Nytran SuPerCharge (SPC) TurboBlotter (Millipore Sigma Cat. No. WHA10416300). RNA was fixed to the membrane by UV crosslinking with a VWR UV crosslinker (VWR; Cat. No. 89131-484).Hybridization and stringency washesMembranes were hybridized in a VWR Hybridization Oven Model 5430 (VWR; Cat. No. 97005-252) at 68 °C. Membranes were prehybridized in 10 mL of DIG Easy Hyb (Roche; 11796895001) for 30 min. The DIG-labeled probes (3.5 μL of 50 ng/μL) were denatured for 5 min at 100 °C and then cooled on ice for 1 min before being added to 3.5 mL of DIG Easy Hyb. The membrane was hybridized with the probe/hybridization mixture for 6 h. The membrane was then washed twice in low stringency wash buffer (2x SSC, 0.1% SDS) at RT for 5 min. The membrane was then washed twice in prewarmed high-stringency wash buffer (0.1x SSC, 0.1% SDS) at 68 °C for 15 min.Immunological detectionImmunological detection was carried out using buffers from the DIG Wash and Block Buffer Set (Roche; Cat. No. 11585762001). The membrane was washed in 1x Washing buffer for 5 min and then blocked in 1x Blocking solution for 30 min. The membrane was then incubated in antibody solution (Anti-Digoxigenin-AP [Roche; Cat. No. 11093274910] diluted 1:10,000 in 1x blocking buffer) for 30 min. The membrane was then washed twice in 1x Washing buffer for 15 min and equilibrated in Detection buffer for 5 min. The membrane was incubated with CDP-Star, ready-to-use solution (Roche; Cat. No. 12041677001) for 5 min before being imaged on an Amersham Imager 600.Stripping and reprobing of RNA blotsMembranes were washed in RNase-free water for 5 min and then incubated in Stripping buffer (50 mM Tris-HCl, pH7.5, 5% SDS 50% Formamide) twice for 60 min at 80 °C. The stripped membrane was washed twice in 2 x SSC at RT for 5 min before reprobing with the SCR1 loading control probe.Western blot analysisYeast strains were grown to log-phase before preparation of whole cell extracts (WCEs) from 5 ODs of cells. Protein was prepared from WCEs by TCA extraction. Western analysis was conducted using a rabbit polyclonal antibody against GFP (Abcam; ab6556; 1/2000 dilution) and a mouse monoclonal antibody against beta-actin (Abcam; ab8224; 1/1000 dilution). All westerns were repeated at least twice on two independently grown cultures for each condition.Quantification and statistical analysisFor northern and western blots, at least two individual replicates were performed to verify that the results were reproducible. Northern and western blots were quantified from 16-bit tif files using Fiji (ImageJ 1.54p). Tif files were converted to 8-bit, inverted, and the background was subtracted using Fiji’s inbuilt subtract background tool. Northern and western blot images are from Amersham Imager 600 generated jpg files.Graphs and Spearman R calculations were done with Igor Pro 9 (Wavemetrics). RNA-seq fragment and ribosome footprint data were processed with custom Python scripts that can be accessed through GitHub as described elsewhere. Gene expression analysis performed as described elsewhere by DESeq2. For display of example traces and some analysis, replicate RNA-seq and ribosome profiling datasets were pooled.Reporting summaryFurther information on research design is available in the Nature Portfolio Reporting Summary linked to this article.Data availabilityThe data supporting the findings of this study are available from the corresponding authors upon request. Raw sequencing data (fastq files) and analyzed data (wig files for gene tracks and full DESeq2 ORF-read outputs) have been deposited in the NCBI GEO database under the accession number GSE308938. Some data also analyzed from previously published repositories GSE145904 and GSE108942, as noted. Source data for the figures and Supplementary Figs. are provided as a Source Data file. Gene track files made from published data and raw imager data are available on Mendeley Data at https://doi.org/10.17632/gk934yd37j.1 Source data are provided in this paper.Code availabilityPython files are available as described throughout on these GitHub repositories: https://github.com/TriLab-bioinf/GUYDOSH_LAB_NMD_Manuscript.githttps://github.com/guydoshlab/deduphttps://github.com/guydoshlab/Yeastcode1ReferencesLosson, R. & Lacroute, F. Interference of nonsense mutations with eukaryotic messenger RNA stability. Proc. Natl. Acad. Sci. USA 76, 5134–5137 https://doi.org/10.1073/pnas.76.10.5134 (1979).Article  CAS  PubMed  PubMed Central  Google Scholar Maquat, L. E., Kinniburgh, A. J., Rachmilewitz, E. A. & Ross, J. Unstable beta-globin mRNA in mRNA-deficient beta o thalassemia. Cell 27, 543–553 https://doi.org/10.1016/0092-8674(81)90396-2 (1981).Article  CAS  PubMed  Google Scholar He, F. & Jacobson, A. Nonsense-mediated mRNA decay: degradation of defective transcripts is only part of the story. Annu. Rev. Genet 49, 339–366 https://doi.org/10.1146/annurev-genet-112414-054639 (2015).Article  CAS  PubMed  Google Scholar Leeds, P., Peltz, S. W., Jacobson, A. & Culbertson, M. R. The product of the yeast UPF1 gene is required for rapid turnover of mRNAs containing a premature translational termination codon. Genes Dev. 5, 2303–2314 https://doi.org/10.1101/gad.5.12a.2303 (1991).Article  CAS  PubMed  Google Scholar Leeds, P., Wood, J. M., Lee, B. S. & Culbertson, M. R. Gene products that promote mRNA turnover in Saccharomyces cerevisiae. Mol. Cell Biol. 12, 2165–2177 https://doi.org/10.1128/mcb.12.5.2165-2177.1992 (1992).Article  CAS  PubMed  PubMed Central  Google Scholar He, F. & Jacobson, A. Identification of a novel component of the nonsense-mediated mRNA decay pathway by use of an interacting protein screen. Genes Dev. 9, 437–454 https://doi.org/10.1101/gad.9.4.437 (1995).Article  CAS  PubMed  Google Scholar Cui, Y., Hagan, K. W., Zhang, S. & Peltz, S. W. Identification and characterization of genes that are required for the accelerated degradation of mRNAs containing a premature translational termination codon. Genes Dev. 9, 423–436 https://doi.org/10.1101/gad.9.4.423 (1995).Article  CAS  PubMed  Google Scholar Lee, B. S. & Culbertson, M. R. Identification of an additional gene required for eukaryotic nonsense mRNA turnover. Proc. Natl. Acad. Sci. USA 92, 10354–10358 https://doi.org/10.1073/pnas.92.22.10354 (1995).Article  CAS  PubMed  PubMed Central  Google Scholar Amrani, N. et al. A faux 3’-UTR promotes aberrant termination and triggers nonsense-mediated mRNA decay. Nature 432, 112–118 https://doi.org/10.1038/nature03060 (2004).Article  CAS  PubMed  Google Scholar Wu, C., Roy, B., He, F., Yan, K. & Jacobson, A. Poly(A)-Binding Protein Regulates the Efficiency of Translation Termination. Cell Rep. 33, 108399 https://doi.org/10.1016/j.celrep.2020.108399 (2020).Article  CAS  PubMed  PubMed Central  Google Scholar Meydan, S. & Guydosh, N. R. Disome and trisome profiling reveal genome-wide targets of ribosome quality control. Mol. Cell 79, 588–602 https://doi.org/10.1016/j.molcel.2020.06.010 (2020).Article  CAS  PubMed  PubMed Central  Google Scholar Mangkalaphiban, K. et al. Transcriptome-wide investigation of stop codon readthrough in Saccharomyces cerevisiae. PLoS Genet. 17, e1009538 https://doi.org/10.1371/journal.pgen.1009538 (2021).Article  CAS  PubMed  PubMed Central  Google Scholar Bharti, N. et al. Translation velocity determines the efficacy of engineered suppressor tRNAs on pathogenic nonsense mutations. Nat. Commun. 15, 2957 https://doi.org/10.1038/s41467-024-47258-9 (2024).Article  CAS  PubMed  PubMed Central  Google Scholar Kolakada, D. et al. Systematic analysis of nonsense variants uncovers peptide release rate as a novel modifier of nonsense-mediated mRNA decay. Cell Genom. 5, 100882 https://doi.org/10.1016/j.xgen.2025.100882 (2025).Article  CAS  PubMed  PubMed Central  Google Scholar Muhlrad, D. & Parker, R. Aberrant mRNAs with extended 3’ UTRs are substrates for rapid degradation by mRNA surveillance. RNA 5, 1299–1307 https://doi.org/10.1017/s1355838299990829 (1999).Article  CAS  PubMed  PubMed Central  Google Scholar Kebaara, B. W. & Atkin, A. L. Long 3’-UTRs target wild-type mRNAs for nonsense-mediated mRNA decay in Saccharomyces cerevisiae. Nucleic Acids Res. 37, 2771–2778 https://doi.org/10.1093/nar/gkp146 (2009).Article  CAS  PubMed  PubMed Central  Google Scholar Decourty, L. et al. Long open reading frame transcripts escape nonsense-mediated mRNA decay in yeast. Cell Rep. 6, 593–598 https://doi.org/10.1016/j.celrep.2014.01.025 (2014).Article  CAS  PubMed  Google Scholar Andjus, S. et al. Pervasive translation of Xrn1-sensitive unstable long noncoding RNAs in yeast. RNA 30, 662–679 https://doi.org/10.1261/rna.079903.123 (2024).Article  CAS  PubMed  PubMed Central  Google Scholar Singh, G., Rebbapragada, I. & Lykke-Andersen, J. A competition between stimulators and antagonists of Upf complex recruitment governs human nonsense-mediated mRNA decay. PLoS Biol. 6, e111 https://doi.org/10.1371/journal.pbio.0060111 (2008).Article  CAS  PubMed  PubMed Central  Google Scholar Hogg, J. R. & Goff, S. P. Upf1 senses 3’UTR length to potentiate mRNA decay. Cell 143, 379–389 https://doi.org/10.1016/j.cell.2010.10.005 (2010).Article  CAS  PubMed  PubMed Central  Google Scholar Le Hir, H., Gatfield, D., Izaurralde, E. & Moore, M. J. The exon-exon junction complex provides a binding platform for factors involved in mRNA export and nonsense-mediated mRNA decay. EMBO J. 20, 4987–4997 https://doi.org/10.1093/emboj/20.17.4987 (2001).Article  PubMed  PubMed Central  Google Scholar Fritz, S. E., Ranganathan, S., Wang, C. D. & Hogg, J. R. The RNA-binding protein PTBP1 promotes ATPase-dependent dissociation of the RNA helicase UPF1 to protect transcripts from nonsense-mediated mRNA decay. J. Biol. Chem. 295, 11613–11625 https://doi.org/10.1074/jbc.RA120.013824 (2020).Article  CAS  PubMed  PubMed Central  Google Scholar Kishor, A., Ge, Z. & Hogg, J. R. hnRNP L-dependent protection of normal mRNAs from NMD subverts quality control in B cell lymphoma. EMBO J. 38, https://doi.org/10.15252/embj.201899128 (2019).Munoz, O., Lore, M. & Jagannathan, S. The long and short of EJC-independent nonsense-mediated RNA decay. Biochem Soc. Trans. 51, 1121–1129 https://doi.org/10.1042/BST20221131 (2023).Article  CAS  PubMed  PubMed Central  Google Scholar Peltz, S. W., Brown, A. H. & Jacobson, A. mRNA destabilization triggered by premature translational termination depends on at least three cis-acting sequence elements and one trans-acting factor. Genes Dev. 7, 1737–1754 https://doi.org/10.1101/gad.7.9.1737 (1993).Article  CAS  PubMed  Google Scholar He, F. et al. Genome-wide analysis of mRNAs regulated by the nonsense-mediated and 5’ to 3’ mRNA decay pathways in yeast. Mol. Cell 12, 1439–1452 https://doi.org/10.1016/s1097-2765(03)00446-5 (2003).Article  CAS  PubMed  Google Scholar He, F., Peltz, S. W., Donahue, J. L., Rosbash, M. & Jacobson, A. Stabilization and ribosome association of unspliced pre-mRNAs in a yeast upf1- mutant. Proc. Natl. Acad. Sci. USA 90, 7034–7038 https://doi.org/10.1073/pnas.90.15.7034 (1993).Article  CAS  PubMed  Google Scholar Oliveira, C. C. & McCarthy, J. E. The relationship between eukaryotic translation and mRNA stability. A short upstream open reading frame strongly inhibits translational initiation and greatly accelerates mRNA degradation in the yeast Saccharomyces cerevisiae. J. Biol. Chem. 270, 8936–8943 https://doi.org/10.1074/jbc.270.15.8936 (1995).Article  CAS  PubMed  Google Scholar Ruiz-Echevarria, M. J. & Peltz, S. W. The RNA binding protein Pub1 modulates the stability of transcripts containing upstream open reading frames. Cell 101, 741–751 https://doi.org/10.1016/s0092-8674(00)80886-7 (2000).Article  CAS  PubMed  Google Scholar Gaba, A., Jacobson, A. & Sachs, M. S. Ribosome occupancy of the yeast CPA1 upstream open reading frame termination codon modulates nonsense-mediated mRNA decay. Mol. Cell 20, 449–460 https://doi.org/10.1016/j.molcel.2005.09.019 (2005).Article  CAS  PubMed  Google Scholar Arribere, J. A. & Gilbert, W. V. Roles for transcript leaders in translation and mRNA decay revealed by transcript leader sequencing. Genome Res 23, 977–987 https://doi.org/10.1101/gr.150342.112 (2013).Article  CAS  PubMed  PubMed Central  Google Scholar May, G. E. et al. Unraveling the influences of sequence and position on yeast uORF activity using massively parallel reporter systems and machine learning. Elife 12, https://doi.org/10.7554/eLife.69611 (2023).Peccarelli, M., Scott, T. D., Steele, M. & Kebaara, B. W. mRNAs involved in copper homeostasis are regulated by the nonsense-mediated mRNA decay pathway depending on environmental conditions. Fungal Genet. Biol. 86, 81–90 https://doi.org/10.1016/j.fgb.2015.12.011 (2016).Article  CAS  PubMed  Google Scholar Gouhier, T., Saveanu, C. & Badis, G. Nonsense-mediated mRNA decay factors target short poly(A)-tailed mRNAs lacking a premature termination codon. Nat. Commun. 17, 5333 https://doi.org/10.1038/s41467-026-72132-1 (2026).Pacheco, M. et al. Genetic screens in Saccharomyces cerevisiae identify a role for 40S ribosome recycling factors Tma20 and Tma22 in nonsense-mediated decay. G3 14, https://doi.org/10.1093/g3journal/jkad295 (2024).Cortazar, M. A., Schmidt, J., Egab, I., Coban-Akdemir, Z. & Jagannathan, S. Genomic stop codon scanning reveals quantitative principles of nonsense-mediated mRNA decay. Preprint athttps://doi.org/10.64898/2025.12.20.695734 (2025).Mabin, J. W. et al. Uncovering the isoform-resolution kinetic landscape of nonsense-mediated mRNA decay with EZbakR. Preprint at https://doi.org/10.1101/2025.03.12.642874 (2025).Xu, Z. et al. Bidirectional promoters generate pervasive transcription in yeast. Nature 457, 1033–1037 https://doi.org/10.1038/nature07728 (2009).Article  CAS  PubMed  PubMed Central  Google Scholar van Dijk, E. L. et al. XUTs are a class of Xrn1-sensitive antisense regulatory non-coding RNA in yeast. Nature 475, 114–117 https://doi.org/10.1038/nature10118 (2011).Article  CAS  PubMed  Google Scholar Wery, M. et al. Nonsense-mediated decay restricts LncRNA lvels in yeast unless blocked by double-stranded RNA structure. Mol. Cell 61, 379–392 https://doi.org/10.1016/j.molcel.2015.12.020 (2016).Article  CAS  PubMed  PubMed Central  Google Scholar Geisler, S., Lojek, L., Khalil, A. M., Baker, K. E. & Coller, J. Decapping of long noncoding RNAs regulates inducible genes. Mol. Cell 45, 279–291 https://doi.org/10.1016/j.molcel.2011.11.025 (2012).Article  CAS  PubMed  PubMed Central  Google Scholar Malabat, C., Feuerbach, F., Ma, L., Saveanu, C. & Jacquier, A. Quality control of transcription start site selection by nonsense-mediated-mRNA decay. Elife 4, https://doi.org/10.7554/elife.06722 (2015).Smith, J. E. et al. Translation of small open reading frames within unannotated RNA transcripts in Saccharomyces cerevisiae. Cell Rep. 7, 1858–1866 https://doi.org/10.1016/j.celrep.2014.05.023 (2014).Article  CAS  PubMed  PubMed Central  Google Scholar Celik, A., Baker, R., He, F. & Jacobson, A. High-resolution profiling of NMD targets in yeast reveals translational fidelity as a basis for substrate selection. RNA 23, 735–748 https://doi.org/10.1261/rna.060541.116 (2017).Article  CAS  PubMed  PubMed Central  Google Scholar Johansson, M. J., He, F., Spatrick, P., Li, C. & Jacobson, A. Association of yeast Upf1p with direct substrates of the NMD pathway. Proc. Natl. Acad. Sci. USA 104, 20872–20877 https://doi.org/10.1073/pnas.0709257105 (2007).Article  PubMed  PubMed Central  Google Scholar Miura, F. et al. A large-scale full-length cDNA analysis to explore the budding yeast transcriptome. Proc. Natl. Acad. Sci. USA 103, 17846–17851 https://doi.org/10.1073/pnas.0605645103 (2006).Article  CAS  PubMed  PubMed Central  Google Scholar Zhang, Z. & Dietrich, F. S. Mapping of transcription start sites in Saccharomyces cerevisiae using 5’ SAGE. Nucleic Acids Res. 33, 2838–2851 https://doi.org/10.1093/nar/gki583 (2005).Article  CAS  PubMed  PubMed Central  Google Scholar Toesca, I., Nery, C. R., Fernandez, C. F., Sayani, S. & Chanfreau, G. F. Cryptic transcription mediates repression of subtelomeric metal homeostasis genes. PLoS Genet. 7, e1002163 https://doi.org/10.1371/journal.pgen.1002163 (2011).Article  CAS  PubMed  PubMed Central  Google Scholar Nagalakshmi, U. et al. The transcriptional landscape of the yeast genome defined by RNA sequencing. Science 320, 1344–1349 https://doi.org/10.1126/science.1158441 (2008).Article  CAS  PubMed  PubMed Central  Google Scholar Pelechano, V., Wei, W. & Steinmetz, L. M. Extensive transcriptional heterogeneity revealed by isoform profiling. Nature 497, 127–131 https://doi.org/10.1038/nature12121 (2013).Article  CAS  PubMed  PubMed Central  Google Scholar Chen, J. et al. Kinetochore inactivation by expression of a repressive mRNA. Elife 6,https://doi.org/10.7554/elife.27417 (2017).Chia, M. et al. Transcription of a 5’ extended mRNA isoform directs dynamic chromatin changes and interference of a downstream promoter. Elife 6,https://doi.org/10.7554/eLife.27420 (2017).Cheng, Z. et al. Pervasive, coordinated protein-level changes driven by transcript isoform switching during meiosis. Cell 172, 910–923 https://doi.org/10.1016/j.cell.2018.01.035 (2018).Article  CAS  PubMed  PubMed Central  Google Scholar Van Dalfsen, K. M. et al. Global proteome remodeling during ER stress involves Hac1-driven expression of long undecoded transcript isoforms. Dev. Cell 46, 219–235 https://doi.org/10.1016/j.devcel.2018.06.016 (2018).Article  CAS  PubMed  PubMed Central  Google Scholar Tresenrider, A. et al. Integrated genomic analysis reveals key features of long undecoded transcript isoform-based gene repression. Mol. Cell 81, 2231–2245 https://doi.org/10.1016/j.molcel.2021.03.013 (2021).Article  CAS  PubMed  PubMed Central  Google Scholar Morse, K., Bishop, A. L., Swerdlow, S., Leslie, J. M. & Unal, E. Swi/Snf chromatin remodeling regulates transcriptional interference and gene repression. Mol. Cell 84, 3080–3097 https://doi.org/10.1016/j.molcel.2024.06.029 (2024).Martens, J. A., Laprade, L. & Winston, F. Intergenic transcription is required to repress the Saccharomyces cerevisiae SER3 gene. Nature 429, 571–574 https://doi.org/10.1038/nature02538 (2004).Article  CAS  PubMed  Google Scholar Houseley, J., Rubbi, L., Grunstein, M., Tollervey, D. & Vogelauer, M. A ncRNA modulates histone modification and mRNA induction in the yeast GAL gene cluster. Mol. Cell 32, 685–695 https://doi.org/10.1016/j.molcel.2008.09.027 (2008).Article  CAS  PubMed  PubMed Central  Google Scholar Thompson, D. M. & Parker, R. Cytoplasmic decay of intergenic transcripts in Saccharomyces cerevisiae. Mol. Cell Biol. 27, 92–101 https://doi.org/10.1128/MCB.01023-06 (2007).Article  CAS  PubMed  Google Scholar Love, M. I., Huber, W. & Anders, S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol. 15, 550 https://doi.org/10.1186/s13059-014-0550-8 (2014).Article  CAS  PubMed  PubMed Central  Google Scholar Young, D. J., Meydan, S. & Guydosh, N. R. 40S ribosome profiling reveals distinct roles for Tma20/Tma22 (MCT-1/DENR) and Tma64 (eIF2D) in 40S subunit recycling. Nat. Commun. 12, 2976 https://doi.org/10.1038/s41467-021-23223-8 (2021).Article  CAS  PubMed  PubMed Central  Google Scholar Spealman, P. et al. Conserved non-AUG uORFs revealed by a novel regression analysis of ribosome profiling data. Genome Res. 28, 214–222 https://doi.org/10.1101/gr.221507.117 (2018).Article  CAS  PubMed  Google Scholar ter Schure, E. G., van Riel, N. A. & Verrips, C. T. The role of ammonia metabolism in nitrogen catabolite repression in Saccharomyces cerevisiae. FEMS Microbiol. Rev. 24, 67–83 https://doi.org/10.1111/j.1574-6976.2000.tb00533.x (2000).Article  PubMed  Google Scholar Rai, R., Genbauffe, F. S. & Cooper, T. G. Structure and transcription of the allantoate permease gene (DAL5) from Saccharomyces cerevisiae. J. Bacteriol. 170, 266–271 https://doi.org/10.1128/jb.170.1.266-271.1988 (1988).Article  CAS  PubMed  PubMed Central  Google Scholar Palanimurugan, R., Scheel, H., Hofmann, K. & Dohmen, R. J. Polyamines regulate their synthesis by inducing expression and blocking degradation of ODC antizyme. EMBO J. 23, 4857–4867 https://doi.org/10.1038/sj.emboj.7600473 (2004).Article  CAS  PubMed  PubMed Central  Google Scholar Farabaugh, P. J., Kramer, E., Vallabhaneni, H. & Raman, A. Evolution of +1 programmed frameshifting signals and frameshift-regulating tRNAs in the order Saccharomycetales. J. Mol. Evol. 63, 545–561 https://doi.org/10.1007/s00239-005-0311-0 (2006).Article  CAS  PubMed  Google Scholar Porrua, O. et al. In vivo SELEX reveals novel sequence and structural determinants of Nrd1-Nab3-Sen1-dependent transcription termination. EMBO J. 31, 3935–3948 https://doi.org/10.1038/emboj.2012.237 (2012).Article  CAS  PubMed  PubMed Central  Google Scholar Kim, T., Xu, Z., Clauder-Munster, S., Steinmetz, L. M. & Buratowski, S. Set3 HDAC mediates effects of overlapping noncoding transcription on gene induction kinetics. Cell 150, 1158–1169 https://doi.org/10.1016/j.cell.2012.08.016 (2012).Article  CAS  PubMed  PubMed Central  Google Scholar Kim, J. H. et al. Modulation of mRNA and lncRNA expression dynamics by the Set2-Rpd3S pathway. Nat. Commun. 7, 13534 https://doi.org/10.1038/ncomms13534 (2016).Article  CAS  PubMed  PubMed Central  Google Scholar Poonia, P. et al. Yeast poly(A)-binding protein (Pab1) controls translation initiation in vivo primarily by blocking mRNA decapping and decay. Nucleic Acids Res. 53, https://doi.org/10.1093/nar/gkaf143 (2025).Blevins, W. R., Carey, L. B. & Alba, M. M. Transcriptomics data of 11 species of yeast identically grown in rich media and oxidative stress conditions. BMC Res. Notes 12, 250 https://doi.org/10.1186/s13104-019-4286-0 (2019).Article  PubMed  PubMed Central  Google Scholar Duncan, C. D. & Mata, J. The translational landscape of fission-yeast meiosis and sporulation. Nat. Struct. Mol. Biol. 21, 641–647 https://doi.org/10.1038/nsmb.2843 (2014).Article  CAS  PubMed  PubMed Central  Google Scholar Luke, B. et al. Saccharomyces cerevisiae Ebs1p is a putative ortholog of human Smg7 and promotes nonsense-mediated mRNA decay. Nucleic Acids Res. 35, 7688–7697 https://doi.org/10.1093/nar/gkm912 (2007).Article  CAS  PubMed  PubMed Central  Google Scholar Yepiskoposyan, H., Aeschimann, F., Nilsson, D., Okoniewski, M. & Muhlemann, O. Autoregulation of the nonsense-mediated mRNA decay pathway in human cells. RNA 17, 2108–2118 https://doi.org/10.1261/rna.030247.111 (2011).Article  CAS  PubMed  PubMed Central  Google Scholar Dehecq, M. et al. Nonsense-mediated mRNA decay involves two distinct Upf1-bound complexes. EMBO J. 37, https://doi.org/10.15252/embj.201899278 (2018).Goetz, A. E. & Wilkinson, M. Stress and the nonsense-mediated RNA decay pathway. Cell Mol. Life Sci. 74, 3509–3531 https://doi.org/10.1007/s00018-017-2537-6 (2017).Article  CAS  PubMed  PubMed Central  Google Scholar Cai, H., Hauser, M., Naider, F. & Becker, J. M. Differential regulation and substrate preferences in two peptide transporters of Saccharomyces cerevisiae. Eukaryot. Cell 6, 1805–1813 https://doi.org/10.1128/EC.00257-06 (2007).Article  CAS  PubMed  PubMed Central  Google Scholar Mort, M., Ivanov, D., Cooper, D. N. & Chuzhanova, N. A. A meta-analysis of nonsense mutations causing human genetic disease. Hum. Mutat. 29, 1037–1047 https://doi.org/10.1002/humu.20763 (2008).Article  CAS  PubMed  Google Scholar Mangkalaphiban, K. et al. Extended stop codon context predicts nonsense codon readthrough efficiency in human cells. Nat. Commun. 15, 2486 https://doi.org/10.1038/s41467-024-46703-z (2024).Article  CAS  PubMed  PubMed Central  Google Scholar Nasif, S. et al. Inhibition of nonsense-mediated mRNA decay reduces the tumorigenicity of human fibrosarcoma cells. NAR Cancer 5, zcad048 https://doi.org/10.1093/narcan/zcad048 (2023).Article  PubMed Central  Google Scholar Pierce, S. E. et al. Prime editing-installed suppressor tRNAs for disease-agnostic genome editing. Nature 648, 191–202 https://doi.org/10.1038/s41586-025-09732-2 (2025).Article  CAS  PubMed  PubMed Central  Google Scholar Albers, S. et al. Engineered tRNAs suppress nonsense mutations in cells and in vivo. Nature 618, 842–848 https://doi.org/10.1038/s41586-023-06133-1 (2023).Article  CAS  PubMed  PubMed Central  Google Scholar Coelho, J. P. L., Yip, M. C. J., Oltion, K., Taunton, J. & Shao, S. The eRF1 degrader SRI-41315 acts as a molecular glue at the ribosomal decoding center. Nat. Chem. Biol. 20, 877–884 https://doi.org/10.1038/s41589-023-01521-0 (2024).Article  CAS  PubMed  PubMed Central  Google Scholar Coller, J. & Ignatova, Z. tRNA therapeutics for genetic diseases. Nat. Rev. Drug Discov. 23, 108–125 https://doi.org/10.1038/s41573-023-00829-9 (2024).Article  CAS  PubMed  Google Scholar Gurzeler, L. A. et al. Drug-induced eRF1 degradation promotes readthrough and reveals a new branch of ribosome quality control. Cell Rep. 42, 113056 https://doi.org/10.1016/j.celrep.2023.113056 (2023).Article  CAS  PubMed  Google Scholar Oltion, K. et al. An E3 ligase network engages GCN1 to promote the degradation of translation factors on stalled ribosomes. Cell 186, 346–362 https://doi.org/10.1016/j.cell.2022.12.025 (2023).Article  CAS  PubMed  PubMed Central  Google Scholar Peltz, S. W., Morsy, M., Welch, E. M. & Jacobson, A. Ataluren as an agent for therapeutic nonsense suppression. Annu Rev. Med. 64, 407–425 https://doi.org/10.1146/annurev-med-120611-144851 (2013).Article  CAS  PubMed  Google Scholar Sharma, J. et al. A small molecule that induces translational readthrough of CFTR nonsense mutations by eRF1 depletion. Nat. Commun. 12, 4358 https://doi.org/10.1038/s41467-021-24575-x (2021).Article  CAS  PubMed  PubMed Central  Google Scholar McGlincy, N. J. & Ingolia, N. T. Transcriptome-wide measurement of translation by ribosome profiling. Methods 126, 112–129 https://doi.org/10.1016/j.ymeth.2017.05.028 (2017).Article  CAS  PubMed  PubMed Central  Google Scholar Langmead, B., Trapnell, C., Pop, M. & Salzberg, S. L. Ultrafast and memory-efficient alignment of short DNA sequences to the human genome. Genome Biol. 10, R25 https://doi.org/10.1186/gb-2009-10-3-r25 (2009).Article  CAS  PubMed  PubMed Central  Google Scholar Cherry, J. M. et al. Saccharomyces Genome Database: the genomics resource of budding yeast. Nucleic Acids Res. 40, D700–D705 https://doi.org/10.1093/nar/gkr1029 (2012).Article  CAS  PubMed  Google Scholar Ng, P. C. et al. Transcriptome visualization and data availability at the Saccharomyces Genome Database. Nucleic Acids Res. 48, D743–D748 https://doi.org/10.1093/nar/gkz892 (2020).Article  CAS  PubMed  PubMed Central  Google Scholar Archer, S. K., Shirokikh, N. E., Beilharz, T. H. & Preiss, T. Dynamics of ribosome scanning and recycling revealed by translation complex profiling. Nature 535, 570–574 https://doi.org/10.1038/nature18647 (2016).Article  CAS  PubMed  Google Scholar Fujioka, Y., Noda, N. N., Nakatogawa, H., Ohsumi, Y. & Inagaki, F. Dimeric coiled-coil structure of Saccharomyces cerevisiae Atg16 and its functional significance in autophagy. J. Biol. Chem. 285, 1508–1515 https://doi.org/10.1074/jbc.M109.053520 (2010).Article  CAS  PubMed  Google Scholar Camacho, C. et al. BLAST+: architecture and applications. BMC Bioinforma. 10, 421 https://doi.org/10.1186/1471-2105-10-421 (2009).Article  CAS  Google Scholar Kumar, S. et al. MEGA12: Molecular evolutionary genetic Analysis Version 12 for Adaptive and Green Computing. Mol. Biol. Evol. 41, https://doi.org/10.15252/embj.201899278 (2024)Saitou, N. & Nei, M. The neighbor-joining method: a new method for reconstructing phylogenetic trees. Mol. Biol. Evol. 4, 406–425 https://doi.org/10.1093/oxfordjournals.molbev.a040454 (1987).Article  CAS  PubMed  Google Scholar Chen, S., Zhou, Y., Chen, Y. & Gu, J. fastp: an ultra-fast all-in-one FASTQ preprocessor. Bioinformatics 34, i884–i890 https://doi.org/10.1093/bioinformatics/bty560 (2018).Article  CAS  PubMed  PubMed Central  Google Scholar Dobin, A. et al. STAR: ultrafast universal RNA-seq aligner. Bioinformatics 29, 15–21 https://doi.org/10.1093/bioinformatics/bts635 (2013).Article  CAS  PubMed  Google Scholar Kent, W. J., Zweig, A. S., Barber, G., Hinrichs, A. S. & Karolchik, D. BigWig and BigBed: enabling browsing of large distributed datasets. Bioinformatics 26, 2204–2207 https://doi.org/10.1093/bioinformatics/btq351 (2010).Article  CAS  PubMed  PubMed Central  Google Scholar Ramirez, F., Dundar, F., Diehl, S., Gruning, B. A. & Manke, T. deepTools: a flexible platform for exploring deep-sequencing data. Nucleic Acids Res. 42, W187–W191 https://doi.org/10.1093/nar/gku365 (2014).Article  CAS  PubMed  PubMed Central  Google Scholar Robinson, J. T. et al. Integrative genomics viewer. Nat. Biotechnol. 29, 24–26 https://doi.org/10.1038/nbt.1754 (2011).Article  CAS  PubMed  PubMed Central  Google Scholar Guydosh, N. R., Kimmig, P., Walter, P. & Green, R. Regulated Ire1-dependent mRNA decay requires no-go mRNA degradation to maintain endoplasmic reticulum homeostasis in S. pombe. Elife 6, https://doi.org/10.7554/eLife.29216 (2017).Rutherford, K. M., Lera-Ramirez, M. & Wood, V. PomBase: a global core biodata resource-growth, collaboration, and sustainability. Genetics 227, https://doi.org/10.1093/genetics/iyae007 (2024).Download referencesAcknowledgementsWe thank Alan Hinnebusch, Tom Dever, Jon Lorsch, and Guydosh lab members for useful feedback during this project. We thank Alan Hinnebusch for providing the transcription terminator construct and wig files from published CAGE-Seq data. We thank Elçin Ünal and Gloria Brar for helpful discussions about LUTIs. We thank Hernan Lorenzi for helpful discussions about computational analysis. We thank Corey Knowles for feedback on the manuscript.FundingThis research was supported by the Intramural Research Program of the NIH, the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) (DK075132 to N.R.G.).This research was supported by the Intramural Research Program of the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) within the National Institutes of Health (NIH). The contributions of the NIH author(s) were made as part of their official duties as NIH federal employees, are in compliance with agency policy requirements, and are considered Works of the United States Government. However, the findings and conclusions presented in this paper are those of the author(s) and do not necessarily reflect the views of the NIH or the U.S. Department of Health and Human Services.Author informationAuthors and AffiliationsLaboratory of Biochemistry and Genetics, National Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of Health, Bethesda, MD, USADavid J. Young, Yuejun Wang & Nicholas R. GuydoshTriLab Bioinformatics Group, National Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of Health, Bethesda, MD, USAYuejun WangAuthorsDavid J. YoungView author publicationsSearch author on:PubMed Google ScholarYuejun WangView author publicationsSearch author on:PubMed Google ScholarNicholas R. GuydoshView author publicationsSearch author on:PubMed Google ScholarContributionsD.J.Y. designed experiments, performed high-throughput sequencing and blotting experiments, analyzed the data, and wrote the paper. Y.W. analyzed RNA-Seq data from related yeast species. N.R.G. designed experiments, analyzed data, and wrote the paper.Corresponding authorCorrespondence to Nicholas R. Guydosh.Ethics declarationsCompeting interestsThe authors declare no competing interests.Peer reviewPeer review informationNature Communications thanks the anonymous reviewers for their contribution to the peer review of this work. A peer review file is available.Additional informationPublisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.Supplementary informationSupplementary information (download PDF )Supplementary Data 1 (download XLSX )Supplementary Data 2 (download XLSX )Supplementary Data 3 (download XLSX )Supplementary Data 4 (download XLSX )Supplementary Data 5 (download ZIP )Reporting Summary (download PDF )Transparent Peer Review file (download PDF )Source dataSource Data (download XLSX )Rights and permissionsOpen Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.Reprints and permissionsAbout this article