We had previously observed upregulation of the Dbp1 RNA helicase in meiotic yeast cells, coincident with downregulation of its paralog, Ded1 (Brar et al., 2012). Ded1 is important for translation initiation, and has been well studied under conditions of mitotic exponential growth (de la Cruz et al., 1997). In contrast, the role of Dbp1 has remained less clear, partially a result of its absent or low expression under commonly studied laboratory conditions. Based on the identity of Dbp1 as an RNA helicase (Jamieson and Beggs, 1991), we predicted that Dbp1 could have a role in regulating gene expression during meiosis. To test this, we deleted DBP1 by replacing the ORF (from start to stop codon) with a standard cassette-encoding Geneticin (G418) resistance, amplified from the pFA6a-kanMX6 plasmid (Longtine et al., 1998), in the SK1 budding yeast strain background (Padmore et al., 1991). We then compared gene expression profiles of wild-type and dbp1Δ::kanMX6 cells during meiosis, using ribosome profiling and mRNA-sequencing (mRNA-seq). To our surprise, MRP51, the gene located directly adjacent to DBP1, exhibited a profound decrease in translation in the dbp1Δ::kanMX6 strain relative to wild-type controls (Figure 1A). Translation of MRP51 was 8.4-fold lower in dbp1Δ::kanMX6 cells despite a 1.8-fold increase in MRP51 mRNA abundance (Figure 1A,B). These changes reflected a 15-fold decrease in MRP51 translation efficiency (TE) in the dbp1Δ::kanMX6 cells compared to the wild-type control (Figure 1C; TE: ribosome footprint RPKM/mRNA RPKM) and led to a reduction in Mrp51 protein level compared to wild-type, as assessed by western blotting (Figure 1D). Because of the close genomic proximity of these genes, we hypothesized that this change in TE could be due to cis- effects from the dbp1Δ::kanMX6 insertion rather than reflecting regulation of MRP51 by Dbp1 protein. Upon closer examination of the mRNA transcripts produced at this locus, we found that replacing the DBP1 ORF with the kanMX6 resistance cassette led to production of a 5′ extended MRP51 mRNA compared to that in wild-type cells. The extended transcript contained 3 AUG-initiated uORFs not present on the wild-type transcript, which were translated at the expense of the MRP51 ORF in dbp1Δ::kanMX6 cells (Figure 1E).

These data suggest that kanMX6 cassette insertion at the DBP1 locus drives the transcription of an aberrant 5′ extended MRP51 mRNA containing repressive uORFs and causes reduced Mrp51 protein production. Thus, the features of MRP51 mis-regulation in this strain mimicked the natural LUTI-based mode of regulation that conditionally downregulates protein synthesis from many genes in yeast (Figure 1F; Chia et al., 2017; Chen et al., 2017; Cheng et al., 2018; Van Dalfsen et al., 2018). Here, the primary difference was that repression of MRP51 expression appeared to occur constitutively as result of kanMX6 insertion at the neighboring DBP1 ORF, rather than resulting from temporally regulated toggling between two TSSs.

We were surprised to observe this dramatic off-target mis-regulation because of the widespread and long-term use of cassette-mediated gene deletion and the detailed quality control measures that were used in construction of our strains to prevent off-target effects caused by background mutations or improper cassette insertion. To confirm aberrant MRP51 regulation was not an artifact of our strain, experimental conditions, or selection cassette, we analyzed a published ribosome profiling and mRNA-seq dataset comparing wild-type and dbp1Δ cells of the S288C budding yeast background. These dbp1Δ cells were generated by replacement of the DBP1 ORF with a cassette-encoding resistance to Hygromycin (dbp1Δ::hphMX4) (Sen et al., 2019). Consistent with our data, insertion of hphMX4 to replace the DBP1 ORF caused mis-regulation of the adjacent gene MRP51. Despite levels of the Mrp51-encoding transcript remaining similar to wild-type in dbp1Δ::hphMX4 cells, translation of MRP51 was decreased 4.4-fold in these conditions (Figure 1—figure supplement 1B,C). The 3.7-fold decrease in TE of MRP51 in dbp1Δ::hphMX4 cells in this study led to the interpretation that the canonical MRP51 transcript is highly dependent on Dbp1 for its translation (Figure 1—figure supplement 1D; Sen et al., 2019). A closer look at the transcripts produced from this locus, however, revealed the presence of a 5′-extended MRP51 LUTI in the dbp1Δ::hphMX4 cells, like the one we had observed. This 5′ extended MRP51 transcript contained competitive AUG-initiated uORFs that were translated in place of the MRP51 ORF, as in our experiments (Figure 1—figure supplement 1E). These findings confirmed that the off-target effects seen in dbp1∆::kanMX6 mutants were not an artifact of our strain background, selection cassette, or experimental conditions.

We next sought to determine whether the mutant phenotypes we observed in dbp1Δ::kanMX6 cells were due to loss of DBP1 function or resistance cassette-dependent mis-regulation of MRP51, a gene encoding a mitochondrial small subunit ribosome protein (mt-SSU) (Green-Willms et al., 1998). Polysome analysis of dbp1Δ::kanMX6 cells during meiosis demonstrated that overall translation rates were diminished relative to wild-type cells, as determined by lower polysome peaks in the mutant (Figure 2A; right, fractions 5 and 6). Consistent with this finding, measurement of radioactive amino acid incorporation rates revealed a 24% decrease in bulk translation in dbp1Δ::kanMX6 cells (Figure 2—figure supplement 1A). Polysome traces also indicated differences in the accumulation of a ribonucleoprotein (RNP) species roughly the size of the cytoplasmic large 60S subunit (Figure 2A; LSU, fraction 3) and a decrease in cytoplasmic small 40S subunit signal (Figure 2A; SSU, fraction 2). We performed label-free mass spectrometry on fractionated wild-type and dbp1Δ::kanMX6 polysomes and used hierarchical clustering of the data to assess global differences in polysome composition. We found that a prominent cluster of proteins enriched in fraction 3 of polysomes from dbp1Δ::kanMX6 cells compared to wild-type was highly enriched for the mitochondrial large ribosome subunit (mt-LSU). This cluster contained 34 of the 42 proteinaceous mt-LSU components quantified in our study (Figure 2A; left). Further analysis of fraction 3 revealed that every mitochondrial 54S protein quantified was enriched in dbp1Δ::kanMX6 cells compared to the wild-type control (Figure 2—figure supplement 1C). These data suggest the accumulated RNP observed in dbp1Δ::kanMX6 cells is the mt-LSU. A previous study from our laboratory observed that when individual cytoplasmic SSU proteins are lost, free cytoplasmic 60S LSUs accumulate, presumably a result of their inability to find an SSU to complex with and form fully assembled 80S species (Cheng et al., 2019). The build-up of free mt-LSU in dbp1Δ::kanMX6 cells, which express lower amounts of mitochondrial small subunit (mt-SSU) component Mrp51 is reminiscent of that effect. Furthermore, we identified a small cluster of cytoplasmically translated mitochondrial proteins that were decreased in dbp1Δ::kanMX6 high polysome fractions compared to wild-type, hinting at effects of broader mitochondrial dysfunction (Figure 2A). These data suggested the cassette-mediated mis-regulation of MRP51 in dbp1Δ::kanMX6 cells may have been responsible for the mutant phenotypes we observed, rather than loss of Dbp1 protein.

Disruption of mitochondrial translation, and ultimately function, would be expected to lead to cellular fitness defects. Based on the apparent mitochondrial defects observed in the mass spectrometry data, we hypothesized that dbp1Δ::kanMX6 cells would grow poorly in conditions in which elevated mitochondrial function is required. Consistently, dbp1Δ::kanMX6 cells exhibited severe growth defects in media containing only non-fermentable carbon sources, such as glycerol or acetate (Figure 2—figure supplement 1D). Additionally, growth defects were evident in dbp1Δ::kanMX6 cells grown in rich media (YEP dextrose or sucrose) during post-diauxic growth stages, after saturated cultures have exhausted fermentable carbon source availability and begin to utilize ethanol through respiration (Figure 2—figure supplement 1E). To determine whether these phenotypes were due to cassette-induced MRP51 mis-regulation or loss of Dbp1 protein, we tested whether they could be rescued by exogenous expression of either Dbp1 or Mrp51. All dbp1Δ::kanMX6 growth defects observed in conditions requiring elevated mitochondrial function remained similar with and without exogenous Dbp1 expression. In contrast, exogenous expression of Mrp51 rescued all previously observed respiratory growth defects of the dbp1Δ::kanMX6 strain, confirming these phenotypes resulted from disrupted expression of Mrp51, rather than lack of Dbp1 protein (Figure 2B; Figure 2—figure supplement 2A,B).

We next assessed whether other observed phenotypes in dbp1Δ::kanMX6 cells were due to off-target mis-regulation of MRP51 or loss of Dbp1 protein. The sporulation defect observed in dbp1Δ::kanMX6 cells was not rescued by exogenous Dbp1 but was fully rescued by exogenous Mrp51 expression (Figure 2C). Based on these results, we propose that this defect results from the dependency of meiotic cells on respiratory function (Jambhekar and Amon, 2008). Finally, polysome profiles of dbp1Δ::kanMX6 cells exogenously expressing Mrp51 looked similar to wild-type polysomes while polysomes from dbp1Δ::kanMX6 cells with or without exogenous expression of Dbp1 were indistinguishable (Figure 2—figure supplement 2C). In light of these data, we conclude that the reduced bulk cytoplasmic translation phenotypes observed in dbp1Δ::kanMX6 cells reflected a cellular response to mitochondrial dysfunction due to off-target Mrp51 mis-regulation rather than loss of Dbp1 protein. Mitochondrial dysfunction and translation rates have both previously been shown to impact the TOR pathway and thus influence global translation (Gao et al., 2016; Jazwinski, 2013; Raught et al., 2001; Topf et al., 2019), consistent with our findings.

To ensure that the inability of Dbp1 to rescue phenotypes in dbp1Δ::kanMX6 cells was not due to a lack of exogenous expression, we compared Dbp1 levels from the endogenous and exogenous loci. Western blot analysis revealed that the level of Dbp1 expressed from the exogenous integration was highly similar to that of the endogenous locus (Figure 2—figure supplement 2D). We conclude that resistance cassette insertion at the DBP1 locus is sufficient to cause widespread off-target phenotypes by mis-regulation of Mrp51. These data raised concerns, given the widespread use of expression cassette insertion in genome editing and were consistent with the phenomena of the NGE (Ben-Shitrit et al., 2012; Atias et al., 2016; Egorov et al., 2021), identified by analysis of deletion collection mutant phenotypes. While these results were disappointing for our study of the function of Dbp1 helicase, they highlighted this locus as a useful context to dissect the mechanism behind, and ideally to fix, these poorly understood cassette-related side effects.

To assess whether cassette-driven off-target effects were common to a specific feature of the kanMX6 and hphMX4 cassettes, we replaced the DBP1 ORF in wild-type cells with two additional cassettes, including one (natMX6; Figure 2D) that shared the pTEF and tTEF regulatory regions with kanMX6 and hphMX4 and one (TRP1; Figure 2D) that did not and is relatively lowly expressed. Insertion of either additional cassette led to a growth defect in non-fermentable carbon source compared to wild-type cells that was similar to that seen with kanMX6 insertion (Figure 2D, Figure 2—figure supplement 2E). As all reported cases of NGE have represented gene replacements with the commonly used kanMX modules, it was previously unclear whether this effect simply represented a problematic feature of kanMX. Our data indicate that the aberrant-transcript-driven mis-regulation of MRP51 is a general consequence of expression cassette insertion rather than a specific result of any sparticular cassette feature .

We reasoned that if the expression cassette inserted at the DBP1 locus drove transcription that altered neighboring gene expression, ‘insulation’ of neighboring genomic regions from transcriptional activity of the cassette should prevent these off-target effects. Toward this end, we placed strong transcription terminator sequences flanking both ends of the resistance cassette (Song et al., 2016). This included placing a portion of the DEG1 terminator 5′ to the TEF promoter, and either the same DEG1 terminator or the CYC1 terminator sequence 3′ to the TEF terminator within the kanMX6 cassette (Figure 3A). We named these plasmids ‘kanMX6-ins1’ and ‘kanMX6-ins2’ and used them to replace DBP1 using the same primers and insertion location as before. Both dbp1Δ::kanMX6-ins strains grew at wild-type rates in the non-fermentable carbon source, glycerol, in contrast to cells housing the original cassette replacement (Figure 3B, Figure 3—figure supplement 1A). Consistent with this finding, levels of Mrp51 protein in the dbp1Δ::kanMX6-ins strains were similar to those in wild-type cells (Figure 3C, D). Finally, 5′ rapid amplification of cDNA ends (5′RACE) confirmed that the MRP51 transcript produced in the dbp1Δ::kanMX6-ins strains was the same as that seen in wild-type cells, as opposed to the synthetic MRP51 LUTI in the initial cassette-inserted strains (Figure 3E, Figure 3—figure supplement 1B,C). To facilitate the use of this strategy and prevent aberrant transcription and neighboring gene mis-regulation in future studies, we created three additional ‘insulated’ cassettes housing the selection markers TRP1, his5⁺ (S. pombe HIS5; which complements S. cerevisiae his3), and nat^r (nourseothricin resistance) within the classic ‘MX6’ backbone (Longtine et al., 1998; Wach et al., 1997; Wach et al., 1994; Figure 3F).

While flanking both ends of the cassette with transcription terminator sequences was sufficient to prevent off-target mutant phenotypes at the DBP1/MRP51 locus, we wished to identify the root cause of neighboring gene mis-regulation. Previous work had suggested that the unusually high expression level of the kanMX6 cassette could cause gross chromatin changes that lead to local changes in transcription, and the NGE (Ben-Shitrit et al., 2012; Egorov et al., 2021). However, the similar off-target respiratory growth phenotype observed when DBP1 was replaced with either the kanMX6 or TRP1 cassettes, driven by the A. gossypii TEF and S. cerevisiae TRP1 promoters, respectively, led us to disfavor this model, as pTRP1 expression is several-fold lower than pTEF. Because bidirectional transcription has been shown to occur from most, if not all, promoters (Neil et al., 2009; Xu et al., 2009; Wei et al., 2011; Teodorovic et al., 2007; Preker et al., 2008; Core et al., 2008; Seila et al., 2008), and the observed aberrant transcript originated from the promoter-proximal end of the cassette, we hypothesized that bidirectional promoter activity from the TEF promoter was the most likely cause of MRP51 LUTI-based repression. Consistent with this model, both S. cerevisiae pTEF1 and pTRP1 display bidirectional activity at their endogenous loci, as determined by tiling array analysis and nascent transcript sequencing (NETseq) (Xu et al., 2009; Churchman and Weissman, 2011). We first tested this model by inserting a kanMX6 cassette in the reverse orientation at the DBP1 locus. This would not be expected to prevent gross chromatin changes caused by extremely high expression of the kan^r gene but would move pTEF away from MRP51 (Figure 3G, Figure 3—figure supplement 2A). Indeed, we found that replacing DBP1 with the reversed kanMX6 cassette did not result in a respiratory growth defect (Figure 3G; Figure 3—figure supplement 2A). Considering these data, and the knowledge that shared homology between cassettes can reduce the efficiency of targeted insertions when used sequentially, we constructed two more ‘insulated’ cassettes containing only a single additional terminator (tDEG1 or tCYC1) (Song et al., 2016) flanking each promoter and sharing no regulatory sequences with each other. Furthermore, to prevent the possibility of recombination between endogenous sequences and these new cassettes, we used the PGK1 promoter and terminator sequence from C. glabrata to drive resistance ORF expression (Figure 3H). Use of these two new insulated cassettes to replace DBP1 in the forward and reverse orientation confirmed that neither cassette caused a respiratory growth defect, in either orientation (Figure 3—figure supplement 2B).

Though severe off-target effects driven by a stable divergent transcript were observed with cassette replacement of DBP1, cassette replacement occurs at many loci without neighboring gene disruption. To better understand how the mis-regulation occurring with cassette insertion at the DBP1 locus compared to effects at other loci, we analyzed mRNA-seq and ribosome profiling data over genomic intervals surrounding cassette insertions from 18 additional cassette-replacement mutant datasets from our laboratory (Cheng et al., 2019), and one case in which only mRNA-seq data were available. All analyzed mutants contained a single ORF replaced with either the kanMX6 or natMX6 resistance cassettes and samples were collected during mitotic exponential growth. mRNA-seq data revealed stable divergent transcripts stemming from cassette-mediated gene replacements at 5 of these 19 loci (Figure 4A–E). For all five of these examples, the aberrant transcript was driven from the 5′ end of the resistance cassette, like in dbp1Δ::kanMX6 cells (Figure 1E; Figure 4A–E), and consistent with a model whereby bidirectional transcription from the cassette promoter drives aberrant neighboring transcripts. However, unlike the strong mis-regulation of MRP51 seen with DBP1 ORF replacement, these divergent transcription events had no obvious effect on the translation of adjacent genes, and a lack of apparent uORF translation in the extended 5′ mRNA regions (Figure 4A–D). For four of these new cases in which aberrant transcription resulted from a cassette-based deletion, the adjacent gene was positioned in the same orientation as the aberrant TSS thus it would be possible for synthetic LUTI-based repression to occur as was observed for MRP51. Although, we also note that for three of these four cases, expression of the adjacent ORF was low, even in wild-type cells, and thus mis-regulation of these genes may be difficult to detect under these conditions (Figure 4B–D). For the last case, in which UPF1 was replaced with the natMX6 cassette, an abundant cassette-induced divergent transcript was observed. However, the adjacent ISF1 gene was oriented antisense to this transcript and there was no evidence of its transcription with or without natMX6 insertion at UPF1 (Figure 4E); hence, there was no baseline expression to disrupt. Together, these data suggest that stable cassette-induced divergent transcripts are fairly common (occurring at 6/20 of loci analyzed here), but that severe phenotypic consequences are more rare.

We noted that the divergent transcript observed with cassette insertion at the UPF1 locus (Figure 4E), which encodes a gene that drives nonsense-mediated RNA decay (NMD) (Hug et al., 2016), was more abundant than any other example we found. This combined with the lack of stable divergent transcripts at ~70% of cassette-inserted loci, led us to wonder if RNA degradation could mask aberrant transcription events. Although NMD might be expected to target a subset of aberrant transcripts that reach the cytosol, nuclear exosome activity seemed likely to be a more major contributor (Hug et al., 2016; Gudipati et al., 2012; Wyers et al., 2005). In fact, one of the studies that initially identified bidirectional transcription as a feature of eukaryotic promoters relied on cells lacking nuclear exosome activity by loss of function of the nuclear catalytic subunit Rrp6 (Xu et al., 2009). It subsequently became clear that an alternative transcription termination mechanism, the ‘NNS’ pathway—named for complex subunits Nrd1, Nab3, and Sen1—leads to termination of most divergent transcripts and their rapid degradation by the Rrp6/exosome (Gudipati et al., 2012; Porrua and Libri, 2015; Wyers et al., 2005; Schulz et al., 2013; Steinmetz et al., 2001).

To test if Rrp6/exosome-mediated degradation was masking divergent transcripts produced from cassette insertion, we analyzed published global mRNA data from strain backgrounds with cassette insertions in an rrp6∆ strain background. For four cassette-replaced loci, control data were also available for the cassette replacement in the presence of RRP6 (Figure 5A–D; Malabat et al., 2015; Wang et al., 2020; Rege et al., 2015). For all of these cases, a divergent transcript from the site of cassette insertion was observed in rrp6∆ cells that either was absent or less abundant in the RRP6 control (Figure 5A–D). Interestingly, this included a case in which UPF1 was deleted by cassette insertion and 5′ transcript ends were sequenced (Figure 5D; Malabat et al., 2015). Here, the divergent transcript’s 5′ end was observed in the RRP6 control, consistent with our data, but was even more abundant in rrp6∆ cells (Figure 4E; Figure 5D). In the three other cases, for which data were available for cassette insertions in a rrp6∆ background but not a RRP6 control strain, including data for a rrp6∆::kanMX6 mutant itself, cassette-induced divergent transcripts were apparent (Figure 5—figure supplement 1A–C; Wang et al., 2020; Schmidt et al., 2012). Thus, for 7/7 (100%) of loci, replaced with a variety of cassettes, divergent transcripts were observed from inserted cassettes when nuclear exosome function was absent (achieved by rrp6∆; Figure 5A–D, Figure 5—figure supplement 1A–C). Although this sample size is not large (n = 7: rrp6∆, n = 20: RRP6), this result supports the model that expression cassette insertion always leads to production of a divergent transcript, driven by bidirectional promoter activity, but that these transcripts are usually terminated and degraded in cells with normal nuclear exosome activity. Together, these data argue that all loci are susceptible to cassette-induced bidirectional-promoter-driven production of divergent transcripts, but additional local characteristics at many loci prevent divergent transcription from disrupting neighboring genes.

Placement of large regions of DNA from other yeast species into S. cerevisiae results in an increased amount of divergent transcription from the foreign promoters, relative to endogenous S. cerevisiae promoters in the same cells (Jin et al., 2017). Because the pTEF version in commonly used budding yeast expression cassettes, including kanMX, is derived from A. gossypii, we wondered if species-specific regulation exacerbates the divergent transcription that we observe with cassette insertion (Wach et al., 1994). To test this, we swapped the pTEF in kanMX6, with the homologus S. cerevisiae pTEF1 region and inserted the new cassette at the DBP1 locus (Figure 5E; short-pTEF1). Insertion of the S. cerevisiae TEF1 promoter caused an even stronger respiratory growth defect than was observed with the original A. gossypii pTEF (Figure 5E, Figure 5—figure supplement 1D). This result is consistent with our finding that the S. cerevisiae TRP1 promoter can drive divergent transcription and neighboring gene disruption (Figure 2D, Figure 2—figure supplement 2E), and argues that off-target effects stemming from expression cassette insertion are not solely a result of inserting a foreign promoter sequence into S. cerevisiae.

Even at its endogenous locus, S. cerevisiae pTEF1 drives a divergent transcript (cryptic unstable transcript 910; ‘CUT910’) that is unstable due to Rrp6/exosome activity (Figure 5E; Xu et al., 2009), which degrades NNS-terminated divergent transcripts (Fox et al., 2015). The early termination and degradation of CUT910 presumably protects the neighboring locus MRL1 from transcription interference (Xu et al., 2009). We thus hypothesized that neighboring gene disruption by divergent transcripts resulting from bidirectional pTEF1 (or pTEF) activity at expression cassette insertion sites is partially a result of the use of a ‘minimal’ promoter region. While this region is competent to drive transcription of a transgene, it is devoid of its larger natural context and evolved failsafe mechanisms, including NNS-based termination signals, which prevent bidirectional transcripts from interfering with neighboring loci (Schulz et al., 2013; Porrua and Libri, 2015). Indeed, we found that when the minimal short-pTEF1 was replaced with a longer region of the endogenous promoter region, including the predicted termination site for CUT910 (Xu et al., 2009; Figure 5E; long-pTEF1), the respiratory growth defect seen with short-pTEF1 insertion was alleviated (Figure 5E). We concluded that the off-target effect from cassette insertion is a result of porting a minimal pTEF1 out of its native context, which leads to unmasking, in some insertion sites, of strong divergent transcript production.

These data suggested that naturally occurring NNS termination sites may be an important contributing factor in determining which loci are sensitive to the cassette-driven NGE. For many loci, the cassette-driven divergent transcripts observed in the rrp6∆ strains were short, likely reflecting termination by the NNS pathway. Though short binding preference motifs have been determined for the NNS complex, predictions of naturally occurring NNS termination signals based on sequence alone is difficult (Porrua and Libri, 2015; Carroll et al., 2004). Therefore, to test the sufficiency of the NNS termination pathway to terminate and prevent cassette-induced neighboring gene disruption, we tested whether a validated NNS termination site flanking pTEF in the kanMX6 cassette could reduce the impact of divergent transcription caused by cassette insertion at DBP1. We found that the ~150 bp intergenic region directly following mature SNR13 RNA, which has been shown to induce NNS-based termination in exogenous transcripts, was sufficient to protect the MRP51 locus from cassette-induced LUTI-based repression (Steinmetz et al., 2001; Schaughency et al., 2014). This was demonstrated by the wild-type growth rates observed in non-fermentable carbon sources when the NNS-kanMX6 cassette was inserted at the DBP1 locus (Figure 5E).

All strains were derived from the Saccharomyces cerevisiae SK1 background (Padmore et al., 1991). Detailed strain genotypes can be found in Key resources table. Strains were diploid MATa/alpha unless otherwise specified as haploid (MATa). Selection-mediated loss of function mutants were created by transforming wild-type haploids with a linear PCR fragment encoding a selection cassette that would replace the ORF of interest via homologous DNA repair mechanisms (Longtine et al., 1998; Wach et al., 1997; Wach et al., 1994; Baudin et al., 1993; Lorenz et al., 1995). Sequences directly upstream and downstream of each replaced ORF were always left intact. Genotypes for all strains were confirmed using PCR and mutants were backcrossed to a wild-type strain of the opposite mating type to ensure clearance of possible background mutations from transformation. The marker-less dbp1Δ was created by transforming a wild-type haploid with a plasmid-encoding Cas9 and a sgRNA targeting the DBP1 ORF, alongside a linear repair fragment that removes the DBP1 ORF while preserving the upstream and downstream adjacent sequence. Strains containing the Mrp51-3v5 allele were created by transforming wild-type or the selection replaced dbp1Δ strain of interest with a plasmid-encoding Cas9 and a sgRNA targeting MRP51, alongside a linear repair template that carboxy-terminally tags Mrp51 with the 3v5 sequence and introduces a synonomous mutation in the guide target sequence to prevent re-editing. Strains expressing exogenous Dbp1 or Mrp51 from the LEU2 locus were created by cloning the ORF of interest as well as ~1 kb of upstream and downstream regulatory sequence (for DBP1), or 500 bp of upstream and downstream regulatory sequence (for MRP51) into a single integration plasmid targeted to LEU2. This plasmid was linearized by digestion with SwaI nuclease (NEB, Ipswich, MA) and transformed into the dbp1Δ::kanMX6 strain.

Yeast were grown for experiments as in Powers and Brar, 2021. Briefly, strains were thawed on YPG plates from glycerol stocks overnight, then grown on YPD plates for a day before use in experiments. All growth experiments were carried out at 30°C using YEP media supplemented with 2% (wt/vol) of the carbon source, except for glycerol which was used at 3% (vol/vol). All growth experiments shown were repeated at least twice with one representative growth curve shown for each. Synchronous sporulating cells were prepared by first growing cells in YEP 2% dextrose for 24 hr at room temperature then diluted into BYTA at OD₆₀₀ = 0.25 and grown overnight at 30°C. Next, cells were pelleted, washed with water, and resuspended in sporulation (SPO) medium: either (0.5% KAc [pH = 7.0] supplemented with 0.02% raffinose) for Figure 1 ribosome profiling and mRNA-seq experiments, or rich SPO medium (2% KAc [pH = 7.0] supplemented with 40 mg/l adenine), 40 mg/l uracil, 20 mg/l histidine, 20 mg/l leucine, 20 mg/l tryptophan for sporulation efficiency, polysome experiments, and ribosome profiling in Figure 5—figure supplement 2. Sporulation efficiency was counted at 24 hr after induction into SPO media and 200 cells were counted per sample for three independent biological replicates. Samples were blinded prior to counting.

Meiotic yeast cells were harvested after 4 or 6 hr in sporulation medium (Figure 1 data: 4 hr; Figure 3—figure supplement 1: 6 hr), by brief treatment with 100 µg/ml cycloheximide then flash freezing. They were subjected to ribosome profiling and mRNA-seq as described in Powers and Brar, 2021. Reads per ORF were determined following mapping of all reads to the SK1 genome. RPKM values for ribosome footprints (FP) and mRNA reads were calculated by dividing the number of raw reads per gene by the total number of million mapped reads per sample and by the length in kilobases of each gene. TE was calculated by dividing the FP RPKM/mRNA RPKM for each gene. Data visualization of genome tracts was made using MochiView (Homann and Johnson, 2010).

Trichloroacetic acid (TCA) extractions were performed to collect total protein from samples as described in Chen et al., 2017; Chen et al., 2017. Briefly, ~2.5 OD₆₀₀ of yeast were treated with 5% TCA at 4°C overnight, washed with acetone, dried then lysed in lysis buffer (50 mM Tris–HCl [pH 7.5], 1 mM ethylenediaminetetraacetic acid (EDTA)), 2.75 mM dithiothreitol (DTT), protease inhibitor cocktail (cOmplete EDTA-free, Roche) for 5 min on a Mini-Beadbeater-96 (Biospec Products). Next, 3× sodium dodecyl sulfate (SDS) sample buffer (187.5 mM Tris [pH 6.8], 6% β-mercaptoethanol, 30% glycerol, 9% SDS, 0.05% bromophenol blue) was added to 1× and the cell lysate was boiled for 5 min. Proteins were run on 4–12% Bis-Tris Bolt gels (Thermo Fisher) then transferred to 0.45 µM nitrocellulose membranes using the 30 min mixed molecular weight protocol on the Trans-Blot Turbo System from Bio-Rad. Membranes were blocked with Intercept (phosphate-buffered saline, PBS) blocking buffer (LI-COR Biosciences, Lincoln, NE) at room temperature then incubated overnight at 4°C with mouse anti-v5 (1:1000, Invitrogen, RRID:AB_2556564) and rat anti-tubulin alpha (1:10,000, Serotec, RRID:AB_325005). Membranes were washed in PBS-0.08% Tween then incubated with an anti-mouse secondary antibody conjugated to IR Dye 800 (RRID:AB_621842) at a 1:15,000 dilution, and either an anti-rat secondary antibody conjugated to IR Dye 680 (RRID:AB_10956590) at a 1:15,000 dilution at room temperature for 1–2 hr (LI-COR Biosciences). Immunoblot images were generated and quantified using the Odyssey system (LI-COR Biosciences).

Cells were treated with 100 µg/ml cycloheximide for 30 s then filter collected, flash frozen, lysed, and prepared for sucrose gradient centrifugation of as in Hughes et al., 2019, with the following exceptions. Samples were thawed just prior to their loading on sucrose gradients, and SUPERase·In was added as for ‘mock digested samples’. Also, the polysome lysis buffer used was modified to contain protease and phosphatase inhibitors and is as follows: 20 mM Tris pH 8, 140 mM KCl, 1.5 mM MgCl₂, 100 µg/ml cycloheximide, 1% Triton X-100, 2 µg/ml Aprotinin, 10 µg/ml Leupeptin, 1 mM phenylmethylsulfonyl flouride (PMSF), 1:100 PIC2, and 1:100 PIC3. Lysates were loaded onto 7–47% sucrose gradients and spun in a SW 41 Ti rotor for 3 hr at 35,000 rcf at 4°C. Following the spin, samples were kept at 4°C and immediately collected using the Gradient Master gradient station from BioComp while monitoring 260 nm absorbance with the Bio-Rad EM-1 Economonitor. Samples were fractionated as shown in Figure 2, flash frozen, and submitted for mass spectrometry analysis.

Proteins were precipitated and desalted using the SP3 method for liquid chomatography with tandem mass spectrometry (LC-MS/MS) as described in Hughes et al., 2019; Cox and Mann, 2008. 50% from each polysome fraction (volume) were processed. Disulfide bonds were reduced with 5 mM dithiothreitol and cysteines were subsequently alkylated with 10 mM iodoacetamide. Proteins were precipitated on 0.5 µg/µl speedBead magnetic carboxylated modified beads (1:1 mix of hydrophobic and hydrophilic beads, cat# 6515215050250, 45152105050250, GE) by addition of 100% ethanol in a 1:1 (vol:vol) sample:ethanol ratio followed by 15 min incubation at 25°C, 1000 rpm. Protein-bound beads were washed in 80% ethanol and proteins were digested off the beads by addition of 0.8 µg sequencing grade modified trypsin (Promega) in 100 mM ammonium bicarbonate, incubated 16 hr at 25°C, 600 rpm. Beads were removed and the resulting tryptic peptides evaporated to dryness in a vacuum concentrator. Dried peptides were further desalted by another round of SP3 precipitation; peptides were reconstituted in 200 µl of 95% acetonitrile (ACN), followed by addition of bead-mix to a final concentration of 0.5 µg/µl and incubated 15 min at 25 °C, 1000 rpm. Beads were subsequently washed in 80% ethanol, peptides were eluted off the beads in 50 µl 2% dimethyl sulfoxide and samples were evaporated to dryness in a vacuum concentrator. Dried peptides were then reconstituted in 3% ACN/0.2% formic acid to a final concentration of 0.5 µg/µl.

About 1 μg of total peptides were analyzed on a Waters M-Class UPLC using a 25 cm Ionopticks Aurora column coupled to a benchtop Thermo Fisher Scientific Orbitrap Q Exactive HF mass spectrometer. Peptides were separated at a flow rate of 400 nl/min with a 190-min gradient, including sample loading and column equilibration times. Data were acquired in data-dependent mode using Xcalibur software. MS1 Spectra were measured with a resolution of 120,000, an AGC target of 3e6 and a mass range from 300 to 1800 m/z. Up to 12 MS2 spectra per duty cycle were triggered at a resolution of 15,000, an AGC target of 1e5, an isolation window of 1.6 m/z and a normalized collision energy of 27.

All raw data were analyzed with MaxQuant software version 1.6.10.43 (de Hoon et al., 2004) using a UniProt yeast database (release 2014_09, strain ATCC 204508/S288c), and MS/MS searches were performed with the following parameters: oxidation of methionine and protein N-terminal acetylation as variable modifications; carbamidomethylation as fixed modification; trypsin/P as the digestion enzyme; precursor ion mass tolerances of 20 ppm for the first search (used for nonlinear mass recalibration) and 4.5 ppm for the main search, and a fragment ion mass tolerance of 20 ppm. For identification, we applied a maximum false discovery rate of 1% separately on protein and peptide level. ‘Match between the runs’ was activated, as well as the ‘LFQ’ (label-free quantification) normalization (at least two ratio counts were necessary to get an LFQ value). We required 1 or more unique/razor peptides for protein identification and a ratio count of 2 or more for label-free protein quantification in each sample. This gave intensity values for a total of 2143 protein groups across both replicates. ‘LFQ’ normalized values were used for all subsequent analyses.

LFQ normalized values were then clustered with Cluster 3.0 (Saldanha, 2004) and visualized with Java TreeView (Song et al., 2016). Proteins not quantified in 2 or more samples were excluded in clustering. GO enrichment was assessed on the clusters specified (June 2022) and compared to the background population of all proteins quantified in the experiment (Boyle et al., 2004).

Cells were transferred to sporulation (SPO) media and incubated at 30°C for 4 hr. To metabolically label the cells 5 µL of EasyTag EXPRESS ³⁵S protein labeling mix (PerkinElmer), was added to 10 ml of SPO cultures and incubated with shaking for 10 min at 30°C. Protein was precipitated by addition of 100 µl 100% TCA to 900 µl SPO culture and incubated at 95°C with shaking. Samples were chilled on ice then pelleted and washed with cold 10% TCA, followed by a wash in cold 100% ethanol. Samples were resuspended in 5 ml of Econo-Safe scintillation fluid (RPI). Scintillation was counted for 2 min and the ³⁵S incorporation rates were derived from counts per minute normalized to cell density between samples and to wild-type measurements.

Based on the previous report that the DEG1 and CYC1 terminators can be used as insulator sequences (Song et al., 2016), we amplified these terminators from wild-type yeast and placed them into the 5′ and 3′ ends of the pFA6a-kanMX6 backbone using gibson assembly. Insulators were placed just inside the standard F1 and R1 primer amplification sites such that they would be useable with any primers designed to the previous system (pFA6a; Longtine et al., 1998). Insulated plasmids were fully sequenced over cloning junctions and the entire region to be amplified and integrated into yeast during transformation. To test the ability of these insulators to prevent aberrant transcription at the DBP1 locus, the kanMX6-ins cassettes were amplified and transformed into yeast exactly as had been done for the previous kanMX6 selection cassettes used in this study. When used to replace DBP1 ORF, Mrp51-3v5 levels and 5′ RACE confirmed the efficacy of the kanMX6-ins cassettes in preventing aberrant LUTI regulation of MRP51. We then replaced kan^r in the insulated plasmids with three additional markers for selection (including S. pombe HIS5 amplified from pFA6a-HIS3MX6): complements S. cerevisiae his3, TRP1 amplified from pFA6a-TRP1, and nat^r such that they can be used conveniently in combination as was done for the original toolkit (Longtine et al., 1998; Wach et al., 1994). To clone kanMX6-ins6 primers were used to amplify the entire kanMX6-ins2 except for the 3′ tCYC1 terminator and the Q5 Site-Directed Mutagenesis Kit was used to circularize the plasmid. For natMX6-ins7 a gBlocks Gene Fragment containing the PGK1 promoter and terminator from C. glabrata expressing nat^r and 5′ flanked by the S. cerevisiae CYC1 terminator was cloned into the pFA6a backbone. Detailed descriptions of these plasmids and can be found in Key resources table.

5′ RACE was performed with total RNA extracted from samples using phenol–chloroform precipitation and the GeneRacer Kit with SuperScript III RT and Topo TA Cloning Kit for Sequencing. For each sample, 4.8 µg of total RNA was used. Random primers were used to reverse transcribe the decapped cDNA library. The 5′ ends of the MRP51 cDNA were then amplified using the GeneRacer 5′ primer and a reverse gene-specific primer (5′ ACCGTCCAAGCAACTCTGCCAATGTC 3′) in a reaction with Platinum Taq High Fidelity DNA Polymerase. Amplified 5′ ends were then cloned into a Topo TA Cloning vector and clones were miniprepped and sequenced using Sanger sequencing. SnapGene was used to visualize and align the sequenced cDNA ends to the corresponding genomic position.

All newly created materials are available upon request. Newly created ‘Insulated’ selection cassettes will also be deposited to Addgene and made publicly available.