Yeast strains, plasmids and oligonucleotides
All the yeast strains used in this study are listed in Supplementary Table 1 and the plasmids are listed in Supplementary Table 2. Strains were cultivated and grown in standard medium at 25 °C. The diploid strain HKY2065 was sporulated and subjected to tetrad dissection followed by analysis of haploid spores for their genetic markers.
FISH
The experiments were essentially carried out as previously described47. To detect poly(A)+ RNA, a Cy3-labelled oligonucleotide d(T)18 probe (Sigma) was used. Cells were grown to mid-logarithmic phase (around 1 × 107 cells per ml) before they were treated as indicated. Cells were fixed by adding formaldehyde to a final concentration of 3.7% for 40 min at room temperature. After washing, permeabilization and pre-hybridization, the Cy3-labelled d(T)18 probes were added and hybridized overnight at 37 °C. DNA was stained with DAPI (Sigma) for 2 min. Microscopy studies were carried out using a Leica AF6000 microscope and an HCX PL APO CS ×63 objective lens. Pictures were obtained using a LEICA DFC360FX camera with a resolution of 1,392 × 1,040 px and LAS AF 2.7.3.9 software (Leica). Images were quantified using Fiji software.
smFISH
The experiment was conducted largely as described above. Cells were grown in 2% raffinose to the logarithmic-growth phase. The expression of PHO85 mRNA and its asRNA were induced by adding 2% galactose. Cells were collected after the indicated times and fixed for 20 min in 3.7% formaldehyde. The probes used are listed in Supplementary Table 5. They were incubated for 3 h at 37 °C. Thereafter, the washing steps with SSC were carried out for 15 min each, as described in the FISH protocol. Quantification of the signal was carried out using Fiji software. To determine the nuclear signal, the DAPI signal was used as a reference. The boundary of the total cell was determined using Nomarski optic. The cytoplasmic signal was calculated by subtracting the nuclear signal from the total signal. The background signal was measured three times per image and subtracted from the measured signal of the cell as follows: integrated density − (selected area) × (mean fluorescence of background readings), which resulted in the final signal strength that was used for all images. For every time point, cells from three independent biological repetitions were quantified.
Immunofluorescence
Cells were grown, collected and treated as described for the FISH experiment. After permeabilization, cells were blocked in ABB (0.1 M Tris, pH 9.0, 0.2 M NaCl, 5% FCS, 0.3% Tween, 500 µg ml−1 transfer RNA) and incubated for 1 h at 37 °C, followed by ABB with the addition of 1/200 µl of the J2 antibody (1 µg µl−1) from Scicons17 and 0.2% Triton for 2 h at 37 °C. The addition of Triton prevented binding to membrane-bound glycan RNA, which was already known to be an antigen of the J2 antibody. Subsequently, cells were washed with 0.5% Triton in 1× PBS for 15 min, twice with 1× PBS for 15 min and finally with ABB for 30 min. The secondary Cy3-conjugated anti-mouse antibody in ABB (1:200) was thereafter incubated for 1 h at room temperature. Subsequently, cells were washed with 0.5% Tween in 1× PBS for 10 min and twice in 1× PBS for 10 min. Nuclei were stained with DAPI (Sigma) and mounting, microscopy and quantification were carried out as described in the FISH experiment.
GFP microscopy
The visualization of GFP-tagged proteins in vivo was essentially done as previously described47. Cells were grown in glucose (2%)-containing medium until the early logarithmic phase (0.5 × 107 cells per ml), washed once with 1 ml sterile H2O, transferred into galactose (2%)-containing medium and grown for 6 h to induce the expression of RNaseIII constructs. Next, cells were fixed with 3.7% formaldehyde for 1 min at room temperature and washed twice with 1 ml P-Solution (0.1 M potassium phosphate buffer, pH 6.5, 1.2 M sorbitol) before adding 20 µl on a polylysine-coated slide for 15 min at room temperature. Permeabilization, DNA staining, microscopy and quantification were carried out as described in the FISH experiment.
Cytoplasmic fractionation
To detect RNAs in the cytoplasm, cells were grown to mid-logarithmic phase (2 × 107 cells per ml), washed once with 1 ml YPD/1 M sorbitol/2 mM DTT and resuspended in YPD/1 M sorbitol/1 mM DTT with the addition of zymolyase (100 mg ml−1) to spheroplast cells. Before cytoplasmic fractionation, 200 µl of the cell suspension was taken for total lysate control. For the analysis shown in Fig. 1a and similar work, after spheroblasting, cells were diluted in 50 ml YPD/1 M sorbitol for 30 min at 25 °C before they were shifted to 37 °C for 1 h. After shifting, 10 ml was taken for total cell lysis. Next, cells were cooled on ice and centrifuged for 5 min at 2,000 rpm. For cytoplasmic fractionation, the cell pellets were resuspended in 500 µl Ficoll buffer (18% Ficoll 400, 10 mM HEPES, pH 6.0) and cells were lysed by adding 1 ml buffer A (50 mM NaCl, 1 mM MgCl2, 10 mM HEPES, pH 6.0) and 1 µl Ribolock RNase Inhibitor (Thermo Fisher). The suspension was vortexed and centrifuged for 10 min at 2,000 rpm. The resulting supernatant reflects the cytoplasmic fraction. To verify correct fractionation, samples were analysed in western blots for the presence of the cytoplasmic Zwf1 (anti-Zwf1 in TBS-T (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 0.1% Tween 20), 1:4,000) and the nuclear proteins Yra1 (anti-Yra1 in TBS-T, 1:1,000) and Nop1 (anti-Nop1 in TBS-T, 1:4,000). RNA was isolated using a Nucleo-Spin RNA Kit (Macherey and Nagel).
J2 RNA co-immunoprecipitation experiment
Yeast strains were grown to mid-logarithmic phase (2 × 107 cells per m) followed by ultraviolet cross-linking with a wavelength of 254 nm for 7 min. Cells were collected and lysed in RIP buffer (25 mM Tris-HCl, pH 7.5, 150 mM NaCl, 2 mM MgCl2, 0.5% (v/v) Triton X-100, 0.2 mM PMSF, 0.5 mM DTT, 10 U RiboLock RNase Inhibitor (Thermo Fisher) and protease inhibitor (Roche)) by using a FastPrep-24 machine (MP Biomedicals) with shaking three times for 30 s at 5.5 m s−1. After centrifugation, 30 µl of the supernatant was taken for input control and the remaining lysate was incubated with or without 3 µl of the J2 antibody (1 µg µl−1)17 from Scicons and the addition of recombinant ShortCut RNaseIII (NEB) for 30 min at 4 °C. After the first incubation, the lysates were transferred to prewashed G-sepharose beads and incubated for another 90 min at 4 °C. The beads were then washed five times with RIP buffer (0.25% Triton). The supernatant was removed and SDS loading dye (125 mM Tris-HCl, pH 6.8, 2% SDS, 10% glycerol, 5% 2-mercaptoethanol, bromphenolblue) was added. Subsequently, samples were incubated at 95 °C for 5 min and loaded onto an SDS gel followed by western blotting and staining with MYC (anti-MYC in TBS-T, 1:,) and Grx4 (anti-Grx4 in TBS-T, 1:4,000).
J2 RIP for RNA-seq
All yeast strains were grown to mid-logarithmic phase (2 × 107 cells per ml). Total RNA was isolated with TRIzol reagent. After the first ethanol precipitation, a DNaseI treatment was conducted followed by a second precipitation overnight. The obtained RNA was eluted in RNase-free water. Then 90 µg RNA and 3 µg J2 antibody were incubated in 500 µl PBST for 120 min at 4 °C (1× PBS, 0.5% Tween-20). After the first incubation, the RNA–antibody mix was transferred to prewashed G-sepharose beads and incubated for another 120 min at 4 °C. The beads were centrifuged for 1 min at 4,000 rpm at 4 °C. The supernatant was transferred a second time, together with 3 µg J2 antibody, to freshly washed beads and incubated for 120 min. Subsequently, these beads were centrifuged and the supernatant was used as the unbound fraction. The beads from the first incubation were washed five times with 1 ml PBST. Between each step, the beads were centrifuged for 1 min at 4,000 rpm at 4 °C. Finally, the RNA was purified from the unbound fraction and from the eluates by TRIzol-chloroform (Ambion RNA by Life Technologies) extraction and forwarded to RNA sequencing. We repeated the experiment three times, and it showed a high reproducibility.
J2 dot-blot
Cells were grown to the logarithmic-phase and shifted, if necessary, as indicated. RNA isolation was carried out with TRIzol reagent. Then 1 µg of the isolated RNA was applied onto a nylon membrane, which was blocked in PBST (1× PBS, 1% Tween-20), 0.05 mg ml−1 ssDNA and 5% (w/v) non-fat dried milk. Subsequently, the J2 antibody (anti-dsRNA in PBST, 1:5,000) was added and incubated for 2 h at room temperature. Finally, there were two washing steps with PBST, each for 15 min at room temperature, before the HRP-coupled goat anti-mouse secondary antibody was added in PBST for 1 h. Finally, the membrane was washed again three times with PBST for 10 min at room temperature, before the ECL detection was carried out with a Fusion FX7 Edge 18.06c (Vilber). Quantification was finalized with the analysis software Bio-1D from Vilber Lourmat.
Protein isolation and purification
Transformed Rosetta 2 E. coli cells were grown in 200 ml LB medium with ampicillin (100 µg ml−1) and chloramphenicol (25 µg ml−1) overnight, diluted to OD600 = 0.1 in 1,200 ml Terrific Broth medium (28.8 g yeast extract, 24 g Trypton, 9 ml 50% glycerin, 17 mM KH2PO4, 72 mM K2HPO4) and 100 µg ml−1 ampicillin. The diluted cells were incubated at 32 °C and 130 rpm for 3 h, followed by 37 °C and 130 rpm for 1 h. For protein induction, 1.2 ml of 1 M IPTG was added and the culture was further incubated at 16 °C and 130 rpm overnight. After induction, cells were washed in 200 ml IMAC loading buffer (50 mM NaH2PO4, 500 mM NaCl, 10 mM Imidazol, pH 7.8) and finally resuspended in 75 ml IMAC loading buffer with Roche complete protease inhibitor (one tablet per 50 ml). Cells were lysed using a microfluidizer with the setting 3 times at 700 bar. Thereafter, the lysate was centrifuged at 15,000g for 90 min. Cleared lysate was loaded onto a 5 ml HisFF column and subsequently washed with IMAC exchange buffer, then 1 M LiCl, again with IMAC exchange buffer, and finally with IMAC loading buffer. The proteins were eluted with IMAC elution buffer (50 mM NaH2PO4, 500 mM NaCl, 400 mM Imidazol, pH 7.8) and dialysed against heparin base buffer (40 mM HEPES KOH, 100 mM KCl, pH 7.5) overnight. After dialysis, the eluate was loaded onto a heparin column and again eluted with heparin elution buffer (40 mM HEPES-KOH, 100 mM KCl, 2 M NaCl, pH 7.5). Finally, the eluate was dialysed in dialysis buffer (30 mM HEPES-KOH, 160 mM KCl, pH 7.6) for 2 days. Protein concentration was determined by measuring the optical density at 280 nm.
EMSA
Either ordered FAM or Cy5-labelled RNAs (Sigma Aldrich) were used. Every RNA contained 36 nucleotides and had the same amount of C, G, T and A (Supplementary Table 4). dsRNAs were formed by incubating 20 µM of the labelled and 20 µM of the reverse complementary non-labelled RNA in dialysis buffer (30 mM HEPES-KOH, 160 mM KCl, pH 7.6) at 65 °C for 5 min and immediate subsequent cooling on ice. Next, 4 µM dsRNAs or ssRNAs were incubated with purified Mex67–Mtr2 and 2 µl Ribolock RNase Inhibitor (Thermo Fisher) in dialysis buffer, resulting in a final volume of 20 µl, at 30 °C for 15 min. For Fig. 2c, Mex67–Mtr2 was added in increasing amounts from 4 µM to 44 µM. For the competition assay depicted in Fig. 2d, 12 µM Mex67–Mtr2 was added to 3 µM substrate RNA, resulting in a molar ratio of 1:4 between substrate RNA and Mex67. The competitor RNA was added after the first incubation and further incubated at 30 °C for 15 min. Finally, a 6× loading dye (10 mM Tris-HCl, pH 7.6, 60% glycerol, 60 mM EDTA, 0.03% bromophenol blue) was added and the samples were loaded onto a 0.5% agarose gel with 1× TAE (40 mM Tris, 1 mM EDTA, 20 mM acetic acid, pH 9.5) running in 1× TAE, pH 9.5. Complexes were separated by running native gels at 300 V and 4 °C for 40 min. In-gel detection was carried out with a Fusion FX7 Edge 18.06c (Vilber) using the filter F-595 YR and Epi-Light module C530, or filter F-710 and Epi-Light module C640, with Evolution-Capt. Edge software.
Export release assay
The mex67-5 xpo1-1 RPS3-GFP strain was grown to the mid-logarithmic phase (2 × 107 cells per ml) and shifted to 37 °C for 2 h. Cells were collected either directly after shifting (0 min) or after shifting them back to 25 °C for 5 min, 10 min, 15 min, 30 min and 60 min. The cell pellets were frozen in liquid nitrogen and subsequent RIP experiments were carried out as described for the J2 RNA co-immunoprecipitation experiments, with the exception that GFP Trap beads were used and no antibody was added. After the final washing step, the beads were split in half for RNA isolation with TRIzol reagent and subsequent qPCRs and for SDS-PAGE and western blot analysis of GFP (anti-GFP in TBS-T, 1:4,000) and Aco1 (anti-Aco1 in TBS-T, 1:2,000). For qPCR measurements, the ssRNAs RPS17A, RPS6A and TDH1 and the dsRNAs FRE5, HPF1 and PRY3 were analysed. dsRNA targets were chosen using three criteria: the asRNA had a higher RPKM (reads per kilobase million) than the sense RNA; they were identified as dsRNA in an RNAi-seq experiment21; and they are enriched in J2 RNA-seq. The ssRNA targets were chosen because of the opposed criteria: the level of the asRNA is less than 1:100 compared with the mRNA and they are not enriched in either RNAi-seq nor J2-seq.
Cell lysis for protein and RNA quantification
Cells (30 ml) were grown overnight in synthetic medium containing 2% raffinose until the logarithmic phase. For asRNA induction, 2% galactose was added. At each indicated time point, a sample of 5 ml was taken and centrifuged at 4,000 rpm and 4 °C for 4 min. Cells were lysed in 400 µl RIP buffer (25 mM Tris-HCl, pH 7.5, 150 mM NaCl, 2 mM MgCl2 0.5% (v/v) Triton X-100, 0.2 mM PMSF, 0.5 mM DTT, 10 U RiboLock RNase inhibitor (Thermo Fisher) and protease inhibitor (Roche)) and divided into two samples. SDS loading dye (125 mM Tris-HCl, pH 6.8, 2% SDS, 10% glycerol, 5% 2-mercaptoethanol, bromphenolblue) was added to one of the samples. Subsequently, samples were incubated at 95 °C for 5 min and loaded onto an SDS gel followed by western blotting and staining of GFP (anti-GFP in TBS-T, 1:4,000) and Hem15 (anti-Hem15 in TBS-T, 1:5,000). The RNA was isolated from the second sample using the NucleoSpin RNA kit (Macherey Nagel) and quantified by qPCR.
Strand-specific cDNA synthesis and qPCR
To exclusively measure either mRNA or asRNA in qPCR, RNA-specific reverse primers were used (Supplementary Table 3) in cDNA synthesis (Nippon Genetics) and two separate samples were created. Each contained either the asRNA primer or the mRNA primer, resulting in separate asRNA and mRNA cDNAs. Furthermore, actinomycin D was added together with the reverse transcriptase because it prevents non-specific transcription from DNA and thereby secures strand-specific transcription, as reported previously48,49. In the qPCRs, the corresponding cDNA of mRNA and asRNA from one gene were measured with the same primer pair.
Drop-dilution analysis
Cells were grown to the logarithmic phase (2 × 107 cells per ml) and diluted to 1 × 106 cells per ml. Then, 10-fold serial dilutions to 1 × 103 cells per ml were prepared and 8 μl of each dilution was spotted onto selective plates. The plates were incubated for 3 days at the indicated temperatures and conditions. Pictures were taken after 2 or 3 days with an Intelli Scan 1600 (Quanto Technology) and the SilverFast Ai program.
RNA-seq
The sequencing of RNA samples was conducted at the NGS-Integrative Genomics Core Unit of the University Medical Center Göttingen. Samples were prepared with the TruSeq RNA Sample Prep Kit v.2, according to the manufacturer’s protocol (Illumina). Single-read (50 bp) sequencing was conducted using a HiSeq 4000 (Illumina). Fluorescence images were transformed to BCL files with the Illumina BaseCaller software (v.3.6.3) and samples were demultiplexed to FASTQ files with bcl2fastq (v.2.17).
Differential gene-expression analysis
Sequences were aligned to the genome reference sequence of Saccharomyces cerevisiae (sacCer3, obtained from UCSC; https://hgdownload.cse.ucsc.edu/goldenPath/sacCer3/bigZips/) using STAR software50 v.2.5, allowing for two mismatches. Subsequently, abundance measurement of reads overlapping with exons or introns was conducted with featureCounts51, subread v.1.5.0-p1, Ensembl (EF4.68) supplemented with the coordinates of UTRs, CUTs, SUTs22,52,53 and XUTs3,29. Data were processed in the R/Bioconductor environment (www.bioconductor.org, R v.3.6.1) using the DESeq2 package54; v.1.24.0). The sequencing data and abundance measurement files have been submitted to the NCBI Gene Expression Omnibus database. For null-hypothesis testing, the Wald test was used with multiple comparison adjustments using the Benjamini and Hochberg method. In downstream analysis, only transcripts with an average count above 40 were considered.
Sense–antisense-pair identification
Overlapping sense–antisense pairs were identified using BEDTools intersect (v.2.3.1)55, requiring overlaps to occur on the opposite strand with a minimum overlap of 0.5. lncRNAs were considered in analysis as SUTs, XUTs or CUTs only if they do not overlap with other transcripts of the other types on the same strand.
RNAi coverage analysis and classification
For gene coverage of RNAi degradation products, reads were trimmed using Cutadapt (v.2.1)56 and aligned to the reference genome with TopHat2 (v.2.1.1)57. For gene coverage, the geneBody_coverage module of the RSeQC package was used (v.2.6.4)58. The input BED file was filtered by lncRNA classes (SUT, CUT or XUT) or by RNA enrichment in RNAi-seq. Overlapping features on the same strand were excluded. To calculate the enrichment in RNAi-seq, the read densities of a transcript in the RNAi strain was divided by its read densities in the wild type. Subsequently, the logarithm to base 2 of this ratio was calculated. For subsequent analyses, transcripts were grouped into ten groups from −5 to 5 without 0, on the basis of their log2-transformed fold change (log2 [RNAi/wild type]). Group 1 contained transcripts with changes between 0 and 1; group 2 contained changes between 1 and 2, and so on. Finally, group 5 contained transcripts that have a log2-transformed fold change above 4. In the negative range, the classification was made in the same way.
Dimethyl sulfate reactivity analysis
The dimethyl sulfate reactivity assay was carried out as previously described45. The dimethyl sulfate reactivity for each transcript was summed in wild type and in dbp2∆. The average reactivity in the wild type was subtracted from the average reactivity in dbp2∆ to obtain the structural change between the strains.
Statistics and reproducibility
Experiments from which a significance was calculated were conducted independently at least three times. In Figs. 1h,i, 2f, 4c,h and Extended Data Figs. 4b, 7b,c,e, 8d,e,g and 9g, data are presented as mean values ±s.d., two-sided t-test P < 0.05*, P < 0.01**, P < 0.001***. In Figs. 1l, 4g and Extended Data Fig. 4a, the box plots are defined by the median as the centre line, the 25th and 75th percentiles as the box boundaries and the 10th and 90th percentiles as the whiskers. Two-sided Welch’s t-test, P < 0.05*, P < 0.01**, P < 0.001***, P < 0.0001****. In Figs. 1a–d, 2c,e, 3a,d, 4c and Extended Data Figs. 4c, 5a,b and 8b, the box plots are defined by the median as the centre line, the 25th and 75th percentiles as the box boundaries and the minimum and maximum values as the whiskers. Two-sided t-test, P < 0.05*, P < 0.01**, P < 0.001***, P < 0.0001****. In Fig. 2e, the centred asterisks show a significant enrichment compared with time point 0. One-sided t-test, P < 0.05*, P < 0.01**, P < 0.001***, P < 0.0001****. Spearman’s rank correlation analysis in Extended Data Fig. 5c and associated and between repetitions in Extended Data Figs. 1d, 2c and 3b were calculated using GraphPad PRISM.
Reporting summary
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.