Common mechanism of transcription termination at coding and noncoding RNA genes in fission yeast

Abstract

Termination of RNA polymerase II (RNAPII) transcription is a fundamental step of gene expression that is critical for determining the borders between genes. In budding yeast, termination at protein-coding genes is initiated by the cleavage/polyadenylation machinery, whereas termination of most noncoding RNA (ncRNA) genes occurs via the Nrd1-Nab3-Sen1 (NNS) pathway. Here, we find that NNS-like transcription termination is not conserved in fission yeast. Rather, genome-wide analyses show global recruitment of mRNA 3' end processing factors at the end of ncRNA genes, including snoRNAs and snRNAs, and that this recruitment coincides with high levels of Ser2 and Tyr1 phosphorylation on the RNAPII C-terminal domain. We also find that termination of mRNA and ncRNA transcription requires the conserved Ysh1/CPSF-73 and Dhp1/XRN2 nucleases, supporting widespread cleavage-dependent transcription termination in fission yeast. Our findings thus reveal that a common mode of transcription termination can produce functionally and structurally distinct types of polyadenylated and non-polyadenylated RNAs.

Fig. 1

S. pombe mRNA 3′ end processing and transcription termination factors are recruited at the 3′ end of independently transcribed snoRNA and snRNA genes. a, b Average ChIP-seq profiles of total RNAPII (Rpb1) relative to mRNA poly(A) site (a n = 4755) and to annotated 3′ end of independently transcribed monocistronic snoRNAs (b n = 31) in WT (solid line, orange) and NNS mutant strains (dotted lines, orange) grown in rich medium (YES). Average ChIP-seq profiles of total RNAPII (Rpb1) in WT and Seb1-deficient cells (Pnmt-seb1) grown for 15 h in thiamine-supplemented minimal medium (EMM) are also shown (gold). Gene coordinates are available in Supplementary Data  1–2. c–e Normalized ChIP-seq signal of RNAPII (Rpb1) and the indicated mRNA 3′ end processing factors across the fba1 mRNA (c), the snR99 snoRNA (d), and the snu5 snRNA (e) genes. f, g Average ChIP-seq profile of the indicated mRNA 3′ end processing factors over the same groups of mRNA (f) and snoRNA (g) genes as a, b

Fig. 2

The mRNA cleavage and polyadenylation complex is required for snoRNA synthesis. a Schematic of the RNase H cleavage assay used in b–e. After RNase H cleavage of the snoRNA at sites of RNA:DNA hybrids in the presence of a sequence-specific DNA oligonucleotide, the 3′ fragment (mature or 3′-extended) is detected by Northern blotting (NB). Addition of oligo d(T) to the RNase H reaction removes heterogenous poly(A) tails, generating discrete products. b, c Total RNA prepared from the indicated strains that were previously grown in thiamine-supplemented medium for 15 h to deplete Pcf11, Seb1, and Dhp1 was treated with RNase H in the presence of DNA oligonucleotides complementary to snR3 (b) and snR99 (c). RNase H reactions were performed in the presence (+) or absence (−) of oligo d(T). The top panel represents a longer exposure of the middle panel to see 3′-extended (3′-ext) cleavage products. The 5S rRNA was used as a loading control. d, e As described in b, c, but using cells that were previously treated with rapamycin to deplete Ysh1 and Rna14 from the nucleus. f Northern blot analysis using total RNA prepared from the indicated strains that were treated with either rapamycin (lanes 1–3) or thiamine (lanes 4–8). The blot was hybridized using DNA probes specific to snR3 and the 18S rRNA. The position of mature snR3, 18S and 25S rRNAs, as well as snR3 read-through (RT) products is indicated on the right

Fig. 3

Independently transcribed snoRNA genes are cleaved and polyadenylated. a Proportion of 13 C/D box and 18 H/ACA box monocistronic snoRNAs with at least one poly(A) site mapped by 3′READS in the WT strain grown in minimal or rich media. b Poly(A) site (PAS) read density (in RPM) profile downstream of the snR99 snoRNA as determined by 3′ READS in the indicated strains grown in either rich (RM) or minimal (MM) media supplemented with thiamine for 15 h to deplete Seb1 and Pcf11. c, d Box plots showing the distribution of the distance calculated between the strongest poly(A) site (PAS) and the annotated snoRNA 3′ end (c), as well as the sum of the read density for all of the poly(A) sites associated to a snoRNA in each condition (d). Center lines correspond to the median and the significance of difference (Wilcoxon signed-rank test) is indicated for comparing groups (^∗p < 0.05; ^∗∗p < 0.01; ^∗∗∗p < 0.005). e Schematic of snR99 constructs with different 3′ flanking sequences used for experiments in f and g. f Northern blot analysis of total RNA prepared from WT (lane 1) and snR99-null (lanes 2–5) cells that were previously transformed with the indicated constructs (EV, empty vector). g Total RNA prepared from snR99-null cells that were previously transformed with snR99 constructs that comprised rps2 (rps2T, lanes 2–5) or trx1 (trx1T, lanes 6–9) terminator sequences was treated with RNase H in the presence of DNA oligonucleotides complementary to snR99

Fig. 4

Tyr1-P and Ser2-P forms of the RNAPII CTD colocalize with 3′ end processing factors at coding and ncRNA genes. a, b Average ChIP-seq profile of total RNAPII (Rpb1) and the indicated CTD modifications in a WT strain across 4755 mRNAs (a) and 24 monocistronic snoRNAs (b) with a mapped poly(A) site in minimal medium. c, d PAS read density (3′READS) and normalized ChIP-seq signal of total RNAPII (Rpb1), the indicated CTD modifications, as well as Pcf11 and Seb1 across the fba1 mRNA (c) and the snR99 snoRNA (d) genes. e, f Average profile of total RNAPII (Rpb1), Ser2-P, Tyr1-P, Pcf11, and Seb1 relative to the poly(A) site of 4755 mRNA (e) and 24 monocistronic snoRNA (f) genes in minimal medium. g Genome-wide pairwise Pearson correlation coefficient matrix at a resolution of 10 bp followed by a hierarchical clustering. Rectangles include the average of a subset of correlation values between 3′ processing factors and CTD modifications

Fig. 5

The torpedo nuclease Dhp1 is required for transcription termination of coding and ncRNA genes. a–c Normalized ChIP-seq signal of RNAPII subunits Rbp1 and Rbp3 in WT and Dhp1-depleted strains across the fba1 mRNA (a), the snR99 snoRNA (b), and the snu5 snRNA (c) genes grown in thiamine-treated minimal medium for 15 h. The dashed-line rectangles highlight delayed transcription termination in Dhp1-deficient cells. d–g Average ChIP-seq profile of Rbp1 and Rbp3 (d, e) or total RNAPII (Rbp1), Ser2-P, and Tyr1-P (f, g) in WT (solid lines) or Dhp1-depleted (dotted lines) cells across 4755 mRNA (d–f) and 24 monocistronic snoRNA (e, g) genes with a mapped p(A) site in thiamine-treated minimal medium

Fig. 6

The endonucleolytic activity of Ysh1 is necessary for termination of snoRNA transcription. a–c Normalized ChIP-seq signal of total RNAPII (Rbp1) in WT (top) and ysh1 mutant (bottom) cells across the fba1 mRNA (a), the snR99 snoRNA (b), and the snu5 snRNA (c) genes grown in rapamycin-treated minimal medium for 4 h. The dashed-line rectangles highlight transcriptional read-through at fba1 and snR99 genes in Ysh1-deficient cells. d, e Average ChIP-seq profile of RNAPII (Rpb1) in WT (solid lines) and Ysh1-deficient (dotted lines) cells across 4755 mRNA (d) and 24 monocistronic snoRNA (e) genes grown in rapamycin-treated minimal  medium for 4 h. f, g RNAPII ChIP-qPCR analysis on the fba1 (f) and snR99 (g) genes using extracts prepared from either wild-type (WT) or ysh1 anchor-away (ysh1-AA) strains containing chromosomally integrated constructs that express the indicated versions of FLAG-tagged Ysh1 (WT, H403F, and H165F) as well as an empty vector (EV) control. Bars above the fba1 and snR99 genes show the positions of PCR products used for ChIP-qPCR analyses. Cells were grown in the presence of rapamycin for 4 h to deplete endogenous Ysh1 from the nucleus. ChIP signals (percent of input) were normalized to region 1. Error bars indicate SD. n = 3 biological replicates from independent cultures. h–j Northern blot analysis of fba1 (h), snR3 (i), and snR99 (j) genes using total RNA prepared from using total RNA prepared from the same WT (lane 1) and ysh1 anchor-away (ysh1-AA, lanes 2–5) strains as in f, g. The position of read-through (RT) transcripts is indicated on the right. The 18S rRNA was used as a loading control

Fig. 7

Transcription termination at lncRNA genes and general model for 3′ end processing and transcription termination of mRNA and snoRNA genes in fission yeast. a–c Normalized ChIP-seq signal of total RNAPII (Rpb1 and Rbp3-HA), the indicated CTD modifications, and 3′ end processing factors across the SPNCRNA.532 (a), the SPNCRNA.491 (b), and the nam1/SPNCRNA.1459 (c) genes. d, e Recruitment of cleavage and polyadenylation factors (CPF) by poly(A) signals (PAS) is a common feature of mRNA and snoRNA genes (see Cleavage). Endonucleolytic cleavage by the CPF-associated Ysh1 nuclease will generate an RNAPII-bound unprotected 5′ end that will be targeted by the 5′–3′ exonuclease Dhp1, contributing to termination of mRNA and snoRNA transcription (see Termination). The RNA-bound CPF complex promotes snoRNA 3′ end maturation, possibly by facilitating Pab2 recruitment to snoRNA precursors. The co-transcriptional recruitment of specific export factors to nascent mRNAs (d) may represent a decisive step that prevents Pab2-dependent exosome-mediated RNA processing in the nucleus (e)