Widespread impact of nonsense-mediated mRNA decay on the yeast intronome

Abstract

Nonsense-mediated mRNA decay (NMD) eliminates transcripts carrying premature translation termination codons, but the role of NMD on yeast unspliced pre-mRNA degradation is controversial. Using tiling arrays, we show that many unspliced yeast pre-mRNAs accumulate in strains mutated for the NMD component Upf1p and the exonuclease Xrn1p. Intron identity and suboptimal splicing signals resulting in weak splicing were found to be important determinants in NMD targeting. In the absence of functional NMD, unspliced precursors accumulate in the cytoplasm, possibly in P-bodies. NMD can also complement RNase III-mediated nuclear degradation of unspliced RPS22B pre-mRNAs, degrades most unspliced precursors generated by a 5' splice site mutation in RPS10B, and limits RPS29B unspliced precursors accumulation during amino acid starvation. These results show that NMD has a wider impact than previously thought on the degradation of yeast-unspliced transcripts and plays an important role in discarding precursors of regulated or suboptimally spliced transcripts.

Figure 1. Tiling array analysis of exons and intron signals in the upf1Δ strain relative to wild-type

A. Dendrograms of all ICG. B. Dendrogram of non-RPGs. C. Dendrogram of RPGs; D. dendrogram of RPGs that show an increase of intronic signal. Exons are represented by boxes, intron by a line. Yellow represents increase, blue, decrease. For genes exhibiting two introns, the first intron is indicated by .1.

Figure 2. Northern blot analysis of ICGs in wild-type and NMD mutant strains

A. and B. Detection of unspliced ICGs in wild-type, upf1Δ and xrn1Δ strains. Shown are the signals obtained by hybridization of northern blots with the indicated probes, which covered exon1 (E1), intron (I), and exon2 (E2), unless indicated otherwise. US, Unspliced pre-mRNA. S, Spliced mRNA. Numbers indicate the fold increase accumulation of unspliced relative to wild-type. SCR1 or G3PDH were used as loading controls. C. Detection of unspliced RPGs in wild-type, upf1Δ, upf2Δ, upf3 Δ strains. The prp2-ts strain was used as a positive control for the detection and migration of unspliced precursors.

Figure 3. Effects of NMD mutations combined with inactivation of Rrp6p, Mlp1 or of Rnt1 stem-loops on the accumulation of unspliced pre-mRNAs

A. Expression of RPGs and detection of unspliced RPGs in wild-type, upf1Δ and rrp6Δ and upf1Δrrp6Δ strains for NMD-sensitive transcripts. For some transcripts, different exposures are shown to properly visualize mRNAs. Legends as in Fig. 2. B. Expression of RPGs and detection of unspliced RPGs in wild-type, upf1Δ and mlp1Δ and upf1Δmlp1Δ strains for NMD-sensitive transcripts. Legends as in Fig. 2. C. Expression of RPGs and lack of detection of unspliced RPGs in wild-type, upf1Δ, mlp1Δ, rrp6Δ, upf1Δmlp1Δ and upf1Δrrp6Δ strains for NMD-insensitive transcripts. Legends as in Fig. 2. D. Inactivation of NMD and of Rrp6 shows synergistic effects on the accumulation of two unspliced pre-mRNAs. Legends as in A. E. Analysis of RPS22B in strains carrying a deletion of the Rnt1p-target stem-loops (SL1,2Δ) and/or carrying mutations of NMD components. Shown are the signals obtained by hybridization of the same membrane with the indicated probes.

Figure 4. NMD targeting is influenced by intron identity and suboptimal splicing signals

A. Detection of spliced and unspliced GOT1 transcripts (GOT1 riboprobe) in wild-type, prp2-ts, upf1Δ, and in a strains where the natural GOT1 intron was replaced by that of RPL19A, and in the same strain where Upf1p was inactivated. B. Similar as A, except that the GOT1 intron was replaced by that of RPS11B. C. Similar as A, except that the GOT1 intron was replaced by that of actin (ACT1) or RPL25. D. Legends as in A. Strains indicated as N contain the natural RPL19A branchpoint (BP) UAACUAAC; Strains indicated as CS contain the consensus sequence UACUAAC.

Figure 5. NMD inactivation exacerbates accumulation of unspliced precursors generated by a 5’ splice site mutation of RPS10B

A. Detection of RPS10B transcripts in strains carrying a wild-type RPS10B gene or with a 5’ splice site (5’SS) mutation (M) in RPS10B, alone or in combination with deletion of UPF1, UPF2 or UPF3. Legends as in Figure 2. B. Legends as in A, with the effects of the inactivation of Rrp6p. C. Legends as A, except that the 5’SS mutation was introduced in the ACT1 intron that was transposed in the GOT1 gene (see Fig. 4B).

Figure 6. Localization of unspliced RPS10B precursors by Fluoresencent In Situ Hybridization (FISH)

A. Shown are the DAPI staining (blue) for nuclear DNA, intron RPS10B FISH (red) and merged images for the wild-type, upf1Δ and upf2Δ strains. B. U14 snoRNA FISH (red) is shown as a control in the same strains.

Figure 7. Detection of unspliced precursors in wild-type and upf1 Δ strains during Camino acid starvation

A. Venn diagram of the overlap between Intron-containing RPGs, ICGs affected by Upf1p inactivation, and ICGs for which splicing is inhibited by amino-acids starvation (Pleiss et al. 2007). B. Indicated strains were shifted for 10 or 20 minutes in a medium containing 50mM 3AT or untreated (mock), and probed for the indicated RPGs. Relative amount of unspliced precursors in the different conditions was standardized to the mock value showing the lowest amount of unspliced precursors.