Multiple transcripts from a 3'-UTR reporter vary in sensitivity to nonsense-mediated mRNA decay in Saccharomyces cerevisiae

Abstract

Nonsense-mediated mRNA decay (NMD) causes accelerated transcript degradation when a premature translation termination codon disrupts the open reading frame (ORF). Although endogenous transcripts that have uninterrupted ORFs are typically insensitive to NMD, some can nonetheless become prone to NMD when translation terminates at out-of-frame premature stop codons. This occurs when introns containing stop codons fail to be spliced, when translation of an upstream ORF (uORF) terminates in the 5'-untranslated region (5'-UTR) or the coding region, or when the 5'-proximal AUG initiation codon is bypassed and translation initiates at a downstream out-of-frame AUG followed by a stop codon. Some 3'-untranslated regions (3'-UTRs) are also known to trigger NMD, but the mechanism is less well understood. To further study the role of 3'-UTRs in NMD, a reporter system was designed to examine 3'-UTRs from candidate genes known to produce NMD-sensitive transcripts. Out of eight that were tested, the 3'-UTRs from MSH4 and SPO16 caused NMD-dependent mRNA destabilization. Both endogenous genes produce multiple transcripts that differ in length at the 3' end. Detailed studies revealed that the longest of six reporter MSH4-3'-UTR transcripts was NMD-sensitive but five shorter transcripts were insensitive. NMD-dependent degradation of the long transcript required Xrn1, which degrades mRNA from the 5' end. Sensitivity to NMD was not associated with extensive translational read-through past the normal stop codon. To our knowledge, this is the first example where multiple transcripts containing the same ORF are differentially sensitive to NMD in Saccharomyces cerevisiae. The results provide a proof of principle that long 3'-UTRs can trigger NMD, which suggests a potential link between errors in transcription termination or processing and mRNA decay.

Figure 1. The 3′UTR of MSH4 and SPO16 produce transcription read-through products in an NMD dependent manner.

(A) A reporter system was constructed on a centromeric plasmid. The reporter consisted of the 3′UTR of genes of interest fused to GFP under control of the copper inducible CUP1 promoter. (B) Eight genes were examined by northern blot in a BY4741 background with both a wild type and Δupf1 strain. The 3′UTR of PGA1 is known to target transcripts for NMD and ACT1 is known to be insensitive to NMD. Both were included as controls . Only the MSH4 3′UTR and SPO16 3′UTR displayed accumulation of a longer transcript, which are indicated by arrows. (C) The relative accumulation of the longer transcript was measured in a W303A background. Relative accumulation was calculated as the percentage of the longer transcript relative to total transcripts. Strain dependent differences in the magnitude of NMD-dependent accumulation in BY4741 and W303A background is consistent with previously published results . All values are the result of averaging a minimum of 3 replicates.

Figure 2. The long transcriptional read-through product from GFP-MSH4-3′UTR accumulates in an NMD-dependent manner.

(A) The accumulation of the long GFP-MSH4 3′UTR construct was examined in a BY4741 background. The longer transcript accumulates to the same level in both a Δupf1 and Δupf2 strain verifying the observed effect is NMD dependent. (B) The endogenous MSH4 gene shows an increase of steady state mRNA in an NMD deficient strain, confirming it as a target of NMD.

Figure 3. The long transcript from GFP-MSH4-3′UTR shows increased stability in NMD-deficient strains.

Representative half-life measurements in a W303A background are shown for both (A) the longer transcript product and (B) the shorter transcript product. Inlay depicts the sample being represented in the graph and is adjacent to the corresponding marker. Only the longer transcript shows increased stability in an NMD deficient strain when transcription is inhibited by the addition of thiolutin.

Figure 4. Five short transcripts are not targets of NMD.

(A) After treatment with RNase H the shortest transcript observed on a 1% agarose gel can be resolved into 5 distinct bands on a 6% polyacrylamide gel. (B) Relative accumulation of the five short transcripts. Values in WT and Δupf1 columns represent each transcript as a percentage of the sum of the first five transcripts. (C) Sizes of all observed transcripts from northern blots performed on polyacrylamide gels (transcripts 1-5) and agarose gels (transcript B). Sizes were determined based on a combination of RT-PCR and estimates based on relative migration in gels to known size markers. (D) Schematic representing the transcription products of the GFP-MSH4 3′UTR construct (not to scale). The plasmid DNA fragment is 710 bases and contains prokaryotic DNA. The HIS3 gene is located on the crick strand relative to the GFP coding region.

Figure 5. GFP-MSH4-3′UTR transcripts are degraded from the 5′ ends.

The steady state mRNA accumulation levels were measured in a W303A background in Δxrn1, Δski7, Δxrn1 Δupf1 and Δski7 Δupf1 strains. Only the Δski7 mutant showed a statistically significant difference in accumulation from a Δupf1 mutant. (Fig 1C) indicating that Upf1 and Xrn1 are both part of the same degradation pathway for 3′UTR-mediated NMD.

Figure 6. 3′UTR targeting is not associated with extensive translational read-through.

(A) A FLAG tag was inserted three codons downstream of the normal stop codon of the GFP-MSH4 3′UTR construct to monitor translational read-through. The normal TAA (stop) codon was also mutated to the sense codon TTA (Leu) as a positive control. The * indicates a stop codon. (B) Northern blots were performed to measure relative accumulation of the NMD sensitive transcript in a wild type and NMD deficient strain. The presence of the FLAG sequence or stop→sense mutation did not affect relative accumulation of the mRNA. (C) Western blot of the GFP-FLAG-MSH4 3′UTR fusion. No translational read-through product was detected in the wild type or NMD deficient strain.