Transcripts containing premature termination codons (PTCs) would produce truncated proteins, were it not for a quality-control process known as nonsense-mediated decay (NMD). This process occurs during the first, or pioneer, round of translation. If a PTC is encountered, the NMD machinery springs into action, catalysing degradation of the mRNA before it can be translated into protein. Numerous studies have defined three hallmarks that distinguish newly synthesized mammalian mRNAs undergoing a pioneer round of translation: a 5′ cap-binding complex comprising CBP20 and CBP80, 3′ poly(A)-binding proteins, and an exon-junction complex (EJC) that remains on the mRNA after splicing ∼20 nucleotides upstream of the exon–exon junction. The EJC contains several factors, known as Upfs, that function in NMD. Once the message has been successfully translated during the pioneer round, the cap-binding complex is replaced by eIF4E and the EJC is remodelled to remove the Upfs, rendering the mRNA refractory to NMD. The mRNA is then accessible for steady-state translation.

Until recently, it was not clear whether the yeast Saccharomyces cerevisiae used a similar mechanism, but a new study from Gao and colleagues has found that yeast NMD does not strictly follow the mammalian paradigm. Yeast also has a cap-binding complex, Cbc1–Cbc2, but a cbc1-deficient strain is competent for NMD, indicating that there is an alternative pathway for cap recognition in yeast NMD. Looking at the association of Cbc1 and eIF4E with unspliced or spliced mRNAs, Gao et al. found that — in contrast to the mammalian case — eIF4E is associated with a significant proportion of unspliced mRNAs, which should not have been exposed yet to NMD.

This raised the question of whether Cbc1 had to bind before eIF4E for NMD to occur. The results indicated that not only did NMD occur on Cbc1-bound mRNAs during a pioneer round of translation, but also that eIF4E can bind to the cap in the absence of Cbc1 and mediate NMD itself. Another difference with the mammalian situation is that mRNAs undergoing the pioneer round of translation do not seem to be associated with Upf3, an NMD factor.

These studies show that Cbc1-bound mRNA can be translated and participate in NMD; however, a previous study by Kuperwasser et al. found that there was no NMD of mRNAs bound by Cbc2. Therefore, the occurrence of NMD on Cbc1–Cbc2-capped mRNAs remains an open question. It is also not clear why, in yeast, eIF4E is able to mediate NMD, when this does not occur in mammalian cells. However, previous studies have shown that in yeast, messages undergoing steady-state translation are also subject to NMD, which would require an ability to function regardless of the cap protein bound to the mRNA.