Messenger RNAs transcribed from yeast linear cytoplasmic plasmids possess unconventional 5’ and 3’ UTRs and suggest a novel mechanism of translation

Václav Vopálenský; Michal Sýkora; Tomáš Mašek; Martin Pospíšek

The structure and function of the cap-binding eIF4E protein from the yeast S. cerevisiae is very well understood. Because pGKL plasmids can be transferred and stably maintained in S. cerevisiae (97), we decided to test possible interaction between the S. cerevisiae eIF4E capbinding protein (S.c.-eIF4E) and pGKL mRNAs in vitro. We produced S.c.-eIF4E as an Nterminal GST fusion protein in E. coli and purified it by glutathione-Sepharose affinity chromatography. To test the possible binding of pGKL-specific mRNAs to the yeast eIF4E, we incubated DNase I-treated total RNA purified from K. lactis IFO1267 with the purified GST-S.c.-eIF4E fusion protein bound to the glutathione Sepharose. After 3 hours of incubation at room temperature, the slurry was extensively washed six times with a great excess of buffer I at room temperature and directly used for RT-PCRs with gene-specific primers. Fig. 5 clearly shows that pGKL mRNAs, represented by mRNA from K2ORF5, do not bind to yeast eIF4E in vitro, whereas cellular mRNAs, represented by mRNA coding for a high-affinity glucose transporter (HGT1), do. This result is further supported by qPCR analysis that revealed comparable abundances of K2ORF5 and HGT1 mRNAs in the K. lactis IFO1267 total RNA (data not shown). Interestingly, K2ORF5 is not one of the pGKL mRNAs with the low frequency of m 7 G cap occurrence at their 5' ends. An explanation could be that the strength of eIF4E binding to mRNA is highly influenced by the first proximal nucleotide following the terminal m 7 G, where the association constant K a is much higher for purines than . CC-BY-NC-ND 4.0 International license author/funder. It is made available under a The copyright holder for this preprint (which was not peer-reviewed) is the . https://doi.org/10.1101/325316 doi: bioRxiv preprint for pyrimidines and slightly higher for guanosine than for adenosine (98,99). Many similar results have been obtained by other researchers; however, cap analogs were used to determine the K a and K d values in most of these studies. Adding more nucleotides after the first proximal nucleotide substantially increased eIF4E binding to the cap while still preserving the differential influence of the first proximal nucleotide on eIF4E binding (99). A short homopolymeric stretch of 5' adenosine nucleotides thus might destabilize eIF4E binding to the mRNA cap. However, decreased eIF4E binding to the pGKL mRNAs does not substantially affect the performance of those mRNAs in translation, which suggests a novel mechanism of translation initiation independent of eIF4E cap-binding protein.