In vitro-transcribed (IVT) mRNAs are modalities that can combat human disease, exemplified by their use as vaccines for severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). IVT mRNAs are transfected into target cells, where they are translated into recombinant protein, and the biological activity or immunogenicity of the encoded protein exerts an intended therapeutic effect^1,2. Modified ribonucleotides are commonly incorporated into therapeutic IVT mRNAs to decrease their innate immunogenicity^3–5, but their effects on mRNA translation fidelity have not been fully explored. Here we demonstrate that incorporation of N¹-methylpseudouridine into mRNA results in +1 ribosomal frameshifting in vitro and that cellular immunity in mice and humans to +1 frameshifted products from BNT162b2 vaccine mRNA translation occurs after vaccination. The +1 ribosome frameshifting observed is probably a consequence of N¹-methylpseudouridine-induced ribosome stalling during IVT mRNA translation, with frameshifting occurring at ribosome slippery sequences. However, we demonstrate that synonymous targeting of such slippery sequences provides an effective strategy to reduce the production of frameshifted products. Overall, these data increase our understanding of how modified ribonucleotides affect the fidelity of mRNA translation, and although there are no adverse outcomes reported from mistranslation of mRNA-based SARS-CoV-2 vaccines in humans, these data highlight potential off-target effects for future mRNA-based therapeutics and demonstrate the requirement for sequence optimization.

The absolute quantification of intracellular protein levels is technically demanding, but has recently become more prominent because novel approaches like systems biology and metabolic control analysis require knowledge of these parameters. Current protocols for the extraction of proteins from yeast cells are likely to introduce artifacts into quantification procedures because of incomplete or selective extraction. We have developed a novel procedure for protein extraction from S. cerevisiae based on chemical lysis and simultaneous solubilization in SDS and urea, which can extract the great majority of proteins to apparent completeness. The procedure can be used for different Saccharomycetes yeast species and varying growth conditions, is suitable for high-throughput extraction in a 96-well format, and the resulting extracts can easily be post-processed for use in non-SDS compatible procedures like 2D gel electrophoresis. An improved method for quantitative protein extraction has been developed that removes some of the sources of artefacts in quantitative proteomics experiments, while at the same time allowing novel types of applications.

Translation of messenger mRNAs makes significant contributions to the control of gene expression in all eukaryotes. Because translational control often involves fractional changes in translational activity, good quantitative descriptions of translational activity will be required to achieve a comprehensive understanding of this aspect of biology. Data on translational activity are difficult to generate experimentally under physiological conditions, however, translational activity as a parameter is in principle accessible through published genome-wide datasets. An examination of the accuracy of genome-wide expression datasets generated for Saccharomyces cerevisiae shows that the available datasets suffer from large random errors within studies as well as systematic shifts in reported values between studies, which make predictions of translational activity at the level of individual genes relatively inaccurate. In contrast, predictions of cell-wide translational activity are possible from such datasets with higher accuracy, and current datasets predict a production rate of about 13,000 proteins per haploid cell per second under fast growth conditions. This prediction is shown to be consistent with independently derived kinetic information on nucleotide exchange reactions that occur during translation, and on the ribosomal content of yeast cells. This study highlights some of the limitations in published genome-wide expression datasets, but also demonstrates a novel use for such datasets in examining global properties of cells. The global translational activity of yeast cells predicted in this study is a useful benchmark against which biochemical data on individual translation factor activities can be interpreted.