Codon optimization depletes stop codons that limit out-of-frame translation in mRNA therapeutics
Investigators analyzed 120 protein-coding nucleic acid therapeutics, including FDA-approved mRNA vaccines, and found that standard codon optimization systematically depletes stop codons in alternative reading frames. This design feature allows out-of-frame translation to produce extended polypeptides averaging 164 amino acids, which is six times longer than those generated by natural human genes. Mass spectrometry confirmed that strategically restoring synonymous stop codons eliminates these unintended translation products without altering the intended therapeutic protein. The findings establish alternative-frame stop codon density as a critical quality parameter for mRNA therapeutic design and analytical characterization.
The original study
Codon optimization depletes stop codons in alternative reading frames of protein-coding nucleic acid therapeutics.
- Authors
- Liu Z, Ying Z, Shen L, Ma W
- Journal
- Proceedings of the National Academy of Sciences of the United States of America
- Type
- Journal Article
- PMID
- 42594282
Original abstract
Out-of-frame translation events, arising from ribosomal frameshifting or noncanonical initiation, are an unavoidable feature of translation. Their consequences depend on the distribution of stop codons in alternative reading frames, which determines the permissiveness of those frames: whether out-of-frame translation terminates quickly or generates extended products. Using quantitative dual-fluorescence reporters, we show that these stop codons function as molecular checkpoints that terminate out-of-frame translation. Genome-wide analysis across 10 organisms reveals that natural coding sequences maintain dense stop codon distributions in alternative frames, with a median spacing of approximately 20 amino acids. Codon optimization, the standard method for enhancing translation, systematically depletes this safeguard. Because all three stop codons (UAA, UAG, UGA) begin with uridine, and optimal human codons exclude uridine from third positions, stop codons in the -1 reading frame become structurally impossible in codon-optimized sequences. Analysis of 120 therapeutic sequences, including FDA-approved COVID-19 messenger RNA (mRNA) vaccines, confirms widespread -1 frame stop codon depletion: out-of-frame products average 164 amino acids, sixfold longer than in natural human genes. Strategic restoration of stop codons through synonymous substitutions eliminates detectable out-of-frame products by mass spectrometry while preserving the intended protein. Although such products and immune responses have been detected in COVID-19 mRNA vaccine recipients, there is no evidence they cause clinical harm; nonetheless, our approach offers a simple way to eliminate them through informed sequence design alone, without changes to manufacturing or regulatory frameworks. Our findings establish stop codon distribution as a critical design parameter for protein-coding nucleic acid therapeutics.