Codon Optimization Introduces Hidden Frameshifting Risk in Protein-Coding Nucleic Acid Therapeutics
Ontology highlight
ABSTRACT: The therapeutic promise of protein-coding genetic medicines depends on accurate translation of designed nucleic acid sequences into functional proteins. Here we reveal that current design practices systematically deplete stop codons from alternative reading frames, creating an unrecognized risk of out-of-frame translation product accumulation. Using quantitative dual-fluorescence reporters, we demonstrate that out-of-frame product accumulation correlates with expression level and that stop codons serve as molecular checkpoints against aberrant translation. Genome-wide analysis shows that natural coding sequences maintain dense stop codon distributions in alternative frames—a safeguard strikingly absent from therapeutic constructs. Analysis of 120 therapeutic sequences, including FDA-approved COVID-19 mRNA vaccines, shows severe stop codon depletion: -1 frame products average 164 amino acids, six-fold longer than in natural human genes. This depletion arises because codon optimization algorithms exclude uridine from third codon positions, thereby preventing -1 frame stop codon formation. We establish stop codon-walking as a method for mapping out-of-frame translation hotspots and develop evolutionary-guided codon optimization that strategically restores stop codons through synonymous substitutions. Mass spectrometry confirms that stop codon restoration eliminates detectable out-of-frame products while preserving intended protein sequences. These findings expose a fundamental design oversight and establish principles for engineering translation fidelity into protein-coding genetic medicines.
ORGANISM(S): Homo Sapiens
SUBMITTER:
Weirui Ma
PROVIDER: PXD073231 | iProX | Sun Jan 18 00:00:00 GMT 2026
REPOSITORIES: iProX
ACCESS DATA