TL;DR
This paper investigates whether the SARS-CoV-2 spike protein sequences used in mRNA and DNA vaccines contain "hidden" overlapping open reading frames (ORFs) that could produce unintended proteins. The researchers found that while the wild-type virus contains several such ORFs, modern codon optimization—specifically in the Moderna vaccine—can effectively eliminate these overlapping genes, though some remain in the Pfizer sequence.
What Problem Does This Paper Address?
In biology, we often learn the "one gene, one protein" rule, but viruses are masters of efficiency. They frequently use "overlapping genes," where a single stretch of RNA codes for different proteins depending on where the cellular machinery starts reading the sequence (the reading frame).
When scientists designed COVID-19 vaccines, they used the genetic sequence for the viral spike protein. The concern addressed here is whether these vaccine sequences accidentally include "hidden" instructions for other viral accessory proteins, such as ORF2b or ORF-Sh. If these unintended proteins are translated in the human body, they could potentially trigger unforeseen biological responses or contribute to side effects.
Methods
The authors employed a suite of computational tools to compare the wild-type (WT) SARS-CoV-2 spike sequence with the publicly available sequences for the Moderna (mRNA-1273) and Pfizer-BioNTech (BNT162b2) vaccines.
- ORF Detection: They used the NCBI ORFfinder to identify potential protein-coding regions within the vaccine sequences, setting a minimum threshold of 75 nucleotides.
- Codon Optimization Analysis: They calculated the Codon Adaptability Index (CAI) to measure how well the vaccine sequences match the "preferred" genetic language of human cells.
- Structural Bioinformatics: Since little is known about the "hidden" proteins ORF2b and ORF-Sh, the team used 3D modeling and structural similarity searches (like TM-score) to predict their likely functions based on their predicted shapes.
Datasets
The study relied on three primary genetic sequences:
- Wild-type SARS-CoV-2: The reference genome (NCBI accession: NC_045512).
- Moderna mRNA-1273: Sourced from publicly available reverse-engineered data.
- Pfizer BNT162b2: Sourced from publicly available reverse-engineered data.
- Protein Data Bank (PDB): Used for structural comparisons of predicted proteins.
Key Insights and Ideas
- The Power of Codon Optimization: Codon optimization is usually done to speed up protein production. However, this paper highlights its role as a "genomic filter." By changing the nucleotide sequence (while keeping the spike protein the same), designers can intentionally or unintentionally break the "start" signals for overlapping ORFs.
- Moderna vs. Pfizer: The Moderna sequence appears more "highly optimized" in this specific context. It contains zero predicted overlapping ORFs on the positive sense strand. In contrast, the Pfizer sequence still contains eight small potential ORFs.
- Potential Function of Hidden Genes: The structural modeling suggested that ORF2b might act as a "viroporin" (a channel-forming protein that sits in the cell membrane), while ORF-Sh resembles a DNA-binding "zinc finger" protein. This suggests that if these were translated, they wouldn't be "junk" peptides but could have active biological roles.
Key Results
- Sequence Identity: The wild-type spike sequence is only 68.7% identical to the Moderna sequence and 45.3% identical to the Pfizer sequence at the nucleotide level, despite coding for nearly the same protein.
- Translation Potential: The Moderna vaccine achieved a CAI of 0.981 (nearly perfect alignment with human codon usage), compared to 0.715 for Pfizer and 0.703 for the wild-type virus.
- ORF Elimination: On the positive strand (the one actually translated from mRNA), the wild-type sequence has several overlapping ORFs. The Moderna sequence successfully eliminated all of them, while the Pfizer sequence retained several small putative ORFs (26–52 residues long).
Limitations and Future Work
The primary limitation is that this study is entirely in silico (computational). While the software predicts these ORFs could be translated, it doesn't prove they are translated in a living human cell.
The authors emphasize the need for experimental validation using techniques like:
- Ribosome Profiling (Ribo-seq): To see exactly which parts of the vaccine mRNA are being read by ribosomes.
- Mass Spectrometry: To detect if these small "hidden" peptides actually exist in the cells of vaccinated individuals.
- DNA Vaccine Scrutiny: The authors note that DNA-based vaccines (like INOVIO’s) need even more scrutiny because they involve a negative strand that could also harbor hidden ORFs.
Summary updated