Oligonucleotide Synthesis Errors Are a Source of Untoward Variation in HDR-Mediated Gene Editing.
Publication Year:
2026
PubMed ID:
42510769
Funding Grants:
- Curing Sickle cell Disease with CRISPR-Cas9 genome editing
- Curing Sickle cell Disease with CRISPR-Cas9 genome editing
- University of California, San Francisco (UCSF) CIRM Alpha Stem Cell Clinic
- University of California, San Francisco (UCSF) Alpha Stem Cell Clinic (ASCC)
- Transplantation of CRISPR-CAS9 Corrected Hematopoietic Stem Cells (CRISPR_SCD001) in Patients with Severe Sickle Cell Disease
Public Summary:
Single-stranded oligonucleotides (ssODNs) are used as donor templates for gene editing by targeted endonuclease cleavage and homology directed repair (HDR). We probed their sequence fidelity by deep sequencing ssODNs from three manufacturers, and genomes derived from editing using these ssODNs as templates for HDR. The ssODNs carry single-nucleotide and small deletion synthesis errors in proportions differing widely among manufacturers; these are propagated into the genome by HDR; therapeutic gene correction strategies will need to account for this additional source of editing errors.
Scientific Abstract:
Background/Objectives: Single-stranded oligonucleotides (ssODNs) are used as donor templates for therapeutic gene editing by CRISPR-Cas9 cleavage and homology-directed repair (HDR). Although ssODN sequence fidelity is critical to the safety and efficacy of editing, standard quality control methods cannot resolve individual nucleotide errors. Methods: We performed deep sequencing of ssODNs from three manufacturers and amplicons from edited hematopoietic stem/progenitor cells. Results: We find that synthesis errors are present in all ssODNs tested at rates that vary more than two-fold among manufacturers, at positions that are dependent on sequence context. These synthesis errors are propagated into the genome by HDR at frequencies proportional to their abundance in the ssODN. In our sickle cell mutation correction protocol, the most prevalent SNEs are predicted to produce benign beta-globin variants, while the less frequent frameshift deletions are predicted to generate beta-thalassemia-like alleles. Conclusions: Current quality control standards are insufficient to detect these errors, and deep sequencing of ssODNs should be incorporated into regulatory submissions for clinical gene editing programs.