Quantification of synthetic errors during chemical synthesis of DNA and its suppression by non-canonical nucleosides.

Substitutions, insertions, and deletions derived from synthetic oligonucleotides are the hurdles for the synthesis of long DNA such as genomes. We quantified these synthetic errors by next-generation sequencing and revealed that the quality of the enzymatically amplified final combined product depends on the conditions of the preceding solid phase chemical synthesis, which generates the initial pre-amplified fragments. Among all possible substitutions, the G-to-A substitution was the most prominently observed substitution followed by G-to-T, C-to-T, T-to-C, and A-to-G substitutions. The observed error rate for G-to-A substitution was influenced by capping conditions, suggesting that the capping step played a major role in the generation of G-to-A substitution. Because substitutions observed in long DNA were derived from the generation of non-canonical nucleosides during chemical synthesis, non-canonical nucleosides resistant to side reactions could be used as error-proof nucleosides. As an example of such error-proof nucleosides, we evaluated 7-deaza-2´-deoxyguanosine and 8-aza-7-deaza-2´-deoxyguanosine and showed 50-fold decrease in the error rate of G-to-A substitution when phenoxyacetic anhydride was used as capping reagents. This result is the first example that improves the quality of synthesized sequences by using non-canonical nucleosides as error-proof nucleosides. Our results would contribute to the development of highly accurate template DNA synthesis technologies.