Malyshev et al. first optimized conditions to increase the extracellular stability of synthetic nucleobases so that the half-lives of both d5SICS and dNaM were increased to ~9 hours. For successful replication, the unnatural nucleobases must be present inside the cell, so the authors then transformed Escherichia coli to contain an algal nucleotide triphosphate transporter (NTT2), which imports d5SICS and dNaM through the cell membrane. They also transformed the cells with an engineered plasmid (pINF) in which the dA·dT base pair at position 505 was replaced with a base pair between dNaM and an analogue of d5SICS, so that later measurements of d5SICS would confirm the in vivo replication of the plasmid.
After 15 hours of growth (~24 doublings) in medium containing the synthetic nucleobases, pINF was recovered from the cells and digested into nucleosides. To measure the level of UBP retention, the ratios of nucleosides (as identified from liquid chromatography–tandem mass spectrometry) were compared, which indicated the presence of ~1 UBP per plasmid. Additionally, PCR and sequencing of the region verified the retention of the UBP in >95% of the E. coli population. Both of these methods independently demonstrate in vivo replication of the semi-synthetic plasmid (the error rate of which was comparable to that seen in some natural systems) and confirm that the UBP had not been excised by a DNA repair pathway, which shows that living cells can replicate with an expanded genetic code.
This is a preview of subscription content, access via your institution
