Improving the Fidelity of Replication of a Six-Letter DNA Alphabet

The Watson-Crick-Franklin (WCF) rules describing nucleobase pairing in antiparallel strands of DNA and RNA can be exploited to create artificially expanded genetic information systems (AEGIS) with as many as 12 independently replicable nucleotides joined by six hydrogen bond pairing schemes. One of these additional pairs joins two nucleotides trivially designated as <b>Z</b> (6-amino-5-nitro-(1<i>H</i>)-pyridin-2-one) and <b>P</b> (2-amino-imidazo-[1,2-<i>a</i>]-1,3,5-triazin-(8<i>H</i>)-4-one). The <b>Z</b>:<b>P</b> pair has supported 6-nucleotide PCR to give diagnostics products, in environmental surveillance kits, and for laboratory <i>in vitro</i> evolution (LIVE) that has generated, <i>inter alia</i>, molecules that inactivate toxins, antibody analogs that bind cancer cells, therapeutic candidates that deliver drugs to those cells, reagents to identify targets on those cells' surfaces, reagents to move cargoes across the blood-brain barrier, and catalysts with ribonuclease activity. However, the <b>Z</b> nucleoside is acidic, with a p<i>K</i>_a of ∼7.8. In its deprotonated form, <b>Z</b>^- forms a WCF pair with G. This leads to the slow replacement of <b>Z:P</b> pairs by C:G pairs during PCR or, in the reverse process, their introduction. Here, we examine analogs of <b>Z</b> that retain the same donor:donor:acceptor hydrogen bonding pattern as earlier generations of the <b>Z</b> heterocycle, still form a WCF pair with <b>P</b>, but have a higher p<i>K</i>_a. Experiments with Taq polymerase show that the rate of loss of <b>Z:P</b> pairs decreases markedly as the p<i>K</i>_a of the <b>Z</b> heterocycle increases. This provides direct support for the hypothesis that <b>Z</b>:<b>P</b> pairs are in fact lost via deprotonated <b>Z</b>^-:<b>G</b> mismatches. Further, it provides a <b>Z</b>:<b>P</b> system that can be replicated with very high fidelity, with >97% retention of the <b>Z</b>:<b>P</b> pairs over 10,000-fold amplification.