Lane 1: Control without template; Lane 2: Amplification of standard template, without dZTP and dPTP; Lane 3: Amplification of standard template, with dZTP and dPTP; Lane 4 and 6: Amplification of synthetic template, without dZTP and dPTP; Lane 5 and 7: Amplification of synthetic template, with dZTP and dPTP. dNTPs (0.1 mM for each), dZTP (0.05 mM) and dPTP (0.6 mM). M: 25 bp marker. See Methods and Materials for PCR conditions.

Standard template (Bsp-G, Table 1) and synthetic template (Bsp-P, Table 1) were amplified using Taq DNA polymerase under 1x Thermopol buffer (pH 8.0), followed by endonuclease digestion(Bsp120I). dA,T,G/TPs = 0.1 mM, dCTP = 0.4 mM, dZTP = 0.05 mM, and dPTP = 0.6 mM.
Lane 1 and 2: Standard template was amplified 10⁴ fold using Taq, without (lane 1) and with (lane 2) dZTP and dPTP;
Lane 3, 4, and 5: Synthetic template, 10³ (lane 3), 10⁴ (lane 4), and 10⁵ (lane 5) fold amplification, with both dZTP and dPTP;
Not digested: indicates the fraction of PCR product retained Z:P pair, therefore, resisted endonuclease digestion;
Digested: indicates the fraction of PCR product was digested. See Methods and Materials for details.

Left column: Matched C:G, Z:P, and T:A pairs all fit the Watson-Crick geometry (a small pyrimidine analog with one ring complements in size a large purine analog with two rings and are all joined by three hydrogen bonds). Electron density presented to the minor groove is represented as shaded green lobes. Right column: Mismatched C:P, Z:G, and T:P pairs. Note clashes between electrons (gray lobes) or hydrogens, which can be mitigated by protonation/deprotonation, respectively.

a. Measuring the “forward” mutation converting C:G pairs into Z:P pairs using digestion with the Bsp120I restriction endonuclease. Standard oligonucleotide (Bsp-G, Table 1) containing the Bsp120I recognition sequence (5′-GGGCCC-3′) were 1000 fold amplified using Taq DNA polymerase at pH = 8.8 or 8.0 with standard dNTPs (0.2 mM of each) with or without dZTP and dPTP. Then, PCR amplicon were digested by endonuclease (Bsp120I).
Lane 1: In the absence of both dZTP and dPTP;
Lane 2: With dPTP (0.2 mM);
Lane 3: With dZTP (0.2 mM);
Lane 4: With both dZTP and dPTP (0.2 mM for each),
Not digested: indicates the fraction of PCR product resisted endonuclease digestion;
Digested: indicates the fraction of PCR product was digested. See Methods and Materials for PCR conditions.
b. Measure of the “reverse” mutation of Z:P to give C:G and T:A using restriction endonuclease. Two complementary synthetic templates (Bsp-Z and Bsp-P, Table 1) containing 5′-GGGCCZ-3′ and 3′-CCCGGP-5′, were 1000 fold amplified using Taq with standard dNTPs (200 μM), dZTP and dPTP (with various concentration, lane 1 (20 μM), lane 2 (10 μM), lane 3 (5 μM)), then, PCR amplicon were digested by endonuclease (Bsp120I). See Methods and Materials for details.
c. Measuring the mutation of Z into C and T (left) and P into A and G (right) using restriction endonuclease digestion.
Single-stranded synthetic oligonucleotide containing either 5′-GGGCCZ-3′ (left, lane 1, Bsp-Z) or 3′-CCCGGP-5′ (right, lane 2, Bsp-P), was 1000 fold amplified using Taq with only four standard dNTPs (0.2 mM) in 1x ThermoPol reaction buffer (pH 8.8 at room temperature). Then, PCR amplicon was digested by endonuclease (Bsp120I). See Methods and Materials for details.
d. Observed pathways of mutation between non-standard nucleotides and standard nucleotides. Conversion of C:G pair to Z:P pair (forward mutation) involves mis-incorporation of dZTP opposite template-G to form Z:G mismatch at high pH (8.8) (Fig. 3a);
Conversion of Z:P pair to C:G pair (reverse mutation) involves two pathways: 1) the most likely pathway, mis-incorporation of dGTP opposite template-Z to form G:Z mismatch at high pH (8.8) (Fig. 3b); 2) mis-incorporation of dCTP opposite template-P to form C:P mismatch (Fig. 3c, lane 2);
Conversion of Z:P pair to T:A pair (reverse mutation) involves two pathways: 1) the most likely pathway, mis-incorporation of dTTP opposite template-P to form T:P mismatch; 2) misincorporation of dATP opposite template-Z to form A:Z mismatch (Fig. 3c).
See Figure 1 for corresponding matched and mismatched base pairs.

(a) The positions of Z and P in an amplicon are inferred by a process that converts Z:P pairs into a mixture of T:A pairs and C:G pairs, followed by standard Sanger sequencing. Comparison of the resulting sequences shows only T:A or C:G pairs at sites where T:A or C:G pairs were present in the initial amplicon, but mixtures of T:A and C:G pairs at sites where Z:P pairs were present in the initial amplicon.
(b) Manipulation of the concentrations of dPTP without dZTP allows stepwise conversion of Z:P pairs into C:G pairs or into T:A pairs.