Construction of l-deoxynucleoside-Bearing Plasmids for DNA Replication.
The aim of this study was to elucidate comprehensively the impact of l-nucleosides on the bypass and fidelity of DNA replication in vitro and in E. coli cells. To this end, we first synthesized a series of site-specific l-nucleoside-carrying ODNs for each type of l-nucleosides and a complement strand COM20 (Table 1). For example, nucleoside T, dG, dC, dA on PstI recognition sequence within STD strand (5'-AATTCTCCTGCAGGCTAGCACTGA-3') was substituted by l-dG, respectively, of which “lesion-containing strands” were named G13 (5'-AATTCTCCGLGCAGGCTAGCACTGA-3'), G14 (5'-AATTCTCCTGLCAGGCTAGCACTGA-3'), G15 (5'-AATTCTCCTGGLAGGCTAGCACTGA-3') and G16 (5'-AATTCTCCTGCGLGGCTAGCACTGA-3'). Then, the complement strand (COM20) was annealing with lesion-containing strands to give original lesion-containing double strand, and a C:C mismatch was inserted around the recognition site of PstI-HF of the double strand, which was to distinguish the replication products of lesion-containing strands from the natural strands as noted previously21. To assess the bypass efficiencies and mutation frequencies of the l-nucleosides, we ligated the aforementioned inserts containing l-nucleosides into pUC19 genome and performed a restriction enzyme–mediated assays to examine how these lesions inhibit DNA replication and induce mutations in vitro and in E. coli cells (Fig. 1). Prior to assessing the cytotoxicity and fidelity induced by l-nucleosides, the digestion products corresponding to 24-mer ODNs cleaved by EcoRI and HindIII restriction endonuclease were detected, indicating successful construction of l-nucleosides-containing genomes during ligation process. Then, we employed two restriction enzymes, i.e. PstI and HindIII, to digest the PCR products of the progeny genome, affording 20mer ODN fragments from the lesion-containing or control genome (STD) (Fig. 1). The released ODNs were subjected to denature PAGE analyses to identify the replication products, as described elsewhere21. As illustrated in Fig. 1, in this vein, if the correct nucleotide opposite l-nucleosides was incorporated during DNA replication, which restore PstI-HF recognition sequence. PstI-HF restriction enzyme was allowed for the selective digestion of the progeny genomes emanating from the replication of the strand that initially contained the l-nucleosides due to C:C mismatch21. In this respect, sequential digestion of l-nucleoside-containing strands with HindIII-HF and PstI-HF would give a 20mer fragment (5'-AGCTTCAGTGCTAGCCTGCA-3') on 20% denature PAGE gel if incorporation of nucleotides opposite l-nucleoside restored the PstI-HF recognition sequence. On the contrary, no band was present corresponding to 20mer fragment due to disrupting PstI-HF recognition sequence. To take a concrete example, replication products (5'-CTGCAG-3') were found to be further digested by PstI-HF when dA was incorporated opposite l-dG within G13 (5'-CGLGCAG-3') during DNA replication. However, the corresponding product harboring an l-dG→dC, dA or dG mutation at the lesion site, i.e. 5'-CCGCAG-3', 5'-CAGCAG-3' or 5'-CGGCAG-3', could not be digested by PstI. Similarly, for G14 (5'-CTGLCAG-3'), G15 (5'-CTGGLAG-3') and G16 (5'-CTGCGLG-3'), restriction digestion would occur when only dC, dG and T were incorporated opposite l-G, respectively. Thus, with the combination of the two enzyme digestion procedures, we were able to distinguish unequivocally the four potential types of replication products, and by monitoring the products from the strand of 20mer ODNs (5'-AGCTTCAGTGCTAGCCTGCA-3'), we could quantify the bypass efficiency and mutation frequencies.
The bypass efficiencies of the l-nucleosides were then calculated from the ratio of the combined intensities of bands observed for the 20mer products from the lesion-containing genome over the intensity of the 20mer product from the control genome STD. With the use of this method, we were able to determine quantitatively the degrees to which the l-dC, l-dG and l-dA lesions inhibit DNA replication and induce mutations in E. coli cells. Although PstI-digestion can be conducted for pUC19 self-ligation plasmid, only 16-mer band on PAGE gel will not interfere assessment of incorporation frequency and bypass efficiency.
The bypass efficiencies and mutation frequency were assessed by our previous method21. Simply, the bypass efficiencies for the l-nucleoside-carrying genomes were then normalized against that for the control genome. Then, the bypass efficiencies of the l-nucleoside were calculated by comparing intensities of 20mer bands observed from the l-nucleoside-containing genome over the intensity of the control genome. The mutation frequency of l-nucleosides was obtained based on above bypass efficiencies, calculating by ratio each bypass efficiency (i.e. l-dG→T) to total bypass efficiency, i.e. l-dG→T, dG, dC and dA. Thus, by monitoring the products and evaluating bypass efficiency and mutation frequency from the replication of l-nucleosides-containing strand, we were able to study the effect of l-nucleosides on DNA replication in vitro and in cells20,21.
Previously, we constructed four l-T-containing recombinants using T13 (5'-CTLGCAG-3'), T14 (5'-CTTLCAG-3'), T15 (5'-CTGTLAG-3') and T16 (5'-CTGCTLG-3') to assess the bypass efficiencies and mutation frequencies of the l-T lesions in vitro and in cells. By a restriction enzyme-mediated assay provided by us, we showed the l-T was bypassed in E. coli cells (99%), and the bypass efficiency in cells was higher than Taq DNA polymerases (71%) and Vent (exo−) DNA polymerases (47%). In addition, the cell replication across l-T lead to l-T→dA (13%), T (22%), dG (19%) and dC (46%) mutations21. Next, we hope to gain a comprehensive understanding of the impact of l-deoxynucleotides on DNA replication in vitro and in E. coli cells.
Impacts of l-dC on the bypass efficiency and mutation frequency of DNA replication in vitro and in E. coli cells
To understand the impact of l-dC lesions on DNA replication in vitro and in E. coli cells, we constructed four l-dC-containing recombinants using C13 (5'-CCLGCAG-3'), C14 (5'-CTCLCAG-3'), C15 (5'-CTGCLAG-3') and C16 (5'-CTGCCLG-3') to assess the bypass efficiencies and mutation frequencies of the l-dC lesions. In this assay, DNA containing l-nucleosides was directly replicated by Taq DNA polymerase and Vent (exo−) DNA polymerases for in vitro study. For study of cell replication, the recombinants containing l-nucleosides were transformed into E. coli cells, following by PAGE analysis for digested PCR products of plasmid genome.
Two restriction enzymes, HindIII and EcoRI, were employed to digest replication products in vitro and in cells from four lesion-containing genomes, only 24mer bands was observed on denature PAGE gel, which indicated successful construction of l-C-bearing plasmids (Fig. 3). In addition, PCR products from recombinants containing l-nucleosides were digested with EcoRI and HindIII, no other bands (> 24-mer or < 24 mer ODNs) were observed on PAGE gel, which indicated no nucleotide insertion and deletions either in vitro or in E. coli cells.
The DNA fragments corresponding to 16-mer restriction products digested by HindIII and PstI were detectable, likely due to fragment generation by the self-ligated pUC19 plasmid. As expected, PCR products of STD (Fig. 2a, lane 2; Fig. 2b, lane 6 and Fig. 2c, lane 2) were digested by PstI and HindIII to produce 20mer fragments. Because the PstI digestion recognition sequence within insert sequence was not scrambled, so it can be digested by PstI and HindIII to generate products correspond to 20mer fragments. In contrast, PstI recognition sequence with one nucleoside substituted (50, 51 and 53) will not be digested by PstI, which can't produce 20mer ODNs (Fig. 2b, lane 2, 4 and 8) on PAGE gel due to disorder of its PstI recognition sequence.
Figure 2a showed that replication products from recombinants, of which T, dG, dC and dA on STD (5'-CTGCAG-3') was replaced by l-dC, was digested by PstI and HindIII under catalysis of Taq DNA polymerase. If dA, dC, dG and T were incorporated opposite l-dC of C13, C14, C15 and C16, respectively, its replication products would be digested by PstI due to restores of its recognition sequences, and thus 20mer band on PAGE gel was observed. Actually, no 20mer fragment was detectable on gel for all the above experimental groups (Fig. 2a, lane 4, 6, 8 and 10), showing dA, dC, dG and T was not incorporated opposite l-dC during DNA replication. These results indicated that the replication of DNA containing l-dC was inhibited by Taq DNA polymerase.
Different from what happened on Taq DNA polymerase, a 20mer oligonucleotide was observed for C15 (Fig. 2b, lane 7) when catalyzed by Vent (exo−) DNA polymerase, which indicated l-dC paired with dG in replication products. No bands corresponding to 20mer fragments were present for C13, C14 and C16 (Fig. 2b, lane 3, 5 and 9), showing no dA, dC and T was incorporated opposite l-dC during DNA replication. Prior to normalizing band intensities of the 20mer fragment to that of the 24-mer fragment, the bypass efficiencies of the l-dC lesions were then calculated from the ratio of the combined intensities of bands observed for the 20mer products from the lesion-containing genome over the intensity of the 20mer product from the control genome (STD). And incorporation frequencies were calculated from the percentage of a mutagenic product relative to all possible products from the control genome. The bypass efficiency of 39% was obtained (Supplementary Table S1 and S2) and l-dC induced no mutagenic replication catalysed by Vent (exo−) DNA polymerase.
For replication in E. coli cells, the result was same as in vitro replication catalysed by Vent (exo−) DNA polymerase. Only one band corresponding to 20mer fragment was obtained for C15 (Fig. 3c, lane 8), which indicated l-dC: dG pairing formed in E. coli cells, with bypass efficiency being 81% (Supplementary Table S1 and S2). No bands corresponding to 20mer fragment were present for C13, C14 and C16 (Fig. 2c, lane 4, 6 and 10), showing no dA, dC and T was incorporated opposite l-dC in cell replication.
Hence, these results revealed that l-dC completely block the DNA replication machinery under catalysis of Taq DNA polymerases, whereas l-dC can be bypassed by Vent (exo−) DNA polymerases or in E. coli cells. The bypass efficiency in E. coli cells (81%) was higher than that obtained under catalysis of Vent (exo−) DNA polymerases (39%). What's more, no mutagenic replication was found for l-dC-containing DNA in vitro and in E. coli cells.
Impacts of L-dG on the bypass efficiency and mutation frequency of DNA replication in vitro and in E. coli cells
To assess the bypass efficiencies and mutation frequencies of the l-dG lesions, four l-G-containing recombinants were parallelly constructed by incorporation of G13 (5'-CGLGCAG-3'), G14 (5'-CTGLCAG-3'), G15 (5'-CTGGLAG-3') and G16 (5'-CTGCGLG-3').
As shown in Fig. 3, 24mer bands on denature PAGE gel corresponding to the digested products by HindIII and EcoRI were present, indicating successful construction of l-G-bearing plasmids. In addition, undesirable 16-mer ODNs from PstI and HindIII digestion was observed, suggesting there was self-ligated pUC19 plasmid. For positive restriction endonuclease assay, 20mer fragments from PstI and HindIII digested PCR products of STD (Fig. 3a, lane 2; Fig. 3b, lane 2 and Fig. 3c, lane 4) were obtained, as expected.
However, for the reaction system catalysed by Taq DNA polymerases, a 20-mer oligonucleotide band was found for G14 (Fig. 3a, lane 6) when the replication products were incubated with PstI and HindIII, which suggested that replication of G14 with replacing G with l-dG resulted in l-dG: dC pairing. Moreover, l-dG induced l-dG→dG substitution with bypass efficiency of 19% (Supplementary Table S1 and S2). By comparation, 20mer band was absent on gel for G13, G15 and G16 (Fig. 3a, lane 4, 8 and 10), in which l-dG was substituted for T, C and A of 5'-CTGCAG-3', respectively. The result indicated that no dA, dG and T was incorporated opposite l-dG during DNA replication. That is, l-dG did not result in a mutation in the reaction system catalysed by Taq DNA polymerases.
In the reaction system catalysed by Vent (exo−) DNA polymerases, 20-mer DNA fragments of digested replication products can be detected in recombinant plasmids G14, where l-dG substituted for G of 5'-CTGCAG-3' (Fig. 3b, lane 6), which indicated that dC was incorporated opposite l-dG. Additionally, 20-mer DNA fragments was observed for G13 (Fig. 3b, lane 4), indicated that dA was incorporated opposite l-dG of 5'-CGLGCAG-3', that is, l-dG induced l-dG→T transversion mutation besides l-dG→G. The bypass efficiency of l-dG was 12% and the mutation frequency of l-dG→T was 65% (Supplementary Table S1 and S2). In contrast, no bands corresponding to 20mer fragments was found for G15 and G16 (Fig. 3b, lane 8 and 10) proved that dG and T was not incorporated opposite l-dG during DNA replication catalysed by Vent (exo−) DNA polymerase.
For replication investigation in E. coli cells, only in the case of G14 was a 20-mer oligonucleotide band found on gel (Fig. 3a, lane 6) when the replication products were incubated with PstI and HindIII. This indicated that replication of DNA containing l-dG resulted in l-dG: dC pairing, with bypass efficiency being 82% (Fig. 3c, lane 5; Supplementary Table S1 and S2), whereas dA, dG and T was not incorporated opposite l-dG during DNA replication (Fig. 3c, lane 3, 7 and 9).
Hence, for study on l-dG-containing DNA replication, Taq DNA polymerase and Vent (exo−) DNA polymerase were inhibited by l-dG, and the bypass efficiencies were decreased to 19% and 12%, respectively. But, in cell replication assay resulted in apparent increase in bypass efficiency (82%) relative to in vitro replication. Mutagenic replication was found in vitro but not in cells, which may attribute to repair machinery and related proteins in cells.
Impacts of l-dA on the bypass efficiency and mutation frequency of DNA replication in vitro and in E. coli cells
We constructed four l-dA-containing recombinants using A13 (5'-CALGCAG-3'), A14 (5'-CTALCAG-3'), A15 (5'-CTGALAG-3') and A16 (5'-CTGCALG-3') to assess the bypass efficiencies and mutation frequencies of the l-dA lesions. Figure 4 indicated successful l-A-bearing plasmids and self-ligated pUC19 plasmid. As expected, PstI and HindIII digestion products corresponding to 20mer ODNs was obtained for STD (Fig. 4a, lane 8; Fig. 4b, lane 2 and Fig. 4c, lane 8) revealed that the PstI digestion recognition sequence within insert sequence was not destroyed. In contrast, no 20-mer ODNs was observed for 58, 59 and 60 (Fig. 4a and 4c, lane 2, 4 and 6) which resulted from destroyed PstI recognition sequence.
The effect of L-dA on DNA replicative bypass and fidelity was investigated by Taq and Vent (exo−) DNA polymerases as well as in cells. Figure 4a showed experimental results in the case of Taq polymerase. Replication products of A13 and A16 were digested by restriction endonuclease PstI-HF and HindIII-HF to produce 20-mer fragments (Fig. 4a, lane 3 and lane 9), whereas no 20mer restriction-digested PCR fragments of A14 and A15 were detected (Fig. 4a, lane 5 and lane 7). This indicated that dA and T rather than dC and dG were incorporated opposite l-dA during DNA replication. Furthermore, Taq DNA polymerase bypassed l-dA with a low bypass efficiency of 16%. l-dA paired with T with frequency being 63% and induced l-dA→T transversion at a frequency of 37% (Supplementary Table S1 and Table S2).
Different from Taq DNA polymerase, only one band corresponding to digested 20-mer DNA fragment of PCR product was detected in recombinant plasmids A13 containing l-dA lesion when catalysed by Vent (exo−) DNA polymerases (Fig. 4b, lane 4), indicate that l-dA was strongly miscoding, and it induced completely l-dA→T transversion mutation catalysed by Vent (exo−) DNA polymerases, with bypass efficiency being 20% (Supplementary Table S1 and Table S2).
In the study of the replication of DNA containing l-dA in cells, PCR products of A14, A15, and A16 were digested by restriction endonuclease PstI-HF and HindIII-HF to produce 20-mer fragments (Fig. 4c, lane 5, lane 7 and lane 9), verifying the l-A→dA, dG, and dC mutation occurred during cell replication, and l-dA did not induce l-dA→T mutation (Fig. 4c, lane 3). In addition, the results showed that bypass efficiency of l-dA in E. coli cells was up to 74%, which was 3–5 times as high as that for Taq DNA polymerases (16%) and Vent (exo−) DNA polymerases (20%). Also, L-A→dA, dG and dC with mutation frequency being 67%, 24% and 9% (see Supplementary Table S1 and Table S2 online).