On the Fidelity of DNA Replication EFFECT OF METAL ACTIVATORS DURING SYNTHESIS WITH AVIAN MYELOBLASTOSIS VIRUS DNA POLYMERASE*

The effect of metal activators on the fidelity of DNA synthesis has been examined. Using the DNA polymerase from avian myeloblastosis virus, the accuracy of CO”~-, Mn”‘-, and Ni’+-activated DNA synthesis was determined with different polynucleotide templates. With poly[d(A-T)l as the template, the error frequency for dCMP incorpora- tion was 1:1400, 1:1100, and 1:600 for Mg”+, Co”+, and Mn”+, respectively, at maximally activating concentrations. The error frequency was invariant with respect to [Mg”+l but increased with greater than activating concentrations of Coz+ and Mn”+. This increase resulted from differential rates of complementary and noncomplementary nucleotide incorporation. The enhanced error frequency was nonspecific as it occurred with all polynucleotide templates and with all noncomplementary deoxy-and ribonucleotides which were tested. Nearest neighbor analyses of the reaction products indicated that the noncomplementary deoxynucleotides were incorporated as single ba>e substitutions. The fidelity of Ni’--activated DNA synthesis was invariant with respect to

and was similar to that obtained using Mg"' . During DNA synthesis with Mg' I, the addition of Co"+, M&l, or NiYi resulted in a decrease in the fidelity of DN4 synthesis. The relationship between decreases in the fidelity of DNA synthesis and metal mutagenesis, or carcmopenesls, or both, is considered.
All of the DNA polymerases which have been isolated and characterized to date require an added divalent metal cation for catalysis (1,2 (10 to 20 dpm/pmol), 50 PM ("HldGTP (10,000 to 50,000 dpmipmol), 1 pg of polyId(A-T)l, and 0.5 pg of AMV DNA polymerase.
Assays were incubated for 60 min at 37".
Assays were incubated for 60 min at 37". Incorporation of the radioactive deoxynucleotides into an acid-insoluble precipitate was determined after repeatedly precipitating the polynucleotide product with 1.0 N perchloric acid and solubilizing with 0.2 M NaOH as previously described (9). The error frequency was defined as the ratio of the noncomplementary to the total complementary deoxynucleotide incorporated.     (Table III). Poly[d(A-T)l-directed products were synthesized using [a-'"PldGTP as the noncomplementary deoxynucleotide and ["HldTTP as the complementary deoxynucleotide. After enzymatic hydrolysis of the Mg'+-directed reaction, 99% of the :FzP from the noncomplementary deoxynucleotides was found to migrate coincident with dTMP after electrophoresis. In contrast, the :'H-labeled thymidine in dTTP was recovered as dTMP. Similar results were obtained with the product from reactions using Co"; less than 1% of the noncomplementary deoxynucleotides were adjacent to another non-   (Table IV). Using the template primer, poly(C) oligo(dG), the accuracy of synthesis was dependent on the Co'+ concentration. These results suggest that the capacity of cobalt to increase mistakes by AMV DNA polymerase also occurred with polyribonucleotide templates and that the primary effect of greater than activating cobalt concentrations was the preferential inhibition of complementary nucleotide incorporation.

Metal
To demonstrate that AMV DNA polymerase incorporated noncomplementary deoxynucleotides into phosphodiester linkage in DNA using a polyribonucleotide template, nearest neighbor analysis was performed on the product reaction using Co'+ or Mg"+ (Table V)  The second product was made in an identical assay mixture which contained 20 ELM dATP and 20 ELM La-:"PldGTP (164 dpmipmol).
The isolation and hydrolysis of the product with micrococcal nuclease and spleen phosphodiesterase were performed as previously described (4)  tion of noncomplementary deoxynucleotides which were introduced into the product as single base substitutions.
The molecular mechanisms which determine the decreased fidelity of DNA synthesis with Co*+ or Mn*+ are unknown. AMV DNA polymerase lacks any detectable 3' -+ 5' proofreading exonuclease activity (4) so that it cannot excise mismatched nucleotides incorporated during polymerization (15,16). Therefore, increases in the error frequencies observed with Co*+ or Mn'+ could not be due to a diminished proofreading capacity. Thus, it may be reasonable to assume that the different metals substitute for Mg*+ at the active site on the enzyme: (a) kinetic analysis of DNA polymerization suggested that AMV DNA polymerase may contain multiple nucleotide binding sites (17) Their results indicate that the enzyme, in the absence of template, adjusts the conformation of the Mn2+-d'M'P so that it is indistinguishable from that of a thymidylate unit in the Watson-Crick double helical structure of DNA. Such a present structure would facilitate complementary base-pairing. In the presence of template, changes in this conformation of the substrate when coordinated with different divalent cations might account for differences in the fidelity of DNA synthesis. In addition, changes in the error frequency at greater than activating metal concentrations could be due to binding of the metal to additional weak sites on the enzyme causing a conformational change in the enzyme structure altering the conformation at the active site. In this regard, it is of interest to note that the interaction of AMV DNA polymerase with Be*+ a nonactivating divalent metal cation, decreased the accuracy of DNA synthesis (19).
It can be argued that the observed increase in error frequency at high metal concentration represented the selective acid precipitation of unincorporated noncomplementary nucleotides. However, nearest neighbor analysis is based on the hydrolysis of the purified product by enzymes of known specificities. These results indicated that the noncomplementary nucleotides were incorporated in phosphodiester linkage and were predominantly distributed in the product as single base substitutions. Also, the high infidelity did not appear to result from selective interaction of metals with particular nucleotides. For example, at high Co2+ concentrations (5 mM) the incorporation of dGTP is markedly inhibited when it is the complementary nucleotide (Table IV, poly(C) as the template), while it is undiminished at the same Co'+ concentration when it is the noncomplementary nucleotide ( Fig. 31, with poly[d(A-T)] as template. These considerations indicated that infidelity using Md+ and Co"+ did not result from selective binary interactions of the metal cations with specific nucleotide substrates.
Physical studies have clearly demonstrated that divalent metal cations interact with the phosphates and bases on polynucleotides (20) suggesting that they can alter base-pairing specificities. Thus, the increased infidelity observed at high concentrations (2 to 10 mm) of Mn2+ and Co'+ could reflect metal interaction with the template facilitating the formation of noncomplementary base pairs during polymerization. However, binding studies do not support this interpretation.
Measurements of the interaction of Mn2+ with activated DNA by electron paramagnetic resonance (21) indicate a stoichiometry of 0.36 Mn"' per DNA phosphorus. For each mole of activated DNA (molecular weight >70,000) there were two tight Mn"+ binding sites and 52 weaker sites having an invariant dissociation constant of 68 PM. The K, for the inhibitory effects of Mn"+ on the incorporation of the correct nucleotide observed in this paper is about 2 mM, a concentration at which all the metal binding sites on the template would be expected to be occupied. Thus, it is unlikely that the high infidelity observed with Mn2+ is brought about by the interaction of metal cations with the template.
Cobalt and nickel are known carcinogens (22). Manganese is a potent mutagen (12,23) and has been recently reported to be carcinogenic (24). It is possible that the alterations in fidelity observed with AMV DNA polymerase using Co'+, Mn2+, or NiZ+ may also occur with eukaryotic DNA polymerases. Furthermore, it may be of interest to note that all of these metals alter the fidelity of DNA synthesis in the presence of saturating concentrations of Mg2+; a situation which could occur in vivo.
during DNA Synthesis