| Abstract: | The proteins coded by bacteriophate T4 replication genes 32, 41, 43, 44, 45, 61, and 62 together can replicate phi X174 DNA templates very efficiently. The fidelity of this in vitro replication reaction has been measured using an infectivity assay. The product molecules have the same specific infectivity as the template DNA. When an amber mutant DNA template is used, no increase in the frequency of revertants is seen even after more than 60 duplications in vitro. By using imbalances in the concentrations of deoxynucleotide substrates, the error rate during DNA replication in vitro can be greatly increased. Control experiments indicate that the increased mutagenesis is not due to the presence of dITP or dUTP as contaminants in the deoxynucleotide substrates used. The increase in the frequency of revertants is linearly related to the ratio of the correct and the incorrect deoxynucleotides. Determination of the DNA sequence of the revertants induced shows that a change in DNA sequence of the amber site predicted from the nucleotide bias occurs. DNA synthesis in vitro resembles in vivo replication in that the error rate depends not only upon the base change required for reversion but also upon the neighboring DNA sequences. The error rate is estimated to be 5 X 10(-6) at am3 site, 6.4 X 10(-7) at am86 site, and less than 2.9 X 10(-7) at am9 site. Comparison of the frequency of G-T and A-C mispairs reveals that most AT leads to GC transition mutations occur through G-T mispairs. Measurement of the frequency of the mispairs required to induce transversion mutations reveals that these occur primarily through purine-purine mispairs. Transition mutations are more frequent than transversion mutations at both the am3 and the am86 sites. These observations support the models for base pairing errors proposed by Watson and Crick ((1953) Nature 171, 964-967) and Topal and Fresco ((1976) Nature 263, 285-289). |
