Sequence determinants for -2 frameshift mutagenesis at NarI-derived hot spots.

The recognition sequence of the NarI restriction enzyme is known to be a strong hot spot for -2 frameshift mutations (G1G2CG3CC-->GGCC) induced by the chemical carcinogen N-2-acetylaminofluorene (AAF). In an attempt to define a "consensus sequence" for this mutation hot spot, we have investigated the role of the bases flanking the central dinucleotide GpC repeat in the NarI sequence (NaGCGCNb) on the mutation frequency induced by the carcinogen. Construction and random modification with AAF of the 16 plasmids resulting from the replacement of Na and Nb by A,T,G and C, respectively, have been undertaken. All 16 sequences tested are found to be -2 frameshift mutation hot spots. Indeed, a level of modification of approximately five AAF adducts per plasmid molecule induces a mutation frequency ranging between 500 and 5000-fold above background. The mutations observed are mainly (90%) deletion of a dinucleotide CpG in the targeted sequence NaGCGCNb. Previous studies on NarI mutagenesis (G1G2CG3CC-->GGCC) have shown that only AAF adducts at G3 in the template for lagging strand synthesis induce -2 frameshift mutagenesis at a high level. When the mutation data obtained in this work are analyzed as originating essentially from adducts to guanines in this strand, we find that it is the nucleotide Nb located 3' to the central dinucleotide GpC repeat that strongly modulates the mutation frequency, while the nucleotide Na located on the 5' side has little effect. Our present model of frameshift mutagenesis at NarI sites involves a template-primer misalignment step with a two-nucleotide slipped mutagenic intermediate. In the context of this model, the modulation of mutagenesis by nucleotide Nb located two nucleotides 3' from the putative adduct site in the template strand can be explained in view of the fact that the replication complex encounters this nucleotide before it encounters the adduct.