Fidelity of replication of repetitive DNA in mutS and repair proficient Escherichia coli

Replication fidelity is not constant among strains within a species or at all genetic loci within a genome. Altered fidelity of replication may affect patterns of pathogenesis and the evolution of these strains. We have been studying replication fidelity in Escherichia coli, both in laboratory attenuated strains and in food-borne pathogens. To understand the altered patterns of mutagenesis at the molecular level, we used a shuttle vector plasmid with a tRNA mutational marker gene which had been altered to include homopolymeric runs of five, seven and nine G:C] pairs, as well as non-repetitive DNA. Replication of the plasmid in mutS strains resulted in a 20-fold increase in mutant progeny plasmids. The mutations were almost all (>90%) frameshift mutations, while base substitution mutations were rare. Most mutations were insertions or deletions of one or two G:C] pairs in the longest homopolymeric runs. Larger deletions (5 to >70bp), also targeted to the repetitive sequence, were likewise common. Mutations increased exponentially with the length of the homopolymeric run. These patterns of mutation, including unexpectedly high levels in repair proficient strains, led to an examination of the E. coli K-12 genome for homopolymeric DNA. This sequence motif was found to be rare, particularly in genes and open reading frames. Amino acid homotrimers were found to avoid usage of homopolymeric codons, even when they are preferred among synonymous codons in E. coli. There appears to be active selection against tandem direct nucleotide repeats in the E. coli genome, correlated with the inability of the organism to accurately replicate such sequence.