Evidence for a unique first position codon-anticodon mismatch in vivo

Abstract

The Ser68(AGC) codon of the beta-lactamase gene was changed to the glycine codons GGA and GGC. With glycine at position 68, beta-lactamase is inactive because it does not have a nucleophilic side-chain to function in the reaction mechanism. The mutant SG68(GGA) allele had no detectable beta-lactamase activity; however, the mutant SG68(GGC) did produce a small amount of activity. Both mutant alleles produce comparable amounts of beta-lactamase protein in a maxi-cell system. To identify why these two "same-sense" beta-lactamase mutants differ phenotypically, we introduced the alleles into Escherichia coli strains with mutations that affect translational fidelity. The rpsD mutation, which decreases fidelity, significantly increased activity with the SG68(GGC) allele, while the rpsL mutation, which increases translational fidelity, had little effect on the beta-lactamase activity. The rpsD and rpsL alleles had no effect on the SG68(GGA) allele. From the allele specificity of the activity produced by the bla mutants, and from the differential effect of translational fidelity on the activity of the SG68(GGC) allele, we infer that tRNA(GCU)Ser, the AGU/C reading tRNA(Ser), mistranslates SG68(GGC) at a frequency of about 0.1%, and subsequently produces active beta-lactamase. This is the first observation of an A/G wobble with a wild-type tRNA at the first position of the codon-anticodon interaction.