An editing-defective aminoacyl-tRNA synthetase is mutagenic in aging bacteria via the SOS response

Abstract

Mistranslation in bacterial and mammalian cells leads to production of statistical proteins that are, in turn, associated with specific cell or animal pathologies, including death of bacterial cells, apoptosis of mammalian cells in culture, and neurodegeneration in the mouse. A major source of mistranslation comes from heritable defects in the editing activities of aminoacyl-tRNA synthetases. These activities clear errors of aminoacylation by deacylation of mischarged tRNAs. We hypothesized that, in addition to previously reported phenotypes in bacterial and mammalian systems, errors of aminoacylation could be mutagenic and lead to disease. As a first step in testing this hypothesis, the effect of an editing defect in a single tRNA synthetase on the accumulation of mutations in aging bacteria was investigated. A striking, statistically significant, enhancement of the mutation rate in aging bacteria was found. This enhancement comes from an increase in error-prone DNA repair through induction of the bacterial SOS response. Thus, mistranslation, as caused by an editing-defective tRNA synthetase, can lead to heritable genetic changes that could, in principle, be linked to disease.

Fig. 1.

MAC of ileS_Ala editing-defective and wild-type E. coli. Frequencies of spontaneous mutants that had acquired resistance to Rif were determined for >60 replicates of each genotype after 1 and 7 days of growth on LB. (A) Shown are the median Rif^R frequencies for day 7 relative to those of day 1 for wild-type and editing-defective strains. (B and C) Each frequency observation was converted to its log value, binned, and counted. Shown is the fraction of observations in each bin. (B) No differences were observed between wild-type and ileS_Ala strains after 1 day (Mann–Whitney U test; n = 67 and 68 for wild-type and editing-defective strains, respectively; P = 0.8). (C) After 7 days, the frequency of spontaneous Rif^R had increased in both wild-type and editing-defective strains. The increase in Rif^R frequency was significantly greater in the editing-defective strain (Mann–Whitney U test; n = 69 for each strain; P < 0.05). See Results for further discussion of the distribution of day-7 Rif^R mutants.

Fig. 2.

Mutations in rpoB. Forty-eight Rif^R mutants of editing-defective and wild-type strains were isolated and sequenced after 1 and 7 days of incubation on LB. Because of isolation of clones from the same plates, the number of unique mutations were 39 and 29 for editing-defective strains and 38 and 30 for the wild-type strains, on days 1 and 7, respectively. (A) The frequencies of each type of transversion and transition are shown, as well as one in-frame codon insertion. The GC → AT mutation was the major type of genetic change identified. GC → TA mutations increased 2-fold at day 7 in both strains. Few differences were observed between the specific mutations isolated after 1 day (Left) and 7 days (Right). (B) The frequencies of amino acid identity changes are shown mapped onto the primary sequence of the RNA pol protein. Bars extending upward represent mutations present after 1 day, whereas those extending downward represent mutations present after 7 days. Although there were no mutations at position R529 after 1 day, the majority of mutations identified after 7 days were at this residue. Included were mutations to C, V, and H in both strains, as well as S in the editing-defective strain.

Fig. 3.

Ci-induced mutations in the editing-defective and wild-type strains. Editing-defective and wild-type strains were exposed to Ci to determine the extent of SOS response-induced mutagenesis in each strain. (A) The cumulative frequency of Ci^R mutants after incubation on minimal media plates plus Ci is shown for editing-defective and wild-type strains. (B) The MICs of clones resistant to Ci from day 2 (preexposure) and days 5–7 (postexposure) reveal a functional difference in Ci^R mutations acquired before and after exposure to Ci. Shown is the relative MIC of preexposure and postexposure mutants [i.e., log₂ (mean preexposure MIC) − log₂ (mean postexposure MIC)]. Error bars represent SEM. (C) The ileS_Ala editing-defective mutation was combined with lexA(wt)-Kan^R and lexA(S119A)-Kan^R. The latter mutation inhibits autocleavage of LexA, making the mutant protein a constitutive repressor and preventing induction of the SOS response. Shown is the accumulation of Ci^R mutants in editing-defective and editing-defective/SOS response-inhibited strains after incubation on minimal media plus Ci.