Global tRNA misacylation induced by anaerobiosis and antibiotic exposure broadly increases stress resistance in Escherichia coli

Abstract

High translational fidelity is commonly considered a requirement for optimal cellular health and protein function. However, recent findings have shown that inducible mistranslation specifically with methionine engendered at the tRNA charging level occurs in mammalian cells, yeast and archaea, yet it was unknown whether bacteria were capable of mounting a similar response. Here, we demonstrate that Escherichia coli misacylates non-methionyl-tRNAs with methionine in response to anaerobiosis and antibiotic exposure via the methionyl-tRNA synthetase (MetRS). Two MetRS succinyl-lysine modifications independently confer high tRNA charging fidelity to the otherwise promiscuous, unmodified enzyme. Strains incapable of tRNA mismethionylation are less adept at growth in the presence of antibiotics and stressors. The presence of tRNA mismethionylation and its potential role in mistranslation within the bacterial domain establishes this response as a pervasive biological mechanism and connects it to diverse cellular functions and modes of fitness.

Figure 1.

Non-methionyl–tRNAs are constitutively misacylated with methionine during anaerobic fermentation. (A) E. coli tRNA microarrays showing tRNA misacylation during anaerobic fermentation, but not during aerobic respiration with glucose. Array map shows the probe locations of E. coli tRNA^fMet and tRNA^eMet. Percent misacylation is calculated by dividing the total signal obtained from a particular tRNA isoacceptor by the total signal from the tRNA^iMet and tRNA^eMet. (B) Semi-quantification of total array spot signals for anaerobic tRNA misacylation to individual tRNA species, averaged from 3 arrays. Variations of individual tRNAs are within 1.5-fold between arrays. The detection limit is ∼0.1% for an individual tRNA. Since the array results are semi-quantitative, we choose to present the data in heat maps rather than precise values with error bars. (C) Array controls to validate ³⁵S-methioine signal from non-methionyl–tRNA probes show exclusion of tRNA^Met cross hybridization with addition of free Met probes, exclusion of ³⁵S tRNA nucleotide thio-modifications by removing aminoacylated ³⁵S-Met through chemical deacylation attached to tRNA, and exclusion of peptidyl–tRNA with aminopeptidase M (APM) treatment of tRNA.

Figure 2.

Other conditions that induce tRNA misacylation and identification of mistranslated peptides. (A) tRNA misacylation during anaerobic respiration with nitrate. (B) Misacylation after 10 μM chloramphenicol was added with the ³⁵S-methioine pulse label. (C) Example fragmentation spectrum indicating leucine-to-methionine mistranslation in peptide TL*LYAINGGVDEK (from TdcE, 2-ketobutyrate/pyruvate-formate lyase), where L2 has been replaced with M, here detected as the sulfoxide. Matched b-ions (red) and y-ions (blue) are indicated by colored peaks in the spectrum and bold numbers in the sequence legend, and site-determining ions for the mistranslation at L2 (b2+ and y11+) are highlighted. (D) Gene names and the peptide locations (in parenthesis) of the Leu-to-Met substituted peptides grouped according to their functions as enzymes, non-enzymes or unknown.

Figure 3.

Succinyl-lysine modifications in E. coli MetRS confer high tRNA charging fidelity to the promiscuous, unmodified enzyme. (A) tRNA microarray showing tRNA misacylation in vitro by the purified, recombinant MetRS. (B) In vitro tRNA charging by purified MetRS that has been pre-incubated with succinyl-CoA, acetyl-CoA or coenzyme A prior to the charging reaction. (C) In vitro tRNA charging by purified MetRS in the presence of succinyl-CoA without preincubation, or in the presence and with preincubation with succinate. (D) Western blot for succinyl-lysine on purified chromosomally his-tagged MetRS that was obtained from E. coli growing aerobically, anaerobically and with antibiotic exposure. MetRS loading control was visualized via Ponceau S straining of the membrane following exposure.

Figure 4.

Succinylation at Lys362 or Lys388 is sufficient to confer high fidelity to E. coli MetRS. (A) Mass spectrometry spectra showing succinylation at Lys362 (top) and Lys388 (bottom) in MetRS during the in vitro succinylation reaction. (B) In vitro tRNA charging by mutant MetRS enzymes containing Glu substitutions at lysine 362 or 388. (C) tRNA charging with Met by WT and mutant MetRSs at 37°C with 100 nM MetRS and 100 μM total E. coli tRNA. (D) In vitro tRNA charging by the mutant MetRS containing alanine substitution at Lys362. (E) E. coli MetRS crystal structure highlighting the regulatory sites 362 (top residue) and 388 (bottom residue) along the putative tRNA binding interface.

Figure 5.

Lysine conservation at the homologous MetRS regulatory sites. Amino acid identity of homologous MetRS regulatory sites 362 and 388 in (A) Gammaproteobacteria and (B) Betaproteobacteria with representative sequence alignments showing E. coli as a reference.

Figure 6.

Strains incapable of tRNA misacylation with Met are less adept at growth in the presence of antibiotics and stressors. (A) Semi-quantitative heatmap of anaerobic tRNA misacylation with Met to individual tRNA species in the WT, misacylation deficient mutant strains and their revertants. The revertants were distinguished from the WT strain through the preservation of the synonymous codon mutations at two adjacent codons to the respective Lys362 or 388 site in the MetRS gene made in the mutants. (B) Mutant growth quantification compared to WT in Biolog phenotype microarrays organized by condition or stress types. All individual compounds and media conditions used can be found under http://www.biolog.com/products-static/phenotype_microbial_cells_use.php. (C) LB plate dilution assay showing growth of WT, mutants and revertants with and without 10 μM chloramphenicol–the same concentration used to induce mismethionylation in Figure 2B.