Innate immune and chemically triggered oxidative stress modifies translational fidelity

Abstract

Translational fidelity, essential for protein and cell function, requires accurate transfer RNA (tRNA) aminoacylation. Purified aminoacyl-tRNA synthetases exhibit a fidelity of one error per 10,000 to 100,000 couplings. The accuracy of tRNA aminoacylation in vivo is uncertain, however, and might be considerably lower. Here we show that in mammalian cells, approximately 1% of methionine (Met) residues used in protein synthesis are aminoacylated to non-methionyl-tRNAs. Remarkably, Met-misacylation increases up to tenfold upon exposing cells to live or non-infectious viruses, toll-like receptor ligands or chemically induced oxidative stress. Met is misacylated to specific non-methionyl-tRNA families, and these Met-misacylated tRNAs are used in translation. Met-misacylation is blocked by an inhibitor of cellular oxidases, implicating reactive oxygen species (ROS) as the misacylation trigger. Among six amino acids tested, tRNA misacylation occurs exclusively with Met. As Met residues are known to protect proteins against ROS-mediated damage, we propose that Met-misacylation functions adaptively to increase Met incorporation into proteins to protect cells against oxidative stress. In demonstrating an unexpected conditional aspect of decoding mRNA, our findings illustrate the importance of considering alternative iterations of the genetic code.

Fig. 1. Induction of tRNA misacylation by viruses

(a) Microarrays showing total tRNAs isolated from uninfected, flu and Adeno virus infected HeLa cells. (b) Large excess of oligonucleotides complementary to tRNA^Mets was included in array hybridization. (c) The Adeno-infected sample was deacylated before array hybridization. (d) Thin-layer chromatography of the Flu-infected sample using biotinylated oligonucleotide probes (longer exposure in inset). (e) Non-denaturing acid gel detection of misacylated tRNAs in the Flu-infected sample. (f) Flu-infected sample plus/minus aminopeptidase. (g) Quantitative comparison of uninfected and virus-infected samples. tRNAs are grouped according to amino acid properties. The detection limit of misacylation was ~0.1% for each probe.

Fig. 2. Misacylated tRNAs are used in translation

(a) Correctly acylated and misacylated tRNAs have the same kinetic properties plus/minus cycloheximide (CHX). Error bars represent s.d. (n=4). (b) 1D SDS-PAGE showing an increase in specific activity of [³⁵S]-Met incorporation upon Vac infection. (c) 2D SDS-PAGE of uninfected and Vac-infected samples. Spot 3 (55–62% of all radioactivity) matches the expected pI of wild-type Ub-HA. Spots 1 and 5 correspond to Lys/Arg-to-Met and Glu/Asp-to-Met substitution, respectively. (d) MALDI-TOF of tryptic digested Ub-HA. Peaks are labeled with their m/z values from the Ub-HA sequence. (e) MS-MS sequencing of 763.87 (m/2z) mass peak by LC-FTMS of tryptic digested Ub-HA.

Fig. 3. tRNA misacylation induced by TLR ligands

(a) Comparison of untreated and LPS, poly-IC-treated HeLa samples. (b) Comparison of immature and poly-IC-matured Bone Marrow Dentritic Cells (DC) including complementary Met-oligos in array hybridization. (c) Thin-layer chromatography of the poly-IC matured sample using biotinylated probes. (d) Quantitative comparison of untreated, LPS and poly-IC-matured DC samples, all AP-treated. The detection limit of tRNA misacylation for these samples was ~0.05% for each probe. (e) Misacylation occurs in vivo. Misacylation for total charged tRNA isolated from mouse liver after 1 min pulse with ³⁵S-Met. Array key shows probe locations for Met-tRNAs (black) and Cys-tRNAs (blue).

Fig. 4. Oxidative stress induces NADPH oxidase (NOX)-dependent RNA misacylation

(a) tRNA misacylation in HeLa cells induced by oxidizing agents H₂O₂ (1h) or arsenite (4h). DPI inhibits aresenite-induced misacylation. (b) Quantitative comparison of tRNA misacylation under oxidative stresses (Arsenite and H₂O₂) and TLR ligand (poly-IC) plus/minus DPI. (c) Percent of all misacylated tRNAs plus/minus DPI when cells were treated under four conditions (100% = all Met-tRNAs). Error bars represent s.d. (n=2). (d) poly-IC induces NOX-dependent RNA misacylation in DC cells. (e) Quantitative comparison of tRNA misacylation in DC cells plus/minus DPI.