Effects of tRNA modification on translational accuracy depend on intrinsic codon-anticodon strength

Abstract

Cellular health and growth requires protein synthesis to be both efficient to ensure sufficient production, and accurate to avoid producing defective or unstable proteins. The background of misreading error frequency by individual tRNAs is as low as 2 × 10(-6) per codon but is codon-specific with some error frequencies above 10(-3) per codon. Here we test the effect on error frequency of blocking post-transcriptional modifications of the anticodon loops of four tRNAs in Escherichia coli. We find two types of responses to removing modification. Blocking modification of tRNA(UUC)(Glu) and tRNA(QUC)(Asp) increases errors, suggesting that the modifications act at least in part to maintain accuracy. Blocking even identical modifications of tRNA(UUU)(Lys) and tRNA(QUA)(Tyr) has the opposite effect of decreasing errors. One explanation could be that the modifications play opposite roles in modulating misreading by the two classes of tRNAs. Given available evidence that modifications help preorder the anticodon to allow it to recognize the codons, however, the simpler explanation is that unmodified 'weak' tRNAs decode too inefficiently to compete against cognate tRNAs that normally decode target codons, which would reduce the frequency of misreading.

Figure 1.

Misreading errors by decrease in the absence of 5-methylaminomethyl-modification of the wobble uridine. The activity of K529 mutants of firefly luciferase expressed in the wild-type (Xac) or mnmE background, which blocks mnm⁵-modification, are compared. The activities of seven mutants, six normally error-prone, were significantly reduced in the mnmE background suggesting that the modification increases the frequency of misreading errors by the tRNA. For the mutants showing a statistically significant change in activity between the wild-type and mnmE backgrounds, the structure is shown of the codon–anticodon complexes required for each of these misreading events (the upper line represents the codon, the lower the anticodon). Vertical lines represent Watson–Crick pairs, filled circles canonical wobble pairs, and open circles pairs requiring a tautomeric shift to form.

Figure 2.

Misreading errors by increase in the absence of 5-methylaminomethyl-modification of the wobble uridine. The activity of E537 mutants of Escherichia coli β-galactosidase expressed in the wild-type (Xac) or mnmE background are compared as in Figure 1.

Figure 3.

Misreading errors by increase in the absence of the wobble queuosine modification. The activity of D201 mutants of Escherichia coli β-galactosidase expressed in the wild-type (Xac) or tgtΔ background, which blocks formation of wobble queuosine, are compared as in Figure 1.

Figure 4.

Misreading errors by decrease in the absence of the wobble queuosine modification. The activity of Y503 mutants of Escherichia coli β-galactosidase expressed in the wild-type (Xac) or tgtΔ background are compared as in Figure 1.

Figure 5.

Misreading errors by decrease in the absence of 2-methylthio-N⁶-isopentenyl modification of adenosine-37, the nucleotide adjacent to the anticodon. The activity of Y503 mutants of Escherichia coli β-galactosidase expressed in the wild-type (Xac) or miaAΔ background are compared as in Figure 1.