Anticodon sequence determines the impact of mistranslating tRNAAla variants

Abstract

Transfer RNAs (tRNAs) maintain translation fidelity through accurate charging by their cognate aminoacyl-tRNA synthetase and codon:anticodon base pairing with the mRNA at the ribosome. Mistranslation occurs when an amino acid not specified by the genetic message is incorporated into proteins and has applications in biotechnology, therapeutics and is relevant to disease. Since the alanyl-tRNA synthetase uniquely recognizes a G3:U70 base pair in tRNA^Ala and the anticodon plays no role in charging, tRNA^Ala variants with anticodon mutations have the potential to mis-incorporate alanine. Here, we characterize the impact of the 60 non-alanine tRNA^Ala anticodon variants on the growth of Saccharomyces cerevisiae. Overall, 36 tRNA^Ala anticodon variants decreased growth in single- or multi-copy. Mass spectrometry analysis of the cellular proteome revealed that 52 of 57 anticodon variants, not decoding alanine or stop codons, induced mistranslation when on single-copy plasmids. Variants with G/C-rich anticodons resulted in larger growth deficits than A/U-rich variants. In most instances, synonymous anticodon variants impact growth differently, with anticodons containing U at base 34 being the least impactful. For anticodons generating the same amino acid substitution, reduced growth generally correlated with the abundance of detected mistranslation events. Differences in decoding specificity, even between synonymous anticodons, resulted in each tRNA^Ala variant mistranslating unique sets of peptides and proteins. We suggest that these differences in decoding specificity are also important in determining the impact of tRNA^Ala anticodon variants.

Keywords: Mistranslation; expanded decoding; genetic code; tRNA biology; tRNAAla.

Figure 1.

Schematic of tRNA^Ala containing a degenerate anticodon and the tetracycline inducible tRNA expression system. (A) The structure of tRNA^Ala with a degenerate anticodon. Bases colored in blue represent the G3:U70 base pair required for recognition and charging by AlaRS. The anticodon is shown in red. (B) tRNA^Ala in the tetracycline inducible system used to regulate tRNA^Ala anticodon variant expression flanked by up- and downstream SUP17 sequence. In the absence of doxycycline, the tet^O promoter is bound by the TetR-VP16 transcriptional activator which represses tRNA expression by driving RNA polymerase II expression across the tRNA gene. In the presence of doxycycline, TetR-VP16 binds doxycycline and dissociates from the promoter allowing the tRNA to be transcribed by RNA polymerase III.

Figure 2.

Impact of tRNA^Ala anticodon variants on yeast growth. (A) Representative growth curves of yeast strain CY8652 containing a plasmid with the control tRNA^Ala_GGC(Ala), a tRNA^Ala_CAU(Met) variant with no growth impact, a tRNA^Ala_GAA(Phe) variant with intermediate growth impact or a tRNA^Ala_AGG(Pro) variant with a severe growth impact. Black, grey and blue lines represent strains grown in media containing 0, 0.01 and 1.0 μg/mL doxycycline, respectively. (B) Yeast strain CY8652 containing plasmids with the control tRNA^Ala_GGC(Ala) or one of the tRNA^Ala anticodon variants were grown for 48 hours at 30°C in medium lacking uracil and leucine, diluted to an OD₆₀₀ of 0.1 in the same medium with 1.0 μg/mL doxycycline and grown for 24 hours at 30°C with agitation. OD₆₀₀ was measured at 15-minute intervals and doubling time was quantified. Relative growth of tRNA^Ala anticodon variants was calculated by multiplying the inverse of each variant doubling time by the average doubling time of the tRNA^Ala_GGC(Ala) control (for raw data and statistical comparisons see Table S5 and S6). For strains containing CGG(Pro) and CAG(Leu) variants area under the curve was used to determine relative growth. Each point represents one biological replicate (n = 3).

Figure 3.

Anticodon sequence affects growth impact. (A) Relative growth for each tRNA^Ala variant as in Figure 2B plotted for each anticodon on Grosjean and Westhof’s alternative representation of the genetic code [58] where ‘strong’ codon:anticodon pairs are at the top of the circle plot and ‘weak’ pairs are at the bottom. The length of each bar is proportional to the relative growth of each variant. (B) Relative growth for mistranslating tRNA^Ala anticodon variants decoding synonymous anticodons differing at base 34. Variants are grouped by number of different possible anticodons decoding a specific amino acid (1-, 2-, 3- or 4-box) with the same base at positions 35 and 36. Amino acid identity is listed below the anticodon. Native and non-native anticodon sequences in yeast are depicted with filled and empty shapes, respectively. D: adenine, guanine or uracil; R: purine; Y: pyrimidine; N: any base.

Figure 4.

Overexpressing some but not all tRNA^Ala variants reduce growth. Strain BY4742 was transformed with 1.0 µg of a URA3 multicopy plasmid containing each tRNA^Ala anticodon variant, plated on medium lacking uracil and grown at 30°C. Transformants were imaged after 48 hours (raw images are in Figure S4) and colony area was quantified using the ImageJ ‘Watershed’ package. Each point represents one colony and horizontal bars at the center of each boxplot represent median colony size. Median colony size of each variant was compared to the control tRNA^Ala_CGC(Ala) (shown in red) using a Wilcoxon rank sum test with Bonferroni correction to determine significance (see Table S7 for p-values). Variants lead to a statistically significant decrease in growth (p < 0.01) unless denoted by ‘ns’.

Figure 5.

Mistranslation detected by mass spectrometry-based analysis of the cellular proteome for strains containing tRNA^Ala anticodon variants. (A) Yeast strain CY8652 containing a tRNA^Ala anticodon variant were diluted into medium containing 1.0 μg/mL doxycycline to induce tRNA expression and harvested at an OD₆₀₀ of ~ 1.0. Cells were lysed in a denaturing buffer and mass spectrometry analysis of the cellular proteome was performed. The proportion of peptides detected as mistranslated, expressed as a percentage, was calculated for each variant from the number of unique peptides where alanine mis-incorporation was observed relative to the number of unique peptides where the wild-type amino acid was observed. Each point represents one biological replicate (n ≥ 3). Variants have statistically elevated levels of mistranslated peptides relative to the control strain containing tRNA^Ala_GGC (Welch’s t-test; Benjamini-Hochberg corrected p < 0.01) unless denoted by ‘ns’. (B) Summed MS1 intensity of mistranslated peptides normalized to the summed MS1 intensities of all detected peptides (normalized mistranslation abundance) plotted against the relative growth for each tRNA^Ala variant for amino acid families with 3 or more anticodons. The grey dashed line represents the linear correlation between the two variables. Note the x-axis is scaled differently for each amino acid plot and does not always start at zero.

Figure 6.

tRNA^Ala variants with leucine NAG anticodons or valine NAC anticodons mistranslate at unique codons leading to different subsets of the proteome experiencing mistranslation. (A) Codon specific proportion of unique mistranslated peptides for the control strain containing tRNA^Ala_GGC(Ala) and four tRNA^Ala variants containing synonymous anticodons decoding leucine codons. Proportion of mistranslated peptides at each codon was calculated from mass spectrometry data as the number of unique mistranslated peptides divided by the number of wild-type peptides where the leucine residue is coded for by the codon indicated. Only peptides containing one leucine residue were considered. Each point represents one biological replicate (n ≥ 3). (B) Venn diagrams showing the overlap between the unique peptides (left) and proteins (right) where mistranslation was detected by mass spectrometry in strains containing tRNA^Ala variants that have one of four synonymous anticodons decoding leucine codons. (C) Codon specific proportion of unique mistranslated peptides for the control strain containing tRNA^Ala_GGC(Ala) and four tRNA^Ala variants containing synonymous anticodons decoding valine codons as in (A). (D) Venn diagrams showing the overlap between the unique peptides (left) and proteins (right) where mistranslation was detected by mass spectrometry in strains containing tRNA^Ala variants that have one of four synonymous anticodons decoding valine codons.

Figure 7.

Profile of mistranslation events detected at synonymous codons for tRNA^Ala variants with G, A, C or U at base 34. For each tRNA^Ala anticodon variant, the proportion of mistranslated peptides identified at each synonymous codon was calculated from the number of unique peptides with a mistranslation event at that codon relative to the number of unique peptides containing a wild-type residue at that specific codon. In all cases, positions 35 and 36 formed Watson-Crick pairs with positions 2 and 1 of the codon, respectively. To confidently localize the mistranslation event and identify the mistranslated codon, only peptides containing a single target amino acid were used in this analysis. Only mistranslation that is statistically above the frequency measured for the same codon in the wild-type strain is shown (Welch’s t-test; Benjamini-Hochberg corrected p < 0.01). Each point represents one biological replicate (n ≥ 3). Note the scale of the y-axis is different for each plot.