Genetic background and mistranslation frequency determine the impact of mistranslating tRNASerUGG

Abstract Transfer RNA variants increase the frequency of mistranslation, the misincorporation of an amino acid not specified by the “standard” genetic code, to frequencies approaching 10% in yeast and bacteria. Cells cope with these variants by having multiple copies of each tRNA isodecoder and through pathways that deal with proteotoxic stress. In this study, we define the genetic interactions of the gene encoding tRNASerUGG,G26A, which mistranslates serine at proline codons. Using a collection of yeast temperature-sensitive alleles, we identify negative synthetic genetic interactions between the mistranslating tRNA and 109 alleles representing 91 genes, with nearly half of the genes having roles in RNA processing or protein folding and turnover. By regulating tRNA expression, we then compare the strength of the negative genetic interaction for a subset of identified alleles under differing amounts of mistranslation. The frequency of mistranslation correlated with the impact on cell growth for all strains analyzed; however, there were notable differences in the extent of the synthetic interaction at different frequencies of mistranslation depending on the genetic background. For many of the strains, the extent of the negative interaction with tRNASerUGG,G26A was proportional to the frequency of mistranslation or only observed at intermediate or high frequencies. For others, the synthetic interaction was approximately equivalent at all frequencies of mistranslation. As humans contain similar mistranslating tRNAs, these results are important when analyzing the impact of tRNA variants on disease, where both the individual’s genetic background and the expression of the mistranslating tRNA variant need to be considered.


Introduction
Mistranslation occurs when an amino acid that differs from what is specified by the standard genetic code is incorporated into a growing polypeptide chain during translation. Mistranslation occurs in all cells but can be enhanced by environmental conditions or mutations in the translational machinery (Lee et al. 2006;Netzer et al. 2009;Ling and Sö ll 2010;Jones et al. 2011;Wiltrout et al. 2012;Reverendo et al. 2014;Schwartz and Pan 2017;Lant et al. 2018). Mutations in tRNAs that cause mistranslation were initially identified as intergenic suppressors that change the meaning of the genetic code Stadler and Yanofsky 1959;Yanofsky and Crawford 1959;Benzer and Champe 1962;Gorini and Beckwith 1966). tRNA Ser variants are particularly prone to mistranslate because the anticodon is not a major identity element for aminoacylation by the cognate serine aminoacyl-tRNA synthetase (Gieg e et al. 1998). Rather, specificity for aminoacylation comes from the long variable arm positioned 3 0 of the anticodon stem (Asahara et al. 1994;Biou et al. 1994;Himeno et al. 1997).
The toxic effects of mistranslating tRNAs are buffered through multiple copies of each tRNA isodecoder (e.g. there are 275 tRNAencoding genes in Saccharomyces cerevisiae; Chan and Lowe 2016) and through protein quality control mechanisms that deal with misfolded protein and protein aggregates (reviewed in . When mistranslation reaches a threshold, protein quality control mechanisms no longer protect the cell and growth is impaired (Berg, Zhu, et al. 2019). The extent of growth impairment is inversely related to the frequency of mistranslation in a linear fashion with yeast growth being arrested when mistranslation approaches approximately 12% .
We previously demonstrated that the negative genetic interactions with mistranslating tRNAs depend on the amino acid substitution (Berg, Zhu, et al. 2021). At similar frequencies of mistranslation, a tRNA variant substituting serine at arginine codons has more genetic interactions than one substituting alanine at proline codons. In this report, we identify the genetic interactions of a mistranslating serine tRNA variant that incorporates serine at proline codons. Using a regulated tRNA expression system, we show that although there is a general correlation between the frequency of mistranslation and impact on growth, the impact of different mistranslation frequencies depends on a strain's specific genetic background. As similar mistranslating tRNAs are found in the human population, these results suggest that genetic background contributes to the impact of tRNA variants on health and disease.

Materials and methods
Yeast strains and growth BY4742 (MATa his3D0 leu2D0 lys2D0 ura3D0; Brachmann et al. 1998) and Y7092 (MATa can1D::STE2pr-SpHIS5 lyp1D his3D1 leu2D0 ura3D0 met15D0) strains are derivatives of S288c. Y7092 was a kind gift from Dr. Brenda Andrews (University of Toronto). Strains from the temperature-sensitive collection are derived from the wild-type MATa haploid yeast strain BY4741 and described in Costanzo et al. (2016). The strains containing the gene expressing tRNA Ser UGG,G26A (CY8613) were made by integrating modified SUP17 and flanking sequence into Y7092 at the HO locus and selecting for the natMX marker, as previously described in Zhu et al. (2020), using the construct described below. The control strain (CY8611) was made by integrating only the natNT2 marker at the HO locus. Transformants were selected on 100 mg/mL nourseothricin-dihydrogen sulfate (clonNAT) and integration was verified by PCR.
Yeast strains were grown at 30 C in yeast peptone media containing 2% glucose (YPD) or synthetic media supplemented with nitrogen base and amino acids, unless otherwise indicated. Growth curves were generated by diluting saturated cultures to OD 600 ¼ 0.1 in synthetic complete media and incubating at 30 C. OD 600 was measured every 15 min for 24 h using a BioTek Epoch 2 microplate spectrophotometer. Doubling time was calculated using the R package "growthcurver" (Sprouffske and Wagner 2016).

DNA constructs
The construct to integrate the gene encoding tRNA Ser UGG,G26A at the HO locus was created using a synthetic DNA containing 200 bp up and downstream of the HO translational start, previously described in Zhu et al. (2020). The construct was cloned into pGEM-T Easy (Promega Corp.) as a NotI fragment to create pCB4386. The natNT2 marker from pFA6-natNT2 was PCR amplified using primers UK9789/UK9790 (Supplementary Table 1) and cloned into pCB4386 as an EcoRI fragment to generate the control SGA integrating vector (pCB4394). The gene encoding tRNA Ser UGG,G26A was PCR amplified from pCB4023 (Berg et al. 2017) using primers UG5953/VB2609 and cloned as HindIII fragments into pCB4394 to create pCB4397.

Synthetic genetic array analysis and validation
The SGA assay was performed as described by Tong et al. (2001) with minor modifications. Strains CY8611 (HO::natMX) and CY8613 (HO::tRNA Ser UGG,G26A -natMX) were crossed to a yeast temperature-sensitive collection (Ben-Aroya et al. 2008;Li et al. 2011;Kofoed et al. 2015;Costanzo et al. 2016) in quadruplicate 1,536 colony array format using a BioMatrix (S&P Robotics Inc.) automated pinning system. In this format, each allele of the temperaturesensitive collection is present in technical quadruplicate on the plate. Double mutants were selected on YPD plates containing 200 mg/L G418 and 100 mg/L clonNAT. Diploids were sporulated on enriched sporulation media and MATa haploid double mutants selected using standard SGA media. The entire SGA procedure was carried out at room temperature, except for the colony scoring in order to minimize growth defect of the temperature-sensitive strains. To identify genetic interactions, double mutants were pinned onto double mutant selection SGA medium and grown at 30 C for 5 days. Images were taken every 24 h to determine colony size computationally. SGATools (Wagih et al. 2013) was used to determine genetic interaction scores using a multiplicative model (e ¼ W AB ÀW A * W B ; where e is the interaction score, W AB is the fitness of the double mutant and W A and W B are the fitness values of each single mutant). Double mutant strains with an average interaction score less than À0.2 and Benjamini-Hochberg corrected P-value less than 0.05 were considered as potential negative genetic interactions.
Double mutants that were identified as negative genetic interactions from the screen were validated by recreating the double mutant strain, starting from the single mutant haploid strains, using the SGA approach. Double mutant strains were grown in liquid media to saturation, cell densities were normalized, and cultures were spotted on SGA media. The temperature-sensitive mutant crossed with the control strain CY8611 and the mistranslating tRNA strain crossed with a control his3D strain were also spotted to determine fitness of the single mutants. Intensity of each spot was measured with ImageJ (Schneider et al. 2012). Expected double mutant growth was calculated based on the growth of the single mutants and compared to the experimental growth of the double mutant. Double mutants that grew more slowly than expected were considered validated negative genetic interactions. Raw and validated data can be found in Supplementary File 2.
Synthetic interactions with various frequencies of mistranslation were assessed by transforming the relevant temperature-sensitive strains with URA3-containing centromeric plasmids expressing tRNA Ser (pCB3076), tRNA Ser UGG,G26A (pCB4023), tRNA Ser UGG,G26A with 5 0 GAL1pr (pCB4568), or tRNA Ser UGG,G26A with 3 0 GAL1pr (pCB4566). At least 3 individual transformants for each plasmid and strain were grown in synthetic complete medium lacking uracil and containing 2% galactose as the carbon source. Cells were grown to confluency, diluted 1:32 in 1Â yeast nitrogen base and 5 lL was spotted onto solid media lacking uracil and containing 2% galactose. Cells were grown at 30 C for 32-56 h, depending on the strain, to achieve a level of growth (for the strain without mistranslation) consistent with the wild type. Mean density of growth was determined with ImageJ (Schneider et al. 2012) and normalized growth for each mistranslating tRNA was calculated as a percent of the wild-type tRNA Ser containing strain.

Heat shock assay
Yeast strains containing the HSE-GFP reporter and a mistranslating tRNA variant were grown to stationary phase in medium lacking uracil and containing 0.6% casamino acids, diluted 1:100 in the same medium and grown for 18 h at 30 C. Cell densities were normalized to OD 600 before measuring fluorescence with a BioTek Synergy H1 microplate reader at an excitation wavelength of 488 nm and emission wavelength of 528 nm. The mean relative fluorescence units were calculated from 3 technical replicates for each biological replicate.

Mass spectrometry
Liquid chromatography-tandem mass spectrometry to identify mistranslation was performed on 5 biological replicates of each strain. Starter cultures of each strain were grown overnight in YPD before being diluted to an OD 600 of 0.1 in the same media and grown to an OD 600 of approximately 1.0. Cells were lysed in a urea lysis buffer (8 M Urea, 50 mM Tris pH 8.2, 75 mM NaCl) and proteins were reduced with dithiothreitol and alkylated with iodoacetamide. Robotic purification and digestion of proteins into peptides were performed on the KingFisher Flex using LysC and the R2-P1 method described in Leutert et al. (2019).
MS/MS spectra were searched against the S. cerevisiae protein sequence database (downloaded from the Saccharomyces Genome Database resource in 2014) using Comet (release 2015.01; Eng et al. 2013). The precursor mass tolerance was set to 50 ppm. Constant modification of cysteine carbamidomethylation (57.0215 Da) and variable modification of methionine oxidation (15.9949 Da) and proline to serine substitution (À10.0207 Da) were used for all searches. A maximum of 2 of each variable modification was allowed per peptide. Search results were filtered to a 1% false discovery rate at the peptide spectrum match level using Percolator (Kä ll et al. 2007). The mistranslation frequency was calculated using the unique mistranslated peptides for which the non-mistranslated sibling peptide was also observed. The frequency is defined as the counts of mistranslated peptides, where serine was inserted for proline, divided by the counts of all peptides containing proline, and expressed as a percentage.

Results and discussion
Synthetic genetic interactions with tRNA Ser

UGG,G26A
Yeast cells expressing tRNA Ser UGG,G26A , which contains a proline anticodon, mistranslate serine at proline codons (Berg, Zhu, et al. 2019;. The G26A mutation is required in combination with the anticodon change to dampen otherwise lethal levels of mistranslation (Berg et al. 2017). To perform the SGA analysis, the gene encoding tRNA Ser UGG,G26A (Fig. 1a), including approximately 300 base pairs of 5 0 and 3 0 flanking sequence and a clonNAT resistance marker, was integrated at the HO locus. A control strain was created with only the clonNAT resistance marker integrated at the HO locus. Mass spectrometrybased analysis of the cellular proteome identified 4.9% proline to serine substitution in the strain expressing tRNA Ser UGG,G26A compared to only 0.6% substitution in the control strain (Fig. 1b). As shown in Fig. 1, c and d, tRNA Ser UGG,G26A reduces cell growth (doubling time of 98 min vs 84 for the control strain) and results in a heat shock response (6.4-times greater than the control strain). We then performed an SGA analysis to identify genetic interactions with tRNA Ser UGG,G26A using a collection of 1,016 temperature-sensitive alleles. The robotic screen identified 125 alleles with negative genetic interactions with tRNA Ser UGG,G26A . Genetic interactions were validated by remaking the double mutant strains, spotting normalized densities of the double mutants and their control strain on selective plates, and measuring growth after 2 days (Supplementary File 2). After validation, 109 alleles representing 91 genes were classified as having a negative genetic interaction with tRNA Ser UGG,G26A (Fig. 2a). To further analyze the network of genes associated with the mistranslating tRNAs, we identified areas of the yeast genetic interaction map (Costanzo et al. 2016) that were enriched for is statistically different compared to the control strain (Welch's t-test; Bonferroni corrected P-value < 0.05). c) Doubling times for the strains described in B were determined from growth curves of the strains diluted to an OD 600 of approximately 0.1 in synthetic complete media containing clonNAT and grown for 24 h. Doubling time was calculated with the R package "growthcurver" (Sprouffske and Wagner 2016). Each point represents 1 biological replicate (n ¼ 4). Doubling time is statistically different between the strain expressing tRNA Ser UGG,G26A and the control strain (Welch's t-test; Bonferroni corrected P-value < 0.05). d) Strains described in B were transformed with a GFP reporter transcribed from a promoter containing heat shock response elements, grown to saturation in media lacking uracil, diluted 1:100 in the same media and grown for 18 h at 30 C. Cell densities were normalized and fluorescence measured. Each point represents 1 biological replicate (n ¼ 5). Relative heat shock induction is statistically different in the strain expressing tRNA Ser UGG,G26A compared to the control strain (Welch's t-test; Bonferroni corrected P-value < 0.05).
negative genetic interactions with the mistranslating tRNA Ser UGG,G26A using SAFE (Baryshnikova 2016;Fig. 2b). Areas of the yeast genetic interaction network annotated with roles in protein turnover, cell polarity, mitosis, DNA replication and repair, transcription, and mRNA processing were significantly enriched.
In a previous screen looking at negative synthetic genetic interactions with tRNA variants that mistranslate alanine at proline codons (tRNA Pro G3:U70 ; Hoffman, Berg, et al. 2017) and serine at arginine codons (tRNA Ser UGG,G26A ; Berg, Zhu, et al. 2019), we identified 10 and 47 negative genetic interactions, respectively (Berg, Zhu, et al. 2021). While these tRNA variants mistranslated at lower frequency ($3%) than tRNA Ser UGG,G26A making specific comparison of genetic interactions difficult, it is possible to compare enriched pathways with a SAFE analysis (Fig. 2c) (Pro!Ala) and tRNA Ser UCU,G26A (Arg!Ser), respectively. The unique genetic interactions identified for Pro!Ala and Pro!Ser mistranslating tRNAs suggest that it is not simply the proline content of the encoded protein that predicts genetic interactions with these tRNA variants. This agrees with our previous result that genes with negative genetic interactions are not enriched for the amino acid being mistranslated (Berg, Zhu, et al. 2021).
Only one strain had a positive genetic interaction with tRNA Ser UGG,G26A . The strain contains a temperature-sensitive allele of eco1, an acetyltransferase required in sister chromatid cohesion. As we demonstrated previously (Zhu et al. 2020), the positive interaction results from tRNA Ser UGG,G26A restoring serine at the S213P mutation of eco1-1.  Fig. 2. Negative genetic interaction network of the mistranslating tRNA Ser UGG,G26A . a) Genes validated as having a negative genetic interactions with tRNA Ser UGG,G26A are arranged according to their predicted function based on gene descriptions in the yeast genome database (www.yeastgenome.org; last accessed February 2022). b) SAFE analysis of genes that have a negative genetic interaction with tRNA Ser UGG,G26A were mapped onto the yeast genetic interaction profile map (Costanzo et al. 2016) using TheCellMap (Usaj et al. 2017). Blue dots represent genes within the local neighborhood of genes validated to have negative genetic interactions with tRNA Ser UGG,G26A . Terms in blue boxes are network regions that are significantly enriched (Bonferroni corrected P-value < 0.05). c) SAFE analysis as performed in B with genetic interactions for tRNA Pro G3:U70 (yellow) and tRNA Ser UGG,G26A (red) from Berg, Zhu, et al. (2021). Terms in boxes represent network regions that are significantly enriched for the, respectively, mistranslating tRNA (Bonferroni corrected P-value < 0.05). et al. 2019; Lant et al. 2019). Due to their ability to generate proteotoxic stress, we and others have suggested that these variants may be genetic modifiers of disease (Reverendo et al. 2014;Berg et al. 2017). In a previous analysis, we demonstrated that, when comparing the same amino acid substitution, there is a near-linear negative correlation between mistranslation frequency and cell growth in a wild-type S. cerevisiae background (Berg, Zhu, et al. 2019). Our goal was to determine if the genetic background of strains having a synthetic interaction with tRNA Ser UGG,G26A changes the impact of different mistranslation frequencies.
To regulate the frequency of mistranslation, we took advantage of our finding that placing a GAL1 promoter (GAL1pr) sequence up or downstream of a tRNA gene represses tRNA expression when cells are grown in galactose . Previously, using mass spectrometry, we determined the frequency of proline to serine mistranslation in strains containing centromeric plasmids expressing wild-type tRNA Ser , tRNA Ser UGG,G26A , tRNA Ser UGG,G26A with 5 0 GAL1pr, and tRNA Ser UGG,G26A with 3 0 GAL1pr in galactose containing medium to be 0.3%, 5.6%, 3.5%, and 0.9%, respectively ( Fig. 3a; . Three frequencies of mistranslation above background are thus achieved: the greatest mistranslation with no flanking GAL1pr, intermediate mistranslation with the GAL1pr upstream of the tRNA, and the least mistranslation with the GAL1pr downstream of the tRNA. To determine the impact of mistranslation frequency, we selected 16 of the strains identified in the SGA analysis, representing different pathways, that had a robust negative genetic interaction but not so severe that there was no dynamic range to investigate the effects of different levels of mistranslation. These and the control strain were transformed with centromeric plasmids containing wild-type tRNA Ser , tRNA Ser UGG,G26A 5 0 GAL1pr-tRNA Ser UGG,G26A , and 3 0 tRNA Ser UGG,G26A -GAL1pr. Triplicate cultures of independent transformants for each strain were grown to confluency, diluted 1:32, spotted onto minimal medium with galactose as the carbon source, and grown at 30 C. Spot plates are shown in Supplementary Fig. 1. The density of the spotted cultures was measured and then expressed as a percentage of the spot density for the same strain background containing wildtype tRNA Ser . For example, tRNA Ser UGG,G26A reduces the growth of the wild-type BY4742 strain to 37% 6 5% of growth seen in BY4742 containing wild-type tRNA Ser . The 5 0 GAL1pr- , and tRNA Ser UGG,G26A [tS(UGG)] used to regulate proline to serine mistranslation frequency. Mistranslation frequencies were measured by mass spectrometry in . b) Wild-type BY4742 or the indicated strains from the temperature-sensitive collection were transformed with the constructs described in A. Strains were grown to confluency in media lacking uracil and diluted 1:32 and spotted on media lacking uracil with galactose as the carbon source. The spot intensity of the strain containing the mistranslating tRNA was divided by the intensity of the strain containing the wild-type tRNA Ser to determine normalized growth. Each point represents 1 biological replicate. tRNA Ser UGG,G26A and 3 0 tRNA Ser UGG,G26A -GAL1pr reduce growth to 55% 6 9% and 71% 6 7%, respectively.
As shown in Fig. 3b and Supplementary Fig. 2, the wild-type strain displays a near-linear decrease with increasing mistranslation, consistent with our previous observations that increasing mistranslation frequency is negatively correlated with effects on growth for the same amino acid substitution (Berg, Zhu, et al. 2019). All the synthetic strains showed a graded response where increased mistranslation results in more severe loss of growth, but interestingly, the pattern of decreased growth in response to changing the frequency of mistranslation differed amongst the strains. This difference suggests that the genetic background influences the impact of mistranslation.
To look at the patterns in more detail, we plotted the normalized growth of each temperature-sensitive strain as a percentage of the growth of the wild-type strain (BY4742) for each of the three tRNA constructs that result in low, medium, and high mistranslation frequency (Supplementary Fig. 3). We note that in these plots, 100% indicates a lack of a negative synthetic effect, not the absence of an impact of mistranslation. Although the patterns appear to represent a continuum, we will focus the analysis on representative examples of four categories (Fig. 4). In the first category are strains where the synthetic interaction with tRNA Ser UGG,G26A increases proportionately with the frequency of mistranslation. In the representative examples ctf8-62 and cdc32-e72g (Fig. 4a), little synthetic interaction is seen at the lowest frequency of mistranslation. Many of the strains, including arc35-6 and las17-14 (Fig. 4b), are in the second category. These show a modest synthetic effect at a low mistranslation frequency, but have a strong negative synthetic interaction at both moderate and high frequencies of mistranslation. The third category, represented best by act1-4, has a nearly equivalent synthetic effect at all three levels of mistranslation (Fig. 4c). The last group includes cdc20-1 and is related to group 3 but appears relatively more impacted by the lowest frequency of mistranslation (Fig. 4d).
Many factors determine the impact of a mistranslating tRNA. Factors intrinsic to the tRNA include the anticodon sequence and its resulting amino acid substitution (Berg, Zhu, et al. 2021), the level of tRNA expression , and the presence of secondary mutations that alter the stability of the tRNA (Berg et al. 2017;Berg, Zhu, et al. 2019). Other factors are extrinsic to the tRNA. These include the number of competing tRNAs that buffer the mistranslating tRNA (Zimmerman et al. 2018), the environment in which cells expressing the mistranslating tRNA is found (Berg, Zhu, et al. 2021), and the genetic background of the organism.
Genetic background contributes to the impact of a mistranslating tRNA in numerous ways. In yeast (this work and Hoffman, Berg, et al. 2017) and Escherichia coli (Ruan et al. 2008), loss of genes that regulate proteotoxic stress increase the severity of mistranslating tRNA variants, most likely because the tRNA variants increase the load of mismade proteins. The impact of an extragenic mutation will depend on the extent to which it disrupts protein quality control or otherwise contributes to proteotoxic stress (Redler et al. 2016). Complexities arise since there are multiple quality control pathways that act independently but ultimately overlap to regulate proteostasis (reviewed in Chen et al. 2011). As such, impairing each pathway has the potential to show a different response to both changing levels of mistranslation and type of amino acid substitution. The nonlinear nature of the response likely arises because growth effects are not observed until a threshold of proteotoxicity is reached; different mutations will approach or exceed this threshold to different extents.
Other  and High: tRNA Ser UGG,G26A [tS(UGG)]g for temperature-sensitive strains expressing ctf8-162 or cdc33-e72g (a), arc35-6 or las17-14 (b), act1-4 (c), and cdc20-1 (d). The growth of the wild-type strain, 100%, is plotted as gray dots. Each point is the average of at least 3 biological replicates as in Fig. 3. function caused by mutation will be compounded by the reduced level of functional protein caused by mistranslation. The closer a genetic mutation brings the protein to the critical level of expression, the more impact the mistranslating tRNA will show when combined with that genetic mutant. Similar to the example above, until a critical threshold is exceeded, mistranslation may have little effect. At a more global level, genes that impact translation or mRNA processing can further limit expression of proteins already reduced by mistranslation.
Factors that alter the gene expression profile will also influence the impact of a mistranslating tRNA. Synonymous codon usage varies across the open reading frames in a genome (reviewed in Liu et al. 2021). Specific tRNA variants will mistranslate at a subset of these synonymous codons as determined by wobble rules and base modifications. tRNA variants only affect genes that are translated, and those that are more highly expressed will lead to greater proteotoxic stress when mistranslated. For yeast, we have shown that one such factor is the environment in which the cells are grown (Berg, Zhu, et al. 2021). As cell type determines gene expression profile in multicellular eukaryotes, different cell types are expected to be impacted differently by a tRNA variant. This argument is relevant to genetic background because mutations alter the internal and potentially external environment of the cell and often invoke a transcriptomic response (Hughes et al. 2000). The altered gene expression will in part determine the impact of a mistranslating tRNA. In addition, the impact of different anticodons decoding the same amino acid should affect the distribution of genetic interactions because anticodon sequence determines the extent and placement of mistranslated residues.
The genetic background could also directly or indirectly alter the expression of the mistranslating tRNA or that of competing endogenous tRNAs, altering the frequency of mistranslation. Genetic mutations that alter tRNA expression could occur in genes involved in regulating tRNA transcription (e.g. by altering chromatin structure) or in genes with roles in tRNA processing, modification, nuclear import, or stability. In our experiments, changes in expression of the mistranslating tRNA could occur if a secondary mutation occurs in a gene regulating transcription from the GAL promoter, though we note that none of the 16 temperature-sensitive alleles we analyzed were in genes with annotated GAL regulatory roles.

Conclusion
The impact of a mistranslating tRNA depends on a cell's genetic background and the expression of the variant as it relates to the frequency of mistranslation. Approximately 20% of humans contain a tRNA variant that potentially mistranslates (Berg, Giguere, et al. 2019). The variants are found within different tRNA isodecoders at different loci, which is highly relevant since tRNA genes in the human genome are expressed at different levels (Torres et al. 2019). Our findings demonstrate that genetic background must be taken into account when determining the contribution of tRNA variants to disease. Indeed, due to genetic or epigenetic differences, some individuals may be particularly sensitive to even low levels of mistranslation.

Data availability
Strains and plasmids are available upon request. The authors affirm that all data necessary for confirming the conclusions of the article are present within the article, figures, and supplemental material. Supplementary File 1 contains all supplemental figures.
Supplementary File 2 contains raw and validated SGA data. The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE (Perez-Riverol et al. 2019) partner repository with the dataset identifiers PDX025934 and PXD032063.
Supplemental material is available at G3 online.