In this article, we present two complementary methods to measure specific rates of translational error and mistranslation in the model mycobacterium, Mycobacterium smegmatis, using gain-of-function reporter systems. The methods can be employed to measure accurate error rates in low throughput or relative error rates in a more high-throughput setting.
The translation of genes into proteins is prone to errors. Although the average rate of translational error in model systems is estimated to be 1/10,000 per codon, the actual error rates vary widely, depending on the species, environment, and codons being studied. We have previously shown that mycobacteria use a two-step pathway for the generation of aminoacylated glutamine and asparagine tRNAs and that this is specifically associated with relatively high error rates due to the modulation of mistranslation rates by an essential component of the pathway, the amidotransferase GatCAB. We modified a previously employed Renilla-Firefly dual-luciferase system that had been used to measure mistranslation rates in Escherichia coli for use in mycobacteria to measure specific mistranslation rates of glutamate at glutamine codons and aspartate for asparagine codons. Although this reporter system was suitable for the accurate estimation of specific error rates, lack of sensitivity and requirements for excessive manipulation steps made it unsuitable for high-throughput applications. Therefore, we developed a second gain-of-function reporter system, using Nluc luciferase and green fluorescent protein (GFP), which is more amenable to medium/high-throughput settings. We used this system to identify kasugamycin as a small molecule that can decrease mycobacterial mistranslation. Although the reporters that we describe here have been used to measure specific types of mycobacterial mistranslation, they may be modified to measure other types of mistranslation in a number of model systems.
The flow of information in molecular biology requires the translation of genetic information to functional proteins. As with all biological systems, gene translation also involves measurable errors. Estimates of the rates of error in translation are typically quoted as approximately 1/10,000 per codon (reviewed by Ribas de Pouplana et al.1). However, error rates vary widely, from fewer than 10-5 to more than 0.05/codon1,2,3,4,5. The wide range of error rates, spanning more than three orders of magnitude, is due to the fact that errors can arise from multiple steps in the translation pathway: from stochastic, mutational, or stress-induced errors in aminoacylation6,7,8,9,10, physiological misacylation of asparaginyl- and glutaminyl-tRNAs5, or ribosomal decoding errors2,3,11. Measurably high error rates, representing over 0.01/codon, suggested that translational errors may perform physiological functions1,12 and that mistranslation may be context-specific13.
We and others have shown that naturally occurring errors in gene translation may be adaptive, especially during environmental stress1,5,12,14,15,16,17,18. In mycobacteria, errors generated by the two-step indirect glutamine/asparagine tRNA aminoacylation pathway19,20 result in a remarkably increased tolerance for the first-line antituberculosis antibiotic rifampicin5. Therefore, we speculated that decreasing mycobacterial mistranslation with a small molecule may potentiate killing by rifampicin. We screened for and identified the naturally occurring aminoglycoside kasugamycin as a compound that could decrease mycobacterial mistranslation, potentiate rifampicin-mediated killing of mycobacteria both in vitro and in vivo21, and limit the emergence of rifampicin resistance21, which threatens the global control of tuberculosis22 — the world's most deadly pathogen.
To study translational error, methods for the measurement of mistranslation have to be employed. There are multiple methods that have been developed for the measurement of mistranslation, each with advantages and disadvantages. Briefly, precision mass spectrometry-based methods have several advantages, the most important of which is that with newer algorithms for the detection of multiple types of translational error, a relatively unbiased measurement of mistranslation can be performed18. However, mass spectrometry is not very suitable for the measurement of deamidation mistranslation events — precisely the type of mistranslation that occurs in mycobacteria due to the error-prone indirect tRNA aminoacylation pathway. This is because of high-frequency nonenzymatic deamidation that occurs in the processing of samples for mass spectrometry23, resulting in an extremely high background signal. Therefore, for the detection of errors in this pathway, genetic gain-of-function reporters offer distinct advantages. Specifically, suitable gain-of-function reporters can have extremely low background rates, allowing the measurement of very low error rates11.
Since performing experiments with pathogenic Mycobacterium tuberculosis requires specialized facilities and extra precautions, we perform most experiments in the nonpathogenic mycobacterium M. smegmatis — and have shown earlier that results between the two species are broadly comparable5,21. To measure mistranslation rates in mycobacteria generated by the indirect tRNA aminoacylation pathway, we modified a Renilla-Firefly dual-luciferase system that had been previously developed to measure ribosomal decoding errors in E. coli11 for use in mycobacteria. We made three specific modifications: the original reporter did not express efficiently in mycobacteria, and therefore, the sequence was codon-optimized, and the C-terminal three amino acids of the firefly luciferase, serine-lysine-leucine, which has been annotated as a trafficking signal in some systems24, was modified to isoleucine-alanine-valine25. The original reporter had a critical lysine residue in the firefly luciferase mutated. Instead, we mutated either a critically conserved aspartate (D120) or glutamate (E144) residue in the Renilla luciferase to asparagine and glutamine, respectively25 (Figure 1). The reporter was subcloned into the episomal tetracycline-inducible plasmid pUV-tetOR (see the Table of Materials). Gain-of-function reporters mutate critically conserved functional residues in enzymes/fluorescent proteins that renders them nonfunctional11,26. Translational (or theoretically, transcriptional) errors that synthesize the functional variant of the protein would result in measurable enzyme activity in a subset of translated proteins. To correct for variation in protein abundance, the mutated reporter is coexpressed with a functional protein that acts as a benchmark and allows for accurate quantification of gain-of-function11. Whilst the Renilla-Firefly dual-luciferase reporter allowed for the accurate measurement of specific mycobacterial mistranslation rates5,25 (Figure 2 and section 1 of the protocol), we quickly realized that it is not suitable for medium/high-throughput screening of molecules that would alter the mistranslation rate. This is due mainly to two reasons, namely, a) the relative lack of potency of Renilla luciferase meant that a minimum of 1 mL of mycobacterial culture/sample was required to measure mistranslation rates, and b) the requirement for lysis of the cells prior to enzyme activity measurement required excessive manual handling: mycobacterial cells have a thick and multi-layered cell wall and envelope that is relatively resistant to lysis. Therefore, we looked to develop a new gain-of-function reporter system that could be used with small volumes (e.g., in a 96-well plate system) and did not require cell-lysis for measurements. We used the highly potent Nluc luciferase and identified a critical aspartate residue that, when mutated to asparagine, resulted in 2 logs loss of function (Figure 1). Furthermore, the small size of Nluc allowed it to have an N-terminal secretion signal tag — from antigen 85A, a major secreted antigen in mycobacteria27 — that would allow Nluc to be secreted into culture supernatant and circumvent the requirement for cell lysis. The benchmark protein, GFP, was expressed from the same promoter as the mutated Nluc, but from an integrated vector (see the Table of Materials), and could be measured in intact cells21 (Figure 3). Despite these advantages, the Nluc/GFP reporter (section 2 of the protocol) also has disadvantages: the relatively modest reduction in Nluc activity (100-fold) with the D140N mutation would not permit the measurement of extremely low mistranslation rates, making the reporter more suitable as a screening tool than for the accurate measurements of translational error. Furthermore, Nluc has no critical glutamate residues; therefore, only asparagine-to-aspartate error rates could be measured. The general principles described in this work should allow researchers to either use these reporters as we have done or modify reporters as appropriate for an accurate and/or facile measurement of other specific translational error rates in their model system of choice.
1. Renilla-Firefly Dual-luciferase Reporter
NOTE: For a visual representation of this method, see Figure 2.
2. Nluc/GFP Reporter
NOTE: For a visual representation of this method, see Figure 3.
A cartoon illustrating the general outline of the two reporter systems used in this work are shown in Figure 1. An overview of section 1 of the protocol is shown in Figure 2 and an overview of section 2 in Figure 3. The effect of kasugamycin on mycobacterial mistranslation is shown in Figure 4, as measured by the Renilla-Firefly reporter system. To show the specificity of the kasugamycin action, the reporter was also expressed in a strain of M. smegmatis in which ksgA was deleted (∆ksgA). KsgA is an rRNA dimethyl transferase, and ksgA-deleted strains are relatively resistant to kasugamycin. Kasugamycin action on mistranslation was also measured using the Nluc/GFP reporter, as shown in Figure 5.
Figure 1: Cartoon illustrating the two gain-of-function reporter systems. (A) The Renilla-Firefly reporter system is comprised of two luciferase enzymes expressed as a fusion protein and under the control of a tetracycline-inducible promoter. Expression of the wild-type dual enzyme results in high measurable activity of both Renilla and firefly luciferases (top). Mutation of a critical aspartate residue, D120, in Renilla to asparagine renders the Renilla enzyme (D120N) inactive, but firefly luciferase is still active. Mistranslation of asparagine to aspartate (in this case, from physiologically misacylated Asp-tRNAAsn tRNA5,21) results in a small proportion of the translated Renilla proteins gaining activity, which can be measured. Comparison of the Renilla/Firefly activity of the mutated reporter compared with the wild-type dual enzyme allows the calculation of the asparagine to an aspartate mistranslation rate. (B) The Nluc/GFP reporter system operates on a similar principle. The mutated Nluc (D140N) gene has an N-terminal secretion signal derived from antigen 85A, a major secreted mycobacterial protein. Both the nluc and gfp genes are expressed from identical tetracycline-inducible promoters, to allow the use of GFP as a relative expression benchmark. Note the lack of wild-type Nluc control strain: this reporter is primarily used in screening, where measurements of relative mistranslation rates are sufficient for the primary screen. Please click here to view a larger version of this figure.
Figure 2: Outline of measuring the mistranslation rate using Renilla-Firefly reporter (section 1 of the protocol). Fresh cultures of M. smegmatis, transformed with a plasmid expressing the dual-luciferase reporters, are grown. The reporter expression is induced, and investigational compounds or conditions are tested. Following 4-6 h of reporter expression, the cultures are pelleted and lysed and the lysates are tested for dual-luciferase activity. Please click here to view a larger version of this figure.
Figure 3: Outline of measuring the mistranslation rate Nluc-GFP reporter (section 2 of the protocol). A fresh culture of M. smegmatis, transformed with the Nluc and GFP reporters, is grown to a sufficient volume for all the plates to be tested (assuming ~10 mL/96-well plate). Immediately prior to pipetting the bacterial culture into each well, induce the expression of the reporters with ATC added to the bulk culture. Incubate the wells by shaking them overnight. To measure the relative mistranslation rates, transfer the cultures to a black plate (to increase the sensitivity of fluorescence detection), measure the GFP fluorescence, and then, spin the plates down to pellet the bacteria. Transfer the supernatant to a white plate for Nluc activity measurement. Please click here to view a larger version of this figure.
Figure 4: Representative result of using Renilla-Firefly reporter measuring the mistranslation rate in the presence of kasugamycin in different strains. Mistranslation rates of aspartate for asparagine in (A) wild-type M. smegmatis and (B) a strain deleted for ksgA (∆ksgA) following treatment with kasugamycin (ksg) as measured by the Renilla-Firefly dual reporter. The cultures were treated with kasugamycin at indicated doses (in micrograms/milliliter) for 6 h prior to cell lysis and measurements. The corrected Ren/FF ratios (y-axes) are indicative of the aspartate for asparagine mistranslation rates. The deletion of ksgA results in a higher baseline mistranslation rate, which is more resistant to modulation by kasugamycin. The bars indicate the mean values, with standard deviation as error. Please click here to view a larger version of this figure.
Figure 5: Representative result of using Nluc-GFP reporter measuring the mistranslation rate in the presence of kasugamycin. Relative mistranslation rates of aspartate for asparagine in wild-type M. smegmatis as measured by the Nluc/GFP reporter, following treatment with kasugamycin (ksg) overnight (16 h). The bars indicate the mean values, with standard deviation as error. Please click here to view a larger version of this figure.
The protocol described here can be adapted for the measurement of mistranslation rates in a wide variety of organisms. There are a number of considerations that should be kept in mind when adapting the protocol to other systems. First, the purpose of the measurement should be considered. For the accurate measurement of mistranslation rates using gain-of-function reporters, what is needed is (a) a robust readout assay (e.g., enzymatic function [in this case, luciferase activity] with a broad linear range). Also, look for (b) a loss of function mutation in a critical residue of the reporter protein that results in a very significant loss of function, below the expected rate of mistranslation. For example, if the expected mistranslation rate to be measured is approximately 10-3/codon, the loss of function mutation needs to be greater than 1,000-fold; otherwise, the lower range of mistranslation events will not be sensitively detected. Lastly, for the accurate measurement of mistranslation rates using gain-of-function reporters, (c) a robust benchmark reporter is needed — in this case, firefly luciferase for section 1 of the protocol. Ideally, the benchmark reporter should be fused to the mutated enzymatic reporter, guaranteeing (for all intents and purposes) that they are both expressed in equimolar ratios. The benchmark (and primary reporter) readout should be robust to perturbations to exposed environments. For example, the fluorescence of wild-type GFP is highly susceptible to perturbation by changes in pH28, making it unsuitable as a benchmark for the measurement of mistranslation rates in phagocytosed mycobacteria in macrophages, which reside in phagosomes that are below pH 729. On the other hand, for applications such as small molecule screening, the most important considerations for a primary screen will be the reproducibility of the measurements (i.e., precision, as opposed to accuracy), the minimization of manual handling, and small culture volumes. The accurate measurements of mistranslation rates are less important and can be performed as secondary assays. Finally, it should be noted that the reporters described in this protocol, based on the enzymatic function of luciferase enzymes, will only measure mean mistranslation rates of a bacterial population but cannot give information about single-cell heterogeneity variation in mistranslation rates. Given the importance of single-cell variability in adaptive phenotypes30,31,32, fluorescent reporters for the measurement of mistranslation events have been developed12,33, including for the measurement of mistranslation in mycobacteria5.
Having considered the requirements for the measurement of mistranslation, it should be noted that the genetic gain-of-function reporters such as described in this protocol can only measure one type of mistranslation (i.e., one amino acid substitution for one codon) per reporter. Therefore, for the measurement of different mistranslation events, multiple reporters are required. For example, for the measurement of mistranslation by ribosomal decoding errors of near-cognate codons by lysyl-tRNA at one position, at least 16 reporters are required2,3,11. High-precision mass spectrometry and bioinformatics allow for the potential identification of many different types of mistranslation events simultaneously18; however, these methods also come with caveats. In general, they are less accurate than gain-of-function reporters. Furthermore, as stated earlier, they are less suitable for the detection of certain types of mistranslation, namely those that could be conflated with nonenzymatic deamidation23. Finally, mass spectrometry is unsuitable as a read-out for medium- or high-throughput screening.
For the adaptation of these reporters and protocols for the measurement of mistranslation in other systems, further considerations include a robust expression of the reporter in the desired model system. We initially attempted to use the dual-luciferase system developed by Farabaugh and colleagues in our mycobacterial system without modifications but had no success, due to a lack of reporter expression. Extensive troubleshooting identified that both codon-optimization and a minor sequence modification of the reporters were required to allow for robust expression25 (and see above). It is conceivable that similar adaptations of the reporters would be required for use in other systems.
Finally, consideration should be given to the type of mistranslation event being measured. For example, if there is a need to measure stop-codon readthrough (nonsense suppression), the stop codon’s loss of function mutation needs to be introduced within a noncritical region of the primary reporter protein34. Otherwise, only specific nonsense suppression events that regain the critical residue (and, hence, reporter function) will be measured by the assay, which would potentially lead to a substantial underestimation of the true mistranslation rates. There is increasing evidence that translational error plays a possible adaptive role in a large number of organisms1,12,13,35. However, the measurement of mistranslation events is still limited to a small, albeit growing number of model species. The adaptation of sensitive methods to measure mistranslation has the potential to further increase scientists’ understanding of the role of translation error in physiology and pathology.
The authors have nothing to disclose.
This work was in part supported by grants from the Bill and Melinda Gates Foundation (OPP1109789), the National Natural Science Foundation of China (31570129), and start-up funds from the Tsinghua University School of Medicine to B.J. B.J. is a Wellcome Trust Investigator (207487/Z/17/Z).
Middlebrook 7H9 | BD Difco | 271310 | |
anhydrotetracycline | Cayman Chemical | 10009542 | |
kasugamycin | sigma | K4013 | |
Dual-luciferase reporter assay system | promega | E1960 | |
Nano-Glo luciferase assay | promega | N1120 | |
Fluoroskan Ascent FL luminometer | Thermo | / | |
Assay Plate, 96 Well White, Flat Bottom High Binding, No Lid | Costar | 3922 | |
Assay Plate, 96 Well Black, Flat Bottom High Biding, No Lid | Costar | 3925 | |
96 Well Cell Culture Cluster, Flat Bottom with Lid | Costar | 3599 | |
Non-commercial reagents (plasmids) | |||
pUV-TetOR-RenFF | NA | NA | episomal shuttle plasmid that allows tetracycline-inducible expression of the wild-type dual luciferase reporter |
pUV-TetOR-Ren-D120N-FF | dual-luciferase reporter with mutated Renilla | ||
pUV-TetOR-Ag85ASec-Nluc-D140N | NA | NA | episomal shuttle plasmid with a tetracyclin-inducible secretable version of mutated Nluc luciferase |
pMC1S-GFP | NA | NA | Mycobacterial integrated (L5 site) vector for tetracycline-inducible expression of GFP |