- Email: [email protected]
Translational misreading in Mycobacterium smegmatis increases in stationary phase
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54
YTUBE1385_proof ■ 21 October 2015 ■ 1/4
Tuberculosis xxx (2015) 1e4
Contents lists available at ScienceDirect
Tuberculosis journal homepage: http://intl.elsevierhealth.com/journals/tube
MOLECULAR ASPECTS
Q3
Translational misreading in Mycobacterium smegmatis increases in stationary phase
Q2
Tianqi Leng 1, 2, Miaomiao Pan 2, Xin Xu, Babak Javid* Collaborative Innovation Centre for the Diagnosis and Treatment of Infectious Diseases, School of Medicine, Tsinghua University, Beijing, 100084, China
a r t i c l e i n f o
s u m m a r y
Article history: Received 9 July 2015 Accepted 20 September 2015
The study of errors in gene translation has largely been confined to a small number of model organisms. We have examined all possible misreading errors at a defined codon in Mycobacterium smegmatis. Using a dual-luciferase gain of function reporter system that employs a mutated essential lysine in firefly luciferase, we accurately quantified mistranslation errors. Overall, accuracy of gene translation was comparable with Escherichia coli at <1/2000 errors/codon during exponential growth. Stationary phase was associated with a dramatic increase in misincorporation errors by Lys-tRNALys CUU at a subset of three codons, each with a single base changed from the AAG lysine codon. The maximum error rate detected was 0.2% with codon AUG. Treatment with streptomycin increased misreading errors at several codons associated in particular with U$U, G$U and C$U codon$anti-codon mismatches, but oxidative stress did not change translational fidelity. Our study is the first comprehensive examination of misreading errors for a defined codon in mycobacteria. © 2015 Published by Elsevier Ltd.
Keywords: Mistranslation Decoding errors Mycobacterium
1. Introduction Accurate gene translation is essential for life, but growing evidence suggests that optimal translational fidelity varies between organisms and organelles, is tunable, and depends on environmental conditions and other stressors [1e6]. In particular, a growing body of evidence suggests that translational error may play a role in the generation of phenotypic diversity and may be adaptive under certain conditions such as under antibiotic stress [1,7]. Although errors in gene translation have been studied for over 50 years [6], studies have largely been limited to a small number of mostly model organisms such as Escherichia coli and Saccharomyces cerevisiae in a limited number of codons, and only a few studies have been initiated in mycobacteria [1,8]. A better understanding of fundamental mycobacterial physiology, such as regulation of translational fidelity is essential, given that mycobacteria likely have a very different requirement for translation than most model organisms [9] and that regulation of translational fidelity may be important for antibiotic resistance and tolerance [1,8].
We had previously investigated rates of translational error, mistranslation, in Mycobacterium smegmatis for glutamate and aspartate using a dual-luciferase based gain of function reporter [1]. We found that mistranslation rates were much higher than E. coliK12 for similar codons [10] at 0.2e0.8%/codon in log phase, rising to 2%/codon in stationary phase [1]. We used the same basic reporter system as our previous study, but with minor modifications. We mutated a critical lysine codon (K529) in the firefly enzyme [11] rather than the Renilla enzyme as previously [1], which allowed us, for the first time, to comprehensively investigate decoding errors at a defined codon in mycobacteria. We mutated codon529 in firefly luciferase to all near-cognate codons for lysine (AAA and AAG) and investigated the rate of misincorporation of Lys-tRNALys to calculate the ribosome misreading rate. We found that decoding in mycobacteria in log phase is fairly accurate, and comparable to those observed in E. coli-K12, but rise significantly in stationary phase, up to 0.2%/codon for misincorporation of Lys-tRNALys CUU for the methionine codon AUG. 2. Materials and methods
* Corresponding author. E-mail address: [email protected] (B. Javid). 1 Current address: Nuffield Department of Medicine, University of Oxford, Oxford, United Kingdom. 2 These authors contributed equally to this work.
2.1. Bacterial culture and strains Wild-type (WT) M. smegmatis mc2-155 was used throughout the study and cultured in Middlebrook 7H9 culture medium
http://dx.doi.org/10.1016/j.tube.2015.09.010 1472-9792/© 2015 Published by Elsevier Ltd.
Please cite this article in press as: Leng T, et al., Translational misreading in Mycobacterium smegmatis increases in stationary phase, Tuberculosis (2015), http://dx.doi.org/10.1016/j.tube.2015.09.010
55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
YTUBE1385_proof ■ 21 October 2015 ■ 2/4
2
T. Leng et al. / Tuberculosis xxx (2015) 1e4
supplemented with albumin, dextrose and salt (ADS), glycerol and 0.05% Tween-20 unless otherwise stated. The plasmids for the dual-luciferase reporters were transformed into M. smegmatis by electroporation and transformants selected on LB-agar supplemented with 50 mg/ml of hygromycin. All cloning was performed in E. coli TOP10 (CW Biotech) or DH5a (Life Technologies) using standard methods. Successfully transformed clones were picked after 3e4 days. All chemicals were from Sigma unless otherwise stated. 2.2. Plasmids and oligonucleotides
peroxide (0.2 mM) e the stressor was added at the same time as ATc. 2.5. Statistical analysis Data are shown as means ± standard deviation. Mistranslation rates were calculated as previously [1,11]. Comparisons between pairs of means were conducted using a two-tailed t-test. All analyses were done using GraphPad Prism® version 6.0 (San Diego, CA). P-values <0.05 (*), <0.01 (**) and <0.001 (***) were considered statistically significant.
Plasmid
Origins and functions
pDONR221-RenFFluc
Gateway®-adapted vector for generation of attL-flanked entry clones containing the dual-luciferase gene, used for site-directed mutagenesis and generation of expression plasmid by Gateway cloning. All reporters were made in this strain prior to LR recombination in the tet-regulated plasmid. Destination shuttle vector used to clone wild-type and mutant dual-luciferase reporters from pDONR221, derived from pVU15tetOR, containing replication of origin for both E. coli and M. smegmatis
pTet(RenFFluc)
2.2.1. Mutagenesis oligonucleotides Only the forward primer is listed. The reverse primer was complementary.
Mutant codon
Mutant amino acid
Forward primer
AAC AAU GAG GAA CAG CAA AGG AGA AUG AUA ACA ACG UAA UAG UGG
Asparagine Asparagine Glutamate Glutamate Glutamine Glutamine Arginine Arginine Methionine Isoleucine Threonine Threonine Stop codon Stop codon Tryptophan
50 -CCTGACCGGCAACCTCGACG-30 50 -CCTGACCGGCAATCTCGACG-30 50 -CCTGACCGGCGAGCTCGACG-30 50 -CCTGACCGGCGAACTCGACG-30 50 -CCTGACCGGCCAGCTCGACG-30 50 -CCTGACCGGCCAACTCGACG-30 50 -CCTGACCGGCAGGCTCGACG-30 50 -CCTGACCGGCAGACTCGACG-30 50 -CCTGACCGGCATGCTCGACG-30 50 -CCTGACCGGCATACTCGACG-30 50 -CCTGACCGGCACACTCGACG-30 50 -CCTGACCGGCACGCTCGACG-30 50 -CCTGACCGGCTAACTCGACG-30 50 -CCTGACCGGCTAGCTCGACG-30 50 -CCTGACCGGCTGGCTCGACG-30
2.3. Site-directed mutagenesis and gateway cloning This was performed using the ‘megaprimer’ strategy using standard methods, as before [1]. The dual-luciferase reporter was cloned into a tetracycline-inducible plasmid using Gateway technology following the manufacturer's instructions. 2.4. Dual luciferase assay This was performed as previously [1] using a similar approach as Kramer and Farabaugh and Grentzmann et al. [11,12] but with a codon- and otherwise optimized dual-luciferase reporter as before. Briefly, 3 colonies of M. smegmatis transformed with the plasmid containing the dual-luciferase reporter were picked for each reporter and grown to stationary phase for two days. Cultures were back-diluted to log phase (1/20 dilution) unless otherwise stated and anhydrotetracyline (ATc) added as an inducer to 50 ng/ml for 6 h before harvesting and lysing the cells and reading luciferase activity with the dual-luciferase assay kit (Promega) using a Fluoroskan Ascent FL luminometer using 1s integration times. For the experiments involving streptomycin (0.2 mg/ml) and hydrogen
3. Results and discussion 3.1. Mistranslation rates vary widely for FFLucK529 nearsynonymous codons in M. smegmatis We wished to investigate the range of mistranslation rates in M. smegmatis at a defined codon. Similar to Kramer and Farabaugh, we made a series of 16 reporter strains expressing a fusion protein of Renilla (Ren) and Firefly (FF) luciferases [11]. The control reporter expressed both WT enzymes. Firefly luciferase (FFluc) has a critical lysine residue at codon529, which when mutated to any other residue loses >99.99% of its activity. Given that lysine can be coded by either AAA or AAG codons, there are 14 possible ‘near-cognate’ codons e i.e. codons that differ from lysine by only one base. We made all 14 near-cognate reporters in FFluc, retaining the Ren luciferase as the WT enzyme. These 14 reporters would have a very low basal FFluc activity, but normal Renluc activity. Variation in FFluc activity would be due to misincorporation of Lys-tRNALys at codon529, since total protein expression under different conditions would be controlled by comparison of Renluc activities. The actual mistranslation rate would be calculated by comparing the FFluc activity in the reporter strain with the WT enzyme, after correcting for variation in Renluc activity [1,11,12]. We reasoned that ribosomal decoding errors are most likely for near-cognate codons (differing from lysine by one base), therefore as a negative control, we chose a non-near-cognate codon, UGG, coding for tryptophan, which differs from lysine codons by at least 2 bases. Under log phase growth (the OD600 of the log phase bacteria increased from ~0.3 to 1.0 for the 6 h of ATc induction), corrected FFluc activity in the reporter strains, representing lysine misincorporation, varied from 0.0025%/codon to 0.072%/codon (Figure 1). The maximal error rates were for the codons AGA (arginine) and AAU (asparagine), and the least error was seen at the stop codons UAA and UAG, which were less than the limit of detection of our reporter system. The maximal error rates, at ~0.07%/codon were comparable to those seen in E. coli (0.04%/ codon for AGA) [11]. In both E. coli and M. smegmatis, the AGA codon is the rarest arginine codon (codon frequency 0.11), and in Farabaugh's study, lack of tRNA availability for rare codons was thought to contribute to misincorporation of lysine tRNA [11]. However, the AAU (asparagine) codon, although it is the less abundant asparagine codon (codon frequency 0.25 compared with
Please cite this article in press as: Leng T, et al., Translational misreading in Mycobacterium smegmatis increases in stationary phase, Tuberculosis (2015), http://dx.doi.org/10.1016/j.tube.2015.09.010
66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
YTUBE1385_proof ■ 21 October 2015 ■ 3/4
T. Leng et al. / Tuberculosis xxx (2015) 1e4
3
3.3. Streptomycin but not oxidative stress increases mistranslation by lysyl-tRNA at a large number of near-synonymous codons
Figure 1. Stationary phase increases mycobacterial ribosomal decoding errors by Lys-tRNALys CUU . The change in FFluc activity in the codon529 mutant reporters is shown in both exponential and stationary phase growth. The mutated codon is shown with the single base substitution in lower case, and the coded amino acid below the codon. The data represent means of triplicates.
1.96 for AAC), is not particularly rare. M. smegmatis does not code for the asparagine tRNA isoacceptor for AAU (anticodon AUU), therefore all AAU codons will be translated by wobble translation by the GUU tRNA isoacceptor. Furthermore, comprehensive transcriptional profiling in Mycobacterium tuberculosis suggests that codon frequency is not a good indication of tRNA abundance [13], therefore although it is possible that Asn-tRNAAsn GUU is limiting, it is likely that the cause of increased translational error at this codon by Lys-tRNALys is due to some other mechanism. 3.2. Stationary phase is associated with increased misincorporation of tRNALys CUU at a subset of near-synonymous codons Under stationary phase culture (OD600 ~6), mistranslation rates for a subset of codons increased dramatically (Figure 1). The misincorporation of lysine for methionine (AUG codon) and glutamine (CAG codon) in particular rose 8.6- and 7.6-fold respectively compared with log phase growth, and the absolute error rate for lysine misincorporation for methionine approached 0.2%/codon e approximately five-fold higher than the maximal errors seen under normal growth in E. coli-K12 using a similar reporter [11], although it should be noted that those studies were performed under log phase growth exclusively. However, prior reports, examining frameshifting and stop-codon read-through errors suggested that E. coli also increased translational error during stationary phase [14], therefore the shift from exponential growth to stationary phase may be associated with a general decrease in translational fidelity in bacteria. Of note, the increased mistranslation in M. smegmatis during stationary phase appeared to be exclusively associated with misincorporation of only one of the two lysine tRNAs e Lys-tRNALys CUU . E. coli-K12 only codes for one tRNA gene ðtRNALys UUU Þ present in 6 copies that functions for both lysine codons AAA and AAG. M. smegmatis on the other hand, has a single gene specifying an AAA decoding lysine tRNA and a single gene specifying an AAG decoding lysine tRNA. During stationary phase growth, a significant increase in mistranslation was observed in 3 of the reporter strains, those with Gln (CAG), Arg (AGG) and Met (ATG) substitutions at position 529 of FFluc. All three codons: CAG, AGG and AUG are near-synonymous only for Lys Lysine-AAG, which would be translated by tRNACUU. Given that amino-acyl tRNA availability may be responsible, at least in part, for the variation in mistranslation of near-synonymous codons [11], it may be that the abundance of tRNALys CUU is specifically upregulated and more available to compete with the cognate aminoacyl-tRNAs for translation during the transition from log to stationary phase.
We also wished to examine whether rates of mistranslation varied according to other stressors in the environment. The small variation in baseline error rates observed between experiments (Figures 1 and 2) suggests that the interaction of mycobacteria with their environment can significantly affect translational fidelity. We first investigated whether the aminoglycoside streptomycin e the first useful anti-tuberculous antibiotic e affected mistranslation rates. It has long been known that streptomycin increases decoding errors, especially stop-codon readthrough. Of the 14 near-cognate codons, there was increased mistranslation in 7 of them with the addition of sub-lethal doses of streptomycin (Figure 2A). Prior work by Farabaugh and colleagues, as well as others suggested that aminoglycosides in particular favoured misreading of U$U in the first or second position, as well as G$U in the second position of the codon [11]. In M. smegmatis, the increase in misreading was most common with U$U mismatches – in all three codon positions. There was also misreading of G$U, C$U and C$C mismatches to a lesser extent. Mismatch position within the codon appeared less important, with misreading at three positions occurring equally frequently (Figure 2A). Due to mycobacteria coding for both lysine tRNAs, as noted above, the increased mistranslation at Asn (AAC and AAU) with streptomycin may have been due to insertion of either lysyl-tRNA. Recent work in both eukaryotic and bacterial systems has suggested that oxidative stress may increase translational error [3,4,15,16]. We investigated whether exposure to hydrogen peroxide may increase misreading of near-synonymous codons by lysl-tRNA using our reporter system. At the tested dose of hydrogen peroxide (0.2 mM), we did not detect any significant increase in mycobacterial mistranslation (Figure 2B). This suggests
Figure 2. Streptomycin but not oxidative stress is associated with widespread ribosomal decoding errors. Change in FFluc activity due to misincorporation by LystRNALys in exponential growth and treatment with 0.2 mg/ml of streptomycin (A) or 0.2 mM hydrogen peroxide (B). The data represent means of triplicates.
Please cite this article in press as: Leng T, et al., Translational misreading in Mycobacterium smegmatis increases in stationary phase, Tuberculosis (2015), http://dx.doi.org/10.1016/j.tube.2015.09.010
66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50
YTUBE1385_proof ■ 21 October 2015 ■ 4/4
4
T. Leng et al. / Tuberculosis xxx (2015) 1e4
that either mycobacteria, unlike E. coli, do not increase mistranslation under oxidative stress, or perhaps more likely, oxidative stress-induced mistranslation is specific for certain types of mistranslation only. In mammalian cells, oxidative stress specifically increased methionine misacylation [4]. In E. coli, oxidative stress also increased mistranslation in a specific manner, either through corruption of threonyl-aminoacyl tRNA synthetase [3,16] or discrimination by phenylalanyl-tRNA synthetase [15]. Since our genetic reporters were only probing mistranslation due to lysyl-tRNAs, they would not have detected these alternative forms of translational error. 4. Conclusions We have shown that translational error at a defined codon by M. smegmatis is relatively low during exponential phase, rising to 0.2%/codon in stationary phase. Although this rate of error is higher than previously reports in E. coli-K12 during log phase growth, it is still a full log lower than the 2%/codon errors that we observed in our previous study examining glutamine and asparagine codons [1]. This may be because there are several alternative routes to mistranslation: ribosomal decoding errors, such as measured in this study, or errors upstream of ribosomal decoding, which may have a larger magnitude. Recent work from our group suggests that the errors at the glutamine and asparagine codons were due to misacylated tRNAs rather than ribosomal decoding errors (BJ, forthcoming publication). Even though it is being recognized that an increasing number of organisms vary translational fidelity, and that mistranslation rates for a small number of amino acids can be extremely high (approaching 10%), error rates of this magnitude across all codons would rapidly result in error catastrophe [6] e a scenario induced by aminoglycoside treatment (Figure 2A). Methods to measure error are fraught with inherent problems in that small changes close to the level of background noise are measured. Methods for protein sequence determination, such as mass-spectrometry are exquisitely sensitive at detecting extremely small differences between two pure protein samples, but have both technical and bioinformatic limitations on detecting generalized error across the proteome (reviewed in reference 6). Genetic gain of function reporters allow for exquisite sensitivity for detection of error, whilst acknowledging that they can only detect one or a limited range of errors at a time [6]. Whilst our study is limited to measuring errors at a defined codon in a genetic reporter, it adds to the growing but small body of investigation of specific regulation of gene translation in mycobacteria e which may be critical for the development of new treatment modalities.
Acknowledgements We thank Melody Toosky for some of the initial cloning work. This work was in part supported by start-up funds from Tsinghua University School of Medicine to BJ. BJ is a Tsinghua-Janssen scholar. The funders had no role in the design, analysis or interpretation of the data, nor in the decision to publish the results. Funding:
None.
Competing interests: Ethical approval:
Q1
None declared. Not required.
References [1] Javid B, et al. Mycobacterial mistranslation is necessary and sufficient for rifampicin phenotypic resistance. Proc Natl Acad Sci U. S. A. 2014;111(3): 1132e7. [2] Li L, et al. Naturally occurring aminoacyl-tRNA synthetases editing-domain mutations that cause mistranslation in mycoplasma parasites. Proc Natl Acad Sci U. S. A. 2011;108(23):9378e83. [3] Ling J, Soll D. Severe oxidative stress induces protein mistranslation through impairment of an aminoacyl-tRNA synthetase editing site. Proc Natl Acad Sci U. S. A. 2010;107(9):4028e33. [4] Netzer N, et al. Innate immune and chemically triggered oxidative stress modifies translational fidelity. Nature 2009;462(7272):522e6. [5] Reynolds NM, et al. Cell-specific differences in the requirements for translation quality control. Proc Natl Acad Sci U. S. A. 2010;107(9):4063e8. [6] Ribas de Pouplana L, et al. Protein mistranslation: friend or foe? Trends Biochem Sci 2014;39(8):355e62. [7] Bezerra AR, et al. Reversion of a fungal genetic code alteration links proteome instability with genomic and phenotypic diversification. Proc Natl Acad Sci U. S. A. 2013;110(27):11079e84. [8] Wong SY, et al. Functional role of methylation of G518 of the 16S rRNA 530 loop by GidB in mycobacterium tuberculosis. Antimicrob Agents Chemother 2013;57(12):6311e8. [9] Cortes T, et al. Genome-wide mapping of transcriptional start sites defines an extensive leaderless transcriptome in mycobacterium tuberculosis. Cell Rep 2013;5(4):1121e31. [10] Manickam N, et al. Studies of translational misreading in vivo show that the ribosome very efficiently discriminates against most potential errors. RNA 2014;20(1):9e15. [11] Kramer EB, Farabaugh PJ. The frequency of translational misreading errors in E. coli is largely determined by tRNA competition. RNA 2007;13(1):87e96. [12] Grentzmann G, et al. A dual-luciferase reporter system for studying recoding signals. RNA 1998;4(4):479e86. [13] Arnvig KB, et al. Sequence-based analysis uncovers an abundance of noncoding RNA in the total transcriptome of mycobacterium tuberculosis. PLoS Pathog 2011;7(11):e1002342. [14] Barak Z, et al. Enhanced ribosome frameshifting in stationary phase cells. J Mol Biol 1996;263(2):140e8. [15] Bullwinkle TJ, et al. Oxidation of cellular amino acid pools leads to cytotoxic mistranslation of the genetic code. Elife 2014;3. [16] Wu J, Fan Y, Ling J. Mechanism of oxidant-induced mistranslation by threonyltRNA synthetase. Nucleic Acids Res 2014;42(10):6523e31.
Please cite this article in press as: Leng T, et al., Translational misreading in Mycobacterium smegmatis increases in stationary phase, Tuberculosis (2015), http://dx.doi.org/10.1016/j.tube.2015.09.010
51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100
YTUBE1385_proof ■ 21 October 2015 ■ 1/4
Tuberculosis xxx (2015) 1e4
Contents lists available at ScienceDirect
Tuberculosis journal homepage: http://intl.elsevierhealth.com/journals/tube
MOLECULAR ASPECTS
Q3
Translational misreading in Mycobacterium smegmatis increases in stationary phase
Q2
Tianqi Leng 1, 2, Miaomiao Pan 2, Xin Xu, Babak Javid* Collaborative Innovation Centre for the Diagnosis and Treatment of Infectious Diseases, School of Medicine, Tsinghua University, Beijing, 100084, China
a r t i c l e i n f o
s u m m a r y
Article history: Received 9 July 2015 Accepted 20 September 2015
The study of errors in gene translation has largely been confined to a small number of model organisms. We have examined all possible misreading errors at a defined codon in Mycobacterium smegmatis. Using a dual-luciferase gain of function reporter system that employs a mutated essential lysine in firefly luciferase, we accurately quantified mistranslation errors. Overall, accuracy of gene translation was comparable with Escherichia coli at <1/2000 errors/codon during exponential growth. Stationary phase was associated with a dramatic increase in misincorporation errors by Lys-tRNALys CUU at a subset of three codons, each with a single base changed from the AAG lysine codon. The maximum error rate detected was 0.2% with codon AUG. Treatment with streptomycin increased misreading errors at several codons associated in particular with U$U, G$U and C$U codon$anti-codon mismatches, but oxidative stress did not change translational fidelity. Our study is the first comprehensive examination of misreading errors for a defined codon in mycobacteria. © 2015 Published by Elsevier Ltd.
Keywords: Mistranslation Decoding errors Mycobacterium
1. Introduction Accurate gene translation is essential for life, but growing evidence suggests that optimal translational fidelity varies between organisms and organelles, is tunable, and depends on environmental conditions and other stressors [1e6]. In particular, a growing body of evidence suggests that translational error may play a role in the generation of phenotypic diversity and may be adaptive under certain conditions such as under antibiotic stress [1,7]. Although errors in gene translation have been studied for over 50 years [6], studies have largely been limited to a small number of mostly model organisms such as Escherichia coli and Saccharomyces cerevisiae in a limited number of codons, and only a few studies have been initiated in mycobacteria [1,8]. A better understanding of fundamental mycobacterial physiology, such as regulation of translational fidelity is essential, given that mycobacteria likely have a very different requirement for translation than most model organisms [9] and that regulation of translational fidelity may be important for antibiotic resistance and tolerance [1,8].
We had previously investigated rates of translational error, mistranslation, in Mycobacterium smegmatis for glutamate and aspartate using a dual-luciferase based gain of function reporter [1]. We found that mistranslation rates were much higher than E. coliK12 for similar codons [10] at 0.2e0.8%/codon in log phase, rising to 2%/codon in stationary phase [1]. We used the same basic reporter system as our previous study, but with minor modifications. We mutated a critical lysine codon (K529) in the firefly enzyme [11] rather than the Renilla enzyme as previously [1], which allowed us, for the first time, to comprehensively investigate decoding errors at a defined codon in mycobacteria. We mutated codon529 in firefly luciferase to all near-cognate codons for lysine (AAA and AAG) and investigated the rate of misincorporation of Lys-tRNALys to calculate the ribosome misreading rate. We found that decoding in mycobacteria in log phase is fairly accurate, and comparable to those observed in E. coli-K12, but rise significantly in stationary phase, up to 0.2%/codon for misincorporation of Lys-tRNALys CUU for the methionine codon AUG. 2. Materials and methods
* Corresponding author. E-mail address: [email protected] (B. Javid). 1 Current address: Nuffield Department of Medicine, University of Oxford, Oxford, United Kingdom. 2 These authors contributed equally to this work.
2.1. Bacterial culture and strains Wild-type (WT) M. smegmatis mc2-155 was used throughout the study and cultured in Middlebrook 7H9 culture medium
http://dx.doi.org/10.1016/j.tube.2015.09.010 1472-9792/© 2015 Published by Elsevier Ltd.
Please cite this article in press as: Leng T, et al., Translational misreading in Mycobacterium smegmatis increases in stationary phase, Tuberculosis (2015), http://dx.doi.org/10.1016/j.tube.2015.09.010
55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
YTUBE1385_proof ■ 21 October 2015 ■ 2/4
2
T. Leng et al. / Tuberculosis xxx (2015) 1e4
supplemented with albumin, dextrose and salt (ADS), glycerol and 0.05% Tween-20 unless otherwise stated. The plasmids for the dual-luciferase reporters were transformed into M. smegmatis by electroporation and transformants selected on LB-agar supplemented with 50 mg/ml of hygromycin. All cloning was performed in E. coli TOP10 (CW Biotech) or DH5a (Life Technologies) using standard methods. Successfully transformed clones were picked after 3e4 days. All chemicals were from Sigma unless otherwise stated. 2.2. Plasmids and oligonucleotides
peroxide (0.2 mM) e the stressor was added at the same time as ATc. 2.5. Statistical analysis Data are shown as means ± standard deviation. Mistranslation rates were calculated as previously [1,11]. Comparisons between pairs of means were conducted using a two-tailed t-test. All analyses were done using GraphPad Prism® version 6.0 (San Diego, CA). P-values <0.05 (*), <0.01 (**) and <0.001 (***) were considered statistically significant.
Plasmid
Origins and functions
pDONR221-RenFFluc
Gateway®-adapted vector for generation of attL-flanked entry clones containing the dual-luciferase gene, used for site-directed mutagenesis and generation of expression plasmid by Gateway cloning. All reporters were made in this strain prior to LR recombination in the tet-regulated plasmid. Destination shuttle vector used to clone wild-type and mutant dual-luciferase reporters from pDONR221, derived from pVU15tetOR, containing replication of origin for both E. coli and M. smegmatis
pTet(RenFFluc)
2.2.1. Mutagenesis oligonucleotides Only the forward primer is listed. The reverse primer was complementary.
Mutant codon
Mutant amino acid
Forward primer
AAC AAU GAG GAA CAG CAA AGG AGA AUG AUA ACA ACG UAA UAG UGG
Asparagine Asparagine Glutamate Glutamate Glutamine Glutamine Arginine Arginine Methionine Isoleucine Threonine Threonine Stop codon Stop codon Tryptophan
50 -CCTGACCGGCAACCTCGACG-30 50 -CCTGACCGGCAATCTCGACG-30 50 -CCTGACCGGCGAGCTCGACG-30 50 -CCTGACCGGCGAACTCGACG-30 50 -CCTGACCGGCCAGCTCGACG-30 50 -CCTGACCGGCCAACTCGACG-30 50 -CCTGACCGGCAGGCTCGACG-30 50 -CCTGACCGGCAGACTCGACG-30 50 -CCTGACCGGCATGCTCGACG-30 50 -CCTGACCGGCATACTCGACG-30 50 -CCTGACCGGCACACTCGACG-30 50 -CCTGACCGGCACGCTCGACG-30 50 -CCTGACCGGCTAACTCGACG-30 50 -CCTGACCGGCTAGCTCGACG-30 50 -CCTGACCGGCTGGCTCGACG-30
2.3. Site-directed mutagenesis and gateway cloning This was performed using the ‘megaprimer’ strategy using standard methods, as before [1]. The dual-luciferase reporter was cloned into a tetracycline-inducible plasmid using Gateway technology following the manufacturer's instructions. 2.4. Dual luciferase assay This was performed as previously [1] using a similar approach as Kramer and Farabaugh and Grentzmann et al. [11,12] but with a codon- and otherwise optimized dual-luciferase reporter as before. Briefly, 3 colonies of M. smegmatis transformed with the plasmid containing the dual-luciferase reporter were picked for each reporter and grown to stationary phase for two days. Cultures were back-diluted to log phase (1/20 dilution) unless otherwise stated and anhydrotetracyline (ATc) added as an inducer to 50 ng/ml for 6 h before harvesting and lysing the cells and reading luciferase activity with the dual-luciferase assay kit (Promega) using a Fluoroskan Ascent FL luminometer using 1s integration times. For the experiments involving streptomycin (0.2 mg/ml) and hydrogen
3. Results and discussion 3.1. Mistranslation rates vary widely for FFLucK529 nearsynonymous codons in M. smegmatis We wished to investigate the range of mistranslation rates in M. smegmatis at a defined codon. Similar to Kramer and Farabaugh, we made a series of 16 reporter strains expressing a fusion protein of Renilla (Ren) and Firefly (FF) luciferases [11]. The control reporter expressed both WT enzymes. Firefly luciferase (FFluc) has a critical lysine residue at codon529, which when mutated to any other residue loses >99.99% of its activity. Given that lysine can be coded by either AAA or AAG codons, there are 14 possible ‘near-cognate’ codons e i.e. codons that differ from lysine by only one base. We made all 14 near-cognate reporters in FFluc, retaining the Ren luciferase as the WT enzyme. These 14 reporters would have a very low basal FFluc activity, but normal Renluc activity. Variation in FFluc activity would be due to misincorporation of Lys-tRNALys at codon529, since total protein expression under different conditions would be controlled by comparison of Renluc activities. The actual mistranslation rate would be calculated by comparing the FFluc activity in the reporter strain with the WT enzyme, after correcting for variation in Renluc activity [1,11,12]. We reasoned that ribosomal decoding errors are most likely for near-cognate codons (differing from lysine by one base), therefore as a negative control, we chose a non-near-cognate codon, UGG, coding for tryptophan, which differs from lysine codons by at least 2 bases. Under log phase growth (the OD600 of the log phase bacteria increased from ~0.3 to 1.0 for the 6 h of ATc induction), corrected FFluc activity in the reporter strains, representing lysine misincorporation, varied from 0.0025%/codon to 0.072%/codon (Figure 1). The maximal error rates were for the codons AGA (arginine) and AAU (asparagine), and the least error was seen at the stop codons UAA and UAG, which were less than the limit of detection of our reporter system. The maximal error rates, at ~0.07%/codon were comparable to those seen in E. coli (0.04%/ codon for AGA) [11]. In both E. coli and M. smegmatis, the AGA codon is the rarest arginine codon (codon frequency 0.11), and in Farabaugh's study, lack of tRNA availability for rare codons was thought to contribute to misincorporation of lysine tRNA [11]. However, the AAU (asparagine) codon, although it is the less abundant asparagine codon (codon frequency 0.25 compared with
Please cite this article in press as: Leng T, et al., Translational misreading in Mycobacterium smegmatis increases in stationary phase, Tuberculosis (2015), http://dx.doi.org/10.1016/j.tube.2015.09.010
66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
YTUBE1385_proof ■ 21 October 2015 ■ 3/4
T. Leng et al. / Tuberculosis xxx (2015) 1e4
3
3.3. Streptomycin but not oxidative stress increases mistranslation by lysyl-tRNA at a large number of near-synonymous codons
Figure 1. Stationary phase increases mycobacterial ribosomal decoding errors by Lys-tRNALys CUU . The change in FFluc activity in the codon529 mutant reporters is shown in both exponential and stationary phase growth. The mutated codon is shown with the single base substitution in lower case, and the coded amino acid below the codon. The data represent means of triplicates.
1.96 for AAC), is not particularly rare. M. smegmatis does not code for the asparagine tRNA isoacceptor for AAU (anticodon AUU), therefore all AAU codons will be translated by wobble translation by the GUU tRNA isoacceptor. Furthermore, comprehensive transcriptional profiling in Mycobacterium tuberculosis suggests that codon frequency is not a good indication of tRNA abundance [13], therefore although it is possible that Asn-tRNAAsn GUU is limiting, it is likely that the cause of increased translational error at this codon by Lys-tRNALys is due to some other mechanism. 3.2. Stationary phase is associated with increased misincorporation of tRNALys CUU at a subset of near-synonymous codons Under stationary phase culture (OD600 ~6), mistranslation rates for a subset of codons increased dramatically (Figure 1). The misincorporation of lysine for methionine (AUG codon) and glutamine (CAG codon) in particular rose 8.6- and 7.6-fold respectively compared with log phase growth, and the absolute error rate for lysine misincorporation for methionine approached 0.2%/codon e approximately five-fold higher than the maximal errors seen under normal growth in E. coli-K12 using a similar reporter [11], although it should be noted that those studies were performed under log phase growth exclusively. However, prior reports, examining frameshifting and stop-codon read-through errors suggested that E. coli also increased translational error during stationary phase [14], therefore the shift from exponential growth to stationary phase may be associated with a general decrease in translational fidelity in bacteria. Of note, the increased mistranslation in M. smegmatis during stationary phase appeared to be exclusively associated with misincorporation of only one of the two lysine tRNAs e Lys-tRNALys CUU . E. coli-K12 only codes for one tRNA gene ðtRNALys UUU Þ present in 6 copies that functions for both lysine codons AAA and AAG. M. smegmatis on the other hand, has a single gene specifying an AAA decoding lysine tRNA and a single gene specifying an AAG decoding lysine tRNA. During stationary phase growth, a significant increase in mistranslation was observed in 3 of the reporter strains, those with Gln (CAG), Arg (AGG) and Met (ATG) substitutions at position 529 of FFluc. All three codons: CAG, AGG and AUG are near-synonymous only for Lys Lysine-AAG, which would be translated by tRNACUU. Given that amino-acyl tRNA availability may be responsible, at least in part, for the variation in mistranslation of near-synonymous codons [11], it may be that the abundance of tRNALys CUU is specifically upregulated and more available to compete with the cognate aminoacyl-tRNAs for translation during the transition from log to stationary phase.
We also wished to examine whether rates of mistranslation varied according to other stressors in the environment. The small variation in baseline error rates observed between experiments (Figures 1 and 2) suggests that the interaction of mycobacteria with their environment can significantly affect translational fidelity. We first investigated whether the aminoglycoside streptomycin e the first useful anti-tuberculous antibiotic e affected mistranslation rates. It has long been known that streptomycin increases decoding errors, especially stop-codon readthrough. Of the 14 near-cognate codons, there was increased mistranslation in 7 of them with the addition of sub-lethal doses of streptomycin (Figure 2A). Prior work by Farabaugh and colleagues, as well as others suggested that aminoglycosides in particular favoured misreading of U$U in the first or second position, as well as G$U in the second position of the codon [11]. In M. smegmatis, the increase in misreading was most common with U$U mismatches – in all three codon positions. There was also misreading of G$U, C$U and C$C mismatches to a lesser extent. Mismatch position within the codon appeared less important, with misreading at three positions occurring equally frequently (Figure 2A). Due to mycobacteria coding for both lysine tRNAs, as noted above, the increased mistranslation at Asn (AAC and AAU) with streptomycin may have been due to insertion of either lysyl-tRNA. Recent work in both eukaryotic and bacterial systems has suggested that oxidative stress may increase translational error [3,4,15,16]. We investigated whether exposure to hydrogen peroxide may increase misreading of near-synonymous codons by lysl-tRNA using our reporter system. At the tested dose of hydrogen peroxide (0.2 mM), we did not detect any significant increase in mycobacterial mistranslation (Figure 2B). This suggests
Figure 2. Streptomycin but not oxidative stress is associated with widespread ribosomal decoding errors. Change in FFluc activity due to misincorporation by LystRNALys in exponential growth and treatment with 0.2 mg/ml of streptomycin (A) or 0.2 mM hydrogen peroxide (B). The data represent means of triplicates.
Please cite this article in press as: Leng T, et al., Translational misreading in Mycobacterium smegmatis increases in stationary phase, Tuberculosis (2015), http://dx.doi.org/10.1016/j.tube.2015.09.010
66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50
YTUBE1385_proof ■ 21 October 2015 ■ 4/4
4
T. Leng et al. / Tuberculosis xxx (2015) 1e4
that either mycobacteria, unlike E. coli, do not increase mistranslation under oxidative stress, or perhaps more likely, oxidative stress-induced mistranslation is specific for certain types of mistranslation only. In mammalian cells, oxidative stress specifically increased methionine misacylation [4]. In E. coli, oxidative stress also increased mistranslation in a specific manner, either through corruption of threonyl-aminoacyl tRNA synthetase [3,16] or discrimination by phenylalanyl-tRNA synthetase [15]. Since our genetic reporters were only probing mistranslation due to lysyl-tRNAs, they would not have detected these alternative forms of translational error. 4. Conclusions We have shown that translational error at a defined codon by M. smegmatis is relatively low during exponential phase, rising to 0.2%/codon in stationary phase. Although this rate of error is higher than previously reports in E. coli-K12 during log phase growth, it is still a full log lower than the 2%/codon errors that we observed in our previous study examining glutamine and asparagine codons [1]. This may be because there are several alternative routes to mistranslation: ribosomal decoding errors, such as measured in this study, or errors upstream of ribosomal decoding, which may have a larger magnitude. Recent work from our group suggests that the errors at the glutamine and asparagine codons were due to misacylated tRNAs rather than ribosomal decoding errors (BJ, forthcoming publication). Even though it is being recognized that an increasing number of organisms vary translational fidelity, and that mistranslation rates for a small number of amino acids can be extremely high (approaching 10%), error rates of this magnitude across all codons would rapidly result in error catastrophe [6] e a scenario induced by aminoglycoside treatment (Figure 2A). Methods to measure error are fraught with inherent problems in that small changes close to the level of background noise are measured. Methods for protein sequence determination, such as mass-spectrometry are exquisitely sensitive at detecting extremely small differences between two pure protein samples, but have both technical and bioinformatic limitations on detecting generalized error across the proteome (reviewed in reference 6). Genetic gain of function reporters allow for exquisite sensitivity for detection of error, whilst acknowledging that they can only detect one or a limited range of errors at a time [6]. Whilst our study is limited to measuring errors at a defined codon in a genetic reporter, it adds to the growing but small body of investigation of specific regulation of gene translation in mycobacteria e which may be critical for the development of new treatment modalities.
Acknowledgements We thank Melody Toosky for some of the initial cloning work. This work was in part supported by start-up funds from Tsinghua University School of Medicine to BJ. BJ is a Tsinghua-Janssen scholar. The funders had no role in the design, analysis or interpretation of the data, nor in the decision to publish the results. Funding:
None.
Competing interests: Ethical approval:
Q1
None declared. Not required.
References [1] Javid B, et al. Mycobacterial mistranslation is necessary and sufficient for rifampicin phenotypic resistance. Proc Natl Acad Sci U. S. A. 2014;111(3): 1132e7. [2] Li L, et al. Naturally occurring aminoacyl-tRNA synthetases editing-domain mutations that cause mistranslation in mycoplasma parasites. Proc Natl Acad Sci U. S. A. 2011;108(23):9378e83. [3] Ling J, Soll D. Severe oxidative stress induces protein mistranslation through impairment of an aminoacyl-tRNA synthetase editing site. Proc Natl Acad Sci U. S. A. 2010;107(9):4028e33. [4] Netzer N, et al. Innate immune and chemically triggered oxidative stress modifies translational fidelity. Nature 2009;462(7272):522e6. [5] Reynolds NM, et al. Cell-specific differences in the requirements for translation quality control. Proc Natl Acad Sci U. S. A. 2010;107(9):4063e8. [6] Ribas de Pouplana L, et al. Protein mistranslation: friend or foe? Trends Biochem Sci 2014;39(8):355e62. [7] Bezerra AR, et al. Reversion of a fungal genetic code alteration links proteome instability with genomic and phenotypic diversification. Proc Natl Acad Sci U. S. A. 2013;110(27):11079e84. [8] Wong SY, et al. Functional role of methylation of G518 of the 16S rRNA 530 loop by GidB in mycobacterium tuberculosis. Antimicrob Agents Chemother 2013;57(12):6311e8. [9] Cortes T, et al. Genome-wide mapping of transcriptional start sites defines an extensive leaderless transcriptome in mycobacterium tuberculosis. Cell Rep 2013;5(4):1121e31. [10] Manickam N, et al. Studies of translational misreading in vivo show that the ribosome very efficiently discriminates against most potential errors. RNA 2014;20(1):9e15. [11] Kramer EB, Farabaugh PJ. The frequency of translational misreading errors in E. coli is largely determined by tRNA competition. RNA 2007;13(1):87e96. [12] Grentzmann G, et al. A dual-luciferase reporter system for studying recoding signals. RNA 1998;4(4):479e86. [13] Arnvig KB, et al. Sequence-based analysis uncovers an abundance of noncoding RNA in the total transcriptome of mycobacterium tuberculosis. PLoS Pathog 2011;7(11):e1002342. [14] Barak Z, et al. Enhanced ribosome frameshifting in stationary phase cells. J Mol Biol 1996;263(2):140e8. [15] Bullwinkle TJ, et al. Oxidation of cellular amino acid pools leads to cytotoxic mistranslation of the genetic code. Elife 2014;3. [16] Wu J, Fan Y, Ling J. Mechanism of oxidant-induced mistranslation by threonyltRNA synthetase. Nucleic Acids Res 2014;42(10):6523e31.
Please cite this article in press as: Leng T, et al., Translational misreading in Mycobacterium smegmatis increases in stationary phase, Tuberculosis (2015), http://dx.doi.org/10.1016/j.tube.2015.09.010
51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100