Engineering of tRNAPro1E2 anticodon stem enhances multiple/consecutive ribosomal incorporation of N-methyl-l-α-amino acids and d-α-amino acids
Ryoichi Hirashima, Takayuki Katoh, Hiroaki Suga
TL;DR
Researchers engineered a tRNA to improve the ribosomal incorporation of nonproteinogenic amino acids, enabling the synthesis of complex peptides with multiple exotic amino acids.
Contribution
Engineering the anticodon stem of tRNAPro1E2 significantly enhances the incorporation of multiple nonproteinogenic amino acids in consecutive ribosomal elongation.
Findings
Eleven tRNAPro1E2 variants increased incorporation of MeLeu and d-Phe by up to 4.5- and 4.4-fold.
Seven variants showed distinct conformations that likely aid in the peptidyl transfer reaction.
A macrocyclic peptide with three consecutive MeAAs and three d-AAs was successfully synthesized.
Abstract
Transfer RNA (tRNA) structures influence the incorporation efficiency of amino acids, particularly when nonproteinogenic amino acids (npAAs), such as N-methyl-l-α-amino acids (MeAAs) and d-α-amino acids (d-AAs), are charged. Such npAAs are generally far poorer substrates than the 20 canonical α-amino acids for translation. However, we have shown that their incorporation efficiencies could be improved by using a chimeric tRNA, termed tRNAPro1E2, bearing optimal T-stem and D-arm. Here we report that the engineering anticodon stem of tRNAPro1E2 further enhances multiple/consecutive elongation of MeAAs and d-AAs. By screening 149 types of anticodon stem mutants, we found eleven tRNAPro1E2 variants capable of enhancing the incorporation of N-methyl-l-leucine (MeLeu) or d-phenylalanine (d-Phe) at six codons by up to 4.5- and 4.4-fold, respectively. Interestingly, 7 out of the 11 variants…
Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.
Click any figure to enlarge with its caption.
Figure 1
Figure 2
Figure 3
Figure 4
Figure 5
Figure 6
Figure 7
Figure 8- —Japan Society for the Promotion of Science10.13039/501100001691
- —JSPS10.13039/501100000646
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsRNA and protein synthesis mechanisms · RNA modifications and cancer · DNA and Nucleic Acid Chemistry
Introduction
Building blocks of peptides in the ordinary translation system are generally limited to 20 proteinogenic amino acids (pAAs) mainly due to the accurate catalytic property of each aminoacyl-transfer RNA (tRNA) synthetase (AARS) that strictly controls the cognate tRNA aminoacylation event. However, it is now well established that if the aminoacylation event is artificially overruled to nonproteinogenic amino acids (npAAs) by a certain method, ribosome—the protein synthesis machinery—accepts npAAs with various sidechains into peptide chain [1–5]. Using in vitro or vivo translation systems, we indeed witnessed that the incorporation of sidechain-altering npAAs with high efficiencies at nonconsecutive or even consecutive positions [6–9]. However, we observed significantly lower efficiencies for backbone-altering npAAs (baAAs), such as N-methyl-l-α-amino acids (^Me^AAs) and d-α-amino acids (d-AAs). Particularly, when they are installed in consecutive positions or alternating positions between npAA and pAA in a row, the incorporation efficiencies are dramatically dropped. Thus, the genetic code reprogramming or expansion had been realized for nonconsecutive or nonalternating incorporation of baAAs until 2013 [9].
The inefficient ribosomal incorporation of the baAAs into peptide chain is likely due to two reasons: (i) slow accommodation of the npAA-tRNA into the ribosome A-site due to weak binding between the baAA and elongation factor thermo unstable (EF-Tu), and (ⅱ) extremely slow peptidyl transfer between A-site baAAs-tRNA and P-site peptidyl-tRNA due to the unusual structure of baAAs, resulting in high frequency of peptidyl-tRNA drop-off than A-site baAAs-tRNA [10–14]. To address the problem of (i), we first paid attention to the fact that certain amino acids, such as Glu bearing a negatively charged sidechain, binds poorly to EF-Tu; but the T-stem of tRNA^Glu^ compensates to the poor affinity, i.e. tRNA^Glu^ has a higher binding affinity to EF-Tu than other tRNAs, giving a nearly uniform level of affinity to all pAA-tRNAs [15]. Because the baAAs are intrinsically poor substrates to EF-Tu like Glu, a tRNA having the T-stem of tRNA^Glu^, referred to as tRNA^GluE2^, engineered to be orthogonal to endogenous AARSs would give a better chance to improve the issue of (i). In fact, the use of baAA-tRNA^GluE2^ prepared by flexizyme (flexible tRNA aminoacylation ribozyme) has led to the ribosomal synthesis of a peptide containing nine different types of ^Me^AA for the first time [11] and 10 consecutive incorporation of d-serine (d-Ser) has also been accomplished using tRNA^GluE2^ under optimized translation factor concentrations [15]. To resolve the issue of (ⅱ), we harnessed Escherichia coli tRNA^Pro1^, whose D-arm is specifically recognized by a translation factor of EF-P [16] and EF-P binds to ribosome E-site and significantly slows down the peptidyl-tRNA drop-off [17, 18]. This EF-P function gives more time for the peptide transfer reaction to occur between Pro and Pro (and also allows for cis-trans isomerization of the peptide bond), resulting in more efficient peptide bond formation. To resolve both issues of (i) and (ii) at the same time, we semi-rationally designed a chimeric tRNA, referred to as tRNA^Pro1E2^, which has the T-stem of tRNA^GluE2^ and the D-arm of tRNA^Pro1^ (Fig. 1A) [16, 19]. Consequently, tRNA^Pro1E2^ drastically improved the incorporation of ^Me^AAs, d-AAs, and β-amino acids (βAAs) in the presence of EF-P [19–25]. For instance, double incorporation of d-alanine (d-Ala) was performed five times more efficiently in the presence of EF-P using tRNA^Pro1E2^ compared to tRNA^GluE2^ [19].
Structures of tRNAPro1E2 and amino acids used in this study. (A) Secondary structure of tRNAPro1E2. The T-stem and D-arm are indicated by yellow and green, respectively. The anticodon stem is indicated by purple, and the numbers of nucleotide bases are indicated in the right figure. (B) Structures of MeAAs, d-AAs, and α,α-disubstituted amino acid (d-AA analogue) for ribosomal incorporation.
However, even if tRNA^Pro1E2^ is used, multiple incorporations of some ^Me^AAs and d-AAs into nascent peptide chains are still challenging, resulting in poor production of desired peptides. Thus, further enhancement of translation efficiency is critical for expanding the range of baAAs utilizable in genetic code reprogramming, particularly regarding the construction of genetically encoded peptide libraries. In addition to the T-stem and D-arm, previous studies have shown that mutations in the anticodon stem can also modulate the translation efficiency of tRNAs [26–30] (Fig. 1A). For instance, mutation of the 27-43 base-pair enhanced the translation efficiency of the amber suppressor su7G36 tRNA^Trp^CUG in response to a mismatched UAG amber codon [29]. Additionally, a recent study demonstrated that translation initiation with npAAs could be drastically improved by anticodon stem engineering of the canonical initiator tRNA^fMet^. The resulting engineered tRNA, referred to as tRNA^iniP^CUU, improved the incorporation of N-acetylproline 1000-fold and led to complete suppression of N-terminal drop-off reinitiation, which dominated when using the original tRNA^fMet^ [30]. We also conducted fine-tuning of the anticodon stem/loop of tRNA^Pro1E2^ for consecutive βAA incorporations as well as alternating incorporations of βAA and pAA [31]. These prior works suggest that the anticodon stem region of tRNA can be another factor affecting the incorporation efficiency and fidelity of npAAs and possibly baAAs. Here we report engineering of the tRNA^Pro1E2^ anticodon stem that improves multiple/consecutive incorporations of baAAs, particularly ^Me^AAs and d-AAs (Fig. 1B)
Materials and methods
Preparation of tRNAs and flexizymes
tRNAs (tRNA^Pro1E2^, tRNA^Pro1E2^ variants, and tRNA^ini^) and flexizymes (eFx and dFx) used in this research were synthesized by in vitro transcription. The DNA templates for transcription were prepared by extension and polymerase chain reaction (PCR) using Taq DNA polymerase (see Supplementary Table S1 for sequences of primers). The PCR products were purified by phenol/chloroform extraction and ethanol precipitation to obtain template DNAs. Transcription of tRNAs and flexizymes were performed at 37°C for overnight under the following conditions: 40 mM Tris-HCl (pH 8.0), 1 mM spermidine, 0.01% (v/v) Triton X-100, 10 mM dithiothreitol (DTT), 22.5 mM MgCl_2_, KOH (30 mM for flexizymes and 22.5 mM for tRNAs), 5 mM (for flexizymes) or 3.75 mM (for tRNAs) nucleoside triphosphates (NTPs), 5 mM guanosine monophosphate (GMP), 0.04 U/μl RNasin RNase inhibitor (Promega), and 0.12 μM T7 RNA polymerase. After the incubation, the reaction mixtures were treated with RQ1 DNase (Promega) at 37°C for 30 min. The resulting tRNAs and flexizymes were purified by isopropanol precipitation and 8% (for tRNAs) or 12% (for flexizyme) denaturing polyacrylamide gel electrophoresis (PAGE) [8% or 12% (w/v) acrylamide/bisacrylamide (19:1), 6 M urea, 44.5 mM Tris-borate, and 2 mM ethylenediaminetetraacetic acid (EDTA)]. The RNA concentration was adjusted to 250 μM with ultra-pure water.
Aminoacylation of tRNAs
3,5-Dinitrobenzylester (DBE) of N-methyl-l-leucine (^Me^Leu), N-methyl-l-isoleucine (^Me^Ile), N-methyl-l-valine (^Me^Val), N-methyl-l-asparagine (^Me^Asn), N-methyl-l-aspartic acid (^Me^Asp), N-methyl-l-glutamic acid (^Me^Glu), d-Ala, d-Ser and 2-aminoisobutylic acid (Aib), cyanomethylester (CME) of d-phenylalanine (d-Phe) and chloroacetyl-d-tyrosine (^ClAc^d-Tyr), and 4-chlorobenzylthioester (CBT) of d-threonine (d-Thr) were synthesized by previously reported methods. ^ClAc^d-Tyr was charged on tRNA^ini^, and the other activated amino acids were acylated on tRNA^Pro1E2^ or its variants. We used dFx for aminoacylation of DBE-activated amino acids and eFx for acylation of CME- or CBT-activated amino acids. Aminoacylation reactions were performed by the following procedure: 83 mM HEPES-KOH (pH 7.5) containing 42 µM tRNA and 42 μM dFx or eFx was heated at 95°C for 2 min and cooled to room temperature for 5 min. 100 mM (for dFx) or 3 M (for eFx) MgCl_2_ was added and the mixture was incubated at 25°C for 5 min. The reaction was initiated by addition of each activated amino acid substrate in dimethyl sulfoxide (25 mM) and incubated on ice for acylation (incubation time: 2 h for d-Ala, Aib, d-Phe and ^ClAc^d-Tyr, 6h for ^Me^Leu, ^Me^Asn, ^Me^Glu and d-Ser and 24 h for ^Me^Ile, ^Me^Val, ^Me^Asp, and d-Thr). The resulting aminoacyl-tRNAs were recovered by ethanol precipitation and washed with 70% ethanol containing 0.1% sodium acetate.
Translation of model peptides
Translation of model peptides (P1–8) was carried out at 37°C for 30 min using the FIT system. The DNA templates encoding mRNAs (mR1–8) were prepared by extension and PCR using the corresponding primers (see Supplementary Table S1 for details). The DNA templates were transcribed into mRNAs by the T7 RNA polymerase included in the FIT system and translated into peptides. For P1, the composition of the FIT system was as follows: 50 mM HEPES-KOH (pH 7.6), 2 mM ATP, 2 mM GTP, 1 mM CTP, 1 mM UTP, 20 mM creatine phosphate, 100 mM potassium acetate, 2 mM spermidine, 12.3 mM magnesium acetate, 1 mM dithiothreitol, 0.1 mM 10-formyl-5,6,7,8-tetrahydrofolic acid, 1.5 mg/ml E. coli total tRNAs, 1.2 μM E. coli ribosome, 0.6 μM methionyl-tRNA formyltransferase, 4 μg/ml creatine kinase, 3 μg/ml myokinase, 0.1 μM inorganic pyrophosphatase, 0.1 μM nucleotide diphosphate kinase, 0.1 μM T7 RNA polymerase, 2.7 μM IF1, 0.4 μM IF2, 1.5 μM IF3, 10 μM EF-Tu/EF-Ts complex, 0.26 μM EF-G, 5 μM EF-P, 0.25 μM RF2, 0.17 μM RF3, 0.5 μM RRF, 0.1 μM DNA template, 0.73 μM AlaRS, 0.03 μM ArgRS, 0.38 μM AsnRS, 0.13 μM AspRS, 0.02 µM CysRS, 0.06 µM GlnRS, 0.23 µM GluRS, 0.09 μM GlyRS, 0.02 μM HisRS, 0.4 μM IleRS, 0.04 μM LeuRS, 0.11 μM LysRS, 0.03 µM MetRS, 0.68 μM PheRS, 0.16 μM ProRS, 0.04 μM SerRS, 0.09 μM ThrRS, 0.03 µM TrpRS, 0.02 μM TyrRS, 0.02 μM ValRS, 0.02 μM ValRS, 0.5 mM Asp, 0.5 mM Lys, 0.5 mM Met, 0.5 mM Tyr, and 25 µM each aminoacyl-tRNA^Pro1E2^ variants. For P2–P8, FIT system was customized for efficient translation: the ARS contained in this modified FIT system are only 0.13 μM AspRS, 0.09 μM GlyRS, 0.11 μM LysRS, 0.03 µM MetRS, and 0.02 μM TyrRS. In addition, 0.5 mM Gly was added and the concentrations of IF2, EF-Tu/EF-Ts complex and EF-G were changed to 3, 20, and 0.1 µM, respectively. For translation of P2, the concentrations of aminoacyl-tRNAs were also changed to 150 µM. For translation of P8, 25 μM of ^ClAc^d-Tyr -tRNA^ini^CAU, 0.02 µM CysRS, and 0.5 mM Cys were also added to the customized FIT system and 10-formyl-5,6,7,8-tetrahydrofolic acid and Met were removed. For radiolabeling of peptides, a mixture of 0.05 mM [^14^C]-Asp and 0.15 mM cold Asp was used in place of 0.5 mM cold Asp.
Tricine sodium dodecyl sulphate-polyacrylamide gel electrophoresis of translated model peptides and their quantification by autoradiography
Translations were conducted in the presence of 0.05 mM of [^14^C]-Asp and 0.15 mM of cold Asp, quenched by adding the same volume of stop solution [0.9 M Tris-HCl (pH 8.45), 8% sodium dodecyl sulphate, 30% glycerol, and 0.001% xylene cyanol], and incubated at 95°C for 5 min. Four microliter of the sample was subjected to 15% tricine sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE) and analyzed by autoradiography using a Typhoon FLA 7000 (Cytiva). The relative translation efficiencies of peptides were estimated by the autoradiographic intensity of the peptide band in comparison with that of the control experiment, whose relative level was defined as 1.
Flag tag purification of translated model peptides
For model peptide P2, P3, and P8, the peptides were purified by using ANTI-FLAG M2 affinity gel (Sigma) prior to mass spectrometric analysis. The translation products were diluted with an equal volume of 2 × TBS [100 mM Tris-HCl (pH 7.6), 300 mM NaCl], mixed with pre-washed 2 × volume of ANTI-FLAG M2 affinity gel, and rotated at r.t. for 1 h. The supernatant was removed, and the gel was washed with 10 × volume of 1 × TBS buffer [50 mM Tris-HCl (pH 7.6), 150 mM NaCl] twice. Then, the peptides were eluted from the gel by adding 2 × volume of 0.2% (v/v) trifluoroacetic acid.
Identifications of translated model peptides by MALDI-TOF mass spectrometry
Translated model peptides were desalted with SPE C-tip (Nikkyo Technos) and cocrystallized with α-cyano-4-hydroxycinnamic acid on a sample plate. Matrix-assisted laser desorption/ionization time-of-flight mass spectroscopy (MALDI-TOF MS) analysis of the samples was carried out using ultrafleXtreame (Bruker Daltonics) by reflector/positive mode. Peptide calibration standard II (Bruker Daltonics) was used for the external mass calibration.
Native and denaturing PAGE analysis of tRNA
tRNA samples for native gel electrophoresis were prepared in 10 μl scale [50 mM HEPES-KOH (pH 7.2), 3 mM MgCl_2_, 10% glycerol, and 12.5 μM tRNA]. Samples were heated to 95°C for 3 min and cooled to room temperature over 10 min. Then, samples were loaded on polyacrylamide gels [20% (w/v) acrylamide/bisacrylamide (19:1), 44.5 mM Tris-borate, and 2 mM EDTA] and run for 4 h at 100 V. For denaturing PAGE, 5 μl of tRNA samples (4 M urea, 1 mM Tris, 1 mM EDTA, and 2.5 µM tRNA) were heated to 95°C for 3 min. Then, the samples were loaded on 8% denaturing gel [8% (w/v) acrylamide/bisacrylamide (19:1), 6 M urea, 44.5 mM Tris-borate, and 2 mM EDTA] and run for 30 min at 180 V. Electrophoresis was conducted in TBE buffer (89 mM Tris-borate, 2 mM EDTA). The gels were stained with ethidium bromide and analyzed using a Typhoon FLA 7000 (Cytiva).
Measurement of the efficiency of flexizyme-catalyzed aminoacylation
For ^Me^Leu, the pellet of ethanol-precipitated 33.3 pmol tRNA pre-charged with ^Me^Leu was dissolved in 0.52 μl of 10 mM sodium acetate (pH 5.2) and mixed with 5.0 μl of acid PAGE loading buffer [83% (v/v) formamide, 150 mM sodium acetate (pH 5.2), and 10 mM EDTA]. The solution was loaded on acid denaturing polyacrylamide gels [12% (w/v) acrylamide/bisacrylamide (19:1), 8M urea, and 50 mM sodium acetate (pH 5.2)] and run for 20 h at 360 V (12 V/cm), 4°C. Electrophoresis was conducted in 50 mM sodium acetate (pH 5.2). The gels were stained with ethidium bromide and analyzed using a Typhoon FLA 7000 (Cytiva). For d-Phe, the pellet of ethanol-precipitated 50 pmol tRNA pre-charged with d-Phe was dissolved in 10 μl of 1 mM sodium acetate (pH 5.2). The biotinylation reaction mixture [60 mM HEPES-KOH (pH 8.0), 6.67 mg/ml Biotin-sulfo-NHS, and 3.33 µM tRNA] was prepared in 3 μl scale and cooled on ice for 1 h. After the reaction, the biotinylated aminoacyl-tRNA was recovered by ethanol precipitation and washed with 70% ethanol containing 0.1% sodium acetate. The resulting pellet of biotinylated aminoacyl-tRNA was dissolved in 1 μl of 1 mM sodium acetate (pH 5.2) and mixed with 4 μl of RNA loading buffer (2 mM Tris, 2 mM EDTA, and 8 M urea) and 3 μl of 2 mg/ml streptavidin. The sample was loaded on 8% denaturing PAGE and run for 30 min at 180 V. Electrophoresis was conducted in TBE buffer. The gels were stained with ethidium bromide and analyzed using a Typhoon FLA 7000 (Cytiva). Aminoacylation efficiency was calculated based on the band intensities of aminoacyl-tRNA (A) and nonacylated tRNA (T) and is presented as (A)/[(A) + (T)] ×100 (%).
Radiolabeling of 3′-terminus of tRNA
The 3′-terminus of a tRNA was radiolabeled as follows: 7.5 μl of 16.7 μM tRNA lacking 3′-terminal adenosine was incubated at 80°C for 3 min, and then put on ice for 10 min. The 12.5 μl of CCA-adding reaction mixture {120 mM Gly-NaOH (pH 9.0), 75 mM MgCl_2_, 30 mM DTT, 5 μM tRNA lacking 3′-terminal adenosine prepared above, 10 mM nonradiolabeled ATP, 0.67 μM [α-^32^P]-ATP, and 200 nM CCA-adding enzyme} was incubated at 37°C for 20 min. The tRNA was purified by phenol/chloroform extraction, Micro Bio-Spin 30 column (Bio-Rad), and ethanol precipitation. The tRNA pellet was dissolved in 2.5 μl water and the concentrations of tRNA were measured by A_260_ of a 100-fold diluted solution.
Quantification of the affinity between EF-Tu and aminoacyl-tRNA (RNase a protection assay)
Aminoacyl-tRNA was prepared as described above except that the acylation reaction was conducted using a mixture of 3′-radiolabeled tRNA (10%) and nonradiolabeled tRNA (90%). The resulting pellet of aminoacyl-tRNA was dissolved in 3.2 μl of 10 mM sodium acetate (pH 5.2). The concentration of aminoacyl-tRNA was adjusted as follows: 0.7 μl of the aminoacyl-tRNA solution was diluted 100-fold with 69.3 μl of 10 mM sodium acetate (pH 5.2) and the concentrations of the total RNA including tRNA and flexizyme were measured by A260 of the diluted solution. The concentration of aminoacyl-tRNA was calculated based on the total RNA concentration and the flexizyme-catalyzed acylation efficiency (Fig. 7B). The concentration of aminoacyl-tRNA was adjusted to 2.0 μM by adding an appropriate volume of 10 mM sodium acetate (pH 5.2). Just prior to use, the concentration of aminoacyl-tRNA was adjusted to 100 nM by mixing 2.0 μM aminoacyl-tRNA solution with buffer A [50 mM HEPES-KOH (pH 7.6), 100 mM KOAc, 12 mM Mg(OAc)2, 1 mM GTP, 1 mM DTT, 20 mM creatine phosphate, 2 mM spermidine, 3 mM phosphoenol pyruvate, and 0.1 μg/μl pyruvate kinase from rabbit muscle (Sigma)].
EF-Tu was incubated in buffer A at 37°C for 30 min in order to be fully converted to a GTP-bound form. Then, 12 μl samples containing different concentrations of EF-Tu, ranging from 1.5 nM to 12.5 μM, were prepared by two-fold serial dilution on ice. After that, 9.6 µl of the EF-Tu solution was mixed with 2.4 μl of 100 nM aminoacyl-tRNA and incubated on ice for 20 min. Under equilibrium binding condition, 10 μl of each solution was mixed with 1 μl of 0.15 mg/ml RNase A (Sigma) on ice to digest the tRNA not bound to EF-Tu. After 20 s, the digestion was quenched by addition of 50 μl of 10% (v/v) trichloroacetic acid (TCA) containing 0.1 mg/ml yeast tRNA (Invitrogen) on ice. The precipitate was filtered using 0.45 μm pore-size nitrocellulose membrane assembled in Bio-Dot microfiltration apparatus (Bio-Rad), and washed five times with 200 μl each 5% (v/v) TCA. The membrane was soaked in 95% (v/v) ethanol for 5 min and dried for 5 min. The dried membrane was analyzed by autoradiography using a Typhoon FLA 7000 (Cytiva). To correct for the background signal derived from any aminoacyl-tRNA that may remain after the 20 s treatment with RNase A, a no EF-Tu control was analyzed in parallel and its radioactivity was subtracted from the experimental data. The resultant radioactivity was converted to the concentration of ternary complex using the conversion factor determined by the calibration aminoacyl-tRNA samples in buffer A containing 20, 4, 0.8, and 0.16 nM tRNA. Equilibrium dissociation constants were determined by fitting the binding data to the following equation (1) using GraphPad Prism 8 (GraphPad Software). As the concentration of aminoacyl-tRNA varied slightly in every experiment due to its multistep preparation, both ∆G and the aminoacyl-tRNA concentration were set as variables in the fitting analysis. The ∆G equation is defined as follows, where aatRNA, EFTu, and ternary complex denote aminoacyl-tRNA, EF-Tu, and their complex, respectively.
\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} \begin{eqnarray*} {{K}_{\mathrm{D}}} &=& \frac{{{{{\left[ {{\mathrm{aatRNA}}} \right]}}_{{\mathrm{eq}}}} \times {{{\left[ {{\mathrm{EFTu}}} \right]}}_{{\mathrm{eq}}}}}}{{{{{\left[ {{\mathrm{ternary\ complex}}} \right]}}_{{\mathrm{eq}}}}}}\\{{K}_{\mathrm{D}}} &=& \frac{{\left( {{{{\left[ {{\mathrm{aatRNA}}} \right]}}_{{\mathrm{input}}}} - {{{\left[ {{\mathrm{ternary\ complex}}} \right]}}_{{\mathrm{eq}}}}} \right) \times \left( {{{{\left[ {{\mathrm{EFTu}}} \right]}}_{{\mathrm{input}}}} - {{{\left[ {{\mathrm{ternary\ complex}}} \right]}}_{{\mathrm{eq}}}}} \right)}}{{{{{\left[ {{\mathrm{ternary\ complex}}} \right]}}_{{\mathrm{eq}}}}}}\\{{\left[ {{\mathrm{ternary\ complex}}} \right]}_{{\mathrm{eq}}}} &=& \frac{{\left( {{{{\left[ {{\mathrm{aatRNA}}} \right]}}_{{\mathrm{input}}}} + {{{\left[ {{\mathrm{EFTu}}} \right]}}_{{\mathrm{input}}}} + {{K}_{\mathrm{D}}} - \sqrt {{{{\left( {{{{\left[ {{\mathrm{aatRNA}}} \right]}}_{{\mathrm{input}}}} + {{{\left[ {{\mathrm{EFTu}}} \right]}}_{{\mathrm{input}}}} + {{K}_{\mathrm{D}}}} \right)}}^2} - 4 \times {{{\left[ {{\mathrm{aatRNA}}} \right]}}_{{\mathrm{input}}}} \times {{{\left[ {{\mathrm{EFTu}}} \right]}}_{{\mathrm{input}}}}} } \right)}}{2}\\{{K}_{\mathrm{D}}} &=& {\mathrm{exp}}\left( {\frac{{\Delta G}}{{RT}}} \right)\\{{\left[ {{\mathrm{ternary\ complex}}} \right]}_{{\mathrm{eq}}}} &=& \ \frac{{\left( {{{{\left[ {{\mathrm{aatRNA}}} \right]}}_{{\mathrm{input}}}} + {{{\left[ {{\mathrm{EFTu}}} \right]}}_{{\mathrm{input}}}} + {\mathrm{exp}}\left( {\frac{{\Delta G}}{{RT}}} \right) - \sqrt {{{{\left( {{{{\left[ {{\mathrm{aatRNA}}} \right]}}_{{\mathrm{input}}}} + {{{\left[ {{\mathrm{EFTu}}} \right]}}_{{\mathrm{input}}}} + {\mathrm{exp}}\left( {\frac{{\Delta G}}{{RT}}} \right)} \right)}}^2} - 4 \times {{{\left[ {{\mathrm{aatRNA}}} \right]}}_{{\mathrm{input}}}} \times {{{\left[ {{\mathrm{EFTu}}} \right]}}_{{\mathrm{input}}}}} } \right)}}{2} \end{eqnarray*}\end{document}Results
Anticodon stem engineering of tRNAPro1E2GAC for MeLeu incorporation
We initiated our investigation on how the mutations in the anticodon stem could affect the ^Me^AA incorporation efficiency. First, ^Me^Leu was chosen as a representative of inefficient ^Me^AA for consecutive incorporation. ^Me^Leu was incorporated into a model peptide P1-^Me^Leu at two consecutive GUU codons of the template mRNA, mR1, by using the parental tRNA^Pro1E2^GAC and 11 anticodon stem variants (Fig. 2A). ^Me^Leu was charged onto these tRNAs using a flexizyme (dFx) and its cognate ^Me^Leu-DBE substrate. Each variant has a mutation in one of the five base-pairs in the anticodon stem region (Fig. 2B and Supplementary Fig. S1). The translation was carried out in the customized FIT system containing a minimal set of amino acids (Met, Tyr, Lys, and Asp), to which [^14^C]-Asp was added for radiolabeling of the peptides. The translation products were subjected to 15% Tricine SDS-PAGE, followed by autoradiographic quantification (Supplementary Fig. S2). The clean translation of P1-^Me^Leu was confirmed by MALDI-TOF MS, where the translation was conducted with cold-Asp in place of [^14^C]-Asp (Supplementary Fig. S3). The relative expression levels of P1-^Me^Leu using the 11 tRNA^Pro1E2^GAC anticodon stem variants were estimated, where that of the parental tRNA^Pro1E2^GAC was defined as 1 (Fig. 2B, single base-pair mutation). Of these single base-pair mutations, five variants (A31U39, C31U39, U31A39, A29U41, and A28U42) showed subtle improvement of the expression levels compared with that of the parental tRNA^Pro1E2^GAC (1.2-, 1.1-, 1.1-, 1.1-, and 1.2-fold, respectively), indicating a need for further improvement of the anticodon stem to enhance the consecutive ^Me^Leu incorporation.
Engineering of the anticodon stem of tRNAPro1E2GAC for MeLeu incorporation. (A, D) Sequences of mRNAs, mR1 and mR2, and the corresponding peptides, P1 and P2, used for MeAA incorporation. The amino acid sequence of ‘flag’ is Asp-Tyr-Lys-Asp-Asp-Asp-Asp-Lys. (B) Relative expression levels of the model peptide P1-MeLeu, where MeLeu was introduced at the GUU codons using the parental tRNAPro1E2GAC and its anticodon stem mutants. The P1-MeLeu level of the parental tRNAPro1E2GAC was defined as 1. Mutations in the anticodon stem are indicated below the bar graph. The variant that showed the highest P1-MeLeu level is indicated by orange bar. Error bars, standard deviation (SD). (n = 3). The improvement of the best variant was confirmed to be significant by Student’s t-test (P = 1.98E-5). See also Supplementary Fig. S2 for the raw data of tricine SDS-PAGE analyses of peptides. (C) Anticodon stem sequences of the parental tRNAPro1E2 and the U31A39/G28C42 variant which showed the best improvement in MeLeu incorporation. (E) MALDI-TOF MS of the model peptide P2, translated with the engineered U31A39/G28C42tRNAPro1E2GAC (left) and the parental tRNAPro1E2GAC (right). ‘Obs.’ and ‘Calc.’ indicate observed and calculated m/z values, respectively.
We next evaluated double and triple base-pair mutations installed in the anticodon stem of tRNA^Pro1E2^GAC (Fig. 2B, see the multiple base-pair mutations). They were designed to have the combinations of four improving mutations (A31U39, U31A39, A29U41, and A28U42) but also two non-improving mutations (G28C42 and U27A43). Out of the nine variants, the U31A39/G28C42 variant of tRNA^Pro1E2^GAC (^U31A39/G28C42^tRNA^Pro1E2^GAC) showed the highest expression levels of P1-^Me^Leu by 1.9-fold compared to the parental one (Fig. 2C).
Even though a 1.9-fold enhancement observed for the single incorporation of ^Me^Leu is rather modest, it might significantly affect to the translation of multiple consecutive ^Me^Leu incorporation. Thus, we performed eight consecutive ^Me^Leu incorporations into a model peptide P2 (Fig. 2D). The MALDI-TOF MS successfully identified the correct peaks of P2 when the U31A39/G28C42 variant was used as the ^Me^Leu-carrier, whereas no peak was detected using the parental tRNA^Pro1E2^GAC (Fig. 2E). Given that ^Me^Leu has been reported as a relatively inaccessible elongator among N-methylated counterparts of 20 amino acids [7], this result shows significant potential of the anticodon stem engineering strategy.
Anticodon stem engineering of tRNAPro1E2 for consecutive MeLeu incorporation at other codons
Encouraged by the above results, we next tested five codons, GAG, AUU, CAG, CAU, and ACU, for consecutive ^Me^Leu incorporation into P1-^Me^Leu using tRNA variants with the corresponding anticodons, CUC, GAU, CUG, GUG, and GGU, respectively (Fig. 3). Since the decoding event via codon-anticodon interaction could be dictated by the anticodon stem in the elongator tRNA^Pro1E2^, we investigated all mutations in a similar manner as it was done for decoding the GUU codon, where we prepared 11 single base-pair mutations and their combinations of choice.
Engineering of the anticodon stem of tRNAsPro1E2 for MeLeu incorporation. MeLeu was precharged onto the parental tRNAsPro1E2 and their anticodon stem mutants bearing CUC, GAU, CUG, GUG, and GGU anticodons and assigned at the GAG (A), AUU (B), CAG (C), CAU (D), and ACU (E) codons of mR1 to translate P1-MeLeu. The mutations in the anticodon stem are indicated below the bar graph. The P1-MeLeu levels of the parental tRNAsPro1E2 were defined as 1. The variants that showed the highest P1-MeLeu levels are indicated by orange bars. Error bars, SD. (n = 3). The improvements of the best variants were confirmed to be significant by Student’s t-test (P <0.001 for all the codons).
For the GAG codon, all tRNA^Pro1E2^CUC containing a single base-pair mutation enhanced P1-^Me^Leu expression (Fig. 3A, single base-pair mutation), while the A28U42 variant (^A28U42^tRNA^Pro1E2^CUC) made the highest impact on increasing the expression level by 4.5-fold. Six combination mutants were also tested (see multiple base-pair mutations), but interestingly none of them could exceed the enhancement observed with the A28U42 variant. This supports our initial assumption that an optimal anticodon stem of tRNA^Pro1E2^ would depend on the codon-anticodon interaction of choice.
Likewise, improved tRNA^Pro1E2^ variants for the consecutive ^Me^Leu incorporation into P1-^Me^Leu peptide were obtained for four other codons: ^U31A39/A29U41/A28U42^tRNA^Pro1E2^GAU for AUU codon by 3.0-fold, ^C31G39^tRNA^Pro1E2^CUG for CAG codon by 1.6-fold, ^C29G41^tRNA^Pro1E2^GUG for CAU codon by 2.9-fold, and tRNA^Pro1E2^GGU for ACU codon by 1.6-fold (Fig. 3B–E). These mutation campaigns of the anticodon stem of tRNA^Pro1E2^ occasionally showed a complete or significant loss of elongation ability of mutants (Fig. 3A, C, and E), presumably due to misfolding of tRNA structure induced by the mutations. Nevertheless, our studies successfully identified tRNA^Pro1E2^ anticodon stem variants assigning to five distinct codons. Even though the observed enhancement is 1.6–3.0-fold, we have expected that such enhancing values would give an appreciable impact to the peptide expression level when multiple npAAs are consecutively incorporated.
Ribosomal incorporation of various MeAAs using the engineered tRNAPro1E2 anticodon stem variants
The tRNA^Pro1E2^ anticodon stem variants to the specific codons motivated us to evaluate their potential for incorporation of multiple kinds of ^Me^AAs at once into a model peptide. First, to verify the applicability of these tRNA^Pro1E2^ variants optimized for ^Me^Leu incorporation to other ^Me^AAs, five kinds of ^Me^AA, ^Me^Val, ^Me^Asn, ^Me^Asp, ^Me^Ile, and ^Me^Glu, were tested for two consecutive incorporations into P1. The combinations of ^Me^AAs and tRNAs are as follows: ^Me^Val on ^C31G39^tRNA^Pro1E2^CUG, ^Me^Asn on ^C31G39^tRNA^Pro1E2^CUG, ^Me^Asp on ^U31A39/G28C42^tRNA^Pro1E2^GAC, ^Me^Ile on ^U31A39/C28G42^tRNA^Pro1E2^GGU, and ^Me^Glu on ^C29G41^tRNA^Pro1E2^GUG (Fig. 4A). These five ^Me^AAs were categorized as inefficient substrates similar to ^Me^Leu in our previous report [7]. The expression levels and identities of P1-^Me^AA were evaluated by autoradiography and MALDI-TOF MS, respectively (Supplementary Fig. S4A and B). Consequently, using the designated ^Me^AA-tRNAs^Pro1E2^ variants, the translation efficiencies were enhanced by 1.6–4.2-fold (Fig. 4A), showing the general validity of the anticodon stem engineering strategy for not only ^Me^Leu but also other ^Me^AAs.
Ribosomal incorporation of multiple types of MeAAs using the parental tRNAsPro1E2 and their anticodon stem variants. (A) Relative expression level of the model peptide P1 bearing various MeAAs. The combinations of codon and MeAA are shown above the graph. CAG, GUU, ACU, and CAU codons are decoded by tRNAs with CUG, GAC, GGU, and GUG anticodons, respectively. The translations using the best tRNA variants engineered for MeLeu incorporation (orange) were compared to their parental tRNAsPro1E2 (gray). The expression levels with the parental tRNAsPro1E2 were defined as 1. Error bars, SD. (n = 3). The improvements were confirmed to be significant by Student’s t-test (P <0.02 for all the MeAAs). See also Supplementary Fig. S4A for the raw data of tricine SDS-PAGE analyses of peptides. (B) mRNA (mR3) and the corresponding peptide sequence (P3) used for MeAAs incorporations. (C) MALDI-TOF MS of the model peptide P3. MeAsn, MeAsp, MeIle, and MeGlu were assigned to CAG, GUU, ACU, and CAU codons in mR3, respectively, and translated with the engineered tRNA set (C31G39tRNAPro1E2CUG, U31A39/G28C42tRNAPro1E2GAC, U31A39/C28G42tRNAPro1E2GGU, and C29G41tRNAPro1E2GUG, top) or the parental tRNAsPro1E2 (bottom). The misincorporation byproducts are indicated with purple arrows. The identities of the byproducts are as follows: (i) Val misincorporation into the CAU codon, (ii) Val into the GUU codon, or Leu or Ile into the CAU codon, (iii) Leu or Ile into the GUU codon, (iv) unidentified.
With these engineered tRNA^Pro1E2^ variants for ^Me^AA incorporation in hand, we demonstrated the simultaneous incorporation of four different ^Me^AAs into a model peptide P3 (Fig. 4B). ^Me^Asn, ^Me^Asp, ^Me^Ile, and ^Me^Glu were flexizyme-charged onto ^C31G39^tRNA^Pro1E2^CUG, ^U31A39/G28C42^tRNA^Pro1E2^GAC, ^U31A39/C28G42^tRNA^Pro1E2^GGU, and ^C29G41^tRNA^Pro1E2^GUG and assigned to the cognate codons of mR3, respectively (Fig. 4B). The resulting peptide was analyzed by MALDI-TOF MS, where the species of the desired P3 was observed as the major product. A few byproduct species derived from misincorporations of l-Val and l-Leu or l-Ile were also observed, but the peak intensities are much smaller than that of the desired P3 (Fig. 4C, top). For comparison, the nonengineered tRNA^Pro1E2^ set was also used for the P3 translation, where the byproduct peaks were observed more intensely (Fig. 4C, bottom). These results showed that tRNA^Pro1E2^ anticodon stem variants enhanced the expression level of encoded peptides and competed out the misincorporations, thereby translating a superior quality of full-length peptide product containing multiple ^Me^AAs.
Anticodon stem engineering of tRNAPro1E2 for consecutive d-α-amino acids
We applied the same anticodon stem engineering to d-AAs to improve their incorporation efficiency. To evaluate anticodon stem mutations, two d-Phe residues were introduced nonconsecutively into a model peptide P4 at any of GAG, AUU, CAG, CAU, ACU, and GUU codons of mR4 using tRNA^Pro1E2^ variants bearing the corresponding anticodons (Fig. 5A–G). The translation levels of P4-d-Phe were quantified for the parental tRNA^Pro1E2^ and 11 variants with single base-pair mutations, followed by the evaluation of variants with combined mutations (Fig. 5B–G). For GAG codon, no tRNA variant showed higher translation levels than the parental tRNA^Pro1E2^ (Fig. 5B). For other codons, we were able to identify “enhancing” variants giving higher translation efficiencies (Fig. 5C–G). The variants that showed the highest translation level for each codon were as follows: ^G28C42^tRNA^Pro1E2^GAU for AUU codon, ^A29U41^tRNA^Pro1E2^CUG for CAG codon, ^G28C42^tRNA^Pro1E2^GUG for CAU codon, ^A28U42^tRNA^Pro1E2^GGU for ACU codon, and ^G28C42^tRNA^Pro1E2^GAC for GUU codon (Fig. 5C–G, indicated by blue bars, 4.4-, 1.4-, 1.9-, 1.7-, and 1.6-fold enhancement, respectively, compared to tRNA^Pro1E2^).
Ribosomal incorporation of d-AAs using the parental tRNAPro1E2 and its anticodon stem variants. (A) mRNAs, mR4, mR5, and mR6, and the corresponding peptide sequences, P4, P5, and P6, used for d-AA incorporations. (B–G) Engineering of the anticodon stem of tRNAsPro1E2 for d-Phe. d-Phe was precharged onto the parental tRNAsPro1E2 and their anticodon stem mutants bearing CUC, GAU, CUG, GUG, GGU, and GAC anticodons and assigned at the GAG (B), AUU (C), CAG (D), CAU (E), ACU (F), and GUU (G) codons of mR4 to translate P4-d-Phe. The mutations in the anticodon stem are indicated below the bar graph. The P4-d-Phe levels of the parental tRNAsPro1E2 were defined as 1. The variants that showed the highest P4-d-Phe levels are indicated by blue bars. Error bars, SD. (n = 3). The improvements of the best variants were confirmed to be significant by Student’s t-test (P <0.005 for all the codons except GAG). (H) Relative expression level of the model peptides P4, P5, and P6 bearing various d-AAs. The codons, mRNAs, and model peptides used for the d-AA incorporation are shown above the graph. The translations using the best tRNA variants with the engineered anticodon stems for d-Phe incorporation (blue) were compared to their parental tRNAPro1E2 (gray). The expression levels with the parental tRNAsPro1E2 were defined as 1. Error bars, SD. (n = 3). The improvements were confirmed to be significant by Student’s t-test (P <0.001 for all the amino acids). See also Supplementary Fig. S5A for the raw data of tricine SDS-PAGE analysis of peptides.
We next tested these tRNA variants for the incorporation of other d-AAs (d-Thr, d-Ala, and d-Ser) as well as an α,α-disubstituted amino acid (Aib). We prepared the respective combinations of d-Thr-^A28U42^tRNA^Pro1E2^GGU, d-Ala-^G28C42^tRNA^Pro1E2^GUG, d-Ser-^A29U41^tRNA^Pro1E2^CUG, and Aib-^A29U41^tRNA^Pro1E2^CUG. d-Thr was subjected to double incorporation into P4 using mR4, while triple incorporation of d-Ala or d-Ser into P5 encoded by mR5 and quintuple incorporation of Aib into P6 encoded by mR6, since these are known to be more efficient substrates than d-Thr for the single incorporation (Fig. 5A and H). The expression levels of the model peptides were quantified by denaturing PAGE autoradiography. Their clean expressions using the respective d-AA-tRNAs were confirmed by MALDI-TOF MS (Supplementary Fig. S5A and B). All combinations of d-AA/engineered tRNA showed higher expression levels of model peptides than did the parental tRNAs^Pro1E2^ (Fig. 5H, 1.8-fold for P4-d-Thr, 1.6-fold for P5-d-Ala, 2.1-fold for P5-d-Ser, and 3.1-fold for P6-Aib). Notably, the largest increase of expression was observed in P6-Aib bearing five npAA residues, again indicating that the effectiveness of anticodon stem engineering is more pronounced as the number of npAAs incorporated increases, even though the enhancement in the incorporation of one or two residues is not remarkable.
We then demonstrated the simultaneous incorporation of three different d-AAs and Aib into a model peptide P7 using the best set of engineered tRNA^Pro1E2^ anticodon stem variants and compared to the parental tRNA^Pro1E2^ set (Fig. 6A). d-Phe, d-Thr, d-Ala, and Aib were assigned to AUU, ACU, CAU, and CAG codons, respectively, using tRNAs bearing the corresponding anticodons (Fig. 5C and H). Both engineered and parental tRNA^Pro1E2^ sets resulted in clean translation of P7, which was confirmed by MALDI-TOF MS analysis (Fig. 6B). The expression levels of P7 were measured by autoradiography, showing that the engineered tRNA set exhibited a 1.5-fold increase in the expression level compared to the parental tRNA^Pro1E2^ set (Fig. 6C). Although the extent of improvement is not conspicuous, this result showed that the enhancement by anticodon stem engineering was exhibited when the different types of d-AA were incorporated simultaneously.
Simultaneous incorporation of multiple types of d-AAs using the parental tRNAsPro1E2 and their anticodon stem variants. (A) mRNA, mR7, and the corresponding peptide sequence, P7, used for simultaneous incorporations of multiple types of d-AAs. (B) MALDI-TOF MS of the model peptide P7. d-Phe, d-Thr, d-Ala, and Aib were assigned to AUU, ACU, CAU, and CAG codons in mR7, respectively, and translated with the engineered tRNA set (G28C42tRNAPro1E2GAU, A28U42tRNAPro1E2GGU, G28C42tRNAPro1E2GUG, and A29U41tRNAPro1E2CUG, top) or the parental tRNAPro1E2 set (bottom). (C) Relative expression level of the model peptide P7. The P7 level of the engineered tRNA set (blue) was compared to that of the parental tRNAsPro1E2 (gray); the latter value was defined as 1. Error bars, SD. (n = 3). The improvement was confirmed to be significant by Student’s t-test (P = 5.83E-3). See also Supplementary Fig. S6 for the raw data of tricine SDS-PAGE analyses of peptides.
The effects of anticodon stem mutations on tRNA conformation, aminoacylation efficiency, and EF-Tu affinity
We further examined the impact of the anticodon stem mutations on the ability of tRNA^Pro1E2^ variants to incorporate baAAs. To assess potential structural alterations, we compared the tertiary structures of the engineered tRNA^Pro1E2^ variants with that of the parental tRNAs^Pro1E2^ using native PAGE analysis. For the CUC, GAU, CUG, and GAC anticodons, the engineered variants exhibited clear differences in electrophoretic mobility relative to the parental tRNAs^Pro1E2^ (Fig. 7A). For example, among the variants of tRNA^Pro1E2^GAU, ^U31A39/A29U41/A28U42^tRNA^Pro1E2^GAU (optimized for ^Me^AAs) migrated less compared to tRNA^Pro1E2^GAU, while ^G28C42^tRNA^Pro1E2^GAU (optimized for d-AAs) shifted downward.
A) Native polyacrylamide gel electrophoresis of the six parental tRNAsPro1E2 and anticodon stem engineered variants. The best variants for MeLeu are labeled in orange, while the best variants for d-Phe are labeled in blue. The inserted red dashed lines represent the mobilities of parental tRNAsPro1E2. (B, C) The aminoacylation efficiency of the parental tRNAsPro1E2 and their variants with MeLeu (B) and d-Phe (C). The error ranges indicate standard deviations (n = 3). (D–I) The affinities of the MeLeu-tRNA to EF-Tu obtained from the RNAse A protection assay for parental tRNAPro1E2 and its best variant for MeLeu incorporation with the CUC (D), GAU (E), CUG (F), GUG (G), GGU (H), and GAC (I) anticodons. Error bars indicate the fitting errors (95% confidence interval).
These observations indicated that the anticodon stem mutations induced detectable changes in the native structural state of these tRNA variants. Although no significant mobility shift was observed for the GUG and GGU anticodons, this does not preclude the possibility of their conformational changes. In contrast, all the tRNAs migrated uniformly as single bands under denaturing conditions (Supplementary Fig. S7A), confirming that the mobility shifts and multiple bands observed in native PAGE reflected conformational heterogeneity rather than sample impurities.
We also analyzed the native PAGE migration patterns of the “second-best” variants for baAA incorporation with > 95% relative activity compared to the best variants: ^U31A39^tRNA^Pro1E2^GUG, ^U31A39/A28U42^tRNA^Pro1E2^GGU, and ^U31A39/A28U42^tRNA^Pro1E2^GAC for ^Me^AAs, and ^C29G41^tRNA^Pro1E2^CUG for d-AAs (Fig. 2B, 3D and E, and 5D). ^U31A39^tRNA^Pro1E2^GUG, ^U31A39/A28U42^tRNA^Pro1E2^GAC, and ^C29G41^tRNA^Pro1E2^CUG shifted in the same direction and exhibited similar mobilities as the best variants (^C29G41^tRNA^Pro1E2^GUG, ^U31A39/G28C42^tRNA^Pro1E2^GAC, and ^A29U41^tRNA^Pro1E2^CUG). In contrast, ^U31A39/A28U42^tRNA^Pro1E2^GGU showed a slight downward shifted compared to the best-engineered ^U31A39/C28G42^tRNA^Pro1E2^GGU (Supplementary Fig. S7B). These observations support the role of anticodon stem-induced conformational changes in enhancing baAA incorporation and suggest that the improved tRNA variants bearing the same anticodon exhibit similar structural characteristics.
We hypothesized that the observed improvements in translation efficiency arose from anticodon stem-induced conformational changes that influenced one or more properties of tRNA: (i) its ability to undergo efficient peptidyl transfer reaction on the ribosome, (ⅱ) flexizyme-mediated aminoacylation efficiency, and (ⅲ) EF-Tu-mediated accommodation of aminoacyl-tRNA. To identify the dominant factor contributing to the enhanced baAA incorporation, we first examined the possibility of (ⅱ) by quantifying aminoacylation efficiency for ^Me^Leu and d-Phe using dFx and eFx, respectively. For ^Me^Leu, the engineered variants exhibited aminoacylation efficiencies comparable to those of the parental tRNAs^Pro1E2^ (∼60%), except for the CUG anticodon, where the aminoacylation of tRNA^Pro1E2^CUG was significantly less efficient than that of other tRNAs (48%; Fig. 7B and Supplementary Fig. S7C). A similar trend was observed for d-Phe (Fig. 7C and Supplementary Fig. S7D), indicating that the anticodon stem mutations exert minimal influence on aminoacylation efficiency in most cases. This observation aligns with previous findings that flexizymes primarily recognize the 3′-terminal CCA consensus sequence of tRNAs [32]. Therefore, (ⅱ) aminoacylation efficiency is very unlikely to be the major determinant of the improved npAA incorporation.
Next, to evaluate the possibility of (ⅲ), we examined whether the anticodon stem mutations affect the formation of the EF-Tu•GTP-aminoacyl-tRNA ternary complex. For the parental tRNAs^Pro1E2^ and their best variants for ^Me^Leu incorporation, EF-Tu binding affinities were quantified using an RNase A protection assay (see Methods for details) [11, 33]. The calculated binding free energies (∆G value) for ^Me^Leu-tRNA ranged from −7.4 to −8.3 kcal·mol⁻¹, which were comparable to the previously reported affinity of ^Me^Leu-tRNA^AsnE2#3^ (−7.9 kcal·mol⁻¹) [11]. This similarity was expected, as tRNA^AsnE2#3^ and tRNA^Pro1E2^ share the same T-stem, a primary determinant of EF-Tu affinity. Importantly, the EF-Tu affinities of the engineered variants differed by <0.8 kcal·mol⁻¹ from those of the parental tRNAs^Pro1E2^ across all the six anticodons (Fig. 7D–I and Supplementary Fig. S8). These results ruled out the possibility that improved baAA incorporation arose from enhanced EF-Tu binding. Since neither aminoacylation efficiency nor EF-Tu affinity were significantly affected by the anticodon stem mutations, we concluded that the observed improvements most likely resulted from promotion of peptidyl transfer due to the anticodon stem-induced conformational changes—the above-mentioned possibility of (i).
Ribosomal synthesis of a macrocyclic model peptide containing three consecutive MeAAs and three d-AAs
Having tRNA^Pro1E2^ variants in the anticodon stem for incorporating both ^Me^AAs and d-AAs, we performed the ribosomal synthesis of a model macrocyclic peptide P8 containing these baAAs. ^Me^Leu, ^Me^Val, d-Ala, and d-Ser were flexizyme-charged onto the engineered tRNAs, ^U31A39/G28C42^tRNA^Pro1E2^GAC, ^U31A39/C28G42^tRNA^Pro1E2^GGU, ^G28C42^tRNA^Pro1E2^GUG, and ^A29U41^tRNA^Pro1E2^CUG, respectively, and assigned to the corresponding codons in mR8 (Fig. 8A). In this model peptide P8, we designed two important features: (i) the N-terminal amino acid was reprogrammed to N-chroloacetyl-d-tyrosine (^ClAc^d-Tyr) for the formation of a thioether bond with the downstream Cys residue to cyclize the peptide backbone, and (ⅱ) three ^Me^AA residues are consecutively installed while three d-AA residues are installed in a scattered manner (Fig. 8A and B). In MALDI-TOF MS analysis, P8 was observed as the major species of product accompanied by very minor misincorporated byproducts (Fig. 8C, top). In contrast, the translation of P8 using the parental tRNA^Pro1E2^ set resulted in more prominent byproducts by the misincorporation of l-Val and l-Leu/l-Ile into GUU and ACU codons, respectively (Fig. 8C, bottom). Again, the result substantiated a remarkable improvement of ribosomal synthesis of macrocyclic peptides containing multiple ^Me^AAs and d-AAs by the use of engineered tRNAs^Pro1E2^ anticodon stem variants.
Ribosomal incorporation of six MeAAs and d-AAs into a single peptide. (A) mRNA, mR8, and the corresponding peptide sequence, P8, used for simultaneous incorporations of multiple types of MeAAs and d-AAs. (B) Structure of the model peptide P8. (C) MALDI-TOF MS of the model peptide P8. MeLeu, MeVal, d-Ala, and d-Ser were assigned to GUU, ACU, CAU, and CAG codons in mR8, respectively, and translated with the engineered tRNA set (U31A39/G28C42tRNAPro1E2GAC, U31A39/C28G42tRNAPro1E2GGU, G28C42tRNAPro1E2GUG, and A29U41tRNAPro1E2CUG, top) and the parental tRNAPro1E2 set (bottom). In addition, ClAcd-Tyr was introduced at the initiator AUG codon using tRNAiniCAU for cyclization. The misincorporation byproducts are indicated with purple arrows. The identities of the byproducts are as follows: (i) Val misincorporation into the ACU and both of GUU codons, (ii) Val into the ACU codon and [Val into the ACU codon, or Leu or Ile into the GUU codon], or 3 × [Val into the ACU codon, or Leu or Ile into the GUU codon], (iii) Val into the GUU codon, or 2 × [Val into the ACU codon, or Leu or Ile into the GUU codon], (iv) Val into the ACU codon, or Leu or Ile into the GUU codon
Discussion
In this work, we have demonstrated that the incorporation of baAAs significantly improved by engineering the anticodon stem of tRNA^Pro1E2^. The engineered anticodon stems had one or two base-pair mutation(s), of which locations and combinations depended on the type of baAAs (^Me^AAs or d-AAs) and codons used for their incorporation. In our previous research, we conducted the anticodon arm engineering for βAAs incorporation in a similar manner, where the fine-tuned anticodon arm sequences also varied from codon to codon [31]. Hence, each combination of amino acid type and codon is considered to have a different optimal anticodon stem.
The anticodon stem engineering not only enhances the peptide yield but also reduces occurrence of misincorporations. Besides, the 7 of the 11 engineered variants showed different mobilities from their parental tRNAs^Pro1E2^ in native PAGE analysis, indicating their tertiary structures have been altered by the anticodon stem mutations. We hypothesize that the sequence of anticodon stem changes a subtle conformational orientation of baAA-tRNA that influences the peptidyl transferase center (PTC) of the ribosome. Since the native translation system evolved to utilize the canonical 20 l-α-amino acids in the peptidyl transfer reaction, the parental tRNA^Pro1E2^, whose anticodon stem is identical to that of the native E. coli tRNA^Pro1^, may not be most suitable to positioning baAAs, such as ^Me^AAs, d-AAs, and βAAs, for the peptidyl bond formation due to their unusual backbone placement.
Preceding studies showed that peptide bond formation between P-site peptidyl-tRNA and A-site N-methylaminoacyl-tRNA (^Me^AA-tRNA) in the PTC is extremely slow compared to the 20 canonical amino acids [13, 34]. Formation of fMet-^Me^Phe (N-methyl-l-phenylalanine) dipeptide is reported to be 8000 times slower than that of fMet-Phe [13]. By contrast, the accommodation of ^Me^Phe-tRNA into A-site is only four times slower than that of Phe-tRNA, clearly showing that slow peptidyl transfer accounts for the inefficiency of ^Me^AA incorporation [13]. Additionally, the use of different types of tRNAs with different anticodon stems altered the rate of ^Me^AA incorporation at the same codon by more than 10-fold, which was also attributed to the acceleration of peptidyl transfer [34]. Therefore, the improvement of ^Me^AA incorporation observed in this study would be attributed to the rearrangement of the α-amino group of the A-site ^Me^AA-tRNA to a preferable position for peptide bond formation, which is mediated by the conformational change of tRNAs induced by the anticodon stem empirical engineering.
d-AAs are also poor substrates for ribosomal elongation. The peptide bond formation between P-site peptidyl-tRNA and A-site d-aminoacyl-tRNA (d-AA-tRNA) in the PTC is known to be sluggish because of the unfavorable direction of the α-amino group for the nucleophilic attack to the carbonyl carbon in the P-site [14, 35, 36]. Similar to the ^Me^AA incorporation, this disadvantageous conformation would be remedied by the anticodon stem mutations introduced in this study, thereby allowing the A-site d-AA-tRNA to be more efficiently utilized for peptide bond formation. Another plausible explanation would be the conformational conflict of the resulting P-site peptidyl-d-AA-tRNA with the ribosome. The P-site peptidyl-d-AA-tRNA shows a significant defect in the ability as a peptidyl transfer donor and arrests the translation with high probability, presumably because it stabilizes the ribosome in a conformation not suitable for peptidyl transfer [35]. The anticodon stem mutations would prevent the P-site peptidyl-d-AA-tRNA from forming such an unfavorable conformation that induces translation arrest.
Although the best-engineered variants with the GGU and GAC anticodons (^C29G41^tRNA^Pro1E2^GUG, ^G28C42^tRNA^Pro1E2^GUG, ^U31A39/C28G42^tRNA^Pro1E2^GGU, and ^A28U42^tRNA^Pro1E2^GGU) did not exhibit clear mobility shifts in native PAGE (Fig. 7A), these variants nonetheless enhanced the incorporations of either ^Me^AAs or d-AAs. Since identical electrophoretic mobility does not necessarily indicate identical tertiary structure, it is plausible that these anticodon stem mutants underwent conformational changes that facilitated peptidyl transfer. Alternatively, the anticodon stem mutations may have increased the flexibility of the resulting tRNAs while maintaining overall conformations similar to that of the parental tRNAs^Pro1E2^. Previous studies have suggested that anticodon stem mutations can influence tRNA conformational flexibility [29, 37]. Enhanced flexibility of tRNA could enable baAAs to adopt orientations more favorable for peptidyl transfer. However, further structural and biochemical analyses of these tRNA^Pro1E2^ anticodon stem variants are required to verify this hypothesis.
The enhancement of incorporation efficiency and quality by anticodon stem engineering was especially pronounced when introducing a larger number of baAAs, such as eight consecutive incorporations of ^Me^Leu and simultaneous incorporation of multiple types of either ^Me^AAs or d-AAs. Furthermore, using the engineered tRNA set, we demonstrated the clean ribosomal synthesis of a model macrocyclic peptide containing three consecutive ^Me^AAs and three d-AAs. These results clearly show that the anticodon stem engineering strategy has great potential for facilitating the ribosomal synthesis of diverse peptides containing many baAAs, leading to the preparation of a random macrocyclic peptide library comprising multiple ^Me^AAs and d-AAs. Since such a library can be readily applied to mRNA display-based screening methodologies, such as the RaPID system [38, 39], anticodon stem engineering is expected to contribute significantly to the development of macrocyclic peptides containing multiple ^Me^AAs and/or d-AAs with high bioactivity.
Supplementary Material
gkag137_Supplemental_Files
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1Noren CJ, Anthony-Cahill SJ, Griffith MC et al. A general method for site-specific incorporation of unnatural amino acids into proteins. Science. 1989;244:182–8. 10.1126/science.2649980.2649980 · doi ↗ · pubmed ↗
- 2Hohsaka T, Sato K, Sisido M et al. Site-specific incorporation of photofunctional nonnatural amino acids into a polypeptide through in vitro protein biosynthesis. FEBS Lett. 1994;344:171–4. 10.1016/0014-5793(94)00381-5.8187878 · doi ↗ · pubmed ↗
- 3Wang L, Brock A, Herberich B et al. Expanding the genetic code of Escherichia coli. Science. 2001;292:498–500. 10.1126/science.1060077.11313494 · doi ↗ · pubmed ↗
- 4Forster AC, Tan Z, Nalam MN et al. Programming peptidomimetic syntheses by translating genetic codes designed de novo. Proc Natl Acad Sci USA. 2003;100:6353–7. 10.1073/pnas.1132122100.12754376 PMC 164450 · doi ↗ · pubmed ↗
- 5Xiao H, Chatterjee A, Choi SH et al. Genetic incorporation of multiple unnatural amino acids into proteins in mammalian cells. Angew Chem Int Ed. 2013;52:14080–3. 10.1002/anie.201308137.PMC 1136023924353230 · doi ↗ · pubmed ↗
- 6Murakami H, Ohta A, Ashigai H et al. A highly flexible t RNA acylation method for non-natural polypeptide synthesis. Nat Methods. 2006;3:357–9. 10.1038/nmeth 877.16628205 · doi ↗ · pubmed ↗
- 7Kawakami T, Murakami H, Suga H. Messenger RNA-programmed incorporation of multiple N-methyl-amino acids into linear and cyclic peptides. Chem Biol. 2008;15:32–42. 10.1016/j.chembiol.2007.12.008.18215771 · doi ↗ · pubmed ↗
- 8Goto Y, Katoh T, Suga H. Flexizymes for genetic code reprogramming. Nat Protoc. 2011;6:779–90. 10.1038/nprot.2011.331.21637198 · doi ↗ · pubmed ↗