Cys-tRNAj as a Second Translation Initiator for Priming Proteins with Cysteine in Bacteria

Abstract

We report the construction of an alternative protein priming system to recode genetic translation in Escherichia coli by designing, through trial and error, a chimeric initiator whose sequence identity points partly to elongator tRNA^Cys and partly to initiator tRNA_f ^Met. The elaboration of a selection based on the N-terminal cysteine imperative for the function of glucosamine-6-phosphate synthase, an essential enzyme in bacterial cell wall synthesis, was a crucial step to achieve the engineering of this Cys-tRNA_j. Iterative improvement of successive versions of Cys-tRNA_j was corroborated in vitro by using a biochemical luciferase assay and in vivo by selecting for translation priming of E. coli thymidylate synthase. Condensation assays using specific fluorescent reagent FITC-Gly-cyanobenzothiazole provided biochemical evidence of cysteine coding at the protein priming stage. We showed that translation can be initiated, by N-terminal incorporation of cysteine, at a codon other than UGC by expressing a tRNA_j with the corresponding anticodon. The optimized tRNA_j is now available to recode the priming of an arbitrary subset of proteins in the bacterial proteome with absolute control of their expression and to evolve the use of xenonucleotides and the emergence of a tXNA_j in vivo.

Figure 1

Metabolic selection for translation initiation ofl-glutamine—D-fructose-6-phosphate aminotransferase at a cysteine codon by a tRNA_j. (A) Schematic representation of translation initiation of GlmS mRNA by tRNA_f^Met. Codons for Cys1, Val2, Glu608, and stop codons of GlmS mRNA are shown. RBS for Ribosome Binding Site is indicated (B) UDP-N-acetyl-d-glucosamine biosynthesis from fructose-6-P involves the three enzymes l-glutamine—D-fructose-6-phosphate aminotransferase, phosphoglucosamine mutase and fused N-acetylglucosamine-1-phosphate uridyltransferase and glucosamine-1-phosphate acetyltransferase respectively encoded by glmS, glmM and glmU. (C) Schematic representation of translation initiation by a tRNA_j of the GlmS mRNA encoded by a met1 deleted version of glmS. The same GlmS mature protein is produced. (D) In a strain lacking both l-glutamine—D-fructose-6-phosphate aminotransferase (glmS) and glucosamine-6-phosphate deaminase (nagB), cell growth can be restored if translation is initiated by a tRNA_j at the met1 deleted allele of glmS or by an exogenous supply of N-acetylglucosamine in the medium.

Figure 2

Growth assay validating the GlmS selection to initiate translation at a cysteine start codon by tRNA_j and cloverleaf depiction of tRNAs. (A) Growth of strains on mineral glucose medium plates supplemented with N-acetylglucosamine (top) and growth of corresponding strains on mineral glucose medium plates (bottom) after 5 days of incubation. “ΔglmS, ΔnagB” denotes the GlmS selection strain (EX160). Other strains on the left plates are EX160 transformed with a plasmid harboring glmS:wt (XE3723), expressing the sycJ1 gene (XE3727) or sycJ1:A73 (XE4039). Right plates: growth of GlmS selection strains expressing the sycJ2 gene (XE3732) or the sycJ2:A73 gene (XE4038). (B) Schematic representations of E. coli initiator tRNA_f^Met: tRNA_f1^Met (G46) or tRNA_f2^Met (A46,), - E. coli elongator tRNA^Cys and the two versions of the engineered tRNA_j: tRNA_j¹ (G46, blue star) and tRNA_j² (A46, blue star).

Figure 3

In vitro aminoacylation assays of tRNA^cys, tRNA_f1^Met, and tRNA_j by CysRS and MetRS. The initial rates of aminoacylation were determined using 2 nM enzyme and the plateau with 1 μM enzyme. tRNA concentration in the 100 μL assays was 5 μM. (A) Initial rate of aminoacylation of tRNA^Cys, tRNA_f1^Met, and tRNA_j by the cysteinyl-tRNA synthetase. (B) Table summarizing in vitro aminoacylation assays with CysRS and MetRS. The initial rates of aminoacylation reactions and aminoacylation plateaus are indicated. Data are presented as mean ± SD (n = 2). The low methionylation plateaus of tRNA^Cys and tRNA_j can be accounted for by contamination of the tRNA preparations with E. coli tRNA_f^Met. Indeed, the initial rates of methionylation correspond to those expected for tRNA_f^Met present at concentrations that can be calculated from the plateau. Thus, the aminoacylation of tRNA_j by MetRS is not detectable. For aminoacylation plateaus, values in parentheses represent the percentage of aminoacylated-tRNA relative to tRNA used in the assay.

Figure 4

Growth assay of GlmS selection strains expressing a tRNA_j³ metZ single mutant on mineral medium and schematic representation of mutations and their impact on tRNA_j³ activity. (A) Positive control, which corresponds to EX160 (ΔglmS ΔnagB) transformed with a plasmid carrying glmS wt gene, is shown in orange. Other selection strains arise from EX160 transformed with two plasmids carrying glmS M1Δand a tRNA_j³ single mutant, respectively. The original tRNA_j² is colored dark blue. For readability, only the names of the mutations of tRNA_j³ are indicated on the graphs. Bacteria were streaked on mineral medium, incubated at 37 °C for 5 days, and results were collected on day 1, day 2, and day 5. Data are presented as mean ± SD (n ≥ 4). (B) Single tRNA_j³ mutations that enhanced the growth of the GlmS selection strain expressing GlmS M1Δcompared with the growth expressing tRNAj³ at the same incubation time. Noncolored positions are not mutated. Dark blue are mutations that lead to increased efficiency all along the assay (days 1 to 5). Light blue are mutations increasing efficiency at the end of the assay (day 2 to 5), dark pink at the beginning of the assay (day 1 to 2) and light pink at day 1. Single mutations in tRNA_j³ that abolish tRNA_j function are highlighted in yellow. tRNA_j³ positions leading to a nonfunctional tRNA_j, whatever the mutation, are shown in red. Single mutations in tRNA_j³ that decrease tRNA_j efficiency are highlighted in purple.

Figure 5

Validation of translation initiation at a cysteine codon by tRNA_j using the Nanoluc assay and thymidylate synthase selection. (A) NanoLuc protein catalyzes the conversion of furimazide to furimamide with light emission in the presence of oxygen. In the presence of an M1C allele of NLuc, protein activity can be restored only if a tRNAj is supplied and active. (B) Normalized bioluminescence of the different strains tested in the Nanoluc assay. All strains are derived from MG1655 expressing wild-type Nanoluc protein (XE4439), or Nanoluc with an additional Cys at position 2 (XE4855), Nanoluc M1C alone (XE4766) or in combination with tRNA_j¹ (XE4785) or tRNA_j² (XE4786). (C) Biosynthesis of dTTP from dUMP involves the three enzymes thymidylate synthase, thymidylate kinase, and nucleoside diphosphate kinase, respectively, encoded by the thyA, tmk, and ndk genes. The nucleobase thymine (T) and the deoxynucleoside thymidine (dT) are salvaged through the action of the recycling enzymes thymidine phosphorylase and thymidine kinase, respectively encoded by the deoA and tdk genes. (D) In a strain devoid of both thymidylate synthase and thymidine phosphorylase, cell growth can be restored if translation is initiated by a Cys-tRNAj at a met deleted allele of thyA or by an exogenous supply of dT in the medium.

Figure 6

In vitro labeling of the N-terminal cysteine of ThyA protein variants with CBT-Gly-FITC. (A) Condensation reaction between the N-terminal Cys of ThyA protein and the CBT-Gly-FITC conducts to a ThyA fluorescent protein. (B) Fluorescence (left) and white light (right) images of a gel loaded with ThyA proteins labeled with CBT-Gly-FITC. ThyA proteins were produced in the E83 selection strain in mineral medium. L: ladder, FL: fluorescent ladder, Lane 1: ThyA wt protein (M1K2), lane 2 ThyA (M1CK2) protein in the presence of the tRNA_j⁴ and ThyA M1C2K3 (ThyA with an additional Cysteine at position 2) as positive control of labeling.

Figure 7

Microbiological assays validate translation initiation at an Ala codon by the tRNA_k. (A) Schematic representation of tRNA_k, which differs from tRNA_j¹ by its UGC anticodon. (B) Growth of the GlmS selection strain named XE6320 (ΔglmS ΔnagB, p₁::glmS M1Δ:C2A p₂::sycK, cys S-T. thermophilus) in mineral glucose medium with a central well containing N-acetyl-glucosamine, S-methyl-l-cysteine or l-cystine, each at a concentration of 100 mM, after 5 days of incubation at 37 °C. (C) Growth of the ThyA selection strains on mineral medium with or without thymidine after 2 days of incubation at 37 °C. EX83 (ΔthyA Δdeo) corresponds to the ThyA selection strain, transformed by a plasmid expressing thyA wt (XE6202), expressing thyA:M1A (XE6454) or expressing thyA:M1A and sycK (XE6410).