In vitro selection of tRNAs for efficient four-base decoding to incorporate non-natural amino acids into proteins in an Escherichia coli cell-free translation system

Abstract

Position-specific incorporation of non-natural amino acids into proteins is a useful technique in protein engineering. In this study, we established a novel selection system to obtain tRNAs that show high decoding activity, from a tRNA library in a cell-free translation system to improve the efficiency of incorporation of non-natural amino acids into proteins. In this system, a puromycin-tRNA conjugate, in which the 3'-terminal A unit was replaced by puromycin, was used. The puromycin-tRNA conjugate was fused to a C-terminus of streptavidin through the puromycin moiety in the ribosome. The streptavidin-puromycin-tRNA fusion molecule was collected and brought to the next round after amplification of the tRNA sequence. We applied this system to select efficient frameshift suppressor tRNAs from a tRNA library with a randomly mutated anticodon loop derived from yeast tRNA CCCG Phe. After three rounds of the selection, we obtained novel frameshift suppressor tRNAs which had high decoding activity and good orthogonality against endogenous aminoacyl-tRNA synthetases. These results demonstrate that the in vitro selection system developed here is useful to obtain highly active tRNAs for the incorporation of non-natural amino acid from a tRNA library.

Figure 1

(A) Schematic illustration of the formation of streptavidin–tRNA fusion using puromycin–tRNA, which contains a puromycin moiety in the place of 3′ terminal aminoacyl-adenosine and a four-base anticodon CCCG. The puromycin–tRNA binds to ribosomal A site and accepts a streptavidin polypeptide chain as an analog of aminoacyl-tRNA in response to a four-base CGGG codon at 3′ terminus of streptavidin mRNA in a cell-free translation. The resulting streptavidin–puromycin–tRNA may be translocated to the P-site. In this case, the next aminoacyl-tRNA binds to the vacant ribosomal A site, but can not accept the polypeptide chain because of the amide bond of puromycin–tRNA. The resulting streptavidin–tRNA fusion is released from the ribosome complex by the addition of EDTA. (B) Schematic illustration of the in vitro selection system of tRNAs. Step 1, a DNA pool encoding tRNAs containing a four-base anticodon CCCG is transcribed by T7 RNA polymerase to tRNA(-CA) pool. Step 2, the tRNA(-CA) pool is ligated with pdCp-Puromycin by T4 RNA ligase to generate puromycin–tRNA. Step 3, a streptavidin mRNA containing a four-base CGGG codon at C-terminus is translated in an E.coli cell-free translation system in the presence of the puromycin–tRNA. Puromycin–tRNAs that successfully decode the CGGG codon form ribosome–mRNA–streptavidin–tRNA complex. Step 4, the streptavidin–tRNA fusion is dissociated from the complex by the addition of EDTA. Step 5, the streptavidin–tRNA fusion is recovered with biotin-coated magnetic beads. Step 6, the streptavidin–tRNA fusion is dissociated from the beads, and then the tRNA moiety is subjected to RT–PCR. Step 7, the tRNA genes are regenerated by overlap-extension PCR with a T7 promoter primer, which are used as template DNAs in the next round of selection.

Figure 2

Formation of streptavidin–tRNA fusion and recovery of the corresponding tRNA gene using yeast ${\text{tRNA}}_{\text{CCCG}}^{\text{Phe}}$. (A) An mRNA sequence used for the formation of streptavidin–tRNA fusion. The streptavidin gene contains T7 tag at N-terminus for the detection by western blot analysis, and five CGGG codons at C-terminus through a linker sequence consisted of GlyGlySerGlyGlySer sequence. (B) Western blot analysis of the formation of streptavidin–tRNA fusion by adding the streptavidin mRNA to a cell-free translation system in the absence of the puromycin–tRNA, in the presence of the puromycin–tRNA, and in the presence of the puromycin-tRNA and after treatment with RNaseA. (C) The amount of tRNA recovered with biotin-coated magnetic beads as streptavidin–tRNA fusion was determined by quantitative PCR analysis. The tRNAs were recovered from the cell-free translation products obtained in the presence of mRNA without CGGG codon, and with five CGGG codons at C-terminus. (D) PAGE analysis of the RT–PCR product of the recovered tRNAs obtained in the presence of mRNA without CGGG codon, and with five CGGG codons at C-terminus. The RT–PCR product of the yeast ${\text{tRNA}}_{\text{CCCG}}^{\text{Phe}}$(-CA) was applied as a positive control.

Figure 3

Selection of tRNA from a random pool library of yeast tRNA^Phe containing NN-CCCG-NN in anticodon loop. (A) Structure of tRNA(-CA), in which N indicates A, G, C or U. (B) The amount of recovered tRNA determined by quantitative PCR in each round of selection. (C) Western blot analysis of the streptavidin–tRNA fusion obtained in the each round of selection.

Figure 4

Evaluation of selected tRNAs. (A) Fluorescence image of SDS–PAGE and western blot analysis of the incorporation of a fluorescently labeled non-natural amino acid into Tyr83 position of streptavidin by using the selected and unselected tRNAs. (B) Relative fluorescence intensity of the band of streptavidin. The data were represented as mean ± SD of six assays.

Figure 5

(A) Western blot analysis of the expression of full-length streptavidin in the presence of non-aminoacylated tRNAs to examine orthogonality of the tRNAs against endogenous E.coli aminoacyl-tRNA synthetases. (B) Relative yield of the full-length streptavidin. The data were represented as means ± SD of three assays.