An anticodon-sensing T-boxzyme generates the elongator nonproteinogenic aminoacyl-tRNA in situ of a custom-made translation system for incorporation

Abstract

In the hypothetical RNA world, ribozymes could have acted as modern aminoacyl-tRNA synthetases (ARSs) to charge tRNAs, thus giving rise to the peptide synthesis along with the evolution of a primitive translation apparatus. We previously reported a T-boxzyme, Tx2.1, which selectively charges initiator tRNA with N-biotinyl-phenylalanine (BioPhe) in situ in a Flexible In-vitro Translation (FIT) system to produce BioPhe-initiating peptides. Here, we performed in vitro selection of elongation-capable T-boxzymes (elT-boxzymes), using para-azido-l-phenylalanine (PheAZ) as an acyl-donor. We implemented a new strategy to enrich elT-boxzyme-tRNA conjugates that self-aminoacylated on the 3'-terminus selectively. One of them, elT32, can charge PheAZ onto tRNA in trans in response to its cognate anticodon. Further evolution of elT32 resulted in elT49, with enhanced aminoacylation activity. We have demonstrated the translation of a PheAZ-containing peptide in an elT-boxzyme-integrated FIT system, revealing that elT-boxzymes are able to generate the PheAZ-tRNA in response to the cognate anticodon in situ of a custom-made translation system. This study, together with Tx2.1, illustrates a scenario where a series of ribozymes could have overseen aminoacylation and co-evolved with a primitive RNA-based translation system.

Graphical Abstract

Figure 1.

Strategies to obtain tRNA 3′-aminoacylation ribozymes. (A) Schematic illustration of the library containing a 40-nt randomized anti-terminator region derived from B. subtilis glyQS T-box riboswitch with its unbound cognate tRNA^Gly. Blue indicates the T-box riboswitch scaffold, and green represents its cognate tRNA. The randomized anti-terminator region is colored in pink, and the A15 linker is in black. (B) Chemical structure of para-azido-l-phenylalanine cyanomethyl ester (Phe^AZ-CME), the amino acid substrate used in this study. (C) General scheme of in vitro selection based on the specific interaction between biotin and SAv. Biotin is colored orange. R. Anti., randomized anti-terminator region. (D) Click reaction-based method using SPAAC between azido group and DBCO-sulfo-biotin to label self-aminoacylated RNA population with biotin for SAv affinity selection. (E) The chemical structure of DBCO-sulfo-biotin used in this study. The click reaction product is shown and abbreviated as a yellow dot. (F) b-tREX method based on NaIO₄ oxidation of RNA 3′-vicinal diols. While 3′-aminoacylated RNAs survive the oxidation and later get a biotin label during the Klenow fragment, exo- extension in the presence of biotin-11-dUTP, free or non-3′-aminoacylated RNAs are oxidized into 3′-dialdehydes and no oxidized RNAs can get a biotin label. (G) Mechanism of NaIO₄ oxidation resulting in RNA 3′-dialdehydes. (H) Chemical structure of biotin-11-dUTP used in this study.

Figure 2.

Outcomes of combined in vitro selection. The RNA libraries from the last round of (A) click reaction-based selection and (B) b-tREX selections were tested for self-aminoacylation activity by SAv-EMSA. (C) NaIO₄ oxidation confirmed that both elT31 and elT32 catalyzed tRNA 3′-aminoacylation. Two elT-boxzymes were also tested for their tRNA anticodon selectivity. AC, anticodon. aa-tRNA, aminoacylated tRNA. β-elm-tRNA, β-eliminated tRNA. N.D., non-detected.

Figure 3.

Characterization of elT49, with enhanced aminoacylation activity, evolved from elT32. (A) A predicted secondary structure of elT49. Coloring follows Figure 1. The numbering follows that of elT32. The ‘t’ placed before numbers stands for tRNA numbering. The brown letter denotes insertion, while the grey denotes deletion. (B) Trans-aminoacylation activity of elT32-mutants and elT32 in the presence of 5 mM Phe^AZ-CME. Negative control (NC) was performed with elT32 without Phe^AZ-CME. aa-tRNA, aminoacylated tRNA. (C) Apparent initial rates of elT49 trans-aminoacylation against tRNAs bearing various anticodons. Red letters stand for mismatched base pairs. Each initial rate was measured under 5 mM Phe^AZ-CME, determined by the mean of n = 3, and normalized to the correct pair's activity, set as 100. Each circle represents a measurement. Error bars show standard deviation.

Figure 4.

Integrating elT32 and elT49 with a custom-made translation system. (A) One-pot in vitro translation carried out in this study. The FIT system contained elT-boxzyme (elT32_GCC or elT49_GCC), tRNA_GGC, Phe^AZ-CME, and a DNA template coding mR2 for the model Phe^AZ-containing peptide, P4. (B) Representative [¹⁴C]-Asp autoradiography of the translation products separated by tricine SDS-PAGE. Lanes 1 and 2 were one-pot in vitro translation in the presence of either elT32 or elT49, while lane 3 was in the absence of elT-boxzyme. Lane 4 is the positive control where full-length P4 was produced in the presence of Phe^AZ-tRNA_GGC prepared by flexizyme (eFx). (C) Quantification of P4 synthesized in lanes 1–3 in (B). The values were determined by the mean of n = 3. Each circle represents a measurement. Error bars show standard deviation. (D) MALDI-TOF MS of the translational product from lane 2. Arrowhead indicates the product P4, where its azido group was reduced into nitrene by laser and resulted in the observation of the peak [M + 1–28]⁺ (61). †, the peak corresponding to the peptide in which alanine remaining in the FIT mixture incorporated into the GCC codon. ‡, the peak corresponding to the peptide dropped off at the GCC codon, whose sequence was FLAG (DYKDDDDK). Calc., calculated; Obs., observed.