Cell-free co-production of an orthogonal transfer RNA activates efficient site-specific non-natural amino acid incorporation

Abstract

We describe a new cell-free protein synthesis (CFPS) method for site-specific incorporation of non-natural amino acids (nnAAs) into proteins in which the orthogonal tRNA (o-tRNA) and the modified protein (i.e. the protein containing the nnAA) are produced simultaneously. Using this method, 0.9-1.7 mg/ml of modified soluble super-folder green fluorescent protein (sfGFP) containing either p-azido-l-phenylalanine (pAzF) or p-propargyloxy-l-phenylalanine (pPaF) accumulated in the CFPS solutions; these yields correspond to 50-88% suppression efficiency. The o-tRNA can be transcribed either from a linearized plasmid or from a crude PCR product. Comparison of two different o-tRNAs suggests that the new platform is not limited by Ef-Tu recognition of the acylated o-tRNA at sufficiently high o-tRNA template concentrations. Analysis of nnAA incorporation across 12 different sites in sfGFP suggests that modified protein yields and suppression efficiencies (i.e. the position effect) do not correlate with any of the reported trends. Sites that were ineffectively suppressed with the original o-tRNA were better suppressed with an optimized o-tRNA (o-tRNA(opt)) that was evolved to be better recognized by Ef-Tu. This new platform can also be used to screen scissile ribozymes for improved catalysis.

Figure 1.

The nnAAs used in this study. (1) p-azido-l-phenylalanine (pAzF), and (2) p-propargyloxy-l-phenylalanine.

Figure 2.

Schematic of the new cell-free platform for site-specific incorporation of nnAAs. The T7 RNA polymerase catalyses the transcription of both the transzyme RNA, which comprises the fusion of the hammerhead ribozyme and the o-tRNA, and the messenger RNA encoding the protein to be modified with the nnAA. Following transcription, the transzyme cleaves itself to release the biologically active o-tRNA needed for nnAA incorporation at the amber stop codon (TAG; depicted by an asterisk). The three-dimensional structure of sfGFP was used to render the modified protein (PDB accession number: 2B3P).

Figure 3.

(A) Titration of linearized o-tRNA and o-tRNA^opt plasmid templates in the CFPS reactions containing pY71sfGFP39TAG. Error bars indicate ±1 standard deviation from six reactions. (*P < 0.01, **P = 0.025, ***P = 0.023) (B) Production of sfGFP containing pPaF at position 23 (sfGFP23pPaF) over 24 h. The ‘o-extract’ curve was generated from CFPS reactions in which the cell extract contained the o-tRNA. The ‘transzyme’ values were obtained from reactions, in which the o-tRNA and sfGFP23pPaF were produced simultaneously. A different set of reactions was incubated for each of the time points. Error bars indicate ±1 standard deviation for three reactions.

Figure 4.

Codon context and nnAA incorporation. (A) sfGFP crystal structure; nnAA incorporation sites are labelled in black. The colour scheme shows the position along the primary sequence, with blue and green indicating the N-terminal regions, and orange and red depicting the C-terminal ones. The chromophore was not given in the crystal structure, hence the absence of Tyr66. (B) Full-length soluble modified protein yields as a function of the incorporation site under different cell-free conditions. ‘No o-tRNA’ and ‘no pPaF’ depict CFPS reactions in which the o-tRNA or the nnAA was omitted from the reaction solutions, respectively. The ‘o-extract’ series includes the reactions in which the orthogonal extract containing the constitutively expressed o-tRNA was used. The ‘transzyme’ and ‘optimized’ reactions contained 200 μg/ml of the linearized transzyme plasmid coding for o-tRNA and o-tRNA^opt, respectively. Suppression efficiency is defined as the ratio of accumulated full-length soluble protein containing the nnAA over its natural counterpart (i.e. full-length soluble sfGFP without the nnAA). The bottom line indicates the nucleotides that lie immediately 3′ to the amber stop codon at indicated positions. Error bars depict ±1 standard deviation for three reactions. The right panel shows the SDS-PAGE autoradiogram of radiolabelled sfGFPs. The position of nnAA incorporation is indicated above each lane. (C) Titration of linearized o-tRNA^opt plasmid templates in the CFPS reactions containing either pY71sfGFP39TAG or pY71sfGFP74TAG. Error bars indicate ±1 standard deviation for nine sfGFP39pPaF reactions, n = 3 for the sfGFP74pPaF reactions.

Figure 5.

Comparison of modified protein yields obtained with the hammerhead ribozyme and with the three Schistosoma ribozyme variants. The four different linearized o-tRNA plasmid templates (each coding for a different ribozyme) were titrated into cell-free reactions at the indicated concentrations. Error bars indicate ±1 standard deviation for three reactions.

Figure 6.

Delaying the protein plasmid template addition. 50, 100 or 200 μg/ml of the linearized o-tRNA^opt plasmid template was added along with the other standard reagents at the beginning of the reaction; the sfGFP39TAG template was added at the indicated times. Error bars indicate ±1 standard deviation for three reactions.