Directed evolution of the fluorescent protein CGP with in situ biosynthesized noncanonical amino acids

Abstract

The incorporation of noncanonical amino acids (ncAAs) into proteins can enhance their function beyond the abilities of canonical amino acids and even generate new functions. However, the ncAAs used for such research are usually chemically synthesized, which is expensive and hinders their application on large industrial scales. We believe that the biosynthesis of ncAAs using metabolic engineering and their employment in situ in target protein engineering with genetic code expansion could overcome these limitations. As a proof of principle, we biosynthesized four ncAAs, O-L-methyltyrosine, 3,4-dihydroxy-L-phenylalanine, 5-hydroxytryptophan, and 5-chloro-L-tryptophan using metabolic engineering and directly evolved the fluorescent consensus green protein (CGP) by combination with nine other exogenous ncAAs in Escherichia coli. After screening a TAG scanning library expressing 13 ncAAs, several variants with enhanced fluorescence and stability were identified. The variants CGP^{V3pMeoF/K190pMeoF} and CGP^{G20pMeoF/K190pMeoF} expressed with biosynthetic O-L-methyltyrosine showed an approximately 1.4-fold improvement in fluorescence compared to the original level, and a 2.5-fold improvement in residual fluorescence after heat treatment. Our results demonstrated the feasibility of integrating metabolic engineering, genetic code expansion, and directed evolution in engineered cells to employ biosynthetic ncAAs in protein engineering. These results could further promote the application of ncAAs in protein engineering and enzyme evolution.

Importance: Noncanonical amino acids (ncAAs) have shown great potential in protein engineering and enzyme evolution through genetic code expansion. However, in most cases, ncAAs must be provided exogenously during protein expression, which hinders their application, especially when they are expensive or have poor cell membrane penetration. Engineering cells with artificial metabolic pathways to biosynthesize ncAAs and employing them in situ for protein engineering and enzyme evolution could facilitate their application and reduce costs. Here, we attempted to evolve the fluorescent consensus green protein (CGP) with biosynthesized ncAAs. Our results demonstrated the feasibility of using biosynthesized ncAAs in protein engineering, which could further stimulate the application of ncAAs in bioengineering and biomedicine.

Keywords: fluorescent protein; genetic code expansion; metabolic engineering; noncanonical amino acids; protein engineering.

Fig 1

Schematic illustration of the integration of metabolic engineering, genetic code expansion, and directed evolution to engineer target proteins with biosynthetic ncAAs.

Fig 2

Chemical structures and suppression efficiency of the ncAAs. (A) Chemical structures of the ncAAs used in this study. (B) Incorporation efficiencies of different ncAAs with corresponding orthogonal aaRS/tRNA pairs. A GFP-based fluorescence assay was employed and the determined values are presented as the means ± standard deviation of three independent experiments.

Fig 3

Metabolic pathway engineering for the biosynthesis of ncAAs. (A) The features of the plasmids used to biosynthesize ncAAs. (B) In-cell fluorescence assay of the metabolic pathway or enzymes expressed by bacteria samples. A GFP^Y151TAG variant was used to evaluate ncAA production by coupling with genetic code expansion. Values are presented as means ± standard deviation of three independent experiments. (C) SDS-PAGE analysis of purified GFP and its ncAA variants. (D) Molecular masses of GFP and its ncAA variants determined by LC-MS.

Fig 4

Screening of the CGP(loop) variant library. (A) Representative flow cytometry results of the expressed ncAA variant libraries with enhanced fluorescence. (B) Representative culture plates of the expressed ncAA variant libraries. The colonies with recovered fluorescence are indicated by red arrows. (C) The fluorescence of bacteria cultures obtained from the plates.

Fig 5

Properties of the CGP ncAA variants. (A) SDS-PAGE analysis of the purified CGP variants. (B) Relative fluorescence of CGP and its ncAA variants. The fluorescence of CGP is defined as 100%. (C) Residual fluorescence of CGP after incubation at 80°C (up panel) and 75°C (down panel) for 30 min. The residual fluorescence is presented as a percentage of the original fluorescence of each protein sample. Values are presented as the means ± standard deviation of three independent experiments.

Fig 6

Molecular masses of CGP and its ncAA variants. (A) In-cell fluorescence of the expressed double variants. (B) SDS-PAGE analysis of the purified double variants. LC-MS analysis of (C) CGP, (D) CGP^{V3pBrF/K190pBrF}, (E) CGP^{V3pMeoF/K190pMeoF}, (F) CGP^{G20pBrF/K190pBrF}, and (G) CGP^{G20pMeoF/K190pMeoF}.

Fig 7

Comparison of the properties between CGP and the double variants. (A) Relative fluorescence of CGP and the double variants. The fluorescence of CGP is defined as 100%. (B) The absorption and emission spectra of CGP and the double variants. The curves were generated by normalizing the peaks of the absorption and emission spectra to 1. (C) The fluorescence of CGP and the double variants at different temperatures determined in an RT-PCR machine. (D) Residual fluorescence of CGP after incubation at 80°C for 30 min. The residual fluorescence is presented as a percentage of the original fluorescence of each protein sample. (E) Melting curves of CGP and the double variants, CGP^{V3pMeoF/K190pMeoF} and CGP^{G20pMeoF/K190pMeoF}, determined using DSC. Values are presented as the means ± standard deviation of three independent experiments.

Fig 8

Molecular dynamic analysis of CGP and the double variant CGP^{V3pMeoF/K190pMeoF} and CGP^{G20pMeoF/K190pMeoF}. (A) RMSD values of CGP, CGP^{V3pMeoF/K190pMeoF}, and CGP^{G20pMeoF/K190pMeoF} over 100 ns. (B) RMSF values of CGP, CGP^{V3pMeoF/K190pMeoF}, and CGP^{G20pMeoF/K190pMeoF}. (C) Overlaid structures of CGP, CGP^{V3pMeoF/K190pMeoF}, and CGP^{G20pMeoF/K190pMeoF}. The structures of CGP, CGP^{V3pMeoF/K190pMeoF}, and CGP^{G20pMeoF/K190pMeoF} are colored green, pink, and blue, respectively. The protruded C-terminus structure of CGP is indicated by a red arrow. (D) Overlaid local structures of CGP and CGP^{V3pMeoF/K190pMeoF} at position 3. The residual side chains at position 3 are shown as sticks. (E) Overlaid local structures of CGP and CGP^{G20pMeoF/K190pMeoF} at position 20. The residual side chains at position 20 are shown as sticks. (F) Overlaid local structures of CGP and the two double variants. The local secondary structures near position 190 showed remarkable differences, as pointed out in the picture.