Optimized directed evolution of E. coli leucyl-tRNA synthetase adds many noncanonical amino acids into the eukaryotic genetic code including ornithine and Nε-acetyl-methyllysine

Abstract

Site-specific incorporation of noncanonical amino acids (ncAAs) into proteins in eukaryotes has predominantly relied on the pyrrolysyl-tRNA synthetase/tRNA pair. However, access to additional easily engineered pairs is crucial for expanding the structural diversity of the ncAA toolbox in eukaryotes. The Escherichia coli-derived leucyl-tRNA synthetase (EcLeuRS)/tRNA pair presents a particularly promising alternative. This pair has been engineered to charge a small yet structurally diverse group of ncAAs in eukaryotic cells. However, expanding the substrate scope of EcLeuRS has been difficult due to the suboptimal yeast-based directed evolution platform used for its engineering. In this study, we address this limitation by optimizing the yeast-based directed evolution platform for efficient selection of ncAA-selective EcLeuRS mutants. Using the optimized selection system, we demonstrate rapid isolation of many novel EcLeuRS mutants capable of incorporating various ncAAs in mammalian cells, including ornithine and N^ε-acetyl-methyllysine, a recently discovered post-translational modification in mammalian cells.

Keywords: Genetic code expansion; aminoacyl-tRNA synthetases; directed evolution; post-translational modifications; protein engineering.

Figure 1.

Structures of the ncAAs used in this work.

Figure 2.

Improving the stringency of the yeast-based selection system. A) The survival-based selection scheme in yeast for engineering aaRSs. B) The established selection system using GAL4–2* is insufficiently stringent: In media lacking uracil, yeast cells expressing the PLRS1/tRNA^EcLeu_CUA pair show robust growth in the presence of Cap (a substrate for PLRS1) as expected. However, weak but significant growth is also observed in the absence of Cap, as well as for yeast cells expressing only tRNA^EcLeu_CUA but no EcLeuRS. C) Introduction of a 3^rd TAG codon in GAL4–2* was explored as a strategy to increase selection stringency and attenuate the leaky survival. D) Evaluating the performance of five GAL4–3* variants during positive selection. Four GAL4–3* variants enable robust growth only in the presence of the PLRS1/tRNA^EcLeu_CUA pair and its substrate Cap. Unlike GAL4–2*, no growth is observed in the absence of Cap or PLRS1. E) Measuring the enrichment of active EcLeuRS mutants, from a defined mixture of active and inactive variants (in a 1:100 ratio), upon positive selection using the GAL4–2* or GAL4–3* selection system. Colony PCR of surviving clones after selection revealed a significantly higher degree enrichment of active EcLeuRS mutant using GAL4–3* (see Figure S2 for more details).

Figure 3.

Engineering EcLeuRS using an optimized selection scheme. A) Structure of the EcLeuRS active site, highlighting the residues randomized in our library and the bound substrate (pink). B) A schematic of the optimized selection workflow using GAL4–3*, where surviving EcLeuRS mutant pool after a selection step is isolated and reintroduced into fresh host cells before the next round of selection. Robust enrichment of ncAA-selective mutants was observed after just three rounds (positive-negative-positive). C) Sequences of unique EcLeuRS variants that exhibited ncAA-dependent LacZ expression in yeast from the Cap selection, and the efficiency of Cap incorporation by these hits in HEK293T cells measured by the expression of the EGFP-39-TAG reporter. D) Sequences of unique EcLeuRS variants that exhibited ncAA-dependent LacZ expression in yeast after the ONBC selection, and the efficiency of ONBC incorporation by these hits in HEK293T cells measured by the expression of the EGFP-39-TAG reporter.

Figure 4.

Genetically encoding Kacme in mammalian cells. A) Kacme has been recently identified as a unique histone H4 PTM. B) Sequences of unique EcLeuRS variants that exhibited ncAA-dependent LacZ expression in yeast after the Kacme selection. C) The efficiency of Kacme incorporation by these hits in HEK293T cells measured by the expression of the EGFP-39-TAG reporter. The results for hits A8, C1, E5, and H1 are magnified in inset. D) Fluorescence images of HEK293T cells co-transfected with EF9RS, tRNA^EcLeu_CUA, and EGFP-39-TAG in the presence and absence of Kacme. E) ESI-MS analysis of purified reporter protein corroborates successful Kacme incorporation. F) Constructs used for Kacme incorporation into histone H4. G) Western blot analysis shows Kacme-dependent expression of H4–5-TAG and H4–12-TAG, supporting successful incorporation of Kacme into two native sites in histone H4 where this PTM is found. GAPDH was used as a loading control.

Figure 5.

Remarkable substrate polyspecificity of two EcLeuRS mutants. A) Incorporation of various ncAAs using B11- and EF9-EcLeuRS in HEK293T cells measured by the expression of EGFP-39-TAG reporter using its characteristic fluorescence in cell-free extract. The lower panel includes various PTMs of lysine. Measurements were performed by co-transfecting pAAV-ITR-CMV-mCherry-1xU6-LeuIGI1 tRNA-plasmid, a pAcBac1-CMV-EGFP-39-TAG reporter plasmid, and a pIDT-CMV-EcLeuRS (B11 or EF9) plasmid into HEK293T cells, in the presence or absence of 10 mM (AcK), or 0.2 mM (for SCOK, TCO*A, TCO4K, BCNK, and ONBO), or 1 mM (all others) ncAA. Expression of EGFP-39-TAG was normalized relative to mCherry (encoded in the tRNA-expressing plasmid), and shown as a % of the wild-type EGFP reporter. Scheme (B) and ESI-MS analysis (C) of purified EGFP-39TAG-ONBO isolated from HEK293T cells using EcLeuRS hit B11 before (left) and after (right) photo-decaging to generate ornithine. Full spectra can be found in Figure S6.