Multiplex suppression of four quadruplet codons via tRNA directed evolution

Genetic code expansion technologies supplement the natural codon repertoire with assignable variants in vivo, but are often limited by heterologous translational components and low suppression efficiencies. Here, we explore engineered Escherichia coli tRNAs supporting quadruplet codon translation by first developing a library-cross-library selection to nominate quadruplet codon–anticodon pairs. We extend our findings using a phage-assisted continuous evolution strategy for quadruplet-decoding tRNA evolution (qtRNA-PACE) that improved quadruplet codon translation efficiencies up to 80-fold. Evolved qtRNAs appear to maintain codon-anticodon base pairing, are typically aminoacylated by their cognate tRNA synthetases, and enable processive translation of adjacent quadruplet codons. Using these components, we showcase the multiplexed decoding of up to four unique quadruplet codons by their corresponding qtRNAs in a single reporter. Cumulatively, our findings highlight how E. coli tRNAs can be engineered, evolved, and combined to decode quadruplet codons, portending future developments towards an exclusively quadruplet codon translation system.

Supplementary Figure 3 | LC-MS/MS analysis of lacZ selection-derived hits.Mass spectra of sfGFP fragments resulting from qtRNA Gly GGGG (a), qtRNA His AGGA (b), qtRNA Thr ACCA (c), qtRNA Glu CGGU (d), and qtRNA Tyr UAGA (e) suppression of cognate quadruplet codon at Y151.Multiple peptides were observed in some cases and are shown for completeness.Summary LC-MS/MS results are reported in Supplementary Table 8.Supplementary Figure 4 | Benchmarking PACE-evolved qtRNA SPs using progressively stringent APs.a) Schematic representation of the accessory plasmid design, wherein either AP copy number was modified (L = wild-type RepA ~4 copies/cell; H = RepA E93K ~27 copies/cell) or the number of quadruplet codons in pIII was progressively increased.In all cases, clonal SPs encoding the indicated engineered or evolved qtRNAs were challenged to form plaques in S3489 cells.For each SP, the threshold for plaque formation is visualized for serine (b), arginine (c), glutamine (d), tryptophan (e), and tyrosine (f).
Supplementary Figure 5 | Analysis of engineered and evolved qtRNAs in bacterial RF1 knockout strains.a) Engineered and evolved UAGA-decoding qtRNAs assayed using an endpoint fluorescence reporter assay using two RF1 knockout strains (C321.ΔA 1 and JX33 2 ) with one RF1+ strains (C321).In all cases, tRNAs were assayed alongside a reporter incorporating the quadruplet codon UAGA at sfGFP position Y151.b) Extension of the sfGFP reporter assay in JX33 and S3489 (control RF+) to all rationally engineered UAGA-decoding qtRNAs (n = 6 biologically independent samples except for Asn, Gly, His, Ile, Phe, Pro, Thr, and Val where n = 5).In all cases, reporter data is normalized to an otherwise wild-type protein.Data represents the mean and standard deviation as appropriate.qtRNA Ser UAGA-Evo1 translates UAGA quadruplet codons at both S357 and S164 more efficiently than when using the host ribosome, especially for reporters with multiple frameshifts (n = 4 biologically independent samples except for S357/S164+tRNA-Ser-UCG where n = 2).c) Orthogonal ribosomes incorporating the described RiboQ1 mutations (U531G/U534A/A1196G/A1197G) 3 show comparable luminescence to the host wildtype ribosome for quadruplet codon translation (n = 4 biologically independent samples except for S357/S164+tRNA-Ser-UCG where n = 2).In all cases, the average wild-type (triplet) LuxAB reporter activity is shown as a dashed line.Data represent the mean and standard deviation as appropriate.OD optical density, AU arbitrary units.

Figure 6 |
Models of engineered and evolved qtRNAs.Cloverleaf models of engineered UAGA qtRNAs and evolved variants: arginine (a), glutamine (b), serine (c), tryptophan (d), and tyrosine (e).In all cases, the engineered UAGA codon is highlighted in gray, and PACE-acquired mutations are highlighted in red.qtRNA Ser UAGA-Evo1 was used to initiate the experiment that produced qtRNA Ser UAGA-Evo2 and qtRNA Ser UAGA-Evo3.Supplementary Figure 7 | LC-MS/MS analysis of engineered and evolved qtRNAs.Mass spectra of the resultant sfGFP fragments from the suppression of UAGA quadruplet codon at sfGFP Y151 by the engineered and subsequently evolved qtRNAs: qtRNA Arg UAGA (a-c), qtRNA Gln UAGA (d-f), qtRNA Ser UAGA (g-j), qtRNA Trp UAGA (k,l), and qtRNA Tyr UAGA (m,n).Multiple peptides were observed in some cases and are shown for completeness.Summary LC-MS/MS results are reported in Supplementary Table 8.Supplementary Figure 8 | Analysis of qtRNA/codon specificity and crosstalk.Evolved UAGA-qtRNAs were tested using mismatched codon reporters to assess instances of decoding crosstalk.LuxAB reporters encoding quadruplet codons with modifications at the third position (a-e) or fourth position (f-j) showcase absolute requirement for guanine at the third position and preference for adenine at the fourth position.k-o) Evolved UAGA-qtRNAs continue to crosstalk with amber (UAG) stop codons, with a moderate preference for purines at the first position of the subsequent codon.In all cases, LuxAB reporter data is normalized to an otherwise wild-type protein.Data represents the mean and standard deviation of 4 biologically independent samples except for Trp-UAGA-Evo1 UAGA/UAGA/UAGC/UAGU and Tyr-UAGA-Evo1 UAAA/UAGC/UAGG/UAGU/UAG.a/UAG.gwhere n = 3 as well as Ser-UAGA-Evo3 UAG.g/UAG.u,Trp-UAGA-Evo1 UAG.c, and Tyr-UAGA-Evo1 UAG.c where n = 2).Supplementary Figure 9 | Translation using orthogonal ribosome.a) Translation of a reporter containing a UAGA codon at either residue 357 or residue 164, in comparison to translation of a luciferase containing UAGA codons at both locations (n = 4 biologically independent samples).b) Using the H3 o-RBS/o-antiRBS pair (5'-AUAUGU/5'-AUGUUC), Figure 10 | LC-MS/MS analysis of evolved qtRNA translating a linker containing adjacent UAGA quadruplet codons.Mass spectra of sfGFP-linked-mCherry fragments resulting from qtRNA Ser UAGA-Evo3 (a) and qtRNA Tyr UAGA-Evo1 (b) suppression of a linker containing six adjacent UAGA quadruplet codons, and qtRNA Gln UAGA-Evo2 (c) suppression of a linker containing five adjacent UAGA quadruplet codons.Mass spectra of the linker fragment resulting from qtRNA Arg UAGA-Evo1 and qtRNA Trp UAGA-Evo1 were unable to be identified, likely due to peptide hydrophobicity limiting chromatographic separation.Multiple peptides were observed in some cases and are shown for completeness.Summary LC-MS/MS results are reported in Supplementary Table 8.Supplementary Figure 11 | LC-MS/MS analysis of qtRNA translating cognate quadruplet codons at positions throughout sfGFP.Mass spectra of sfGFP fragments resulting from qtRNA His AGGA suppression of its cognate quadruplet codon at H148 (a), qtRNA Gly GGGG suppression of its cognate quadruplet codon at G174 (b), qtRNA Ser UAGA-Evo3 suppression of its cognate quadruplet codon at S202 (c), and qtRNA Glu CGGU suppression of its cognate quadruplet codon at E213 (d).Multiple peptides were observed in some cases and are shown for completeness.Summary LC-MS/MS results are reported in Supplementary Table 8 .

Table 2 | Sequences of all natural E. coli tRNA scaffolds used for qtRNA engineering
. In all cases, tRNA sequences are shown in magenta, and the anticodon is shown in purple.Flanking sequences (black) were included in vector design to ensure efficient qtRNA maturation.All coordinates derive from E. coli DH10B genome.

Table 3 | Doubling time analysis for all natural, engineered, and evolved qtRNAs.
All doubling time analyses used S3489 cells with tRNA expression plasmids encoding the shown tRNA under induced conditions.Data represents the mean and standard deviation of 4 -8 biological replicates.
Supplementary Table 4 | Amino acid abundance at position Y151 of sfGFP in response toUAGA quadruplet codon translation.Mutations are indicated for each variant using universal tRNA nomenclature.AA: amino acid.