Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2015 Feb 5;518(7537):89-93.
doi: 10.1038/nature14095. Epub 2015 Jan 21.

Recoded organisms engineered to depend on synthetic amino acids

Alexis J Rovner  1 , Adrian D Haimovich  1 , Spencer R Katz  1 , Zhe Li  1 , Michael W Grome  1 , Brandon M Gassaway  2 , Miriam Amiram  1 , Jaymin R Patel  1 , Ryan R Gallagher  1 , Jesse Rinehart  2 , Farren J Isaacs  1
Affiliations

Recoded organisms engineered to depend on synthetic amino acids

Alexis J Rovner et al. Nature. .

Erratum in

Abstract

Genetically modified organisms (GMOs) are increasingly used in research and industrial systems to produce high-value pharmaceuticals, fuels and chemicals. Genetic isolation and intrinsic biocontainment would provide essential biosafety measures to secure these closed systems and enable safe applications of GMOs in open systems, which include bioremediation and probiotics. Although safeguards have been designed to control cell growth by essential gene regulation, inducible toxin switches and engineered auxotrophies, these approaches are compromised by cross-feeding of essential metabolites, leaked expression of essential genes, or genetic mutations. Here we describe the construction of a series of genomically recoded organisms (GROs) whose growth is restricted by the expression of multiple essential genes that depend on exogenously supplied synthetic amino acids (sAAs). We introduced a Methanocaldococcus jannaschii tRNA:aminoacyl-tRNA synthetase pair into the chromosome of a GRO derived from Escherichia coli that lacks all TAG codons and release factor 1, endowing this organism with the orthogonal translational components to convert TAG into a dedicated sense codon for sAAs. Using multiplex automated genome engineering, we introduced in-frame TAG codons into 22 essential genes, linking their expression to the incorporation of synthetic phenylalanine-derived amino acids. Of the 60 sAA-dependent variants isolated, a notable strain harbouring three TAG codons in conserved functional residues of MurG, DnaA and SerS and containing targeted tRNA deletions maintained robust growth and exhibited undetectable escape frequencies upon culturing ∼10(11) cells on solid media for 7 days or in liquid media for 20 days. This is a significant improvement over existing biocontainment approaches. We constructed synthetic auxotrophs dependent on sAAs that were not rescued by cross-feeding in environmental growth assays. These auxotrophic GROs possess alternative genetic codes that impart genetic isolation by impeding horizontal gene transfer and now depend on the use of synthetic biochemical building blocks, advancing orthogonal barriers between engineered organisms and the environment.

PubMed Disclaimer

Figures

Extended Data Figure 1
Extended Data Figure 1. Comprehensive map of synthetic auxotrophs
Circos plot summarizing synthetic auxotrophs generated in this study. Red and green genes reflect knockouts and insertions, respectively. Outermost ticks indicate genomic location, inner blue ticks indicate locations where TAG codons were converted to TAA in the GRO, and green ticks reflect locations of 303 E. coli essential genes. The shaded grey inner circle contains essential TAG loci in synthetic auxotrophs, where yellow ticks represent amino-terminal insertions, blue ticks represent tolerant substitutions, and red ticks represent functional site substitutions. Innermost links represent unique combinations of TAGs in higher-order synthetic auxotrophs. Links of a single color correspond to a single strain.
Extended Data Figure 2
Extended Data Figure 2. rpsD.Q54R is sufficient for loss of pAcF-dependence in SecY.Y122α
a, Plate map with genotypes of strains shown in b and c. Top half of plate: SecY.Y122α.E1 (upper right quadrant) contains the rpsD.Q54R mutation and is an EM of pAcF-auxotroph, SecY.Y122α (upper left quadrant). Lower half of plate: The rpsD.Q54R mutation was introduced into SecY.Y122α (lower right quadrant), resulting in a loss of pAcF-dependence, and reverted to wild type in SecY.Y122α.E1 (lower left quadrant), restoring pAcF-dependence. The amino acid present at residue 54 within RpsD is indicated at the perimeter of the plate, where red signifies that the given mutation was introduced into the genotype by MAGE to demonstrate the causal mechanism of escape. b, Growth on solid permissive media demonstrates growth of all strains. c, Growth on solid nonpermissive media. Introduction of the rpsD.Q54R mutation into the synthetic auxotroph SecY.Y122α results in loss of containment (lower right quadrant). Reverting the mutation to wild type in SecY.Y122α.E1 results in restoration of containment (lower left quadrant). Together, these data demonstrate that the rpsD.Q54R mutation is sufficient for loss of pAcF-dependence in SecY.Y122α.
Extended Data Figure 3
Extended Data Figure 3. Quantitative assessment of amino acid tolerance in higher-order pIF auxotrophs
Representative assay surveying tolerance of one of three essential TAG loci to the twenty amino acids in different synthetic auxotrophs and expressed as log10 of total cell survival. The + symbol indicates the presence of a TAG codon at the specified locus in the background strain and – indicates the wild type codon. Blue and yellow indicate high and low tolerance to substitution, respectively. Substitutions DnaX.Y113W and SecY.Y122Q are tolerated but yielded a lower percentage of survival on nonpermissive media in a background with two TAGs, an effect that was pronounced in a background with three TAGs. While DnaX.Y113, SecY.Y122 and LspA.Y54 are permissive for most natural amino acids, strains with more than one of these essential TAGs are less prone to survive in the event that any one TAG is compromised. SecY.Y122Q and DnaX.Y113W were tolerated substitutions also observed in real EMs of these strains (Supplementary Table 7). Refer to the methods for a complete description of this experiment.
Extended Data Figure 4
Extended Data Figure 4. Deletion of tyrT and tyrV restores pIF-dependence and fitness of rEc.β.dC.12′.E7
a, Plate map with genotypes of strains in parts b and c. rEc.β.dC.12′.E7 is an EM of its sAA-dependent ancestor (rEc.β.dC.12′) and contains a tyrT ochre suppressor mutation (supC). The fitness of rEc.β.dC.12′.E7 in permissive media is impaired relative to rEc.β.dC.12′, with DTs of 91.74 (± 1.49) and 61.81 (± 0.65) minutes, respectively. Tyrosine tRNA redundancy was eliminated (ΔtY) in both strains by λ-Red mediated replacement of tyrT and tyrV with chloramphenicol acetyltransferase (cat), rendering the resulting strains (rEc.β.dC.12′.ΔtY and rEc.β.dC.12′.E7.ΔtY) dependent on tyrU for tyrosine incorporation during normal protein synthesis. Elimination of tyrosine redundancy reduced the EF of rEc.β.dC.12′ from 2.17×10-9 (Fig. 2e) to <4.85×10-12 (no EMs were observed upon plating 2.06×1011 cells) and restored pIF-dependence in rEc.β.dC.12′.E7 to <4.73×10-12 (no EMs were observed upon plating 2.12×1011 cells). EMs were not detected for either strain up to seven days after plating on nonpermissive media (Fig. 3d and Supplementary Table 11). Tyrosine tRNA deletion also restored the fitness of the EM to approximately that of its sAA-dependent ancestor (60.66 ± 0.12 minutes). Taken together, these results establish tyrT as the causal mechanism of escape in rEc.β.dC.12′.E7. b, Growth on solid permissive LB media. c, Growth on solid nonpermissive LB media. All reported DTs are averages, where n=3 technical replicates, and error bars are ±s.d. Refer to the methods for a complete description of EFs.
Extended Data Figure 5
Extended Data Figure 5. Growth profiles of strains expressing phenylalanine or tryptophan amber suppressor tRNAs
Growth was assessed for rEc.γ.dC.46′.ΔtY and rEc.β.dC.12′.ΔtY in the presence of amber suppression by either pTech-supU (blue), pTech-supPhe (red), or in the absence of plasmid-based amber suppression (black). Cells were washed twice with dH2O and re-suspended in the same volume of 1× PBS. Washed cells were normalized by OD600 to inoculate roughly equal numbers of cells per well. Growth profiles are shown for a, b, rEc.γ.dC.46′.ΔtY and c, d, rEc.β.dC.12′.ΔtY in permissive (+sAA/+l-arabinose, solid lines) and nonpermissive (-sAA/-l-arabinose, dashed lines) LB liquid media. DTs are shown for the ancestral strain (black) in permissive media and suppressor-containing strains (red and blue) in nonpermissive media where growth was observed. Plasmid containing strains were always grown in the presence of zeocin for plasmid maintenance. Growth was never observed for the contained ancestors in nonpermissive media (black, dashed lines). In the presence of tryptophan suppression, growth of rEc.γ.dC.46′.ΔtY was not observed and growth of rEc.β.dC.12′.ΔtY was severely impaired (380 minute DT) with a 6.24-fold increase in DT relative to the contained ancestor grown in permissive media. In the presence of phenylalanine suppression, growth of rEc.β.dC.12′.ΔtY was not observed and growth of rEc.γ.dC.46′.ΔtY was severely impaired (252 minute DT) with a 3.90-fold increase in DT relative to the contained ancestor grown in permissive media. Representative growth profiles and DTs are reported. These results repeated at least three times in individual experiments.
Extended Data Figure 6
Extended Data Figure 6. Long term growth of rEc.γ.dC.46′.ΔtY in liquid LB media relative to rEc.γ
Approximately 1011 cells of strain rEc.γ.dC.46′.ΔtY (▲) was inoculated into 1 L of permissive (+sAA/+l-arabinose, blue) or nonpermissive (-sAA/l-arabinose, red) LB media and incubated with agitation at 34°C for 20 days. Results from the equivalent experiment with the non-contained ancestor rEc.γ (◆) are also shown. Cultures were frequently monitored by a, OD600 and quantification of CFUs on solid b, permissive (+sAA/+l-arabinose) and c, nonpermissive (-sAA/-l-arabinose) LB media. CFUs are plotted as the average of three replicates. Open symbols indicate that no CFUs were observed. Symbols for rEc.γ.dC.46′.ΔtY are not visible because CFUs were never observed from either permissive or nonpermissive liquid cultures plated on nonpermissive solid media. At the end of the 20-day growth period, both cultures containing rEc.γ.dC.46′.ΔtY were interrogated for the presence of a single EM by plating each 1 L culture across 30 nonpermissive solid media plates. CFUs were not observed and remained below the limit of detection for the following seven-day observation period. We hypothesize that the decrease in CFU counts obtained on permissive solid media for the permissive culture of rEc.γ.dC.46′.ΔtY reflects pAzF degradation at ≥ 6 days. Reported CFUs are averages, where n=3 technical replicates, and error bars are ±s.d. Refer to the methods for a complete description of this experiment.
Extended Data Figure 7
Extended Data Figure 7. Dose-dependent growth of rEc.γ.dC.46′.ΔtY in pAzF and l-arabinose compared to the non-contained ancestor
Growth in LB media supplemented with different concentrations of pAzF and l-arabinose. Growth profiles for rEc.γacross a gradient of pAzF concentrations in the presence of a, 0%, b, 0.002%, c, 0.02%, and d, 0.2% l-arabinose. Growth profiles for rEc.γ.dC.46′.ΔtY across a gradient of pAzF concentrations in the presence of f, 0%, g, 0.002%, h, 0.02%, and i, 0.2% l-arabinose. e and j, Growth profiles illustrated in parts a-d and f-i are depicted as heat maps in parts e and j, respectively, where the maximum OD600 was obtained from the average of three replicates and plotted in MATLAB. Reported growth profiles and heat map values are averages, where n=3 technical replicates, and error bars are ±s.d.
Extended Data Figure 8
Extended Data Figure 8. Dose-dependent growth of rEc.β.dC.12′.ΔtY in pIF and l-arabinose
Growth in LB media supplemented with different concentrations of pIF and l-arabinose. Growth profiles for rEc.β.dC.12′.ΔtY across a gradient of pIF concentrations in the presence of a, 0%, b, 0.002%, c, 0.02%, and d, 0.2% l-arabinose. e, Growth profiles illustrated in parts a-d are depicted as a heat map, where the maximum OD600 was obtained from the average of three replicates and plotted in MATLAB. Reported growth profiles and heat map values are averages, where n=3 technical replicates, and error bars are ±s.d.
Extended Data Figure 9
Extended Data Figure 9. Proximity-dependent complementation of biotin auxotrophy
Wild type E. coli K-12 substr. MG1655 and three strains auxotrophic for biotin, EcNR2, rEc.γ (a non-contained GRO with an integrated pAzF OTS), and rEc.γ.dC.46′ (also a synthetic auxotroph) were grown either adjacent or separately on rich-defined solid media. EcNR2 grew on biotin-deficient media when plated in close proximity to wild type E. coli, suggesting cross-feeding of the essential metabolite. The pAzF auxotroph only grew on media supplemented with biotin, pAzF, and l-arabinose.
Figure 1
Figure 1. Strategy used to engineer GROs to require sAAs for growth
a, Approaches used to identify suitable loci within essential proteins for synthetic amino acid (sAA, blue) incorporation. b, MAGE was used for site-specific incorporation of TAG codons into essential genes of a genomically recoded organism (GRO) lacking all natural TAG codons (ΔTAG) and release factor 1 (ΔprfA), and containing an orthogonal translation system (OTS, green) consisting of the M. jannaschii aminoacyl-tRNA synthetase (aaRS) and cognate UAG-decoding tRNA. c, Synthetic auxotrophs that depend on sAAs for growth were isolated.
Figure 2
Figure 2. Characterization of strains dependent on sAA incorporation in essential proteins
a, Doubling time (DT) ratios for the non-contained ancestor to synthetic auxotroph containing one TAG. b, Escape frequencies (EFs) of strains from a. c, Superimposed MS/MS spectra for DnaX peptides from DnaX.Y113α (red) and the non-contained ancestor rEc.α (blue). Overlapping peaks are purple and a mass shift relative to rEc.α identifies Y113 as the pAcF incorporation site in DnaX.Y113α; see methods. d, EFs for strains with multiple TAG codons e, and/or functional mismatch repair (prime, mutS+). For all plots, average values of 3 technical replicates are plotted with error bars, ±s.d.
Figure 3
Figure 3. Characterization of strains dependent on sAA incorporation at active and dimerization sites in essential proteins
a, Doubling time (DT) ratios for the non-contained ancestor to pAzF auxotroph with one or more TAGs at functional loci calculated from growth in 5 mM pAzF and 0.2% l-arabinose. b, Escape frequencies (EFs) of strains in part a; bars represent EFs below the detection limit; average EFs are plotted. c, Representative assay surveying tolerance of TAG loci to 20 amino acids in different synthetic auxotrophs and expressed as log10 of total cell survival; + indicates a TAG codon at the locus in the background strain and – indicates the wild type codon; blue and yellow indicate high and low tolerance to substitution, respectively; see methods. d, Representative escape assay monitoring EFs up to seven days after plating on solid nonpermissive media; hollow symbols/dashed lines, no observed EMs; see methods. e, Temporal monitoring of permissive (P, blue) and nonpermissive (NP, red) cultures inoculated with ∼109, 1010, or 1011 cells of rEc.γ.dC.46′.ΔtY by OD600 and f, associated CFUs as sampled on P (solid lines) or NP (dashed lines) solid media; CFUs were never observed on NP solid media; hollow data points indicate no observed CFUs. g, Maximum OD600 values during growth in LB across a concentration gradient of pAzF and l-arabinose. For all plots, average values of 3 technical replicates are plotted with error bars, ±s.d.
Figure 4
Figure 4. Investigating the viability of synthetic auxotrophs on diverse media types
Rescue by cross-feeding shown through spotting on diverse media types +/- pAzF/l-arabinose and biotin supplementation; EcNR2, rEc.γ, and rEc.γ.dC.46′ are auxotrophic for biotin (ΔbioA/B) and rEc.γ.dC.46′ is also a pAzF auxotroph.
See this image and copyright information in PMC

Comment in

  • Synthetic biology: GMOs in lockdown.
    Nunes-Alves C. Nunes-Alves C. Nat Rev Microbiol. 2015 Mar;13(3):125. doi: 10.1038/nrmicro3443. Epub 2015 Feb 9. Nat Rev Microbiol. 2015. PMID: 25659321 No abstract available.
  • Synthetic biology: GMOs in lockdown.
    Nunes-Alves C. Nunes-Alves C. Nat Rev Genet. 2015 Mar;16(3):127. doi: 10.1038/nrg3909. Nat Rev Genet. 2015. PMID: 25690387 No abstract available.

Similar articles

See all similar articles

Cited by

See all "Cited by" articles

References

    1. Way JC, Collins JJ, Keasling JD, Silver PA. Integrating Biological Redesign: Where Synthetic Biology Came From and Where It Needs to Go. Cell. 2014;157:151–161. doi: 10.1016/j.cell.2014.02.039. - DOI - PubMed
    1. Schmidt M, de Lorenzo V. Synthetic constructs in/for the environment: managing the interplay between natural and engineered Biology. FEBS letters. 2012;586:2199–2206. doi: 10.1016/j.febslet.2012.02.022. - DOI - PMC - PubMed
    1. Moe-Behrens GH, Davis R, Haynes KA. Preparing synthetic biology for the world. Frontiers in microbiology. 2013;4:5. doi: 10.3389/fmicb.2013.00005. - DOI - PMC - PubMed
    1. Pieper DH, Reineke W. Engineering bacteria for bioremediation. Current opinion in biotechnology. 2000;11:262–270. - PubMed
    1. Steidler L. Genetically engineered probiotics. Best practice & research Clinical gastroenterology. 2003;17:861–876. - PubMed

Publication types

MeSH terms

Associated data