Expanded genetic code
An expanded genetic code is an artificially modified
The key prerequisites to expand the genetic code are:
- the non-standard amino acid to encode,
- an unused codon to adopt,
- a tRNAthat recognises this codon, and
- a tRNA synthetasethat recognises only that tRNA and only the non-standard amino acid.
Expanding the genetic code is an area of research of synthetic biology, an applied biological discipline whose goal is to engineer living systems for useful purposes. The genetic code expansion enriches the repertoire of useful tools available to science.
In May 2019, researchers, in a milestone effort, reported the creation of a new
Introduction
It is noteworthy that the genetic code for all organisms is basically the same, so that all living beings use the same 'genetic language'.
In order to incorporate a novel amino acid into the genetic code several changes are required. First, for successful translation of a novel amino acid, the codon to which the novel amino acid is assigned cannot already code for one of the 20 natural amino acids. Usually a nonsense codon (stop codon) or a four-base codon are used.[6] Second, a novel pair of tRNA and aminoacyl tRNA synthetase are required, these are called the orthogonal set. The orthogonal set must not crosstalk with the endogenous tRNA and synthetase sets, while still being functionally compatible with the ribosome and other components of the translation apparatus. The active site of the synthetase is modified to accept only the novel amino acid. Most often, a library of mutant synthetases is screened for one which charges the tRNA with the desired amino acid. The synthetase is also modified to recognize only the orthogonal tRNA.[6] The tRNA synthetase pair is often engineered in other bacteria or eukaryotic cells.[9]
In this area of research, the 20 encoded proteinogenic amino acids are referred to as standard amino acids, or alternatively as natural or canonical amino acids, while the added amino acids are called non-standard amino acids (NSAAs), or unnatural amino acids (uAAs; term not used in papers dealing with natural non-proteinogenic amino acids, such as phosphoserine), or non-canonical amino acids.
Non-standard amino acids
The first element of the system is the amino acid that is added to the genetic code of a certain strain of organism.
Over 71 different NSAAs have been added to different strains of E. coli, yeast or mammalian cells.[10] Due to technical details (easier chemical synthesis of NSAAs, less crosstalk and easier evolution of the aminoacyl-tRNA synthase), the NSAAs are generally larger than standard amino acids and most often have a phenylalanine core but with a large variety of different substituents. These allow a large repertoire of new functions, such as labeling (see figure), as a fluorescent reporter (e.g. dansylalanine)[11] or to produce translational proteins in E. coli with Eukaryotic post-translational modifications (e.g. phosphoserine, phosphothreonine, and phosphotyrosine).[10][12]
The founding work was reported by Rolf Furter, who singlehandedly used yeast tRNAPhe/PheRS pair to incorporate p-iodophenylalanine in E.coli.[13]
Unnatural amino acids incorporated into proteins include heavy atom-containing amino acids to facilitate certain x-ray crystallographic studies; amino acids with novel steric/packing and electronic properties; photocrosslinking amino acids which can be used to probe protein-protein interactions in vitro or in vivo; keto, acetylene, azide, and boronate-containing amino acids which can be used to selectively introduce a large number of biophysical probes, tags, and novel chemical functional groups into proteins in vitro or in vivo; redox active amino acids to probe and modulate electron transfer; photocaged and photoisomerizable amino acids to photoregulate biological processes; metal binding amino acids for catalysis and metal ion sensing; amino acids that contain fluorescent or infra-red active side chains to probe protein structure and dynamics; α-hydroxy acids and D-amino acids as probes of backbone conformation and hydrogen bonding interactions; and sulfated amino acids and mimetics of phosphorylated amino acids as probes of post-translational modifications.[14][15][16]
Availability of the non-standard amino acid requires that the organism either import it from the medium or biosynthesize it. In the first case, the unnatural amino acid is first synthesized chemically in its optically pure L-form.[17] It is then added to the growth medium of the cell.[10] A library of compounds is usually tested for use in incorporation of the new amino acid, but this is not always necessary, for example, various transport systems can handle unnatural amino acids with apolar side-chains. In the second case, a biosynthetic pathway needs to be engineered, for example, an E. coli strain that biosynthesizes a novel amino acid (p-aminophenylalanine) from basic carbon sources and includes it in its genetic code.[16][18][19] Another example: the production of phosphoserine, a natural metabolite, and consequently required alteration of its pathway flux to increase its production.[12]
Codon assignment
Another element of the system is a codon to allocate to the new amino acid.[citation needed]
A major problem for the genetic code expansion is that there are no free codons. The genetic code has a non-random layout that shows tell-tale signs of various phases of primordial evolution, however, it has since frozen into place and is near-universally conserved.[20] Nevertheless, some codons are rarer than others. In fact, in E. coli (and all organisms) the codon usage is not equal, but presents several rare codons (see table), the rarest being the amber stop codon (UAG).
Codon | Amino acid | Abundance (%) |
---|---|---|
UUU | Phe (F) | 1.9 |
UUC | Phe (F) | 1.8 |
UUA | Leu (L) | 1.0 |
UUG | Leu (L) | 1.1 |
CUU | Leu (L) | 1.0 |
CUC | Leu (L) | 0.9 |
CUA | Leu (L) | 0.3 |
CUG | Leu (L) | 5.2 |
AUU | Ile (I) | 2.7 |
AUC | Ile (I) | 2.7 |
AUA | Ile (I) | 0.4 |
AUG | Met (M) | 2.6 |
GUU | Val (V) | 2.0 |
GUC | Val (V) | 1.4 |
GUA | Val (V) | 1.2 |
GUG | Val (V) | 2.4 |
UCU | Ser (S) | 1.1 |
UCC | Ser (S) | 1.0 |
UCA | Ser (S) | 0.7 |
UCG | Ser (S) | 0.8 |
CCU | Pro (P) | 0.7 |
CCC | Pro (P) | 0.4 |
CCA | Pro (P) | 0.8 |
CCG | Pro (P) | 2.4 |
ACU | Thr (T) | 1.2 |
ACC | Thr (T) | 2.4 |
ACA | Thr (T) | 0.1 |
ACG | Thr (T) | 1.3 |
GCU | Ala (A) | 1.8 |
GCC | Ala (A) | 2.3 |
GCA | Ala (A) | 0.1 |
GCG | Ala (A) | 3.2 |
UAU | Tyr (Y) | 1.6 |
UAC | Tyr (Y) | 1.4 |
UAA | Stop | 0.2 |
UAG | Stop | 0.03 |
CAU | His (H) | 1.2 |
CAC | His (H) | 1.1 |
CAA | Gln (Q) | 1.3 |
CAG | Gln (Q) | 2.9 |
AAU | Asn (N) | 1.6 |
AAC | Asn (N) | 2.6 |
AAG | Lys (K) | 3.8 |
AAA | Lys (K) | 1.2 |
GAU | Asp (D) | 3.3 |
GAC | Asp (D) | 2.3 |
GAA | Glu (E) | 4.4 |
GAG | Glu (E) | 1.9 |
UGU | Cys (C) | 0.4 |
UGC | Cys (C) | 0.6 |
UGA | Stop | 0.1 |
UGG | Trp (W) | 1.4 |
CGU | Arg (R) | 2.4 |
CGC | Arg (R) | 2.2 |
CGA | Arg (R) | 0.3 |
CGG | Arg (R) | 0.5 |
AGU | Ser (S) | 0.7 |
AGC | Ser (S) | 1.5 |
AGA | Ser (S) | 0.2 |
AGG | Ser (S) | 0.2 |
GGU | Gly (G) | 2.8 |
GGC | Gly (G) | 3.0 |
GGC | Gly (G) | 0.7 |
GGA | Gly (G) | 0.9 |
Amber codon suppression
The possibility of reassigning codons was realized by Normanly et al. in 1990, when a viable mutant strain of E. coli read through the UAG ("amber") stop codon.[22] This was possible thanks to the rarity of this codon and the fact that release factor 1 alone makes the amber codon terminate translation. Later, in the
The amber codon is the least used codon in Escherichia coli, but hijacking it results in a substantial loss of fitness. One study, in fact, found that there were at least 83 peptides majorly affected by the readthrough[26] Additionally, the labelling was incomplete. As a consequence, several strains have been made to reduce the fitness cost, including the removal of all amber codons from the genome. In most E. coli K-12 strains (viz.
Rare sense codon reassignment
In addition to the amber codon, rare sense codons have also been considered for use. The AGG codon codes for arginine, but a strain has been successfully modified to make it code for 6-N-allyloxycarbonyl-lysine.[30] Another candidate is the AUA codon, which is unusual in that its respective tRNA has to differentiate against AUG that codes for methionine (primordially, isoleucine, hence its location). In order to do this, the AUA tRNA has a special base, lysidine. The deletion of the synthase (tilS) was possible thanks to the replacement of the native tRNA with that of Mycoplasma mobile (no lysidine). The reduced fitness is a first step towards pressuring the strain to lose all instances of AUA, allowing it to be used for genetic code expansion.[31]
E. coli strain Syn61 is a variant where all uses of TCG (Ser), TCA (Ser), TAG (STOP) codons are eliminated using a synthetic genome (see § Recoded synthetic genome below). By removing the unneeded tRNA genes and RF1, strain Syn61Δ3 was produced. The three freed codons then become available for adding three special residues, as demonstrated in strain "Syn61Δ3(ev4)".[32]
Four base (quadruplet) codons
While triplet codons are the basis of the genetic code in nature, programmed +1 frameshift is a natural process that allows the use of a four-nucleotide sequence (quadruplet codon) to encode an amino acid.
tRNA/synthetase pair
Another key element is the tRNA/synthetase pair.
The orthologous set of synthetase and tRNA can be mutated and screened through directed evolution to charge the tRNA with a different, even novel, amino acid. Mutations to the plasmid containing the pair can be introduced by error-prone PCR or through degenerate primers for the synthetase's active site. Selection involves multiple rounds of a two-step process, where the plasmid is transferred into cells expressing chloramphenicol acetyl transferase with a premature amber codon. In the presence of toxic chloramphenicol and the non-natural amino acid, the surviving cells will have overridden the amber codon using the orthogonal tRNA aminoacylated with either the standard amino acids or the non-natural one. To remove the former, the plasmid is inserted into cells with a barnase gene (toxic) with a premature amber codon but without the non-natural amino acid, removing all the orthogonal syntheses that do not specifically recognize the non-natural amino acid.[6] In addition to the recoding of the tRNA to a different codon, they can be mutated to recognize a four-base codon, allowing additional free coding options.[42] The non-natural amino acid, as a result, introduces diverse physicochemical and biological properties in order to be used as a tool to explore protein structure and function or to create novel or enhanced protein for practical purposes.
Orthogonal sets in model organisms
The orthogonal pairs of synthetase and tRNA that work for one organism may not work for another, as the synthetase may mis-aminoacylate endogenous tRNAs or the tRNA be mis-aminoacylated itself by an endogenous synthetase. As a result, the sets created to date differ between organisms.
This section is missing information about codon used and new AA; consider tRNATyrCUA formatting.(March 2019) |
Pair | Source | E. coli | Yeast | Mammals | Notes and references |
---|---|---|---|---|---|
tRNATyr-TyrRS | Methanococcus jannaschii |
Yes | No | No | |
tRNALys–LysRS | Pyrococcus horikoshii | Yes | No | No | [43] |
tRNAGlu–GluRS | Pyrococcus horikoshii | Yes | No | No | [44] |
tRNALeu–LeuRS | tRNA: mutant Halobacterium sp. RS: Methanobacterium thermoautotrophicum |
Yes | No | No | [45] |
tRNAAmber-PylRS | Methanosarcina mazei |
Yes | Yes | Yes | [46] |
tRNAAmber-3-iodotyrosyl-RS | RS: variant Methanocaldococcus jannaschii aaRS |
Yes | No | No | [47] |
tRNATyr/Amber-TyrRS | Escherichia coli | No | Yes | No | Reported in 2003,[48] mentioned in 2014 LeuRS[49] |
tRNAiMet-GlnRS | tRNA: human RS: Escherichia coli |
No | Yes | No | Switched to Amber codon.[50] |
tRNAifMet-TyrRS | tRNA: Escherichia coli RS: S. cerevisiae |
Yes | Yes | No | Switched to Amber codon.[50] |
tRNALeu/Amber-LeuRS | Escherichia coli | No | Yes | Yes | Reported in 2004 and mutated for 2-Aminooctanoic acid, o-methyl tyrosine, and o-nitrobenzyl cysteine.[49] Evolved in yeast for 4,5-dimethoxy-2-nitrobenzyl serine,[51][52] tested in mice and mammalian cells with photosensitive 4,5-dimethoxy-2-nitrobenzyl-cysteine.[53][54] |
tRNATyr-TyrRS | Bacillus stearothermophilus |
No | No | Yes | [9] |
tRNATrp-TrpRS | Bacillus subtilis, RS modified | No | No | Yes | New AA is 5-OH Trp.[55] |
In 2017 a mouse engineered with an extended genetic code that can produce proteins with unnatural amino acids was reported.[56]
Orthogonal ribosomes
Similarly to orthogonal tRNAs and aminoacyl tRNA synthetases (aaRSs), orthogonal ribosomes have been engineered to work in parallel to the natural ribosomes. Orthogonal ribosomes ideally use different mRNA transcripts than their natural counterparts and ultimately should draw on a separate pool of tRNA as well. This should alleviate some of the loss of fitness which currently still arises from techniques such as Amber codon suppression. Additionally, orthogonal ribosomes can be mutated and optimized for particular tasks, like the recognition of quadruplet codons. Such an optimization is not possible, or highly disadvantageous for natural ribosomes.
o-Ribosome
In 2005 three sets of ribosomes were published, which did not recognize natural mRNA, but instead translated a separate pool of orthogonal mRNA (o-mRNA).
Ribo-X
In 2007 the group of Jason W. Chin presented an orthogonal ribosome, which was optimized for Amber codon suppression.[58] The 16S rRNA was mutated in such a way that it bound the release factor RF1 less strongly than the natural ribosome does. This ribosome did not eliminate the problem of lowered cell fitness caused by suppressed stop codons in natural proteins. However through the improved specificity it raised the yields of correctly synthesized target protein significantly (from ~20% to >60% percent for one amber codon to be suppressed and from <1% to >20% for two amber codons).
Ribo-Q
In 2010 the group of Jason W. Chin presented a further optimized version of the orthogonal ribosome. The Ribo-Q is a 16S rRNA optimized to recognize tRNAs, which have quadruplet anti-codons to recognize quadruplet codons, instead of the natural triplet codons.[37] With this approach the number of possible codons rises from 64 to 256. Even accounting for a variety of stop codons, more than 200 different amino acids could potentially be encoded this way.
Ribosome stapling
The orthogonal ribosomes described above all focus on optimizing the 16S rRNA. Thus far, this optimized 16S rRNA was combined with natural large-subunits to form orthogonal ribosomes. If the 23S rRNA, the main RNA-component of the large ribosomal subunit, is to be optimized as well, it had to be assured, that there was no crosstalk in the assembly of orthogonal and natural ribosomes (see figureX B). To ensure that optimized 23S rRNA would only form into ribosomes with the optimized 16S rRNA, the two rRNAs were combined into one transcript.[59] By inserting the sequence for the 23S rRNA into a loop-region of the 16S rRNA sequence, both subunits still adopt functioning folds. Since the two rRNAs are linked and thus in constant proximity, they preferably bind each other, not other free floating ribosomal subunits.[citation needed]
Engineered peptidyl transferase center
In 2014 it was shown that by altering the peptidyl transferase center of the 23S rRNA, ribosomes could be created which draw on orthogonal pools of tRNA.[60] The 3' end of tRNAs is universally conserved to be CCA. The two cytidines base pair with two guanines the 23S rRNA to bind the tRNA to the ribosome. This interaction is required for translational fidelity. However, by co-mutating the binding nucleotides in such a way, that they can still base pair, the translational fidelity can be conserved. The 3'-end of the tRNA is mutated from CCA to CGA, while two cytidine nucleotides in the ribosomes A- and P-sites are mutated to guanidine. This leads to ribosomes which do not accept naturally occurring tRNAs as substrates and to tRNAs, which cannot be used as substrate by natural ribosomes.
To use such tRNAs effectively, they would have to be aminoacylated by specific, orthogonal aaRSs. Most naturally occurring aaRSs recognize the 3'-end of their corresponding tRNA.[61][62] aaRSs for these 3'-mutated tRNAs are not available yet. Thus far, this system has only been shown to work in an in-vitro translation setting where the aminoacylation of the orthogonal tRNA was achieved using so called "flexizymes". Flexizymes are ribozymes with tRNA-amino-aclylation activity.[63]
Applications
With an expanded genetic code, the unnatural amino acid can be genetically directed to any chosen site in the protein of interest. The high efficiency and fidelity of this process allows a better control of the placement of the modification compared to modifying the protein post-translationally, which, in general, will target all amino acids of the same type, such as the thiol group of cysteine and the amino group of lysine.[64] Also, an expanded genetic code allows modifications to be carried out in vivo. The ability to site-specifically direct lab-synthesized chemical moieties into proteins allows many types of studies that would otherwise be extremely difficult, such as:
- Probing protein structure and function: By using amino acids with slightly different size such as O-methyltyrosine or dansylalanine instead of tyrosine, and by inserting genetically coded reporter moieties (color-changing and/or spin-active) into selected protein sites, chemical information about the protein's structure and function can be measured.
- Probing the role of post-translational modifications in protein structure and function: By using amino acids that mimic post-translational modifications such as phosphoserine, biologically active protein can be obtained, and the site-specific nature of the amino acid incorporation can lead to information on how the position, density, and distribution of protein phosphorylation effect protein function.[65][66][67][68]
- Identifying and regulating protein activity: By using photocaged aminoacids, protein function can be "switched" on or off by illuminating the organism.
- Changing the mode of action of a protein: One can start with the gene for a protein that binds a certain sequence of DNA and, by inserting a chemically active amino acid into the binding site, convert it to a protein that cuts the DNA rather than binding it.
- Improving immunogenicity and overcoming self-tolerance: By replacing strategically chosen tyrosines with p-nitro phenylalanine, a tolerated self-protein can be made immunogenic.[69]
- Selective destruction of selected cellular components: using an expanded genetic code, unnatural, destructive chemical moieties (sometimes called "chemical warheads") can be incorporated into proteins that target specific cellular components.[70]
- Producing better protein: the evolution of T7 bacteriophages on a non-evolving E. coli strain that encoded 3-iodotyrosine on the amber codon, resulted in a population fitter than wild-type thanks to the presence of iodotyrosine in its proteome[71]
- Probing protein localization and protein-protein interaction in bacteria.[72]
Future
The expansion of the genetic code is still in its infancy. Current methodology uses only one non-standard amino acid at the time, whereas ideally multiple could be used. In fact, the group of Jason Chin has recently broken the record for a genetically recoded E.coli strain that can simultaneously incorporate up to 4 unnatural amino acids.[73] Moreover, there has been development in software that allows combination of orthogonal ribosomes and unnatural tRNA/RS pairs in order to improve protein yield and fidelity.[73]
Recoded synthetic genome
One way to achieve the encoding of multiple unnatural amino acids is by synthesising a rewritten genome.[74] In 2010, at the cost of $40 million an organism, Mycoplasma laboratorium, was constructed that was controlled by a synthetic, but not recoded, genome.[75] The first genetically recoded organism was created by a collaboration between George Church's and Farren Isaacs' labs, when the wild type E.coli MG1655 was recoded in such a way that all 321 known stop codons (UAG) were substituted with synonymous UAA codons and release factor 1 was knocked out in order to eliminate the interaction with the exogenous stop codon and improve unnatural protein synthesis.[28] In 2019, Escherichia coli Syn61 was created, with a 4 megabase recoded genome consisting of only 61 codons instead of the natural 64.[3][2] In addition to the elimination of the usage of rare codons, the specificity of the system needs to be increased as many tRNA recognise several codons[74]
Expanded genetic alphabet
Another approach is to expand the number of nucleobases to increase the coding capacity.
An unnatural base pair (UBP) is a designed subunit (or
In 2012, a group of American scientists led by Floyd Romesberg, a chemical biologist at the
The successful incorporation of a third base pair into a living micro-organism is a significant breakthrough toward the goal of greatly expanding the number of amino acids which can be encoded by DNA, thereby expanding the potential for living organisms to produce novel proteins.[84] The artificial strings of DNA do not encode for anything yet, but scientists speculate they could be designed to manufacture new proteins which could have industrial or pharmaceutical uses.[86]
In May 2014, researchers announced that they had successfully introduced two new artificial
Related methods
Selective pressure incorporation (SPI) method for production of alloproteins
There have been many studies that have produced protein with non-standard amino acids, but they do not alter the genetic code. These protein, called
These studies rely on the natural
In this context, the SPI method generates recombinant protein variants or alloproteins directly by substitution of natural amino acids with unnatural counterparts.[99] An amino acid auxotrophic expression host is supplemented with an amino acid analog during target protein expression.[100] This approach avoids the pitfalls of suppression-based methods[101] and it is superior to it in terms of efficiency, reproducibility and an extremely simple experimental setup.[102] Numerous studies demonstrated how global substitution of canonical amino acids with various isosteric analogs caused minimal structural perturbations but dramatic changes in thermodynamic,[103] folding,[104] aggregation[105] spectral properties[106][107] and enzymatic activity.[108]
in vitro synthesis
The genetic code expansion described above is in vivo. An alternative is the change of coding in vitro translation experiments. This requires the depletion of all tRNAs and the selective reintroduction of certain aminoacylated-tRNAs, some chemically aminoacylated.[109]
Chemical synthesis
There are several techniques to produce peptides chemically, generally it is by solid-phase protection chemistry. This means that any (protected) amino acid can be added into the nascent sequence.[citation needed]
In November 2017, a team from the
See also
- Bioengineering
- Directed evolution
- Hachimoji DNA
- List of genetic codes
- Nucleic acid analogue
- Non-proteinogenic amino acids
- Protein labelling
- Protein methods
- Synthetic biology
- Xenobiology
References
- PMID 16260173.
- ^ a b Zimmer C (15 May 2019). "Scientists Created Bacteria With a Synthetic Genome. Is This Artificial Life? - In a milestone for synthetic biology, colonies of E. coli thrive with DNA constructed from scratch by humans, not nature". The New York Times. Retrieved 16 May 2019.
- ^ PMID 31092918.
- PMID 29030023.
- PMID 28671771.
- ^ S2CID 6702011.
- ^ ISBN 978-0-8153-4105-5.
- PMID 10704480.
- ^ PMID 12409460.
- ^ PMID 20307192.
- PMID 16785423.
- ^ PMID 24646179.
- PMID 9521119.
- PMID 16689635.
- PMID 20147747.
- ^ a b "The Peter G. Schultz Laboratory". Schultz.scripps.edu. Archived from the original on 2018-07-12. Retrieved 2015-05-05.
- PMID 16515468.
- PMID 12537491.
- ^ "context :: 21-amino-acid bacteria: expanding the genetic code". Straddle3.net. Retrieved 2015-05-05.
- PMID 19117371.
- ISBN 978-0-87969-453-1.
- PMID 2141650.
- doi:10.1021/ja000595y. Archived from the original(PDF) on 2011-09-27. Retrieved 2010-12-02.
- PMID 11866580.
- PMID 12154230.
- PMID 25378305.
- PMID 21764749.
- ^ PMID 24136966.
- PMID 25607366.
- PMID 25044341.
- PMID 24433543.
- PMID 34083482.
- ^ Atkins, J. F.; Bjoerk, G. R. "A gripping tale of ribosomal frameshifting: extragenic suppressors of frameshift mutations spotlight P-site realignment." Microbiol. Mol. Biol. Rev. 2009, 73, 178-210.
- ^ Anderson, J. C.; Wu, N.; Santoro, S. W.; Lakshman, V.; King, D. S.; Schultz, P. G. "An expanded genetic code with a functional quadruplet codon." Proc. Natl. Acad. Sci. U. S. A. 2004, 101, 7566-7571.
- ^ Neumann, H.; Wang, K.; Davis, L.; Garcia-Alai, M.; Chin, J. W. "Encoding multiple unnatural amino acids via evolution of a quadruplet-decoding ribosome." Nature 2010, 464, 441-444.
- ^ Niu, W.; Schultz, P. G.; Guo, J. "An expanded genetic code in mammalian cells with a functional quadruplet codon." ACS Chem. Biol. 2013, 8, 1640-1645.
- ^ S2CID 4390989.
- PMID 30262649.
- ^ Niu, W., Schultz, P. G., and Guo, J. (2013) An expanded genetic code in mammalian cells with a functional quadruplet codon. ACS Chem Biol 8, 1640-1645.
- PMID 34588441.
- ^ Chen, Y., Wan, Y., Wang, N., Yuan, Z., Niu, W., Li, Q., and Guo, J. (2018) Controlling the Replication of a Genomically Recoded HIV-1 with a Functional Quadruplet Codon in Mammalian Cells. ACS Synth. Biol. 7, 1612-1617.
- PMID 18397770.
- PMID 15138302.
- PMID 14627803.
- PMID 12911301.
- PMID 20925334.
- PMID 24375139.
- S2CID 2376187.
- ^ PMID 15521721.
- ^ PMID 11226228.
- PMID 17965709.
- S2CID 215757172.
- PMID 24139041.
- PMID 33826797.
- PMID 15187228.
- PMID 28220771.
- S2CID 37181098.
- S2CID 19683574.
- PMID 26465656.
- PMID 24907900.
- S2CID 46222849.
- PMID 382994.
- PMID 17884697.
- PMID 19318213.
- PMID 21868676.
- PMID 26350765.
- PMID 26350500.
- PMID 26030730.
- PMID 21768354.
- PMID 19555063.
- PMID 24487692.
- PMID 27775882.
- ^ S2CID 237271721.
- ^ S2CID 10152595.
- PMID 20488990.
- S2CID 22055476.
- S2CID 6494156.
- PMID 19073696.
- PMID 22121213.
- S2CID 23329867.
- PMID 22773812.
- ^ PMID 24805238.
- ^ Callaway E (May 7, 2014). "Scientists Create First Living Organism With 'Artificial' DNA". Nature News. Huffington Post. Retrieved 8 May 2014.
- ^ a b Fikes BJ (May 8, 2014). "Life engineered with expanded genetic code". San Diego Union Tribune. Archived from the original on 9 May 2014. Retrieved 8 May 2014.
- ^ Sample I (May 7, 2014). "First life forms to pass on artificial DNA engineered by US scientists". The Guardian. Retrieved 8 May 2014.
- ^ Pollack A (May 7, 2014). "Scientists Add Letters to DNA's Alphabet, Raising Hope and Fear". The New York Times. Retrieved 8 May 2014.
- ^ Pollack A (May 7, 2014). "Researchers Report Breakthrough in Creating Artificial Genetic Code". The New York Times. Retrieved May 7, 2014.
- S2CID 86967999. Retrieved May 7, 2014.
- ^ Amos J (8 May 2014). "Semi-synthetic bug extends 'life's alphabet'". BBC News. Retrieved 2014-05-09.
- PMID 3045813.
- PMID 24939187.
- PMID 17272838.
- PMID 15782218.
- S2CID 9439946.
- PMID 11514527.
- PMID 6413975.
- PMID 26136259. NIHMSID: NIHMS711205
- PMID 20391526.
- PMID 15578784.
- PMID 14662389.
- PMID 22493661.
- PMID 28085996.
- PMID 16782415.
- PMID 18301757.
- PMID 19416900.
- PMID 18854410.
- PMID 12729742. Archived from the originalon 2017-08-11. Retrieved 2017-08-11.
- S2CID 86352672.
- PMID 24959531.
- ^ 'Unnatural' microbe can make proteins. BBC News. 29 November 2017.
- ^ PMID 29189780.
- ^ Howgego J (February 2014). "On stranger nucleotides". Chemistry World.
- PMID 24152106.