Incorporation of unnatural amino acids for synthetic biology

Abstract

The challenge of synthetic biology lies in the construction of artificial cellular systems. This requires the development of modular “parts” that can be integrated into living systems to elicit an artificial, yet programmed, response or function. The development of methods to engineer proteins bearing unnatural amino acids (UAAs) provides essential components that may address this challenge. Here we review the emerging strategies for incorporating UAAs into proteins with the endgame of engineering artificial cells and organisms.

---

1. Introduction

Synthetic biology aims to engineer artificial biological systems with the ultimate goal of programming novel cell and organism behaviour.^1–5 This is being achieved via a bottom-up approach in which the key “parts” are nucleic acids, metabolites and proteins. With the advances in nucleic acid synthesis and recombinant DNA technology driving down the cost of constructing genes,^6,7 the fabrication of cells to stimulate pattern formation,⁸ produce drugs,⁹ target cancer cells¹⁰ and disperse biofilms¹¹ (among many others) have been realized. To meet the challenge inherent in expanding the function of biological systems, the development of orthogonal “parts” capable of functioning in the context of living matter is becoming increasingly important.^12,13

Proteins are essential components of synthetic cellular networks. Synthesis of cellular pathways with the ultimate goal of creating functional artificial cells can benefit greatly from the generation of proteins with desirable properties that can be effectively regulated. Proteins capable of functioning independently of the host organism’s endogenous circuitry require further developments in “parts” design.

The recent advances in reassigning the genetic code to unnatural amino acids (UAAs)^12,14–16 expand the molecular toolbox of protein “parts”. In fact, introduction of UAA can impart new functions that are difficult or impossible to create with proteins comprised of the natural 20 amino acid building blocks such as photo-induced switching,¹⁷IR probe active¹⁸ and redox sensitive¹⁹proteins. Moreover, proteins bearing UAAs may result in improved properties such as hyperstability^20,21 and protease resistance.²² Such features may be useful for tailoring orthogonal proteins in the context of synthetic biological networks. Here we discuss approaches that focus on protein engineering and that can potentially expand synthetic biology methods by UAA incorporation for the creation of artificial proteins.

2. Proteins in synthetic biology

Proteins play an important role in synthetic biology especially since the cellular pathways are mediated by protein–nucleic protein–protein and protein–small molecule interactions. The earliest example of synthetic genetic circuits relied on stitching various protein–DNA regulators in specified arrangements.^23,24 Elowitz and Leibler engineered a repressilator comprised of a series of 3 well-studied transcriptional repressors P_Llac01, P_Ltet01 and λP_R in a single construct that resulted in oscillations akin to a biological clock.²³ The Collins group also exploited the use of repressors and promoters to assemble a genetic toggle switch in which each promoter was inhibited by the repressor produced from the opposing promoter (Fig. 1A).²⁴ This allowed the use of inducers to control the switch. The role of proteins as regulatory elements in genetic networks designs remain important for oscillating systems and genetic toggle switch engineering (Fig. 1A).^23–25 Changes in half-life of proteins and cooperative regulatory protein–DNA as well as protein–protein (vida infra) interactions represent valuable aspects for tuning oscillations in synthetic genetic systems.²⁵

Fig. 1 Proteins in synthetic biology. (A) Repressilator²³ (left) is a synthetic network under transcriptional regulation where repressor protein, LacI, inhibits the transcription of tetR repressor gene. The product of the tetRgene, in turn, inhibits the transcription of λcIgene. To complete the cycle, λCI inhibits lacI expression. TetR also inhibits transcription of the gfp reporter gene. Oscillations of the repressilator network are observed in E. coli cells by monitoring GFP fluorescence. Genetic toggle switch²⁴ (right) is constructed of two repressors and two constitutive promoters, resulting in oscillations produced by transient inducers. In the network, promoters are inhibited by repressors that are under transcriptional control of the opposing promoter, i.e. Rep 1 inhibits expression of repressor 2gene. (B) Non-native protein–protein interactions (top)²⁸ are designed to rewire cellular pathway. Ste5 mutant contains mutations that destroy recruitment of Ste7 kinase resulting in non-functional mating pathway in yeast. Replacement of defective recruitment interaction with a designed heterologous protein–protein interactions rewires the signal transduction pathway and alleviates mating deficiency. A single input protein switch²⁶ (bottom) is turned on and off by introduction of competitive ligand (CL). In the “on” position, N-WASP domain is able to carry out actin polymerization. (C) The RBP receptor is computationally designed to bind non-biological ligand (TNT)³² and initializes signal transduction pathway through interaction with transmembrane histidine kinase, Trz, resulting in autophosphorylation and phosphorylation of transcription regulator OmpR. Phosphorylated OmpR activates the lacZ reporter gene expression. (D) Signalling metabolite acetyl phosphate is used to regulate an oscillating synthetic circuit.³³ The circuit is comprised of two enzymes that are under the transcriptional control of acetyl phosphate and two metabolites (acetyl CoA and acetyl phosphate) to generate oscillations.

In addition to protein–nucleic acid interactions, protein–protein interactions have also been manipulated to re-wire signal transduction cascades.^26–31 As the first example of engineering protein–protein networks, Lim et al. inserted a heterologous protein interaction within the yeast MAP kinase pathway and demonstrated that simple tethering can produce the appropriate function or output.²⁸ Known mutations that disabled the kinase function in the mating pathway in Saccharomyces cerevisiae were rescued by introduction of alternative interactions in the scaffold. Most remarkably, a synthetic pathway was designed where native interactions were destroyed by mutagenesis and non-native interactions were established through tethering a different kinase set, allowing for a non-natural input-output scheme (Fig. 1B).²⁸ This confirmed that proteins could form functional pathways that were “modular and flexible”.²⁸ Moreover, using the actin regulatory protein N-WASP as the model system, Lim and coworkers designed a synthetic switch where the output of N-WASP was essentially spliced into a PDZ domain-ligand pair.²⁶ This resulted in the switch being in the “off” state and upon addition of PDZ ligand, the switch was turned “on” (Fig. 1B). To further expand the ability of these switches to process multiple inputs, a combinatorial library was synthesized.

Four parameters were varied: (1) domain type; (2) domain–ligand interaction; (3) linker length and (4) domain architecture.²⁶ The library resulted in synthetic switches with a variety of novel gating behaviours to non-physiological inputs. In order to engineer artificial pathways and functions like the ones described,^26–31,34 a method to systematically identify common protein motifs and protein–protein interactions is needed. Programs and databases are being developed to recognize, classify, and organize productive protein–protein interactions with the overall objective for construction and reprogramming of intracellular interaction for synthetic biosystems.^35–37

Signal transduction cascades as well as metabolic pathways often rely on protein–small molecule interactions. Key elements of molecular recognition for effective receptor–ligand interaction can be extracted for building highly specific cellular pathways. An initial example of engineering proteins for new ligands focused on employing computational methods to alter the specificities of periplasmic binding protein (PBP) receptors that were experimentally tested for binding artificial ligands (Fig. 1C).³² When the re-engineered receptors for trinitrotoluene (TNT) and seratonin were integrated into a 2-component synthetic signal transduction pathway with a reporter gene in Escherichia coli, expected dose response signals were observed.³²

In the case of metabolic pathway engineering,^9,33,38 the small molecule and how it interacts with various proteins in a pathway are critical. The Liao group designed an oscillatory circuit that coupled metabolite inputs and outputs with transcriptional regulation.³³ The metabolite acetyl coenzyme A (AcCoA) was designed to be converted into acetyl phosphate (AcP) by phosphate acetyltransferase (Pta) and acetate by acetate kinase (Ack). The enzyme acetyl coenzyme A synthetase (Acs) could then process acetate back into AcCoA (Fig. 1D). This two-metabolite system was connected to transcription by engineering the acsgene under the glnAp2 promoter, which could be activated by AcP. The glnAp2 promoter was also engineered to control the lac repressor to repress the ptagene.³³ The oscillatory circuit demonstrated the importance of small molecules in the network connectivity, in addition to protein–protein and protein–DNA interactions within the network.

Because of the integral role proteins play in synthetic biology as described above, protein engineering methods are increasingly in demand for the design of transcription networks,^39–41 novel input-output signal transduction cascades,^27,42 and metabolic pathways.^38,43 As the applications for synthetic biology become focused on generating an “artificial” or “non-natural” function or product, approaches that can enhance orthogonality within the system will become vital. Protein modifications such as UAA incorporation provides a valuable mode of integrating orthogonality and may enable tuning of a complex circuit beyond the constrains of nature.⁴⁴

3. Existing cellular pathways of unnatural amino acid incorporation

3.1 Selenocysteine and pyrrolysine

Until very recently, the standard genetic code was thought to encode for a set of 20 amino acids. Two amino acids that have always existed in natural organisms, selenocysteine and pyrrolysine, have been the latest additions to this set.⁴⁵Selenocysteine is genetically encoded in prokaryotes and eukaryotes by the UGA stop codon.⁴⁶ A selenocystenyl-tRNA^[Ser]Sec is aminoacylated with serine by the seryl-tRNA synthetase and serine is converted into selenocysteine by selenocysteine synthase (SelA) in the presence of a selenium monophosphate donor provided by SelD^46,47 (Fig. 2). Incorporation of selenocysteine depends on the existence of a selenocysteine insertion sequence (SECIS) adjacent to the UGA codon in prokaryotes or in the 3′-untranslated region of the selenoprotein mRNA in eukaryotes.⁴⁶ The presence of the SECIS element allows suppression of UGA and cotranslational selenocysteine incorporation. In addition, selenocysteine incorporation requires the presence of several specialized factors, such as selenocysteinyl tRNA^[Ser]Sec-specific elongation factor (SelB) in prokaryotes.⁴⁶ The EFsec (eSelB), selenocysteinyl tRNA^[Ser]Sec-specific elongation factor and the SECIS binding protein (SBP2) together form a functional equivalent of SelB in eukaryotes.⁴⁶ An additional factor, L30 ribosomal protein (rpL30), enhances Sec incorporation into proteins in eukaryotes.⁴⁶

Fig. 2 Cotranslational insertion mechanism of selenocysteine and pyrrolysine.^46,47 (A) An opal suppressor tRNA is charged with serine by seryl-tRNA synthetase. In the presence of selenium monophosphate, serine is converted into selenocysteine in a reaction catalyzed by selenocysteine synthase (SelA). SelB mediates selenocysteine insertion at the UGA codon adjacent to the SECIS element on the mRNA. (B) Pyrrolysine is charged to amber suppressor tRNA by pyrrolysyl-tRNA synthetase and inserted at the amber codon next to the PYLIS element. EF-Tu aids in the cotranslational insertion.

Pyrrolysine is determined to be genetically encoded by Methanosarcinaceae by the UAG stop codon in the methylamine methyltransferasegenes required for methylamine metabolism as well as the trimethylamine methyltransferase homolog from the gram-positive Desulfitobacterium hafniense.^48,49 Similar to selenocysteine, pyrrolysine incorporation is reported to depend on a specific structural element, termed pyrrolysine insertion sequence (PYLIS) in the mRNA.⁴⁸ Whereas selenocysteine is synthesized on the tRNA, pyrrolysine is directly charged to tRNA^Pyl (pylT) by the designated pyrrolysyl-tRNA synthetase (PylS) (Fig. 2).⁴⁸ Elongation factor EF-Tu binds tRNA^Pyl and is able to decode UAG as pyrrolysine as long as the pylT and pylSgenes are also provided.⁴⁸ Because both selenocysteine and pyrrolysine utilize a stop codon for incorporation, it has been a challenge to identify these amino acids until recently.

4. Unnatural amino acid incorporation methods

4.1 Site-specific incorporation

Methods designed specifically for UAA incorporation in vitro and in vivo rely on engineering changes into the cellular protein translation pathways. Based on our understanding of the selenocysteine and pyrrolysine pathways, the use of stop codonvia nonsense suppression appears to be a suitable strategy that can be applied to other UAAs. In particular, UAA incorporation by this method employs: (i) introduction of a stop codon at the position of interest into mRNA; (ii) isolation of a suppressor tRNA that recognizes the stop codon; and (iii) selection of aminoacyl-tRNA synthetase (AARS) to charge the suppressor tRNA with the particular analog. This would allow for the site-specific incorporation of amino acid analogs.

Although the first two steps can be achieved by site-directed mutagenesis and use of suppressor tRNAs, the aminoacylation of the tRNA with UAA requires more effort. As an alternative to employing the AARS, tRNA aminoacylation with UAA can be accomplished with T4 RNA ligasein vitro.⁵⁰ The UAA can be aminoacylated chemically upon reaction with pdCpA dinucleotide and then ligated to the truncated supressor tRNA lacking pCpA dinucleotide. Such aminoacylated tRNAs have been initially used in a cell-free translation mixture to incorporate UAA.⁵¹In vitro translation systems use cellular extracts that contain all the required components for efficient translation. The site of UAA incorporation is based on a readthrough of an amber (UAG) codon in mRNA by a suppressor tRNA aminoacylated with the analog. From initial in vitro experiments, it was recognized that the suppressor tRNA should not be charged by the endogenous AARSs from the cellular extracts since this would result in reduced efficiency of UAA incorporation. Thus, S. cerevisiae tRNA_CUA^Phe is utilized in the in vitro E. coli expression system because S. cerevisiaetRNAs are orthogonal and not recognized by E. coli AARS.⁵¹ Although this allows for efficient incorporation of UAA, the chemically aminoacylated tRNA is not regenerated, limiting the amount of protein that can be produced.

4.1.1 In vivo UAA incorporation in Escherichia coli

Engineered AARS/suppressor tRNA pairs have been developed in which the AARS is optimized to charge suppressor tRNA with a particular UAA.⁵² Efficient enzymatic charging of the tRNA requires optimization or engineering of the AARS for specific recognition of a particular UAA. Initial experiments relied on using the nonessential suppressor tRNA_CUA^Phe/phenylalanyl-tRNA synthetase pair from S. cerevisiae (SctRNA_CUA^Phe/PheRS) by Furter for site-specific incorporation of p-fluorophenylalanine in E. coli.⁵³ Since then, tremendous progress had been made in improving the AARS/tRNA pair so that they are completely non-cross reactive and orthogonal to the endogenous AARS/tRNA pairs within the cellular translational machinery. The Schultz group used tRNA_CUA^Tyr/tyrosyl-tRNA synthetase pair from Methanococcus jannaschii (MjtRNA_CUA^Tyr/TyrRS) for UAA incorporation in E. coli.^18,54–61 Orthogonality of the pair was established through rounds of negative and positive screening of the MjtRNA_CUA^Tyr mutant library^15,61 (Fig. 3A). Suppressor tRNA variants in which eleven nucleotides of MjtRNA_CUA^Tyr, that did not directly interact with MjTyrRS, were targeted by mutagenesis. For negative selection, the suppressor MjtRNA_CUA^Tyr library was co-transformed into E. coli with a plasmid encoding the barnase gene bearing TAG codons at permissive sites within the gene. Cell death occurred when the suppressor MjtRNA_CUA^Tyr was recognized and charged by endogenous E. coli AARS, resulting in barnase production. Truncated nonfunctional barnase was produced due to inability of endogenous AARSs to charge MjtRNA_CUA^Tyr, yielding orthogonal or nonfunctional tRNAs. For positive selection, a plasmid containing the MjtRNA_CUA^Tyrgene was purified from cells that survived negative selection and transformed into E. coli already bearing a plasmid encoding the cognate MjTyrRS and β-lactamase with an in-frame TAG codon. The transformed cells were grown on ampicillin allowing selection of survivors with β-lactamase transcribed, indicating successful charging of MjtRNA_CUA^Tyr by cognate MjTyrRS. Cells died when cognate MjTyrRS could not charge MjtRNA_CUA^Tyr or suppressor MjtRNA_CUA^Tyr was not functional.

Fig. 3 Selection strategies of orthogonal AARS/tRNA pairs for in vivo, site-specific incorporation using nonsense codons. (A) Negative and positive selection using barnase and β-lactamase as reporter genes for orthogonal amber suppressor MjtRNA in E. coli.^15,61 When MjtRNA is charged by endogenous AARS, barnase is produced and cells die. Orthogonal MjtRNA cannot be recognized by cellular AARS and cells survive. Cognate AARS was identified by ability of cells to grow on ampicillin. (B) Positive and negative selection scheme for MjTyrRS specific for UAA. The selected mutRNA_CUA^Tyr as shown in (A) is used for the following selections.⁵⁴MjTyrRS library variants capable of activating UAA survive on chloramphenicol in presence of UAA, but cannot survive in media containing no UAA due to read through of the catgene.

MjTyrRS variants that are able to charge suppressor MjtRNA_CUA^Tyr with a particular UAA were identified from a library.^{18,54,57–61} Five residues in the active site of the MjTyrRS were randomized via saturation mutagenesis to generate a library of 1.6 × 10⁹ variants.^15,61 The residues were identified from the crystal structure of a TyrRS homolog from Bacillus stearothermophilus. A MjTyrRS variant capable of charging the suppressor MjtRNA_CUA^Tyr with a selected UAA with high fidelity was identified through several rounds of positive and negative selection (Fig. 3B).

As a positive selection, the MjTyrRS library was transformed into E. coli bearing a plasmid with a gene for the orthogonal suppressor MjtRNA_CUA^Tyr and chloramphenicol acetyltransferase (CAT) with an in-frame TAG codon. Surviving cells grown on chloramphenicol in the presence of UAA indicated that MjTyrRS was able to charge MjtRNA_CUA^Tyr with UAA or natural amino acids. This positive selection was followed by negative selection, where plasmid containing MjTyrRS variants from survived cells was transformed into E. coli bearing a plasmid encoding barnase with several in-frame nonsense codons. Cells were grown in the media depleted of UAA. Cells expressing MjTyrRS variants capable of charging suppressor MjtRNA_CUA^Tyr only with UAA survived because barnase was not produced. Genes for MjTyrRS variants that charged MjtRNA_CUA^Tyr with UAA could be isolated and serve as a base for a new library created by DNA shuffling. The cycle of positive and negative selections could be repeated several times to identify the most effective and efficient MjTyrRS variants for UAA incorporation. Another selection method for synthetase specificity involved combination of chloramphenicol resistance and a fluorescent reporter allowing use of a single E. coli strain for positive and negative screens.^54–55,63 The MjtRNA_CUA^Tyr/TyrRS pair was used to incorporate a variety of UAA into proteins in E. coli^{15,18,54–60,63–65} (Table 1). In addition, other orthogonal tRNA/AARS pairs were developed for in vivo incorporation in E. coli^66,67 (Table 1). Recently, a single plasmid selection system for the evolution of AARSs was reported by Schultz and coworkers.⁶⁸ The fusion genechloramphenicol acetyltransferase (cat) and uracil phosphoribosyltransferase (uprt) was used as a positive/negative selection marker.⁶⁸ Cells that were able to incorporate UAA in response to the amber codon were selected for in the presence of chloramphenicol and cells that harbored AARS variants that incorporated endogenous amino acids were selected against by adding 5-fluorouracil.⁶⁸ The creation of orthogonal pairs further expands the molecular toolbox of UAAs that can be cotranslationally inserted into proteins.

Table 1 Orthogonal amber suppressor tRNA/AARS pairs used for site- specific UAA incorporation in vivo

Organism	tRNA/AARS pair	Derived from	Analog
E. coli	tRNACUA^Phe/PheRS	S. cerevisiae	p-fluorophenylalanine ^53
	tRNACUA^Tyr/TyrRS	M. jannaschii	p-aminophenylalanine ^55,60,63
			naphthylalanine ^55
			p-benzoylphenylalanine ^55,65
			p-methyoxyphenylalanine^55
			p-aminomethylphenylalanine ^55
			p-methylphenylalanine ^55
			p-trifluoromethylphenylalanine^55,56
			p-nitrophenylalanine ^55
			p-isopropylphenylalanine ^63
			O-allyl-tyrosine ^63
			p-acetylphenylalanine ^57,65
			p-azidophenylalanine ^54,65
			p-iodophenylalanine ^58,65,68
			β-N-acetylglucosamine-serine^59
			p-cyanophenylalanine ^18
			p-hydroxyphenyllactic acid ^64
			O-methyltyrosine^74
			phenyl-selenocysteine ^75
			(2,2′-bipyridin-5-yl)alanine^76,77
			(8-hydroxyquinolin-3-yl)alanine ^78
			p-iodophenylalanine ^68
			2-amino-3-(8-hydroxyquinolin-3-yl)propanoic acid ^78
	tRNACUA^Pyl/PylRS	M. maize	N ^ε-cyclopentyloxycarbonyl-L-lysine^79
			O-nitrobenzyl-oxycarbonyl-N^ε-L-lysine^79
	tRNAAAA^Phe/PheRS T415G	S. cerevisiae	3-(2-naphthyl)alanine ^80
	tRNACUA^Phe/PheRS T415G	S. cerevisiae	p-bromophenylalanine ^67
			benzothienylalanine^66
			6-bromotryptophan ^66
			6-chlorotryptophan ^66
	tRNACUA^Pyl/PylRS	M. barkeri	N ^ε-acetyllysine^81
S. cerevisiae	tRNACUA^Tyr/TyrRS	E. coli	p-benzoylphenylalanine ^62,69
			p-azidophenylalanine ^62,69
			p-acetylphenylalanine ^62,69
			p-propargyloxyphenylalanine^69
			O-methyltyrosine^62
			p-iodotyrosine ^62
			4,5-dimehtoxy-2-nitrobenzylserine^62
	tRNACUA^Leu/LeuRS	E. coli	methionine and cysteine homologs^82
			2-aminocaprylic acid^82
			2-aminononanoic acid ^82
			2-aminodecanoic acid ^82
P. pastoris	tRNACUA^Tyr/TyrRS	E. coli	p-benzoylphenylalanine ^83
			p-azidophenylalanine ^83
			p-(propargyloxy)-phenylalanine^83
			p-methoxyphenylalanine^83
			p-iodophenylalanine ^83
	tRNACUA^Leu/LeuRS	E. coli	4,5-dimethoxy-2-nitrobenzylserine ^83
			dansylalanine^83
Mammalian cells	tRNACUA^Tyr/TyrRS	B. stearothermophilus	3-iodotyrosine ^70
		E. coli	p-acetylphenylalanine ^73
			p-benzoylphenylalanine ^73
	tRNACUA^Trp/TrpRS	B. subtilis	5-hydroxytryptophan ^71
	tRNACUA^Tyr/TyrRS	E. coli	O-methyltyrosine^72
			p-benzoylphenylalanine ^72
	tRNACUA^Leu/LeuRS	E. coli	2-amino-3-(5-(dimethylamino)naphthalene-1-sulfoamido)
			propanoic acid ^72
	tRNACUA^Pyl/PylRS	M. maize	o-nitrobenzyl-oxycarbonyl-N^ε-L-lysine^79
			N ^ε-tert-butyloxycarbonyl-L-lysine^84
			N ^ε-benzyloxycarbonyl-L-lysine^84
			N ^ε-acetyllysine^84

---

4.1.2 In vivo UAA incorporation in Saccharomyces cerevisiae. Employing a similar strategy and integrating exogenous components to complement the existing translation pathway,⁶² site-specific UAA incorporation was achieved in yeast by Schultz et al. The orthogonal tRNA_CUA^Tyr/TyrRS and tRNA_CUA^Leu/LeuRS pairs from E. coli^62,69 were shown to function efficiently in S. cerevisiae. A library of EcTyrRS variants, created on the basis of known structure of a structural homolog from B. stearothermophilus, was subjected to several rounds of negative and positive selection for its ability to charge EctRNA_CUA^Tyr with UAA of interest in yeast^62,69 (Fig. 4). A S. cerevisiaehistidine and uracil auxotrophic strain was employed to express the transcriptional activatorprotein GAL4 in which two codons were substituted by amber stop codons. For positive selection, cells were grown in media containing UAA and 3-aminotriazole (3-AT) or depleted of uracil (URA) and histidine. Successful suppression of TAG codons in GAL4 produced functional protein that allowed transcription of reporter genesHIS3, URA3, and lacZ leading to cell survival. The presence of the lacZgene allowed the colorimetric evaluation of selection results in the presence of X-gal. When GAL4 was not expressed, cells died in presence of 3-AT, a competitive inhibitor of the HIS3 promoter, without uracil and histidine. Negative selection eliminated EcTyrRS variants that charged EctRNA_CUA^Tyr with natural amino acid. Since 5-fluoroorotic acid (5-FOA) is converted to toxic product by URA3, growing cells in media depleted of UAA and containing 5-FOA, resulted in cell death if natural amino acid was utilized by the EcTyrRS. In contrast, cells that expressed EcTyrRS variants incorporating only UAA survived (Fig. 4). By this approach p-benzoylphenylalanine, p-azidophenylalanine, p-acetylphenylalanine, p-propargyloxyphenylalanine, O-methyltyrosine and p-iodotyrosine were incorporated in yeast (Table 1).^62,69

Fig. 4 Positive and negative selection of orthogonal EcTyrRS/mutRNA_CUA^Tyr pair specific for UAA in S. cerevisiae.⁶² Only S. cerevisiae cells carrying genes for orthogonal EcTyrRS/mutRNA_CUA^Tyr pair that mediate UAA incorporation can survive on 3-AT or without uracil when grown in presence of UAA. In the absence of UAA, cells that contain active EcTyrRS/mutRNA_CUA^Tyr pair cannot survive in presence of 5-FOA. When GAL4 amber mutations are suppressed with natural amino acids, URA3 is expressed and cells are converting 5-FOA to a toxic product.⁶²

4.1.3 In vivo UAA incorporation in mammalian cells. Recently cotranslational incorporation of UAA in mammalian cells was reported (Table 1).^70–73 The earliest example from the Yokoyama group employed the Bacillus stearothermophilus amber suppressor tRNA_CUA^Tyr and E. colityrosyl-tRNA synthetase variant pair, previously used in a wheatgerm cell-free translation system. This pair was used to incorporate 3-iodotyrosine into test protein in CHO cells where stable TyrRS expression was optimized with 95% fidelity.⁷⁰ Subsequently, Schultz and coworkers demonstrated the incorporation of 5-hydroxytryptophan (5-HTPP) into proteins expressed in HK293T cells using Bacillus subtilistryptophanyl opal suppressor tRNA (mutRNA_UCA^Trp) and corresponding synthetase (BsTrpRS) engineered for 5-HTPP incorporation.⁷¹ The site for mutation of the BsTrpRS was determined by analysis of the active site structure of a homologous synthetase from B. stearothermophilus. One of the 19 mutants tested, V144P BsTrpRS selectively aminoacylated mutRNA_UCA^Trp with 5′-HTPP. In both cases, a structure-based approach was used for identifying orthogonal AARS variants capable of accepting the UAAs.

Exploiting the orthogonal pairs from previous selections, the Wang group utilized the EctRNA_CUA^Tyr/TyrRS pair selected from yeast.⁷² Several expression constructs were tested in HeLa cell lines . In order to overcome the challenge of expressing sufficient quantities of functional tRNAs, the H1 promoter was utilized to drive the functional biosynthesis of EctRNA_CUA^Tyr. Moreover, the 3′-flanking sequence of human tRNA^fMet was also needed for efficient expression of the EctRNA_CUA^Tyr in mammalian cells. The genes for EctRNA_CUA^Tyr and EcTyrRS was engineered on a single tRNA/AARS expression plasmid. This plasmid was coexpressed with another plasmid containing the GFP gene with an in-frame TAG codon. Those EctRNA_CUA^Tyr that were expressed and correctly processed was aminoacylated, yielding a fluorescent cell through suppression of the TAG codon in GFP. Orthogonality of EctRNA_CUA^Tyr was tested by removing EcTyrRS gene from the expression construct. Transfection of the resultant plasmid into HeLa cells did not change fluorescent intensity, indicating inability of the endogenous synthetases to aminoacylate the EctRNA_CUA^Tyr. In addition, they proved that the H1 promoter could efficiently express E. colitRNAs by demonstrating fluorescence using a different orthogonal EctRNA_CUA^Leu/EcLeuRS pair. For UAA incorporation, the EcTyrRS gene in the tRNA/AARS expression plasmid were replaced by the OmeTyrRS gene, a synthetase specific for O-methyltyrosine (OmeTyr) previously evolved in yeast.⁷²

Successful suppresssion of TAG in GFP was demonstrated in HeLa cells grown in media containing UAA, with a 41% incorporation efficiency. Expression of BraRS, a synthetase specific for p-benzoylphenylalanine (Bra) evolved in yeast, led to 13% incorporation of the UAA. EctRNA_CUA^Leu and mutant synthetase specific for 2-amino-3-(5-(dimethylamino)-naphthalene-1-sulfonamido)propanoic acid (DanAla) were expressed in HeLa cells with a 13% incorporation efficiency. Similarly, incorporation of OmeTyr and Bra was achieved in mouse hippocampal neurons. While OmeTyr was incorporated into position 19 of the inactivation peptide of potassium channel Kv1.4 expressed in HEK293T cells using the EctRNA_CUA^Tyr/OmeTyrRS pair, DanAla was incorporated using the EctRNA_CUA^Leu/DanAlaRS pair. The presence of different sized UAAs allowed insight into the inactivation mechanism of Kv1.4.⁷² RajBhandary, Sakmar and coworkers employed amber suppressor tRNA derived from B. stearothermophilus tRNA_CUA^Tyr, and E. coli TyrRS variants developed previously to incorporate p-acetylphenylalanine (Acp) and p-benzoylphenylalanine (Bzp) into CCR5 and rhodopsin in HEK293T cells.⁷³ In this case, the firefly luciferase reporter was used to screen for orthogonality and specificity of UAA incorporation. The luciferase reporter contained a TAG codon at position Y70, which in the presence of the appropriate orthogonal tRNA/TyrRS pair and UAA produced full length luciferase.

4.1.4 Microinjection methods for UAA incorporation. Synthetically aminoacylated tRNA and mutant mRNA have been microinjected into Xenopusoocytes for UAA-labeled protein expression in eukaryotes^85–88 (Fig. 5A, Table 2). Suppressor tRNA for oocytes expression system was prepared based on the yeast suppressor tRNA_CUA^Phe bearing G20U mutation in the D stem loop and A73G mutation at the acceptor stem to minimize aminoacylation by endogenous XenopustRNAsynthetases .⁸⁶ The suppressor tRNA chemically acylated with 3-N-(7-nitrobenz-2-oxa-1,3-diazol-4-yl)-2,3-diaminopropionic acid (NBD) was injected into oocytes along with an mRNA encoding neurokinin-2 receptor (NK2) containing stop codons at two positions: 103 and 248 (Fig. 5A). Suppression efficiency was calculated to be 18% for this system.⁸⁶ Functionally active NK2 with fluorescent NBD was produced, allowing the investigation of ligand-receptor interactions. Nonsense and amber suppressor tRNAs from Tetrahymena thermophila G73 (THG73) were used to incorporate a variety of tyrosine and phenylalanine analogs for studying the role of tyrosine residues in binding and gating of the 5-hydroxytryptamine 3 receptor by expression in oocytes.⁸⁷ Additional improvements, such as the acceptor stem mutations on THG73, showed reduced aminoacylation by endogenous AARSs allowing for the production of highly orthogonal amber suppressor tRNAs.^90,91T. thermophilaglutamine amber and opal tRNA suppressors were developed and demonstrated suppression efficiency roughly 50% relative to those of THG73 in oocytes.^90,91

Fig. 5 Alternative approaches for site-specific incorporation. (A) UAA incorporation through nonsense suppression in Xenopusoocytes.⁸⁸ Designed tRNA with a chemically ligated UAA is introduced into Xenopusoocytes by microinjection and translates the mRNA bearing the stop codon. (B) Frameshift method for site-specific incorporation of UAA using four-base codons in cell-free expression systems. When CGGG is translated by UAA-charged tRNA, full length protein is produced with the UAA inserted at the CGGG site.⁸⁹ When CGGG is read as CGG by endogenous tRNAprotein synthesis is terminated.

Table 2 UAA incorporation in Xenopusoocytes and in cell-free expression systems

Expression system	tRNA	Analog
Xenopus oocytes	S. cerevisiae tRNA CUA ^Phe	3-N-(7-nitrobenz-2-oxa-1,3-diazol-4-yl)2,3-diaminopropionic acid^86
	T. thermophila tRNACUA	over 100^88
	T. thermophila tRNACUA^Gln	5-fluorotryptophan ^90
	T. thermophila tRNAACU^Gln	5-fluorotryptophan ^85
	S. cerevisiae tRNACCCG^Phe	5-fluorotryptophan ^85
	S. cerevisiae tRNAACCC^Phe	α-aminobutyric acid^85
		norvaline ^85
E. coli	Archae tRNAAGGA^Lys/LysRS	L-homoglutamine^93
	Archae tRNAAGGA^Lys/LysRS and Mj tRNACUA^Tyr/TyrRS	L-homoglutamine and O-methyltyrosine^93
E. coli cell-free expression system	S. cerevisiae tRNAACCU^Phe	p-biphenylalanine^94
	S. cerevisiae tRNAACCC^Phe	p-benzoylphenylalanine ^94
	S. cerevisiae tRNAACCG^Phe	p-phenylazophenylalanine^94
		p-naphthylalanine ^94
	S. cerevisiae tRNACCCG^Phe	BODIPY FL p-aminophenylalanine^95
	S. cerevisiae tRNAUACCG^Phe	p-nitrophenylalanine ^96
		ε-nitrobenzoxadiazolyllysine^96
Eukaryotic rabbit reticulocyte cell-free expression system	S. cerevisiae tRNAACCG^Phe	p-nitrophenylalanine ^97
	S. cerevisiae tRNAAGCG^Phe	p-nitrophenylalanine ^97
	S. cerevisiae tRNAAGGG^Phe	p-nitrophenylalanine ^97
	S. cerevisiae tRNAAGAG^Phe	p-nitrophenylalanine ^97
	S. cerevisiae tRNAAUAG^Phe	p-nitrophenylalanine ^97
	S. cerevisiae tRNAACCC^Phe	p-nitrophenylalanine ^97
Sf21 insect cell-free expression system	M. acetivorans tRNA^PylCCCG or	p-nitrophenylalanine ^98
	M. acetivorans tRNAPyl^UCCG	1-naphthylalanine ^98
		2-naphthylalanine ^98

---

4.1.5 Frameshift codons. Competition from release factors reduces the efficiency of UAA incorporation into a protein at the designated stop codon site.⁹² To eliminate this problem, investigations to use alternative codons distinct from stop codons have been undertaken. First implemented by Sisido and coworkers in vitro, frameshift suppression uses tRNAs with four-base anticodons that deliver the UAA in response to quadruplet codons⁸⁹ (Fig. 5B, Table 2). In cell-free expression systems, a chemically aminoacylated tRNA and mRNA with a four-base codon at specific position are added to cellular extracts containing all the required protein expression components. In vitro frameshift UAA incorporation has been demonstrated in E. coli S30, rabbit reticulocyte, and Spodoptera frugiperda 21(Sf21) cell-free translation systems.^89,94,97,98 Four-base codons derived from rarely used codons (CGG and GGG) are especially useful in the E. coli-based in vitro system because they efficiently compete with three-base decoding by endogenous tRNAs⁹⁵ (Table 3). Frameshift suppression is not limited to four-base codons as exemplified by the incorporation of p-nitrophenylalanine and ε-nitrobenzoxadiazolyl-L-lysine into streptavidin at Y54 in response to the five-base CGGUA codon in E. coli cell-free expression system.⁹⁶

Table 3 Residue specific UAA incorporation

Cells	AARS	Analog
E. coli	ProRS	per-thiaproline^113
		3-fluoroproline ^115
		4-fluoroproline ^114
		3,4-dehydroproline^114
		4-hydroxyproline ^114
		4,4-difluoroproline ^114
		azetidine-2-carboxylic acid ^114
	TyrRS	6-fluorotryptophan ^124,125
		5-hydroxytryptophan ^125
		7-azatryptophan^125
		4-fluorotryptophan ^122
		4-aminotryptophan ^123
		5-aminotryptophan ^123
	MetRS	selenomethionine ^106,107
		azidohomoalanine ^120
		homopropargylglycine^127
		2-amino-5-hexenoic acid ^110
		2-amino-5-hexynoic acid ^110,111
		difluoromethionine ^108
		trifluoromethionine ^109
		trans-crotylglycine^134,135
		cis-crotylglycine^134,135
		2-aminoheptanoic acid ^134,135
		norvaline ^134,135
		norleucine ^112
		methoxinine^112
		2-butynylglycine ^134,135
		allylglycine ^134,135
	PheRS	o-fluorophenylalanine ^122
		m-fluorophenylalanine ^122
		p-fluorophenylalanine ^122
	IleRS	trifluorovaline ^116,119
		trifluoroisoleucine^116,117
		2-amino-3-methyl-4-pentenoic acid ^118
	ValRS	trifluorovaline ^119
	LeuRS	trifluoroleucine^127,142
		hexafluoroleucine^21
	ProRS C443G	piperidine-2-carboxylic acid ^113
	MetRS L13G	azidonorleucine ^103
	MetRS L13S, Y260L, H301L	6,6,6-trifluoronorleucine ^104
	PheRS A294G	p-iodophenylalanine ^140
		p-cyanophenylalanine ^140
		p-ethynylphenylalanine ^127,140
		p-azidophenylalanine ^140,141
		2-, 3-, 4-pyridylalanine^140
		p-chlorophenylalanine ^138
		p-bromophenylalanine ^139,140
	PheRS T251G, A294G	p-acetylphenylalanine ^143
	ValRS T792P	aminobutyrate ^144
	LeuRS T252Y	norvaline ^145
		norleucine ^145
		allylglycine ^145
		homoallylglycine^145
		homoporpargylglucine^145
		2-butynylglycine ^145
Mammalian cells	MetRS	azidohomoalanine ^130
		homopropargylglycine^131
		photomethionine^132
	LeuRS	photoleucine^132

---

A selection system for efficient frameshift suppressor tRNAs with high decoding activity was developed recently.⁹⁵ Synthetic tRNAs were ligated with 5′-phosphodeoxycytidyl-phospho-puromycin (pdCp-puromycin) by T4 RNA ligase. These puromycin-tRNAs were added to the cell-free translation system with streptavidin-linked mRNA that contained the CGGG codon at the C-terminus. Puromycin-tRNAs that bound to the ribosome produced steptavidin-tRNA fusions that could be released from the ribosome complex through reaction with EDTA and recovered with biotin-covered beads. To create a DNA library that encoded functional tRNAs, recovered steptavidin-tRNAs were subjected to RT-PCR. Several rounds of selection were performed through reaction with EDTA and recovered with biotin-covered beads. To create a DNA library that encoded functional tRNAs, recovered steptavidin-tRNAs were subjected to RT-PCR. Several rounds of selection were carried out to identify frameshift tRNA_CGGG with high efficient decoding activity in E. coli cell-free expression system.⁸⁵ Frameshift suppression was also used in vivo to introduce L-homoglutamine in response to the four-base codon AGGA in E. coli.⁹³ The orthogonal LysRS/tRNA pair derived from Archaea was used in the system.⁹³ Simultaneous incorporation of two amino acids (L-homoglutamine and O-methyltyrosine) at two different positions in myoglobin was demonstrated in vivo in E. coli using both the developed frameshift suppression (Archae tRNA_AGGA^Lys/LysRS) and amber suppression (MjtRNA_CUA^Tyr/TyrRS).⁹³

In vivo UAA incorporation in Xenopusoocytes using frameshift suppression was demonstrated with SctRNA_CCCG^Phe for 5-fluorotryptophan and SctRNA_ACCC^Phe for α-aminobutyric acid or norvaline.⁸⁵ Simultaneous incorporation of 5-fluorotryptophan and α-aminobutyric acid (or norvaline) was done by microinjection of both tRNA_CUA (THG73, described above) and SctRNA_CCCG^Phe (or SctRNA_ACCC^Phe) with a neuroreceptor mRNA. For incorporation of three amino acids, three tRNAs were microinjected into oocytes: 5-fluorotryptophan-tRNA_CUA/THG73, α-aminobutyric acid-SctRNA_CCCG^Phe and norvaline-tRNA_ACCC^Phe. Expression levels were adequate for electric current measurements of the functional neuroreceptor.⁸⁹

4.1.6 Development of orthogonal ribosomes. To further avoid competition of release factor which results in limited protein yields, improvement in UAA incorporation was achieved using orthogonal ribosome (O-ribosome)/orthogonal mRNA (O-mRNA) pairs.^99–102 Positive and negative selections of O-ribosome/O-mRNA pairs was accomplished by the Chin group through use of a construct combining genes for chloramphenicol acetyltransferase (CAT) and uracil phosphoribosyltransferase (UPRT) creating the cat-upp fusion⁹⁹ (Fig. 6A). O-ribosomes that recognize ribosome binding sites (rbs) of an O-mRNA library enabled cells to grow on chloramphenicol by translation of the cat-upp fusion reporter gene. Cells in which endogenous ribosomes recognize and bind to O-mRNA resulted in UPRT production that converted 5-fluorouracil into a thymidylate synthaseinhibitor, causing cell death. To generate a O-ribosome/O-mRNA pair that can facilitate UAA incorporation, an O-ribosome library was created by saturation mutagenesis on the A-site (the 530 loop) on the E. coliribosome to identify variants that effectively carried out amber suppression and prevented release factor (RF1) binding ¹⁰⁰ (Fig. 6B).

Fig. 6 Orthogonal O-mRNA/O-ribosome selection in E. coli for improved UAA incorporation.⁹⁹ (A) The ribosome binding site (RBS) on the mRNA is modified to eliminate endogenous ribosome binding. Uracil phosphoribosyltransferase (UPRT) is encoded by uppgene. 5-fluorouracil is converted by UPRT into an inhibitor of thymidylate synthase, causing cell death. Translation of cat allows O-ribosome containing cells survive on chloramphenicol. (B) Selection of O-ribosome with improved UAGA decoding in E. coli.¹⁰⁰ Cells are transformed with O-rRNA library and the reporter plasmido-cat. Survival of cells on chloramphenicol indicates efficient UAGA decoding.

The ribosome library was co-transformed with a reporter plasmid bearing the catgene with 2 UAGA stop codons fused downstream of the orthogonal ribosome-binding site (o-cat) to isolate ribosomes capabe of translating the O-mRNA.¹⁰⁰ Cells containing o-cat (UAGA103,UAGA146)/tRNA^ser2_UCUA and O-ribosome library were grown in the presence of chloramphenicol. The ribosome binding site sequence of the O-mRNA was altered so that it was not recognized by the endogenous ribosome.^99,100 Through this selection U531G, U534A 16S rRNA mutant ribosome was identified for efficient UAGA decoding. This resulted in an O-ribosome variant that could only carry out translation of the O-mRNA, and not the cellular mRNAs.¹⁰⁰ The O-ribosome/O-mRNA pair was capable of functioning in parallel to the cellular translation.¹⁰⁰ The system offered the unique advantage of not interrupting cellular translation that could consequently cause cell death.¹⁰⁰

Using a previously developed p-benzoylphenylalanyl-tRNA synthetase/tRNA_CUA (BraRS/tRNA_CUA) pair evolved from the MjTyrRS/tRNA_CUA pair in E. coli,⁶³ in conjunction with the evolved O-ribosome with corresponding O-mRNA, they reported an increase in efficiency of 62% for site-specific p-benzoylphenylalanine (Bpa) incorporation into protein expressed in E. coli^100,102Bpa incorporation at two amber stop codon sites in the test protein resulted in a slightly lower efficiency of 38%.¹⁰⁰ Thus O-ribosome/O-mRNA pair was shown to effectively read through the amber codon for UAA incorporation. Essentially, the evolved O-ribosome decreased the functional interaction with E. coli RF-1, allowing suppressor tRNA to compete more efficiently for A-site binding in the presence of a UAG codon on the mRNA.^100,102

4.2 Residue-specific incorporation

Residue-specific UAA incorporation employs the organism’s cellular translational system to replace a natural amino acid with an analog. Since all the amino acids are substituted with an UAA, this often results in alteration of the global properties of proteins.¹⁰⁵ Initially, UAAs have been incorporated into proteins expressed in E. coli strains auxotrophic for a particular amino acid. The first experiments to demonstrate this was performed by Cowie and Cohen in 1957 in which methionine was replaced with selenomethonine.^106,107 This provided a heavy atom replacement that would be used subsequently as a tool for X-ray crystallography. By residue-specific incorporation, structurally similar analogs to the natural amino acids such as difluoromethionine,¹⁰⁸trifluoromethionine,¹⁰⁹2-amino-5-hexenoic acid,¹¹⁰2-amino-5-hexynoic acid,^110,111norleucine,^110,112 structurally diverse set of proline analogs,^113–115 trifluoroisoleucine,^116,117 2-amino-3- methyl-4-pentenoic acid,¹¹⁸trifluorovaline,^116,119azidohomoalanine,^120,121fluorophenylalanine isomers,¹²²aminotryptophan isomers,¹²³fluorotryptophan isomers,^124–1265-hydroxytryptophan,¹²⁶ 7-azatryptophan,¹²⁶ homopropargylglycine^121,127 and trifluoroleucine^128,129 was incorporated in vivo into proteins expressed in E. coli (Table 3). Expression of proteins bearing UAA such as azidohomoalanine,^130,131 homopropargylglycine,¹³² photo-eucine¹³³ and photo-methionine¹³³ in mammalian cells has been accomplished as well (Table 3).

In order to expand the structural diversity of UAA incorporated globally, modifications to the cellular biosynthetic machinery have been made. Of the components involved in protein translational, the AARS serves as the gatekeeper for accepting or rejecting a particular amino acid to be coupled with its tRNA. Specifically, the methods implemented to shepherd UAA incorporation relies on: (i) introducing additional copies of the natural AARS; (ii) enlargening the AARS binding pocket; and (iii) shrinking the editing domain of the AARS.

4.2.1. Engineering AARS to expand UAA incorporation. By engineering an extra copy of the AARS under a constitutive promoter, previous UAAs that were not incorporated by the endogenous AARS were being integrated into proteins (Table 3). Overexpression of MetRS was shown to enable activation of trans-crotylglycine, cis-crotylglycine, 2-aminoheptanoic acid, norvaline, 2-butynylglycine, and allylglycine for protein synthesis in E. coli.^134,135 Moreoever, LeuRS overexpression was shown to sustain incorporation of the previously refractory UAA, hexafluoroleucine, in E. coli.²¹

Substrate recognition specificity and editing function of the AARS have been modified in order to expand incorporation of non-isosteric amino acid analogs. In order to introduce bulkier UAAs, Henneke and coworkers created an A294G mutation in the E. coliphenylalanyl-tRNA synthetase (PheRS).^136,137 This increased the size of the binding pocket to allow charging of tRNA^Phe with the para-halogenated phenylalanine analogs: p-chlorophenylalanine and p-bromophenylalanine (Table 3).^136–139 The same mutant PheRS was used by the Tirrell group to incorporate additional phenylalanine analogs p-azidophenylalanine, p-iodophenylalanine, p-ethynylphenylalanine and p-cyanophenylalanine in E. coli (Table 3).^127,140,141 The introduction of a T251G in addition to the A294G mutation in PheRS enabled the incorporation of ketone functionality of p-acetylphenylalanine into recombinant dihydrofolate reductase (DHFR) (Table 3).¹⁴³ As an alternative to altering the binding pocket of the AARS, others have made changes to disable the editing domain to facilitate UAA incorporation. First demonstrated by Schimmel and coworkers, a single amino acid substitution in the editing pocket of valyl-tRNA synthetase (ValRS) from threonine to proline resulted in the misacylation of tRNA^Val.¹⁴⁴ This enabled aminobutyrate incorporation in response to valinecodons with 20% efficiency.¹⁴⁴ Subsequently Tang and Tirrell illustrated that substitution of T252 in the editing pocket of leucyl-tRNA synthetase (LeuRS) with larger amino acids led to misacylation of tRNA^Leu, allowing the incorporation of norvaline, norleucine, allylglycine, homoallylglycine, homopropargylglycine and 2-butynylglycine (Table 3).¹⁴⁵

Recently, the development of screening methods to identify variants of AARS capable of charging amino acid analogs has allowed for the rapid expansion of residue-specific UAA incorporation.^103,104,120 A saturation mutagenesis library of MetRS variants was screened to identify mutants that enable azidonorleucine incorporation by Tirrell et al.¹⁰³ (Fig. 7A, Table 3). The library was cloned into a vector that also contained the gene for outer membrane protein C (OmpC). Reactive UAA, such as azidohomoalanine or azidonorleucine, was introduced into OmpC in E. coli. Since the OmpC was expressed on the outer surface of the E. coli, cells displaying UAA could be covalently biotinylated by a Cu-catalyzed azide-alkyne ligation and labeled with fluorescent avidin.¹²⁰Fluorescence-activated cell sorting (FACS) was employed to sort and differentiate labeled from unlabeled cells. Selected cells could undergo another round of sorting viaFACS to identify MetRS variants able to charge the azide-bearing UAAs. Through this screen a MetRS variant with a L13G mutation was identified to incorporate azidonorleucine.¹²⁷

Fig. 7 Screening methods of AARS mutants for residue-specific incorporation of UAAs. (A) Identification of MetRS variants in E. coli for incorporation of azidonorleucine by exploiting its reactive side chain via bio-orthogonal 3 + 2 cycloaddition.¹⁰³ Cells are transformed with plasmid bearing a MetRS library and the reporter OmpCgene. OmpC expression is induced in presence of UAA. Successful incorporation of UAA results in cell surface display of reactive side chain that forms covalent bond with reactive biotin in the presence of Cu(II). Biotin on the cell surface is subsequently bound to fluorescent avidin. Employing fluorescence-activated cell sorting (FACS), labelled cells can be isolated. Sorted cells can be analyzed to identify MetRS mutants or subject to additional screening. (B) Screening for MetRS variants that activate trifluoronorleucine for residue-specific incorporation in E. coli.¹⁰⁴ Cells transformed with a construct containing a MetRS library and the GFP reporter gene are grown in media containing 19 amino acids in the presence (positive screening) or absence (negative screening) of UAA. Cells collected in negative screen are subjected to further positive and negative screening to identify cells that grow in the presence of UAA and exhibit high fluorescence.

For this type of screen, the UAA must be limited to a bioorthogonal chemical handle for reaction with the label. To overcome this limitation, OmpC was replaced with a GFP reporter (Fig. 7B).¹⁰⁴ All the methionine residues in the β-barrel region of GFP were randomized first to generate a mutant that retained its fluorescence even upon methionine replacement. A GFP mutant in which all the methionine residues were replaced and exhibited fluorescence comparable to wild type was identified. Five methionine residues were then introduced into this mutant at positions that did not alter fluorescence of the folded GFP. This resulting mutant GFP was used for the screening process. Cells were transformed with a plasmid bearing the MetRS library and mutant GFP were grown in media supplemented with 19 amino acids in the presence of 6,6,6-trifluoronorleucine (Tfn) and sorted by FACS as a positive screen (Table 3). A negative screen was perfomed in which the selected cells were expressed in 19 amino acids in the absence of methionine to remove any MetRS variants that could activate natural amino acids other than methionine The postive and negative screening was repeated several times to enrich for clones active towards UAA incorporation. The resulting MetRS variant (L13S, Y260L, H301L) was identified to demonstrate Tfn incorporation into the recombinant test protein DHFR with near complete replacement.

4.3 Multi-site-specific incorporation

In the prior cases of site-specific incorporation, the ability to include a second or additional UAAs in response to the same stop codon is limited due to lower read through and competition with release factors. Residue-specific incorporation, on the other hand, does not suffer from low read through or yield but is restricted to complete replacement of a designated natural amino acid with UAAs. Although the development of orthogonal ribosomes and mRNAs is improving the suppression efficiency, an alternative approach to accomplish multi-site-specific incorporation relies on exploiting the degeneracy of codons.

Reassignment of degenerate sense codons was used to incorporate L-3-(2-naphthyl)alanine (Nal) at five positions in DHFR by Tirrell and coworkers⁸⁰ (Table 1, Fig. 8). Similar to site-specific incorporation, an orthogonal SctRNA/PheRS pair was introduced into the E. coli expression system. Rather than utilizing a suppressor tRNA, they employed the tRNA_GAA^Phe from yeast and altered the anticodon to produce tRNA_AAA^Phe. This allowed for exact Watson–Crick pairing with the UUU minor codon such that Nal could be incorporated in response to their presence. The yeast PheRS T415G mutant was selected to accommodate the bulky Nal. This mutant ScPheRS/ tRNA_AAA^Phe pair was used for the expression of DHFR in an E. coliphenylalanine auxotrophic strain.⁸⁰ Although the tRNA_AAA^Phe was shown to compete with endogenous tRNA for UUU decoding, the codons were assigned to Nal with 80% efficiency.¹⁰⁵ The efficiency of incorporation was comparable to the efficiency of site-specific incorporation.⁵⁵ Moreover, by this method, five residues were replaced with the UAA.

Fig. 8 Reassignment of degenerate sense codonsin vivo in E. coli.^80,105 A modified orthogonal tRNA_AAA/PheRS pair from S. cerevisiae are introduced into E. coli that enable incorporation of naphthylalanine in response to the UUU codon. At the same time, endogenous E. colitRNA/PheRS pair mediates phenylalanine incorporation in response to the UUC codon.

5. Organismal UAA incorporation

In all the aforementioned work, the UAAs have been incorporated into target proteins that are overexpressed^15,16 or are being newly synthesized and later isolated or tagged.^130,131,133 In addition to tailoring specific proteins in the cell, the ability to design most or all proteins with UAA in the cell would advance synthetic biology by providing a whole organism comprised of unnatural “parts”. This may lead to the development of truly artificial cells that will potentially possess different properties from their natural counterparts. Initial experiments were performed by Rennert and Anker in 1963.¹⁴⁶ They demonstrated that E. coli was adaptable to growth on medium containing the UAA by the use of a chemostat. Normal growth rates and 5′,5′,5′-trifluoroleucine incorporation into all the proteins in E. coli were reported.¹⁴⁶ More recently, E. coli variants capable of surviving in media containing 4-fluorotryptophan were evolved by Bacher and Ellington.¹⁴⁷ Instead of using a chemostat, they gradually increased the amount of amino acid analog in the growth media until complete replacement was accomplished. This yielded global UAA incorporation throughout E. coli proteome. To identify mutations that contributed to the ability of E. coli to grow on 4-fluorotryptophan (Flt), genes from evolved clones were compared to the parental strain. They found that three proteins were mutated: tryptophanyl-tRNA synthetase (Q109P), aromatic amino acid permease (T30S, I47V, V365G), and tyrosine repressor (R380P).¹⁴⁷ This suggested that these proteins and their mutations were critical for the organism to grow and integrate the Flt into E. coli. Moreover through this process, one might be able to generate strains capable of surviving and functioning in the presence of other UAAs.

6. Engineered pathways for biosynthesis of UAA

In the previous case, the UAA is added externally by combining it with the medium that is used to culture the cells. Without UAA supplementation the cells are not capable of surviving, which may be problematic for synthetic biology applications that aim to interface artificial cell circuits with natural organisms and tissue. Schultz and coworkers engineered E. coli capable of biosynthesizing and incorporating p-aminophenylalanine (pAF) into proteins in response to amber nonsense codons⁶⁰ (Fig. 9). The papA (4-amino-4-deoxychorismate synthase), papB (chorismate mutase), papC (prephenate dehydrogenase) genes involved in the biosynthetic pathway for pAF production from S. venezuelae were introduced into E. coli, along with a plasmid containing the MjtRNA_CUA^Tyr/pAFRS pair previously isolated by rounds of positive and negative selection from MjTyrRS.⁶⁰ Cellular synthesis of pAF and substitution into the recombinant sperm whale myoglobin in response to TAG codon (position 4) was detected by sequencing and mass spectroscopy analysis.⁶⁰

Fig. 9 Biosynthesis and site-specific incorporation of p-aminophenylalanine (pAF) in E. coli.⁶⁰ The genespapA, papB, and papC from S. venezuelae were engineered into the constructs to enable the in vivo production of pAF.

7. Directed evolution of proteins bearing UAA

In order to achieve fully artificial organisms comprised of unnatural protein “parts”, methodology that enables the design of functional proteins bearing UAA is critical. Although the incorporation of UAA has led to functional artificial proteins with improved properties, in many cases, a reduction in function ensues especially when the UAA is integrated without rational design.^129,148 Directed evolution provides a solution for the identification of functional proteins and can be applied to proteins comprised of UAAs.^129,148 A diverse library of protein variants can be created by error-prone PCR or DNA shuffling and then screened for a particular function.^{150–152,129,148} In contrast to rational design where mutants are constructed based on known structural information of the target protein, directed evolution relies on the construction of a library of mutants in the absence of structure.^{150–152,129,148}

Sequence optimization experiments were reported for GFP containing UAA at one position by Sisido et al. (Fig. 10).¹⁴⁹ These experiments were carried out in cell-free E. colitranscription system. Specifically, p-aminophenylalanine (pAF) was incorporated into GFP at Y66 in response to the frameshift CGGG codon, resulting in reduced fluorescence intensity as compared to wild type GFP. Saturation mutagenesis was used to create library of mutants with amino acid substitutions at S65 and Y145 (Fig. 10). From this library, 42 colonies were subjected to colony PCR, in vitro translation and fluorescence measurements. Three mutants, Y145L, Y145F and Y145M, were identified to exhibit improved fluorescence intensity with pAF at position 66. By combining site-specific incorporation with directed evolution, they were able to enhance the fluorescent property of the GFP bearing the UAA.

Fig. 10 Identification of a variant with improved fluorescence of GFP bearing as site-specific inserted UAA.¹⁴⁹ GFP library, created by saturation mutagenesis, was transformed into E. coli. Selected colonies were isolated for GFP gene amplification . These genes were translated in cell-free E. coli expression system in the presence of UAA and expressed variants were analyzed by fluorescence measurements.

Residue-specific incorporation and directed evolution was integrated to recover or enhance the function of proteins bearing UAAs (Fig. 11). The incorporation of 5′,5′,5′-trifluoroleucine (TFL) into chloramphenicol acetyltransferase (CAT) yielded a loss in thermostability.^129,148,156 In the case of CAT, Montclare and Tirrell created random mutagenesis libraries by error-prone PCR, transformed them into E. colileucine auxotrophic strains and expressed them in the presence of TFL in 96-well plates.¹²⁹

Fig. 11 Scheme of directed evolution protocol for CAT¹²⁹ and GFP¹⁴⁸proteins with UAA incorporated in residue-specific manner expressed in E. coli. Gene library generated by error-prone PCR was transformed in E. coli and grown in media containing UAA. Variants were screened for desirable properties: activity for CAT, fluorescence intensity for GFP. The best variants were chosen for additional rounds of directed evolution.

A high-throughput colorimetric screening method was developed for cell lysates bearing the fluorinated variants.¹²⁹ Most active clones were identified for the additional rounds of random mutagenesis and screening. A fluorinated CAT variant with enhanced thermal stability containing three mutations (K46M, S87N and M142I) was identified.¹²⁹ Recently, this screen was also employed by the Montclare and coworkers to obtain an activity and thermostability profile of single-TFL-to-isoleucine mutants.¹⁵⁷ This protocol of applying directed evolution to proteins bearing UAA was performed on GFP using FACS as a screening method. The incorporation of TFL into GFP resulted in a loss of fluorescence.¹⁴⁸ To identify GFP variants with enhanced folding character, a random library was generated via error-prone PCR similar to the CAT work but with a slightly higher mutation rate.^129,148 Cells underwent additional rounds of directed evolution resulting in the isolation of a fluorinated GFP mutant bearing twenty mutations with a total of fifteen fluorinated analogs.¹⁴⁸ These experiments demonstrate that through combination of residue-specific UAA incorporation and directed evolution, functional proteins can be achieved.^129,148

Recently site-specific incorporation in combination with phage display was employed to select a highly functional artificial antibody containing UAA by the Schultz and Smider groups.¹⁵³ Multivalent hyperphage display was employed in which the C-terminus of the antibody scFv fragment was fused to pIII. Phagemid-encoded scFv was engineered to bear an amber codon within its DNA sequence, resulting in the incorporation of UAA at position 111 in the protein.¹⁵³ Read-through of the amber codon yielded expression of functional pIII leading to phage assembly and scFv display on the phage surface¹⁵³ (Fig. 12). By optimizing expression system, the bias of low yield of UAA containing scFv was eliminated and directed evolution experiments were carried out.¹⁵³ The phage displayed scFv bearing the UAA sulfotyrosine (sTyr) was selected for gp 120 binding.¹⁵³ Enhanced binding affinity and specificity of gp120 binding by the identified antibody was further confirmed and characterized by ELISA. These results demonstrate that directed evolution of proteins with UAA can offer a selective advantage through the specific functional role it provides beyond the natural set of 20 amino acids.

Fig. 12 Illustration of phage display integrated with site-specific UAA incorporation.^153–155 Phage produced by E.coli displayed scFv antibody that undergoes directed evolution. After cleavage with protease and elution, scFv antibodies were assayed for gp120 binding, and the best variants were selected to undergo another round of directed evolution.

Conclusions

Tremendous strides have been made in the construction of unnatural protein “parts” and much of the advances described here have emanated from exploiting the principle of orthogonality.^13,15 The development of elegant screens or selections for molecular evolution has not only accelerated the ability to incorporate UAA, but also has enabled the creation of tailor-made artificial proteins.^148,149 This, coupled to advances in organismal evolution to allow for survival on UAA¹⁴⁷ as well as engineering pathways for UAA biosynthesis,⁶⁰ will facilitate the integration of such unnatural proteins into living matter. With these molecular tools now available, the tantalizing prospect of engineering wholly synthetic biological cells and systems may soon be realized.

Acknowledgements

We are grateful for support from the NYU:Poly Seed Grant, Wechsler Award, Othmer Institute for Interdisciplinary Studies, National Science Foundation (CBET-072242 and MRSEC program DMR-0820341) and the Air Force Office of Scientific Research Young Investigator Program (FA-9-550-07-1-0060).

Notes and references

1. E. Andrianantoandro, S. Basu, D. K. Karig and R. Weiss, Mol. Syst Biol., 2006, 2, 0028.
2. D. Baker, B. F. Group, G. Church, J. Collins, D. Endy, J. Jacobson, J. Keasling, P. Modrich, C. Smolke and R. Weiss, Scientific American, 2006, 294, 44–51.
3. D. A. Drubin, J. C. Way and P. A. Silver, Genes Dev., 2007, 21, 242–254.
4. E. L. Haseltine and F. H. Arnold, Annu. Rev. Biophys. Biomol. Struct., 2007, 36, 1–19.
5. J. W. Chin, Nat. Chem. Biol., 2006, 2, 304–311.
6. J. Tian, H. Gong, N. Sheng, X. Zhou, E. Gulari, X. Gao and G. M. Church, Nature, 2004, 432, 1050–1054.
7. P. A. Carr, J. S. Park, Y. J. Lee, T. Yu, S. Zhang and J. M. Jacobson, Nucleic Acids Res., 2004, 32, e162.
8. S. Basu, Y. Gerchman, C. H. Collins, F. H. Arnold and R. Weiss, Nature, 2005, 434, 1130–1134.
9. D. K. Ro, E. M. Paradise, M. Ouellet, K. J. Fisher, K. L. Newman, J. M. Ndungu, K. A. Ho, R. A. Eachus, T. S. Ham, J. Kirby, M. C. Y. Chang, S. T. Withers, Y. Shiba, R. Sarpong and J. D. Keasling, Nature, 2006, 440, 940–943.
10. J. C. Anderson, E. J. Clarke, A. P. Arkin and C. A. Voight, J. Mol. Biol., 2005, 335, 619–627.
11. T. K. Lu and J. J. Collins, Proc. Natl. Acad. Sci. U. S. A., 2007, 104, 11197–11202.
12. N. Budisa, Engineering the Genetic Code. Expanding the Amino Acid Repertoire for the Design of Novel Proteins, edn, Wiley-VCH, Weinheim, 2005.
13. J. W. Chin, Curr. Opin. Struct. Biol., 2006, 16, 551–556.
14. J. Xie and P. G. Schultz, Curr. Opin. Chem. Biol., 2005, 9, 548–554.
15. J. Xie and P. G. Schultz, Nat. Rev. Mol. Cell Biol., 2006, 7, 775–782.
16. R. E. Connor and D. A. Tirrell, Journal of Macromolecular Science, Part C: Polymer Reviews, 2007, 47, 9–28.
17. M. Bose, D. Groff, J. M. Xie, E. Brustad and P. G. Schultz, J. Am. Chem. Soc., 2006, 128, 388–389.
18. K. C. Schultz, L. Supekova, Y. Ryu, J. Xie, R. Perera and P. G. Schultz, J. Am. Chem. Soc., 2006, 128, 13984–13985.
19. L. Alfonta, Z. W. Zhang, S. Uryu, J. A. Loo and P. G. Schultz, J. Am. Chem. Soc., 2003, 125, 14662–14663.
20. Y. Tang, G. Ghirlanda, W. A. Petka, T. Nakajima, W. F. DeGrado and D. A. Tirrell, Angew. Chem., Int. Ed., 2001, 40, 1494–1496.
21. Y. Tang and D. A. Tirrell, J. Am. Chem. Soc., 2001, 123, 11089–11090.
22. H. Meng and K. Kumar, J. Am. Chem. Soc., 2007, 129, 15615–15622.
23. M. B. Elowitz and S. Leibler, Nature, 2000, 403, 335–338.
24. T. S. Gardner, C. R. Cantor and J. J. Collins, Nature, 2000, 403, 339–342.
25. M. R. Atkinson, M. A. Savageau, J. T. Myers and A. J. Ninfa, Cell, 2003, 113, 597–607.
26. J. E. Dueber, B. J. Yeh, K. Chak and W. A. Lim, Science, 2003, 301, 1904–1908.
27. J. E. Dueber, E. A. Mirsky and W. A. Lim, Nat. Biotechnol., 2007, 25, 660–662.
28. S. H. Park, A. Zarrinpar and W. A. Lim, Science, 2003, 299, 1061–1064.
29. K. E. Prehoda, J. A. Scott, R. D. Mullins and W. A. Lim, Science, 2000, 290, 801–806.
30. A. Zarrinpar, S. H. Park and W. A. Lim, Nature, 2003, 426, 676–680.
31. J. E. Dueber, B. J. Yeh, R. P. Bhattacharyya and W. A. Lim, Curr. Opin. Struct. Biol., 2004, 14, 690–699.
32. L. L. Looger, M. A. Dwyer, J. J. Smith and H. W. Hellinga, Nature, 2003, 423, 185–190.
33. E. Fung, W. W. Wong, J. K. Suen, T. Bulter, S. G. Lee and J. C. Liao, Nature, 2005, 435, 118–122.
34. T. Schrader and S. Koch, Mol. BioSyst., 2007, 3, 241–248.
35. H. Kitano, Nature, 2002, 420, 206–210.
36. J. S. Bader, A. Chaudhuri, J. M. Rothberg and J. Chant, Nat. Biotechnol., 2004, 22, 78–85.
37. J. B. Pereira-Leal, A. J. Enright and C. A. Ouzounis, Proteins: Struct., Funct., Bioinf., 2004, 54, 49–57.
38. S. K. Lee, H. Chou, T. S. Ham, T. S. Lee and J. D. Keasling, Curr. Opin. Biotechnol., 2008, 19, 556–563.
39. N. Rosenfeld, M. B. Elowitz and U. Alon, J. Mol. Biol., 2002, 323, 785–793.
40. A. E. Friedland, T. K. Lu, X. Wang, D. Shi, G. Church and J. J. Collins, Science, 2009, 324, 1199–1202.
41. K. Brenner, D. K. Karig, R. Weiss and F. H. Arnold, Proc. Natl. Acad. Sci. U. S. A., 2007, 104, 17300–17304.
42. C. J. Bashor, N. C. Helman, S. Yan and W. A. Lim, Science, 2008, 319, 1539–1543.
43. S. Atsumi, T. Hanai and J. C. Liao, Nature, 2008, 451, 86–89.
44. J. B. Lucks, L. Qi, W. R. Whitaker and A. P. Arkin, Curr. Opin. Microbiol., 2008, 11, 567–573.
45. J. F. Atkins and P. V. Baranov, Nature, 2007, 448, 1004–1005.
46. D. L. Hatfield, B. A. Carlson, X. M. Xu, H. Mix and V. N. Gladyshev, Prog. Nucleic Acid Res. Mol. Biol., 2006, 81, 97–142.
47. C. Baron and A. Bock, in tRNA: Structure, Biosynthesis, and Function, ed. D. a. R. Soll and L. Uttam, ASM Press, Washington, D.C., 1995.
48. J. A. Krzycki, Curr. Opin. Microbiol., 2005, 8, 706–712.
49. M. Ibba and D. Soll, Curr. Biol., 2002, 12, R464–466.
50. S. M. Hecht, B. L. Alford, Y. Kuroda and S. Kitano, J. Biol. Chem., 1978, 253, 4517–4520.
51. C. J. Noren, S. J. Anthony-Cahill, M. C. Griffith and P. G. Schultz, Science, 1989, 244, 182–188.
52. M. C. Hartman, K. Josephson and J. W. Szostak, Proc. Natl. Acad. Sci. U. S. A., 2006, 103, 4356–4361.
53. R. Furter, Protein Sci., 1998, 7, 419–426.
54. J. W. Chin, S. W. Santoro, A. B. Martin, D. S. King, L. Wang and P. G. Schultz, J. Am. Chem. Soc., 2002, 124, 9026–9027.
55. J. C. Jackson, S. P. Duffy, K. R. Hess and R. A. Mehl, J. Am. Chem. Soc., 2006, 128, 11124–11127.
56. J. C. Jackson, J. T. Hammill and R. A. Mehl, J. Am. Chem. Soc., 2007, 129, 1160–1166.
57. L. Wang, Z. Zhang, A. Brock and P. G. Schultz, Proc. Natl. Acad. Sci. U. S. A., 2003, 100, 56–61.
58. J. Xie, L. Wang, N. Wu, A. Brock, G. Spraggon and P. G. Schultz, Nat. Biotechnol., 2004, 22, 1297–1301.
59. Z. Zhang, J. Gildersleeve, Y. Y. Yang, R. Xu, J. A. Loo, S. Uryu, C. H. Wong and P. G. Schultz, Science, 2004, 303, 371–373.
60. R. A. Mehl, J. C. Anderson, S. W. Santoro, L. Wang, A. B. Martin, D. S. King, D. M. Horn and P. G. Schultz, J. Am. Chem. Soc., 2003, 125, 935–939.
61. L. Wang and P. G. Schultz, Chem. Biol., 2001, 8, 883–890.
62. J. W. Chin, T. A. Cropp, J. C. Anderson, M. Mukherji, Z. W. Zhang and P. G. Schultz, Science, 2003, 301, 964–967.
63. S. W. Santoro, L. Wang, B. Herberich, D. S. King and P. G. Schultz, Nat. Biotechnol., 2002, 20, 1044–1048.
64. J. Guo, J. Wang, J. C. Anderson and P. G. Schultz, Angew. Chem., Int. Ed., 2007, 47, 722–725.
65. Y. H. Ryu and P. G. Schultz, Nat. Methods, 2006, 3, 263–265.
66. I. Kwon and D. A. Tirrell, J. Am. Chem. Soc., 2007, 129, 10431–10437.
67. I. Kwon, P. Wang and D. A. Tirrell, J. Am. Chem. Soc., 2006, 128, 11778–11783.
68. C. E. Melancon, 3rd and P. G. Schultz, Bioorg. Med. Chem. Lett., 2009, 19, 3845–3847.
69. S. Chen, P. G. Schultz and A. Brock, J. Mol. Biol., 2007, 371, 112–122.
70. K. Sakamoto, A. Hayashi, A. Sakamoto, D. Kiga, H. Nakayama, A. Soma, T. Kobayashi, M. Kitabatake, K. Takio, K. Saito, M. Shirouzu, I. Hirao and S. Yokoyama, Nucleic Acids Res., 2002, 30, 4692–4699.
71. Z. Zhang, L. Alfonta, F. Tian, B. Bursulaya, S. Uryu, D. S. King and P. G. Schultz, Proc. Natl. Acad. Sci. U. S. A., 2004, 101, 8882–8887.
72. W. Wang, J. K. Takimoto, G. V. Louie, T. J. Baiga, J. P. Noel, K. F. Lee, P. A. Slesinger and L. Wang, Nat. Neurosci., 2007, 10, 1063–1072.
73. S. Ye, C. Kohrer, T. Huber, M. Kazmi, P. Sachdev, E. C. Yan, A. Bhagat, U. L. RajBhandary and T. P. Sakmar, J. Biol. Chem., 2008, 283, 1525–1533.
74. L. Wang, A. Brock, B. Herberich and P. G. Schultz, Science, 2001, 292, 498–500.
75. J. Guo, J. Wang, J. S. Lee and P. G. Schultz, Angew. Chem., Int. Ed., 2008, 47, 6399–6401.
76. H. S. Lee and P. G. Schultz, J. Am. Chem. Soc., 2008, 130, 13194–13195.
77. J. Xie, W. Liu and P. G. Schultz, Angew. Chem., Int. Ed., 2007, 46, 9239–9242.
78. H. S. Lee, G. Spraggon, P. G. Schultz and F. Wang, J. Am. Chem. Soc., 2009, 131, 2481–2483.
79. P. R. Chen, D. Groff, J. Guo, W. Ou, S. Cellitti, B. H. Geierstanger and P. G. Schultz, Angew. Chem., Int. Ed., 2009, 48, 4052–4055.
80. I. Kwon, K. Kirshenbaum and D. A. Tirrell, J. Am. Chem. Soc., 2003, 125, 7512–7513.
81. H. Neumann, S. Y. Peak-Chew and J. W. Chin, Nat. Chem. Biol., 2008, 4, 232–234.
82. E. Brustad, M. L. Bushey, A. Brock, J. Chittuluru and P. G. Schultz, Bioorg. Med. Chem. Lett., 2008, 18, 6004–6006.
83. T. S. Young, I. Ahmad, A. Brock and P. G. Schultz, Biochemistry, 2009, 48, 2643–2653.
84. T. Mukai, T. Kobayashi, N. Hino, T. Yanagisawa, K. Sakamoto and S. Yokoyama, Biochem. Biophys. Res. Commun., 2008, 371, 818–822.
85. E. A. Rodriguez, H. A. Lester and D. A. Dougherty, Proc. Natl. Acad. Sci. U. S. A., 2006, 103, 8650–8655.
86. G. Turcatti, K. Nemeth, M. D. Edgerton, U. Meseth, F. Talabot, M. Peitsch, J. Knowles, H. Vogel and A. Chollet, J. Biol. Chem., 1996, 271, 19991–19998.
87. D. L. Beene, K. L. Price, H. A. Lester, D. A. Dougherty and S. C. Lummis, J. Neurosci., 2004, 24, 9097–9104.
88. D. A. Dougherty, Curr. Opin. Chem. Biol., 2000, 4, 645–652.
89. K. Yamanaka, H. Nakata, T. Hohsaka and M. Sisido, J. Biosci. Bioeng., 2004, 97, 395–399.
90. E. A. Rodriguez, H. A. Lester and D. A. Dougherty, RNA, 2007, 13, 1703–1714.
91. E. A. Rodriguez, H. A. Lester and D. A. Dougherty, RNA, 2007, 13, 1715–1722.
92. M. Sisido and T. Hohsaka, Appl. Microbiol. Biotechnol., 2001, 57, 274–281.
93. J. C. Anderson, N. Wu, S. W. Santoro, V. Lakshman, D. S. King and P. G. Schultz, Proc. Natl. Acad. Sci. U. S. A., 2004, 101, 7566–7571.
94. N. Muranaka, T. Hohsaka and M. Sisido, Nucleic Acids Res., 2006, 34, e7.
95. H. Taira, T. Hohsaka and M. Sisido, Nucleic Acids Res., 2006, 34, e44.
96. T. Hohsaka, Y. Ashizuka, H. Murakami and M. Sisido, Nucleic Acids Res., 2001, 29, 3646–3651.
97. H. Taira, M. Fukushima, T. Hohsaka and M. Sisido, J. Biosci. Bioeng., 2005, 99, 473–476.
98. M. Taki, Y. Tokuda, T. Ohtsuki and M. Sisido, J. Biosci. Bioeng., 2006, 102, 511–517.
99. O. Rackham, K. Wang and J. W. Chin, Nat. Chem. Biol., 2006, 2, 254–258.
100. K. H. Wang, H. Neumann, S. Y. Peak-Chew and J. W. Chin, Nat. Biotechnol., 2007, 25, 770–777.
101. O. Rackham and J. W. Chin, J. Am. Chem. Soc., 2005, 127, 17584–17585.
102. O. Rackham and J. W. Chin, Biochem. Soc. Trans., 2006, 34, 328–329.
103. A. J. Link, M. K. Vink, N. J. Agard, J. A. Prescher, C. R. Bertozzi and D. A. Tirrell, Proc. Natl. Acad. Sci. U. S. A., 2006, 103, 10180–10185.
104. T. H. Yoo and D. A. Tirrell, Angew. Chem., Int. Ed., 2007, 46, 5340–5343.
105. A. J. Link and D. A. Tirrell, Methods, 2005, 36, 291–298.
106. G. N. Cohen and D. B. Cowie, C R Hebd Seances Acad. Sci., 1957, 244, 680–683.
107. D. B. Cowie and G. N. Cohen, Biochim. Biophys. Acta, 1957, 26, 252–261.
108. P. Walasek and J. F. Honek, BMC Biochem., 2005, 6, 21.
109. H. S. Duewel, E. Daub, V. Robinson and J. F. Honek, Biochemistry, 2001, 40, 13167–13176.
110. J. C. van Hest, K. L. Kiick and D. A. Tirrell, J. Am. Chem. Soc., 2000, 122, 1282–1288.
111. K. L. Kiick and D. A. Tirrell, Tetrahedron, 2000, 56, 9487–9493.
112. C. Wolschner, A. Giese, H. A. Kretzschmar, R. Huber, L. Moroder and N. Budisa, Proc. Natl. Acad. Sci. U. S. A., 2009, 106, 7756–7761.
113. W. Kim, A. George, M. Evans and V. P. Conticello, ChemBioChem, 2004, 5, 928–936.
114. N. Budisa, C. Minks, F. J. Medrano, J. Lutz, R. Huber and L. Moroder, Proc. Natl. Acad. Sci. U. S. A., 1998, 95, 455–459.
115. W. Kim, K. I. Hardcastle and V. P. Conticello, Angew. Chem., Int. Ed., 2006, 45, 8141–8145.
116. S. Son, I. C. Tanrikulu and D. A. Tirrell, ChemBioChem, 2006, 7, 1251–1257.
117. P. Wang, Y. Tang and D. A. Tirrell, J. Am. Chem. Soc., 2003, 125, 6900–6906.
118. M. L. Mock, T. Michon, J. C. van Hest and D. A. Tirrell, ChemBioChem, 2006, 7, 83–87.
119. P. Wang, A. Fichera, K. Kumar and D. A. Tirrell, Angew. Chem., Int. Ed., 2004, 43, 3664–3666.
120. A. J. Link and D. A. Tirrell, J. Am. Chem. Soc., 2003, 125, 11164–11165.
121. A. Wang, N. Winblade Nairn, R. S. Johnson, D. A. Tirrell and K. Grabstein, ChemBioChem, 2008, 9, 324–330.
122. C. Minks, R. Huber, L. Moroder and N. Budisa, Anal. Biochem., 2000, 284, 29–34.
123. M. Rubini, S. Lepthien, R. Golbik and N. Budisa, Biochim. Biophys. Acta, Proteins Proteomics, 2006, 1764, 1147–1158.
124. Q. S. Zhang, L. Shen, E. D. Wang and Y. L. Wang, J. Protein Chem., 1999, 18, 187–192.
125. J. F. Parsons, G. Xiao, G. L. Gilliland and R. N. Armstrong, Biochemistry, 1998, 37, 6286–6294.
126. M. Katragadda and J. D. Lambris, Protein Expression Purif., 2006, 47, 289–295.
127. K. E. Beatty, F. Xie, Q. Wang and D. A. Tirrell, J. Am. Chem. Soc., 2005, 127, 14150–14151.
128. Y. Tang, G. Ghirlanda, W. A. Petka, T. Nakajima, W. F. DeGrado and D. A. Tirrell, Angew. Chem., Int. Ed., 2001, 40, 1494–1496.
129. J. K. Montclare and D. A. Tirrell, Angew. Chem., Int. Ed., 2006, 45, 4518–4521.
130. D. C. Dieterich, A. J. Link, J. Graumann, D. A. Tirrell and E. M. Schuman, Proc. Natl. Acad. Sci. U. S. A., 2006, 103, 9482–9487.
131. D. C. Dieterich, J. J. Lee, A. J. Link, J. Graumann, D. A. Tirrell and E. M. Schuman, Nat. Protoc., 2007, 2, 532–540.
132. K. E. Beatty, J. C. Liu, F. Xie, D. C. Dieterich, E. M. Schuman, Q. Wang and D. A. Tirrell, Angew. Chem., Int. Ed., 2006, 45, 7364–7367.
133. M. Suchanek, A. Radzikowska and C. Thiele, Nat. Methods, 2005, 2, 261–267.
134. K. L. Kiick, J. C. van Hest and D. A. Tirrell, Angew. Chem., Int. Ed., 2000, 39, 2148–2152.
135. K. L. Kiick, R. Weberskirch and D. A. Tirrell, FEBS Lett., 2001, 502, 25–30.
136. M. Ibba and H. Hennecke, FEBS Lett., 1995, 364, 272–275.
137. M. Ibba, P. Kast and H. Hennecke, Biochemistry, 1994, 33, 7107–7112.
138. P. Kast and H. Hennecke, J. Mol. Biol., 1991, 222, 99–124.
139. N. Sharma, R. Furter, P. Kast and D. A. Tirrell, FEBS Lett., 2000, 467, 37–40.
140. K. Kirshenbaum, I. S. Carrico and D. A. Tirrell, ChemBioChem, 2002, 3, 235–237.
141. I. S. Carrico, S. A. Maskarinec, S. C. Heilshorn, M. L. Mock, J. C. Liu, P. J. Nowatzki, C. Franck, G. Ravichandran and D. A. Tirrell, J. Am. Chem. Soc., 2007, 129, 4874–4875.
142. J. K. Montclare, S. Son, G. A. Clark, K. Kumar and D. A. Tirrell, ChemBioChem, 2009, 10, 84–86.
143. D. Datta, P. Wang, I. S. Carrico, S. L. Mayo and D. A. Tirrell, J. Am. Chem. Soc., 2002, 124, 5652–5653.
144. V. Doring, H. D. Mootz, L. A. Nangle, T. L. Hendrickson, V. de Crecy-Lagard, P. Schimmel and P. Marliere, Science, 2001, 292, 501–504.
145. Y. Tang and D. A. Tirrell, Biochemistry, 2002, 41, 10635–10645.
146. O. M. Rennert and H. S. Anker, Biochemistry, 1963, 2, 471–476.
147. J. M. Bacher and A. D. Ellington, J. Bacteriol., 2001, 183, 5414–5425.
148. T. H. Yoo, A. J. Link and D. A. Tirrell, Proc. Natl. Acad. Sci. U. S. A., 2007, 104, 13887–13890.
149. D. Kajihara, T. Hohsaka and M. Sisido, Protein Eng., Des. Sel., 2005, 18, 273–278.
150. C. M. Yuen and D. R. Liu, Nat. Methods, 2007, 4, 995–997.
151. H. Zhao, Biotechnol. Bioeng., 2007, 98, 313–317.
152. F. H. Arnold, P. L. Wintrode, K. Miyazaki and A. Gershenson, Trends Biochem. Sci., 2001, 26, 100–106.
153. C. C. Liu, A. V. Mack, M. L. Tsao, J. H. Mills, H. S. Lee, H. Choe, M. Farzan, P. G. Schultz and V. V. Smider, Proc. Natl. Acad. Sci. U. S. A., 2008, 105, 17688–17693.
154. S. Rondot, J. Koch, F. Breitling and S. Dubel, Nat. Biotechnol., 2001, 19, 75–78.
155. M. Pastrnak and P. G. Schultz, Bioorg. Med. Chem., 2001, 9, 2373–2379.
156. T. Panchenko, W. W. Zhu and J. K. Montclare, Biotechnol. Bioeng., 2006, 94, 921–930.
157. N. Voloshchuk, M. X. Lee, W. W. Zhu, I. C. Tanrikulu and J. K. Montclare, Bioorg. Med. Chem. Lett., 2007, 17, 5907–5911.

---