Unnatural Amino Acid Incorporation into Proteins

• Introduction
• Current Research
  • In Vitro Incorporation of Unnatural Amino Acids into Proteins
  • In Vivo Incorporation of Unnatural Amino Acids
  • Developing tRNAs to Suppress Four and Five Base Codons
  • Expanding the Genetic Code of E.coli
• Contacts
  

Introduction

Proteins are at the crossroads of virtually every biological process, from photosynthesis and vision to signal transduction and the immune response. These complex functions result from a polyamide based polymer consisting of twenty relatively simple building blocks arranged in a defined primary sequence. One important tool for understanding the relationship between the structure of these building blocks and the functions they encode is site-directed mutagenesis, in which a specific amino acid in a protein can be replaced with any of the other nineteen common amino acids. This method also has made possible the engineering of proteins with enhanced properties including stability, catalytic function and binding specificity. However, in contrast to small molecule synthesis where virtually any change in structure can be made, changes in protein structure are limited to the twenty encoded amino acids. Ideally, one would like to tailor changes in a protein (the size, acidity, nucleophilicity, hydrogen-bonding or hydrophobic properties of amino acids) to fulfill a specific structural or functional property of interest. The ability to incorporate such unnatural amino acids into proteins would greatly expand our ability to rationally and systematically manipulate the structures of proteins, both to probe protein function and create proteins with new properties.

Current Research

In Vitro Incorporation of Unnatural Amino Acids into Proteins

We developed a biosynthetic approach that allows for the relatively facile site-specific incorporation of unnatural amino acids into proteins (Fig 1). This method takes advantage of the fact that the genetic code contains three stop codons. Because only one stop codon is needed for translational termination, the other two can in principle be used to encode nonproteinogenic amino acids. A suppressor tRNA is prepared that recognizes the stop codon UAG (but is not recognized by the naturally occurring aminoacyl-tRNA synthetases) and chemically aminoacylated with the desired unnatural amino acid. Conventional site-directed mutagenesis is used to introduce the stop codon UAG at the site of interest in the protein gene. When the acylated suppressor tRNA and the mutant gene are combined in an in vitro transcription/translation system, the unnatural amino acid is incorporated in response to the UAG codon to give a protein containing that amino acid at the specified position.

Although in vitro protein synthesis cannot be carried out on the same scale as in vivo synthesis, in vitro methods can yield hundreds of micrograms of purified protein containing an unnatural amino acid. We have produced several such proteins in quantities sufficient for their characterization using circular dichroism (CD), nuclear magnetic resonance (NMR) spectrometry, and X-ray crystallography. This methodology has been used to investigate the role of hydrophobicity, packing, side chain entropy and hydrogen bonding in determining protein stability and folding. It is also being used to probe catalytic mechanism, signal transduction and electron transfer in proteins. In addition, the properties of proteins have been modified using this methodology. For example, photocaged proteins have been generated that can be activated by photolysis, and novel chemical handles have been introduced into proteins for the site specific incorporation of optical and other spectroscopic probes. Over 100 noncoded amino acids have been introduced into proteins to date.

Gating of K+ channels by a dynamic selectivity filter

K+ channels use diverse mechanisms to control gating, ranging from the voltage-sensitive S4 transmembrane helix and N-terminal inactivation "ball" to intracellular polyamines and the C-terminal inactivation gate. We have shown that the ion selectivity filter, an outer pore region critical for K+ selectivity and permeation, functions as a gate in the inward rectifier K+ (Kir) channel Kir2.1. At constant negative membrane potentials, single Kir2.1 channels open and close spontaneously and exhibit multiple and distinct subconductance levels. The gating behavior changes dramatically when the backbone carbonyls of two conserved glycine residues lining the selectivity filter are mutagenized from amides to esters, reducing their electronegativity. The single-channel gating behavior is also altered when K+ is replaced by Tl+, a permeant K+ ion substitute which induces conformational changes in the selectivity filter. These results demonstrate a tight coupling between gating and ion permeation and suggest that the electrostatics and conformational dynamics of the selectivity filter directly contribute to the spontaneous full-level open-closed gating as well as the sublevel gating of Kir2.1 channels.

In Vivo Incorporation of Unnatural Amino Acids

The development of a general approach for the site-specific incorporation of unnatural amino acids into proteins in vivo, directly from the growth media, would greatly enhance the power of unnatural amino acid mutagenesis. For example, the ability to synthesize large quantities of proteins containing heavy atoms would facilitate protein structure determination, and the ability to site-specifically substitute fluorophores or photocleavable groups into proteins in living cells would provide powerful tools for studying protein function in vivo. Alternatively, one might be able to enhance the properties of proteins by providing building blocks with new functional groups, such as a keto-containing amino acid. Using a multistage approach, we have been successful for the first time to engineer living cells that are capable of site-specifically incorporating unnatural amino acids into proteins.

Our overall strategy consists of four elements:

1. the design and synthesis of the unnatural amino acid;
2. the construction of an orthogonal amber suppressor tRNA (O-tRNA_CUA) that is not a substrate for any of the naturally occurring aminoacyl-tRNA synthetases and which will eventually be used to deliver our unnatural amino acid in response to a UAG codon in the mRNA encoding the protein of interest;
3. the selection of an aminoacyl-tRNA synthetase (from a library of mutants) that recognizes the O-tRNACUA but does not recognize any endogenous tRNAs; and
4. the screening of a library of mutants of this tRNA aminoacyl synthetase for the ability to acylate the O-tRNA_CUA with the unnatural amino acid but not a common amino acid.

Development of orthogonal tRNA/synthetase pairs

Initially we tried to make an orthogonal tRNA/synthetase pair out of E. coli tRNA^Gln/GlnRS pair.Three sites at which mutations were expected to modulate the ability of GlnRS to acylate the tRNA ("knobs") were selected, and tRNAs bearing mutations at each site (and in all possible combinations) were generated. Based on in vitro aminoacylation with GlnRS and in vitro suppression studies, the amber suppressor tRNA with all three know mutations, O~tRNA^Gln(CUA), was found to be orthogonal in E.coli. However, we were not able to isolate an orthogonal GlnRS mutant that charges this orthongal tRNA exclusively. The best mutant GlnRS evolved still acylates the wild-type tRNAGln substrate 9-fold better than the O

We then showed that the amber suppressing derivative of SctRNAGln (O-SctRNAGln(CUA)) and ScGlnRS constitute an orthogonal tRNA/synthetase pair in E. coli, as evidenced by in vitro aminoacylation assay and in vitro suppression studies. When characterized in vivo using a b-lactamase assay, cells expressing the O-SctRNAGln(CUA) and an inactive mutant of ScGlnRS exhibited an IC50 of about 20 mg ml-1 ampicillin, indicating virtually no acylation by endogenous synthetases. With an active ScGlnRS, cells exhibit an IC50 of about 500 mg ml-1 ampicillin, indicating that the ScGlnRS acylates the O~-ctRNAGln(CUA) in E. coli.

More recently, we demonstrated that the amber surppressing derivative of Methanococcus jannaschii tRNATyr and M. jannaschii TyrRS constitutes an orthogonal pair in E. coli. Using the same b-lactamase assay, O-MjtRNATyr(CUA) expression alone confers an IC50 of about 55 mg ml-1, but co-expression with the MjTyrRS confers resistance to an IC50 of about 1,200 mg ml-1. This indicates both that the O-MjtRNATyr(CUA) is slightly less orthogonal to endogenous E. coli synthetases than O~SctRNAGln(CUA) and that the MjTyrRS is much more active than the ScGlnRS under the expression conditions examined.

The orthogonality of this MjtRNATyr(CUA) was improved with a selection strategy (Fig. 3). A tRNA mutant library was first passed through a negative selection, which removed tRNA variants that can be acylated by endogenous E. coli tRNAs. The products of this selection were then passed through a positive selection in the presence of the MjTyrRS. Survival of the cells, grown in the presence of antibiotic, required that the MjTyrRS acylate the tRNA variant to support translation of an antibiotic resistance gene. The resulting O-mutMjtRNATyr(CUA) supported survival on ampicillin in the b-lactamase suppression assay at an IC50 of only 12.4 mg mL-1, and was still acylated sufficiently by MjTyrRS to support survival at an IC50 of 436 mg mL-1 ampicillin. This pair was successfully used in later evolution of a number of synthetases with unnatural amino acid specificities (see last section).

Fig. 3 A general selection for suppressor tRNAs that are poor substrates for the E. coli synthetases and charged efficiently by a cognate synthetase of interest

Later we found that O-SctRNA^Asp(CUA), a derivative from yeast tRNA^Asp, together with a yeast AspRS mutant (Asp188-Lys) constituted another orthogonal pair in E. coli, albeit with weak activity. Expression levels of the synthetase and tRNA were increased and an RF-1 deficient strain of E. coli was employed to make this pair useful.

Methods for Generating Aminoacyl-tRNA synthetases With Altered Amino Acid Specificities

number of different strategies are being developed to generate aminoacyl synthetases with altered specifities. These include:

(i) in vivo selections based on the suppression of amber codons in b- lactamase (positive selection) and barnase (negative selection);
(ii)screens based on suppression of an OmpA epitope or a GFP based reporter system; and
(iii)phage based screens for synthetases with altered amino acid specificity.

We have developed a selection for mutant aminoacyl-tRNA synthetases capable of charging any ribosomally accepted molecule, including unnatural amino acids, a-hydroxy acids, and b-amino acids onto an orthogonal suppressor tRNA. The paradox of selecting for acceptance of an unnatural substrate which is not required for any cellular function was resolved by using both positive and negative selections. In the positive selection, an amber nonsense codon is placed in b-lactamase at a nonessential position (i.e., one which tolerates any amino acid but lies before one or more critical residues). Cells expressing mutant synthetase enzymes are selected in the presence of a library of unnatural amino acids added to the growth media for their ability to insert any natural or unnatural amino acid in response to this amber codon. The vast majority of survivors express synthetase enzymes which maintain their ability to charge cognate natural amino acids onto the suppressor tRNA. Rarely, a mutant synthetase enzyme may emerge which acylates one or more unnatural amino acids onto the orthogonal suppressor (Figure 4). All cells surviving the positive selection are then subjected to a negative selection in which an amber nonsense codon is placed in barnase at a nonessential position. Cells expressing mutant synthetases are now selected for their ability to survive in the absence of unnatural amino acids. Those cells expressing synthetase enzymes capable of charging natural amino acids onto the orthogonal suppressor will produce full-length toxic protein, inhibiting or terminating cell growth (Figure 4). Any cell that expresses a mutant synthetase capable of exclusively acylating a unnatural, but not natural, amino acid onto the orthogonal tRNA will survive both selections. Cells harboring synthetases with mixed substrate specificities will experience growth advantages proportional to their ability to charge unnatural amino acids and inversely proportional to their ability to charge natural amino acids.

Fig. 4. A general selection for mutant aaRS enzymes that charge any ribosomally accepted small molecule onto an orthogonal tRNA.

Several selectants have been isolated after a few rounds of selection whose growth in the b-lactamase assay is profoundly dependent upon the presence of the mutant YGlnRS, YO-tRNA and one or more unnatural amino acids. Of these, two in particular have been characterized with respect their ability to insert unnatural amino acids at a specific position in a protein. Preliminary LCMS results indicate that approximately 0.2% of the DHFR N-terminal peptide isolated from expression with these synthetases contains the unnatural amino acid S-ethyl-cysteine. Control experiments with the wild-type YGlnRS grown with and without S-ethyl-cysteine have no detectable S-ethyl-containing-containing peptide in the resultant DHFR. This suggests that the selection in its current form is highly sensitive to even extremely low levels of unnatural amino acid insertion.

Two of the most significant problems that have arisen out of controls and selections are slow elimination of wild-type glutamine-inserting activity and reversion of the amber mutation in the b-lactamase. To solve these problems, we have generated mutants of YGlnRS with alanines at critical residues that bind the glutamine side-chain. These inactive mutants are then the starting point for either direct shuffling with randomization by addition of oligonucleotides corresponding to these residues, or shuffling preceded by PCR-based direct randomization of these residues. In order to accommodate the wide array of amino acids we have chosen, we are extensively randomizing the active site at nearly all the residues that bind the glutamine side-chain. Another strategy, however, is to choose a more conservative analog of the original cognate amino acid, and use less extensive mutagenesis as guided by the crystal structure. For example D291 (D66 in E. coli) binds the a-amine of glutamine. We are attempting to insert the a-hydroxy analog of glutamine by randomization of this residue and more conservative mutation of the rest of the active site (since the side chain is the same). Similarly, N-substituted derivatives of the carboxamide, g-substituted derivatives and N-Cg cyclic derivatives of glutamine can be used with other concentrated libraries. Ideally, we would like to identify analogs that are as similar to glutamine as possible but neither toxic to the cell nor accepted by the E. coli GlnRS or YGlnRS.

Another approach for selecting the mutant synthetase capable of acylating a suppressor tRNA with a particular unnatural amino acid involves cell-surface display. Cell-surface display relies on the expression of a particular epitope, in this case a poliovirus C3 peptide fused to an outer membrane porin OmpA, on the surface of the E. coli cell. The C3 epitope is displayed on the cell surface only when a stop codon in the C3 message is suppressed during translation. The displayed C3 peptide then contains the amino acid recognized by one of the mutant aminoacyl-tRNA synthetases in the library, and the cell containing the corresponding synthetase gene can be isolated with antibodies raised against C3 peptides containing specific unnatural amino acids.

In addition to the methods outlined above, we are developing a sensitive, high-throughput screening system, based on suppression of a green fluorescent protein (GFP) reporter system, to evolve aminoacyl-tRNA synthetases that can selectively acylate an orthogonal suppressor tRNA with an unnatural amino acid in E. coli. Using fluorescence-activated cell sorting (FACS), yeast glutaminyl-tRNA synthetase variants that are present within an initial library of ~108 will be selected on the basis of their ability to aminoacylate a coexpressed orthogonal suppressor tRNA derived from the S. cerevisiae tRNAGln with an unnatural amino acid. In order to evolve aminoacyl-tRNA synthetases that accept as substrates unnatural but not natural amino acids, it will be necessary to apply selection pressure not only for acylation with unnatural amino acids but also for the lack of acylation with natural amino acids. This will be accomplished by an iterative positive/negative selection strategy in which cells will first be sorted after growth in the presence of an unnatural amino acid and then again after growth in the absence of unnatural amino acids (Fig. 5B). Cells grown in the presence of unnatural amino acids will be selected only if they have a fluorescence intensity above a particular value, while those grown in the absence of unnatural amino acids will be selected only if they have a fluorescence intensity below a particular value. Following each round of positive/negative selection, selected synthetases will be collected and shuffled to produce a new library for beginning the next round of selection.

Fig. 5. FACS-based evolution strategy. (A) Plasmids for expression of aaRS library and O-tRNA (library plasmid) and for the T7 RNA polymerase/GFP reporter system (reporter plasmid). "TAG" corresponds to suppressible amber stop codons. (B) Positive/negative selection scheme. Gray and green circles correspond to fluorescing and non-fluorescing cells, respectively.

Mutant synthetases isolated using any of the above methods will be analyzed with respect to their kinetics (kcat, KM) and specificity and will be structurally characterized to determine the molecular basis for the altered specificity in collaboration with Ray Stevens (TSRI)

An alternative strategy towards aminoacyl-tRNA synthetases with altered amino acid specificities is based on selection for binding to unnatural amino acids. We are currently displaying yeast glutaminyl- and aspartyl-tRNA synthetase and M. jannaschii tyrosyl-tRNA synthetase on bacteriophage M13 and T7. Displayed protein libraries will be sorted by binding to unnatural amino acids and aminoacyl-adenylate analogues immobilized on solid support. After several rounds of selection and amplification, selected binders will be tested in vitro and in vivo for charging the corresponding suppressor tRNAs.

Developing tRNAs to Suppress Four and Five Base Codons

Fig 1. Expansion of the genetic code with 4-base codon suppressing tRNAs

In addition to tRNAs isolated from Salmonella, tRNAs generated synthetically have been found to suppress +1 frameshift mutations by decoding four bases at a time, instead of the canonical three bases. Indeed, using in vitro translation systems, both Hardesty and Sisido have independently created tRNAs with extended anticodon loops that suppress 1-base insertions after relatively rare codons (like arginine's AGG or the amber stop codon TAG).

In an effort to (1) improve the suppression efficiencies of these systems, (2) expand the genetic code with many more than a single codon capable of suppression and (3) develop further tRNAs for the in vivo unnatural amino acid mutagenesis methodology, we searched for the best 4-base codon and suppressor tRNAs with library methods. Using a b-lactamase reporter selection (with a library of +1 frameshift mutants wherein a single codon of b-lactamase was replaced with NNNN), we selected from among tRNAs with randomized, extended anticodon loops.

Our selectants included both known and novel suppressible 4-base codons and resulted in a set of very efficient, non-cross-reactive tRNA/4-base codon pairs for AGGA, UAGA, CCCU and CUAG. The most efficient 4-base codon suppressors had Watson-Crick complementary anticodons, and the sequences of the anticodon loops outside of the anticodons varied with the anticodon. Additionally, 4-base codon reporter libraries were used to identify "shifty"; sites at which +1 frameshifting is most favorable in the absence of suppressor tRNAs in E. coli.

We next set our sites on larger codons. Since tRNAs with 8 nt anticodon loops were the best suppressors of 4-base codons, we examined the suppression of 2-, 3-, 4-, 5- and 6-base codons with tRNAs containing 6 to 10 nt in their anticodon loops. We found that the E. coli translational machinery tolerates codons of 3 to 5 bases, and that tRNAs with 6 to 10 nt anticodon loops can suppress these codons. However, N-length codons were found to prefer N + 4 length anticodon loops, and sequence preferences in the loops were evident, including the requirement of Watson-Crick complementarity to the codon.

We are currently adapting the frameshift suppression of 4-base codons for the site-specific incorporation of unnatural amino acids in vivo. We have identified an orthogonal leucyl-tRNA synthetase from an archaebacterium. We have identified orthogonal AGGA suppressor tRNAs that are substrates for the synthetase, and we are currently developing selections to identify mutant synthetases capable of charging unnatural amino acids. Ultimately, we hope to be able to incorporate two or more different unnatural amino acids into a single polypeptide using a combination of frameshift and amber supression.

Expanding the Genetic Code of E. coli

To expand the number of genetically encoded amino acids, we have developed a strategy that makes it possible to site-specifically incorporate unnatural amino acids directly into proteins in living cells. The orthogonal O-mutMjtRNATyr(CUA)/MjTyrRS was used in this study, and the amino acid specificity of the orthogonal TyrRS was altered as follows: a library of TyrRS mutants with five active site residues mutated was generated and screened. A positive selection was applied that is based on suppression of an amber stop codon at a nonessential position in the chloramphenicol acetyltransferase (CAT) gene in the presence of the unnatural amino acid. Cells surviving on chloramphenicol were then grown in the presence of chloramphenicol and in the absence of the unnatural amino acid. Those cells that did not survive were isolated from a replica plate supplemented with the unnatural amino acid. The mutant TyrRS genes were isolated from these cells, recombined in vitro by DNA shuffling, and transformed back into E. coli for further rounds of selection with increasing concentrations of chloramphenicol.

After two rounds of selection and DNA shuffling, a clone was evolved whose survival in chloramphenicol was dependent on the addition of 1mM O-methyl-L-tyrosine to the growth media. To demonstrate that the observed phenotype is due to the site-specific incorporation of O-methyl-L-tyrosine by the O-mutMjtRNATyr(CUA)/mutant TyrRS pair in response to an amber stop codon, an O-methyl-L-tyrosine mutant of dihydrofolate reductase (DHFR) was generated and characterized. The third codon of the E. coli DHFR gene was mutated to TAG. When the mutant TyrRS was expressed in the presence of O-mutMjtRNA^Tyr(CUA) and 1 mM O-methyl-L-tyrosine, full length DHFR was produced. In the absence of either O-methyl-L-tyrosine, O-mutMjtRNA^Tyr(CUA) or mutant TyrRS, no DHFR (<; 0.1% by densitometry) was observed by analysis with SDS-PAGE and silver staining. The identity of the amino acid inserted in response to the TAG codon was confirmed to be O-methyl-L-tyrosine by mass analysis of both the intact protein and tryptic fragments. No indication of the incorporation of tyrosine or other amino acids at that position was observed. Analysis of the sequence of the mutant TyrRS revealed the following mutations: Tyr32-;Gln, Asp158-Ala, Glu107-Thr, and Leu162-;Pro. Kinetics of adenylate formation of O-methyl-L-tyrosine and tyrosine with ATP catalyzed by the mutant TyrRS were analyzed in vitro usinga pyrophosphate-exchange assay. The value of k_cat/K_m of the mutant TyrRS for O-methyl-L-tyrosine is about 100 fold higher than that of tyrosine.

Contacts:

• John Christopher Anderson
  
• Jason Chin
  
• Shigeo Matsuda
  
• Eric Meggers
  
• Ryan "Luke Warm" Mehl
  
• Steve "Stone Cold" Santoro
  
• Lei Wang
  
• Jianming Xie
  
• Jonathan Zhang
  
• Nicole Zimmermann