Masumi Taki

A non‐natural amino acid for efficient incorporation into proteins as a sensitive fluorescent probe

A small and highly fluorescent non‐natural amino acid that contains an anthraniloyl group (atnDap) was incorporated into various positions of streptavidin. The positions were directed by a CGGG/CCCG four‐base codon/anticodon pair. The non‐natural mutants were obtained in excellent yields and some of them retained strong biotin‐binding activity. The fluorescence wavelength as well as the intensity of the anthraniloyl group at position 120 were sensitive to biotin binding. These unique properties indicate that the atnDap is the most suitable non‐natural amino acid for a position‐specific fluorescent labeling of proteins that is highly sensitive to microenvironmental changes.

Fission yeast Ubr1 ubiquitin ligase influences the oxidative stress response via degradation of active Pap1 bZIP transcription factor in the nucleus.

Cells adapt to oxidative stress by transcriptional activation of genes encoding antioxidants and proteins of other protective roles. A bZIP transcription factor, Pap1, plays a critical role in this process and overexpression of Pap1 confers resistance to various oxidants and drugs in fission yeast. Pap1 temporarily enters the nucleus upon oxidative stress but returns to the cytoplasm once cells adapt to the stress, suggesting that cellular localization regulates Pap1 function. We report here an additional regulatory mechanism that Ubr1 ubiquitin ligase‐dependent degradation lowered the Pap1 protein levels. ubr1 cells were causally resistant to hydrogen peroxide because of the increment of Pap1 levels. Pap1 was preferentially degraded in the nucleus where Ubr1 was consistently enriched. Proteolysis was critical to downregulate Pap1 especially when its activation persisted, as constitutively nuclear Pap1 severely inhibited growth in ubr1 mutants. Inactive mutations in the bZIP DNA binding domain stabilized Pap1 but rescued the lethality caused by constitutively active Pap1 in ubr1 mutants. These findings indicate that either nuclear export or Ubr1‐mediated proteolysis must be operative to prevent uncontrolled Pap1 function. Coincidental dysfunction in both inhibitory pathways causes lethality because of prolonged activation of Pap1. Ubr1 is a critical regulator for the homeostasis of oxidative stress response.

Practical Tips for Construction of Custom Peptide Libraries and Affinity Selection by Using Commercially Available Phage Display Cloning Systems

Phage display technology is undoubtedly a powerful tool for affinity selection of target-specific peptide. Commercially available premade phage libraries allow us to take screening in the easiest way. On the other hand, construction of a custom phage library seems to be inaccessible, because several practical tips are absent in instructions. This paper focuses on what should be born in mind for beginners using commercially available cloning kits (Ph.D. with type 3 vector and T7Select systems for M13 and T7 phage, respectively). In the M13 system, Pro or a basic amino acid (especially, Arg) should be avoided at the N-terminus of peptide fused to gp3. In both systems, peptides containing odd number(s) of Cys should be designed with caution. Also, DNA sequencing of a constructed library before biopanning is highly recommended for finding unexpected bias.

Synthesis of a cyclic peptide/protein using the NEXT-A reaction followed by cyclization

By using the NEXT-A reaction, we introduced a non-natural amino acid at the N-terminus of a peptide/protein that contained a cysteine unit. The side chain of the introduced amino acid spontaneously reacted with the cysteine to afford a cyclic peptide/protein.

Expanding the genetic code in a mammalian cell line by the introduction of four-base codon/anticodon pairs.

Position-specific incorporation of non-natural amino acids into proteins by using orthogonal four-base codon/anticodon pairs in vitro allowed us to introduce special functions into threedimensional polypeptide frameworks. 3] Extension of this methodology to cellular systems will add a new dimension to the application of non-natural amino-acid mutagenesis. Expression of non-natural mutants in mammalian cells is advantageous, because the proteins are expected to fold correctly under physiological conditions and undergo post-translational modifications. In this work, we employed Chinese hamster ovary (CHO) cells, which are commonly used for expressing recombinant proteins for medicinal applications, 9] because they provide proper mammalian post-translation modifications. For the position-specific incorporation of non-natural amino acids in mammalian cells, three key steps must be achieved. First, one must design a tRNA that cannot be aminoacylated by any aminoacyl-tRNA synthetases (ARSs) in the cell, but can work as the amino-acid carrier once it is aminoacylated by some other means. Second, aminoacylation of the orthogonal tRNA with an non-natural amino acid must be carried out inside a cell or the aminoacylated tRNA must be imported into the cell. Third, codon/anticodon pairs must be expanded for assigning an non-natural amino acid. In short, when an non-natural amino acid is linked to an orthogonal tRNA, it should be introduced into proteins at positions directed by the expanded codon/ anticodon pairs by ribosomal-mediated synthesis. Orthogonal tRNAs have been selected either as tRNA/ARS pairs or just as single tRNAs. Aminoacylation with non-natural amino acids has been successfully carried out by mutant ARSs in E. coli [12] or mammalian cells. Non-natural aminoacylation has also been successful in vitro by using ribozymes or peptide nucleic acids (PNA) linked with an amino-acid thioester were used. Although not actually achieved yet due to some technical limitations, ribozyme and PNA methods will also probably work in cells. Compared with the former two methods, studies on the expansion of codon/anticodon pairs in cells are still in an early stage. Codon expansion has been achieved by using amber (UAG), opal 21] (UGA), and ochre 22] (UAA) stop codons with the corresponding suppressor tRNAs in prokaryotic and mammalian cell lines. Here, we demonstrate for the first time the expansion of the genetic code in mammalian cells using four-base codons. The advantages of four-base codons rather than stop codons for assigning non-natural amino acids have been demonstrated with the E. coli S30 in vitro protein-synthesis system. In particular, we have introduced two independent non-natural amino acids into single proteins by using several orthogonal sets of four-base codons; these include (CGGG, AGGU), (CGGG, GGGU), and (CGGG, GCCC). However, since these orthogonal sets were established in the E. coli system, we had to search for new sets of four-base codons that could work in mammalian cell lines. In this study, we examined the efficiencies of UAGN (N=A, U, G, and C) four-base codons in CHO cells. Since the UAGN codons are derived from the amber-stop codon, competition with three-base sense codons is avoided, and more importantly, lethal damage to the cell can be minimized. Another advantage of the UAGN four-base codons rather than the UAG codon is that we can avoid production of readthrough proteins that do not contain non-natural amino acids. Stop-codon readthrough usually takes place when the stop codon is misread by some near-cognate tRNAs; this allows the synthesis of a full-length protein that does not contain an non-natural amino acid. Readthrough is commonly observed in eurkaryotes : up to ~2% of stops-codon encounters result in a readthrough in mouse cells, ~5% in plant, and ~55% in yeast cells. When we use the UAGN codons instead of a UAG codon, readthrough with some endogenous tRNAs causes a frameshift that produces a nonsense polypeptide that is often truncated by an encounter with one of the stop codons. As a consequence, the full-length target protein is produced only when the UAGN codons are successfully translated to an non-natural amino acid. We manipulated the gene for enhanced green fluorescent protein (EGFP) to contain the UAGN four-base codons at Tyr66 (UAGN). We also constructed tRNAncua genes that contained the ncua (n=a, u, g, and c) four-base anticodons from human tRNA. These genes were coexpressed in CHO cells, and the decoding efficiency of the UAGN codon by tyrosyl–tRNAncua was evaluated from the fluorescence intensity of full-length EGFP. We also searched for new sets of four-base codon/anticodon pairs that were derived from rarely used three-base codons in mammalian cells. The reporter system contained a gene for EGFP that carried one of the TAGN four-base codons instead of a conserved TAC codon for Tyr66. It is known that the phenolate anion of Tyr66 plays a crucial role in green fluorescence emission; replacement of this amino acid by other ones either results in a complete loss of emission or changes the emission maximum so that it lies outside the monitored wavelengths. Thus the intensity of EGFP emission provides a simple measure of the decoding efficiency of UAGN four-base codons by tyrosinated tRNAncua molecules. Recombination of the CHO genome with an EGFP66TAGN vector is needed for strict regulation of mutant EGFP mRNA expression levels. In order to obtain successful recombination and avoid the problems encountered by random insertion of [a] Prof. Dr. M. Taki, J. Matsushita, Prof. Dr. M. Sisido Department of Bioscience and Biotechnology Faculty of Engineering, Okayama University 3-1-1 Tsushimanaka, Okayama 700-8530 (Japan) Fax: (+81)86-251-8219 E-mail : sisido@cc.okayama-u.ac.jp Supporting information for this article is available on the WWW under http://www.chembiochem.org or from the author.

Chemoenzymatic Transfer of Fluorescent Non-natural Amino Acids to the N Terminus of a Protein/Peptide

Leucyl/phenylalanyl-tRNA–protein transferase (L/F transferase) from E. coli is known to catalyze the transfer of hydrophobic amino acids from an aminoacyl-tRNA to the N terminus of proteins that possess lysine or arginine as the N-terminal residue. We have reported that substrates of the enzymatic reaction can be expanded to include non-natural amino acids by using a tRNA aminoacylated with various types of non-natural amino acids. 5] This should become a new and versatile tool for N-terminal-specific incorporation of specialty amino acids into proteins and peptides. However, it is difficult to transfer large non-natural amino acids to the N terminus by using wild-type L/F transferase. Here, we engineered L/F transferase to allow the incorporation of large fluorescent nonnatural amino acids. Another possible problem with this technique was the tedious and costly transcription process of a tRNA that lacks the 3’-terminal CpA unit. To solve this problem, we minimized the tRNA structure as substrate for the transferase and found that even micro-RNAs can mediate aminoacyl transfer. Recently, the crystal structure of wild-type L/F transferase complexed with an aminoacyl-tRNA analogue (puromycin) was reported (PDB ID: 2DPT). Judging from the crystal structure, we postulated that there are steric conflicts between large amino acids on a bound aminoacyl-tRNA and hydrophobic residues in the binding pocket of the wild-type transferase. To expand the substrate specificity so as to accommodate nonnatural amino acids with large side groups, we designed and expressed four different mutant transferases (M144A, F173A, F177A, I185A) in which hydrophobic amino acids in the binding pockets were replaced by alanines to minimize the steric conflicts with amino acid side groups bound to aminoacyltRNAs. By using these mutant transferases, successful transfers were observed for fluorescent non-natural amino acids that possessed acridonyl and benzoacridonyl groups (Figure 1A and B). Acrydonylalanine (acdAla; 1) and benzoacrydonylalanine (badAla; 2) are known to be highly fluorescent and photodurable when excited with a blue-laser at the wavelength of about 405 nm or with a visible laser at approximately 450– 500 nm, respectively. Thus the attachment of these fluorescent amino acids should enable the visualization of peptides and proteins. tRNAs charged with these fluorescent amino acids at the 3’ ends were prepared according to the chemical aminoacylation method developed by Hecht and co-workers. Key intermediates for chemical aminoacylation (aminoacyl pdCpAs) were synthesized as described before. The resulting aminoACHTUNGTRENNUNGacyl-tRNA was added to a reaction mixture that contained a target peptide (H-KRRPPGFSPFR-OH=Lys-bradykinin) and each mutant transferase. Products of the aminoacyl transfer to the peptide were identified by MALDI-TOF mass spectroscopy. As described in the Supporting Information, all mutant transferases successfully transferred the large fluorescent amino acids to Lys-bradykinin. Aminoacyl transfer to a target protein (Lys-SoCBM13) was also successful. Fluorescence images of SDS-PAGE experiments are shown in Figure 1. The images indicate that the fluorescent amino acids were transferred to Lys-SoCBM13 when the reaction was performed with some of the mutant transferases. The product of the aminoacyl transfer to the Lys-SoCBM13 protein was also identified by MALDI-TOF mass spectroscopy. The average mass of the substrate Lys-SoCBM13 (m/z found 17000) was found to be shifted by 313 (m/z found 17313). This shift corresponds to the mass of a single badAla unit (314). Interestingly, amino acid preferences were somewhat different for different mutant enzymes: the F173A mutant preferred badAla rather than acdAla (lane 3), whereas acdAla was favored by I185A in preference to badAla (lane 5). After the enzymatic reaction, N-terminal sequences of resultant proteins were analyzed by Edman degradation by using an amino-acid sequencer. The efficiency of aminoacyl transfer of acdAla with I185A was 22%, and that of badAla with F173A was 41%, under optimized conditions. 5] Changing the molar ratio of the reagents or increasing the reaction time did not enhance the efficiency. Favorable mutation positions seem to depend on the side-chain structures of amino-acid substrates. For example, p-aminophenylalanine was more efficiently introduced to Lys-bradykinin by the I185A mutant than either by the F173A mutant or wild-type transferase. In the case of 6-methoxy-2-naphthylalanine, successful introduction was observed with the F173A mutant, but not with the I185A mutant or wild-type transferase. Next, we made an attempt to minimize the RNA structures as carriers of the amino acid. A previous report has suggested that the tRNA acceptor stem region is important for substrate recognition. We isolated the acceptor stem region of E. coli tRNA and aminoacylated the resulting microhelix with 1naphthylalanine (napAla). The RNA was chemically synthesized (Japan Bioservice, Saitama, Japan) and charged with napAla at [a] Prof. Dr. M. Taki, H. Kuroiwa, Prof. Dr. M. Sisido Department of Bioscience and Biotechnology Faculty of Engineering, Okayama University 3-1-1 Tsushimanaka, Okayama 700-8530 (Japan) Fax: (+81)86-251-8219 E-mail : taki@biotech.okayama-u.ac.jp sisido@cc.okayama-u.ac.jp [b] Prof. Dr. M. Taki Division of Biology, 147-75 California Institute of Technology (CALTECH) 1200 E. California Boulevard, Pasadena, CA 91125 (USA) E-mail : taki@caltech.edu Supporting information for this article is available on the WWW under http://www.chembiochem.org or from the author.

N-Terminal Specific Point-Immobilization of Active Proteins by the One-Pot NEXT-A Method

Immobilization of proteins onto surfaces of gels or solid supports is of great importance in many applications, such as drug screening and medical diagnostics. 2] Proteins are most commonly immobilized onto solid supports with amineor thiol-reactive reagents. However, the resulting immobilized proteins become heterogeneous in their link positions and spatial orientations, which can change or even cause the loss of their native activities. Optimization of immobilization conditions is time consuming and often unsuccessful. This problem is more serious when large numbers of different proteins have to be immobilized systematically. Site-specific covalent immobilization is an alternative approach. It has been reported that a variety of functional groups can be introduced at a tag or C terminus of a protein in a chemoselective or chemoenzymatic manner. 3, 4] In contrast to tag or C-terminal specific immobilization, N-terminal specific immobilization is rarely reported. We have recently reported that non-natural amino acids possessing functional groups or fluorophores can be enzymatically introduced at the basic N-terminal position of various proteins or peptides. The enzyme L/F-transferase is known to catalyze the transfer of hydrophobic amino acids from an aminoacyl-tRNA to the N terminus. 10] We combined this L/F-transferase-mediated functionalization with tRNA-aminoacylation using engineered ARS in the same test tube and named the whole method NEXT-A (N-terminal extension of protein by transferase and aminoacyl-tRNA synthetase). The advantages of the NEXT-A method over conventional chemical or chemoenzymatic introduction of functional groups are as follows. It can be conducted in a very fast, mild, and perfectly N-terminal specific manner. It allows us to introduce various kinds of amino acids including functional groups and fluorophores. This method also works in the presence of other proteins or even in crude protein mixtures, as in the case of the previously reported L/Ftransferase system. When a reactive group is introduced at the N terminus of a protein by the NEXT-A method, it must be bioorthogonal to other functional groups on the same protein, but must react with another bioorthogonal counterpart on a solid support. Among the bioorthogonal reactions reported so far, Cu-catalyzed 1,3-dipolar Huisgen-type cycloaddition—also called click chemistry—is most frequently employed to immobilize proteins onto solid supports because of its effectiveness and simplicity. However, proteins are often degraded in the presence of the copper ion, 13] and we have also been confronted with this bothersome problem. Thus, copper-free fast 1,3-diACHTUNGTRENNUNGpolar cycloaddition is desirable for protein immobilization. Here we report a practical method that combines the NEXTA reaction and noncatalytic 1,3-dipolar cycloaddition for imACHTUNGTRENNUNGmobilizing any kind of active protein onto supports with an ACHTUNGTRENNUNGordered orientation. Typically, a sugar-binding protein (lectin EW29Ch) was successfully immobilized onto an acrylamidebased gel. We monitored stoichiometry of the reaction during the immobilization. The lectin-immobilized gel was subjected to frontal affinity chromatography and sugar–lectin interACHTUNGTRENNUNGactions on the gel were measured. Direct immobilization of a target protein in a cell lysate without purification was also demonstrated. First, we compared several 1,3-dipolar cycloadditions for their reactivity and bioorthogonality. Phenylazide is known to react efficiently with DIB derivative or NOR in aqueous solution without catalyst. We carried out the NEXT-A reaction to introduce azdPhe or fazdPhe into a model peptide (Lys-AlaAMC), and then the terminal azide groups were subjected to the cycloaddition with DIB–biotin or NOR–biotin. The NEXT-A reaction was quantitative in the case of azdPhe, whereas the introduction of fazdPhe was not perfect. Judging from the HPLC and TOF-MS, both the azdPhe and fazdPhe reacted spontaneously with DIB–biotin in a quantitative yield. In contrast, NOR–biotin seldom reacted with either azdPhe or fazdPhe. Thus, the NEXT-A introduction of azdPhe and the subsequent cycloaddition with DIB were the most effective. With this optimum combination, we introduced a biotin unit to a lectin protein possessing Lys at the N terminus (Lys-EW29Ch) using DIB– biotin. To monitor the stoichiometry of the reaction, a single cysteine unit in the Lys-EW29Ch was modified by a fluorophore by using Alexa488-maleimide. Quantitative introduction of azdPhe into the Lys-EW29Ch-Alexa488 was achieved within 30 min at 37 8C, as evidenced by TOF-MS and the Edman degradation. The azdPhe-Lys-EW29Ch-Alexa488 was mixed with DIB–biotin. TOF-MS data indicated that the azdPhe-LysEW29Ch-Alexa488 disappeared and the biotinylated azdPheLys-EW29Ch-Alexa488 appeared instead. The biotinylated EW29Ch was quantitatively immobilized onto streptavidin–sepharose gel (Figure 1 A); this again indicates that the bioorthogonal biotinylation occurred spontaneously after the NEXTA reaction. In contrast, neither DIB–biotin-treated Phe-Lys[a] K. Ebisu, H. Kuroiwa, K. Kawakami, M. Ikeuchi, Prof. M. Sisido, Prof. M. Taki Department of Bioscience and Biotechnology Faculty of Engineering, Okayama University 3-1-1 Tsushimanaka, Okayama 700-8530 (Japan) Fax: (+ 81) 86-251-8021 E-mail : sisido@cc.okayama-u.ac.jp taki@biotech.okayama-u.ac.jp [b] Dr. H. Tateno, Prof. J. Hirabayashi National Institute of Advanced Industrial Science and Technology (AIST) 1-1-1 Higashi, Tsukuba Science City 305-8562 (Japan) Supporting information for this article is available on the WWW under http ://dx.doi.org/10.1002/cbic.200900430.

Specific N-Terminal Biotinylation of a Protein In Vitro by a Chemically Modified tRNAfmet can Support the Native Activity of the Translated Protein.

Biotinylation of a protein generally involves chemical modification of a translated protein. Using this methodology, however, biotinylation at a specific position remains difficult. We investigated whether it would be possible to use an Escherichia coli initiator tRNA(fmet) aminoacylated with methionine biotinylated at the alpha-amino group to introduce a biotin tag specifically at the N terminus. We report here that a biotin tag could be incorporated into the green fluorescent protein (GFP) at the N-terminal site, in the presence of an E. coli initiator tRNA(fmet) aminoacylated with methionine biotinylated at the alpha-amino group. The biotinylated GFP was purified by simple monomeric streptavidin-agarose affinity column chromatography. Based on the total amount of GFP molecules, the purification yield and the biotin labelling efficiency of this system were approximately 7% and 10-20%, respectively, according to the densitometric analysis of Western blots. Judging from the results of a fluorescence imaging experiment, almost all the purified GFP molecules retained the native fluorescence activity. Importantly, the present results support the hypothesis that the E. coli initiator tRNA(fmet) aminoacylated with a relatively large substituent can be recognized by an E. coli ribosome and adequately placed at the P site to initiate translation.

Gp10 based-thioetherification (10BASEd-T) on a displaying library peptide of bacteriophage T7

The site-specific introduction of a haloacetamide derivative into a designated cysteine on a displaying peptide on a capsid protein (gp10) of bacteriophage T7 has been achieved. This easiest gp10-based thioetherification (10BASEd-T) is carried out in one-pot without side reactions or loss of phage infectivity.

Kinetic analysis of the leucyl/phenylalanyl‐tRNA‐protein transferase with acceptor peptides possessing different N‐terminal penultimate residues

The introduction of non‐natural amino acids at the N‐terminus of peptides/proteins using leucyl/phenylalanyl‐tRNA‐protein transferase (L/F‐transferase) is a useful technique for protein engineering. To accelerate the chemoenzymatic reaction, here we systematically optimized the N‐terminal penultimate residue of the acceptor peptide. Positively charged, small, or hydrophilic amino acids at this position show positive effects for the reaction. Kinetic analysis of peptides possessing different penultimate residues suggests that the side chain of the residue affects peptide‐binding affinity towards the L/F‐transferase. These findings also provide biological insight into the effect of the penultimate amino acid on substrate specificity of natural proteins to be degraded via the N‐end rule pathway.