Creation of Functional Peptides by Evolutionary Engineering with

Chapter 11

Creation of Functional Peptides by Evolutionary Engineering with Bioorthogonal Incorporation of Artificial Components Seiichi Tada,1 Takanori Uzawa,1,2 and Yoshihiro Ito*,1,2 1Nano

Medical Engineering Laboratory, RIKEN, 2-1 Hirosawa, Wako, Saitama 351-0198, Japan 2Emergent Bioengineering Materials Research Team, RIKEN Center for Emergent Matter Science, 2-1 Hirosawa, Wako, Saitama 351-0198, Japan *E-mail: [email protected].

Molecular evolutionary engineering is a method used to harness the power of natural selection to evolve proteins or RNA with desirable properties not found in nature. The possibility of evolving a commonly existing biomolecule into a variety of functional biomolecules has now been realized in the form of aptamers through the development of in vitro selection. In addition to their high affinity and high specificity for desired targets, aptamers are easily synthesized chemically and can be modified for downstream applications. Although aptamers were originally selected from a library containing only natural components, the past decade has seen a wealth of new aptamers selected from libraries containing unnatural components to provide new aptamer functions artificially. Here we highlight this transition (the shift between selection from natural components and selection from unnatural components) and the applications of the selected aptamers. For the selection, functional molecule was attached to tRNA through amino acid and the misacylated tRNA was incorporated into in vitro selection process.

© 2015 American Chemical Society

Introduction Biomacromolecules, including peptides and proteins, are among the most effective Green Polymers because of their variety of functions provided by different pattern of sequences and biodegradability. Thus far, a large number of peptides and proteins have been utilized for chemical engineering, clinical treatment and other applications. For further improvement of their function, such as binding affinity and catalytic activity, in vitro selection method is a quite promising methodology (1). Phage display, one of the most famous methods, was developed as the first example of in vitro selection method (2), and other methods including ribosome display (3), mRNA display (4, 5) and in vitro compartment method (6) were reported. These methods have provided a number of peptides and proteins with various functions through the selection of better sequences from a library of random-sequence peptides or mutant proteins. This methodology of selection is also called as “evolutionary molecular engineering”, because of the similarity to Darwinian evolution in the natural world. Recently evolutionary molecular engineering has evolved onto the next generation of artificial peptide aptamers selected from a library including unnatural amino acids (Figure 1). Although the original in vitro selection method has a limitation on usage of only the 20 natural amino acids as the components of random peptide library, incorporation of unnatural, or artificial, amino acid into the random library was achieved by using misacylated tRNA (7, 8) and nonliving translation system (3–5). The resulting peptide aptamer with artificial component should be expected to have special functions which are not seen in natural peptides consisting of canonical amino acids.

Figure 1. Schematic image of in vitro selection of functional peptides modified with unnatural amino acid. Reproduced with permission from reference (1). Copyright 2013 The Royal Society of Chemistry. 170

In this review, first we discuss the method of peptide aptamer selection from unnatural amino acid-containing random sequence library. Subsequently, we introduce our recent studies of functional peptides, such as fluorogenicity (9), photoresponsivity (10, 11) and better inhibition effect against an enzyme (12), with artificially modified amino acids.

In Vitro Selection of Peptide Aptamer Incorporating Unnatural Amino Acid The introduction of a novel, unnatural amino acid into the desired position of a peptide sequence was accomplished by using unnatural amino acid-conjugated tRNA. Noren et al. synthesized misacylated suppressor tRNA, which recognized amber stop codon, by modifying unnatural amino acid and applied that tRNA to cell-free translation system (7). The obtained protein had the unnatural residue at amber codon position (Figure 2).

Figure 2. Incorporation of an unnatural amino acid (Uaa) with misacylated tRNAamber. Reproduced with permission from reference (1). Copyright 2013 The Royal Society of Chemistry. This technique can be utilized for peptide aptamer selection with ribosome display and mRNA display methods. In these in vitro selection methods, cell-free translation system is used for association of mRNA and translated product of peptide, therefore unnatural amino acids can be incorporated into a peptide sequence of mRNA/peptide complex by just adding unnatural amino acid-conjugating tRNA to translation system solution (3, 13). Based on this selection method, we selected various functional peptide aptamers containing unnatural residues including fluorophore, photoresponsive molecule and enzyme inhibitor, respectively (9–12). These peptide aptamers were selected by ribosome display method with each artificial component-conjugated tRNA (Figure 3). Briefly, we prepared DNA templates encoding random peptide sequences including a stop-codon (UAG) at fixed position for incorporation of 171

unnatural amino acid, but any other stop codons were removed from peptide sequence. These DNA templates were transcribed into an mRNA pool (library) and subsequently translated with cell-free translation system containing tRNA with unnatural amino acid. This translation process provided the complex of mRNA, translated peptide and ribosome because of lack of any stop codons, except UAG codon for incorporation of unnatural amino acid. We incubated this complex solution with target molecule-immobilized magnetic beads and collected high-affinity peptide complexes to the target molecule. After purification of mRNA and subsequent RT-PCR, DNA templates of high-affinity peptide sequences were amplified and used for the next selection round. We repeated this selection process several times and obtained DNA templates were analyzed their sequences. We synthesized some of selected peptide sequences and finally decided the sequence of a functional peptide aptamer by estimating the function of selected peptide candidates.

Figure 3. Selection process of ribosome display from random peptide library with unnatural amino acids.

We have reported three kinds of functional peptide aptamers with unnatural amino acids by using ribosome display and misacylated tRNA. First example is a fluorogenic peptide aptamer containing NBD fluorophore, which bound to Ca2+containing form of calmodulin molecule and increase its fluorescence intensity (9). The second example is a photoresponsive peptide aptamer containing azobenzene molecule (10, 11). This peptide changes its binding affinity to target molecules by irradiation of UV light. The third example is a “superinhibitor” peptide containing a small inhibitor molecule, which provide higher inhibiting effect than that of normal inhibitor molecule (12). We discuss these functional peptides in following parts. 172

Fluorogenic Peptide Aptamer Molecular probes that can both bind to a target and indicate a signal of the presence of the target protein have been developed for the easier detection of specific proteins. For this purpose, enzymatic degradation of a signaling substrate was used to detect specific enzymes (14–16). However, these strategies cannot be applied to any arbitrarily selected target proteins, in contrast to the strategies for creating antibodies and aptamers. Therefore we endeavored to generate a signaling peptide with the characteristics of both aptamers and signaling molecular probes. We planned to select a functional peptide aptamer, which emits fluorescence upon specific binding to an arbitrarily selected protein, from a pool of random sequence peptides containing a fluorogenic amino acid (9). We chose 7-nitro-2,1,3-benzoxadiazole (NBD; Figure 4) as the fluorogenic probe, because its fluorescence intensity increases drastically only in a hydrophobic environment. We expected that the NBD would directly contribute to the interaction between the peptide and the target protein by forming part of the molecular recognition site, and upon binding, emit fluorescence.

Figure 4. 7-nitro-2,1,3-benzoxadiazole (NBD) conjugated aminophenylalanine.

We prepared misacylated tRNA of NBD-conjugated aminophenylalanine and performed ribosome display selection of fluorogenic peptide aptamers using that tRNA. In this peptide selection, a calcium-signal transducer protein calmodulin (CaM) was used as a target molecule because its mechanism of binding to various peptides and proteins has been extensively studied (17, 18). Upon Ca2+ binding, its N- and C-terminal lobe-domains, which are connected by a central helical linker region, undergo a conformational change that exposes a hydrophobic patch for interaction with other peptides and proteins. After six rounds of selection and subsequent functional analysis, we finally obtained a fluorogenic peptide aptamer sequence C5 (YWDKIKDXIGG, where X is NBD-aminophenylalanine). C5 peptide synthesized with solid phase method showed specific binding to Ca2+-bound CaM and significant increase of fluorescence signal from NBD, whereas in the absence of Ca2+ the fluorescence intensity increased only minimally (Figure 5). Based on the fluorescence titration, the dissociation constant (Kd) was determined to be 1.03 ± 0.16 µM, which is supported by an observation using surface plasmon resonance (850 nM). 173

Figure 5. Fluorescence change of C5 peptide upon specific binding to Ca2+-bound CaM. (A) The fluorescence intensity of the C5 peptide increased upon addition of increasing concentrations of CaM while in the presence of Ca2+. (B) The relative fluorescence intensity change at 535 nm clearly shows that the C5 peptide only emits fluorescence in the presence of Ca2+-bound CaM. Reproduced with permission from reference (9). Copyright 2014 The Royal Society of Chemistry.

We succeeded selection of a fluorogenic peptide against Ca2+-bound CaM using a peptide pool incorporating an artificial fluorogenic probe. To the best of our knowledge, this peptide is the first example of the successful selection of a fluorogenic peptide from a random sequence pool. This technology could prove to be useful for the development of separation-free immunoassays and bio-imaging analyses.

Photoresponsive Peptide Aptamer Artificial control of dynamic molecular recognition is one of the important and promising research fields (19–21). Control by light irradiation has been utilized as a dominant external stimulus against dynamic molecular recognition (22, 23), Although many photoresponsive molecules have been reported, such as photoresponsive crown ethers (24, 25), a photo-chemically driven molecular machine (26), a light-powered molecular pedal (27) and so on, these examples were based on a rigorous molecular design (commonly termed “rational design”). Therefore the design of the photo-responsive host molecule for arbitrary targets, which is structurally complicated or unknown, would be difficult and was not achieved. 174

We proposed an approach to achieve the photocontrol of dynamic molecular recognition with an arbitrary target by using an in vitro selection method. We isolated a photo-responsive peptide aptamer which recognizes a target by using ribosome display (10). An azobenzene molecule was chosen for application of photoresponsive property to peptide aptamer and coupled with aminoacyl tRNA. This azobenzeneconjugated tRNA was used for ribosome display selection of a photoresponsive peptide aptamer binding to a target material, streptavidin-conjugated magnetic beads. We first incubated mRNA-peptide complex library containing azobenzene with streptavidin beads for 30 min under dark conditions, in which almost azobenzene took trans-form. Subsequently, the beads were irradiated with ultraviolet light for 10 min. In this step, peptides with cis-form azobenzenes can be released from the beads, since conformational changes of the peptides by the trans–cis isomerization of azobenzenes are accompanied by the changes in affinity to streptavidin. After five rounds of these selection processes, finally a high binding affinity sequence LA81 (GVTXRRFIXYV, where X is azobenzene-aminophenylalanine) was obtained. This peptide sequence showed better binding affinity to streptavidin beads than other selected peptide sequences (Figure 6A). The dissociation constant of the LA81 peptide bound to the microbead was calculated to be 6.31 µM by curve-fitting to a Langmuir isotherm. In addition, LA81 adsorbed onto the microbead under visible light irradiation but this adsorption was significantly reduced by UV irradiation (Figure 6B). In contrast, a scrambled sequence peptide (control) did not bind and showed no photo-responsiveness. In order to confirm the wide utility of this selection strategy to various target molecules, we selected another photoresponsive peptide aptamer, which bound to different target material, glutathione-immobilized microbeads (11). After eight rounds of ribosome display selection process using azobenzene-carrying aminoacyl tRNA and glutathione beads, two peptide sequences B09 and B69 (RNGXSSGRHGD and KDGXGGEEGET, respectively, where X is azobenzene-aminophenylalanine) were chosen as photoresponsive peptide aptamers. Both of these peptides bound to the target beads, but B09 showed higher affinity compared to B69 peptide (Figure 7A). The dissociation constants of the B09 and B69 peptides bound to the microbeads were calculated as 5.21 and 1.19 µM, respectively, by curve fitting to a Langmuir isotherm. One of the peptides, B09, adsorbed onto the microbeads under visible light irradiation but this adsorption was significantly reduced by UV irradiation, although the B69 peptide did not show significant photoresponsive differences in adsorption behavior (Figure 7B). This less photoresponsiveness of the B69 peptide is likely caused by a lack of some components in the ribosome display. Additionally, streptavidin beads-binding peptide LA81 showed larger affinity change than glutathione beads-binding peptide B09. This difference seemed to be due to the number of azobenzene molecule incorporated to peptide sequence. LA81 carried two azobenzene molecules in its sequence, resulting in larger comformational change after UV irradiation.

175

Figure 6. Binding analysis on selected photoresponsive peptides binding to streptavidin beads. (A) Binding of peptides containing LA37, LA40 and LA81 sequences (VLIMVAVXASS, HSCXVTIDVFF and GVTXRRFIXYV, respectively, where X represents azobenzene-aminophenylalanine). The bound microbead was stained with an FITC-labeled anti-FLAG-antibody and its fluorescent intensity was quantified by a microplate reader. (B) Photo-responsive binding of peptides containing LA81 or a scrambled sequence. White bar: under the visible light irradiation conditions; grey bar: under the UV irradiation conditions. Reproduced with permission from reference (10). Copyright 2012 The Royal Society of Chemistry.

We succeeded the selection of photo-responsive peptide aptamers against two kinds of target materials using in vitro selection combined with photo-manipulation. The photo-responsive peptide sequence will be available as a protein-tag for the purification of biologically significant protein. Moreover, the photo-responsive peptide that binding nanomaterials, such as carbon nanotube, could also be applicable in particular photo-switching nanodevices.

Superinhibitor Consisting of Small Molecule and Peptide Aptamer The discovery and development of selective inhibitors against protein kinase is highly demanded for the treatment of cancer and a number of neurological, immunological, metabolic, and infectious diseases (28, 29). Threfore, direct in vitro selection of peptides with inhibitory activity has been reported (30–32). Especially, Li and Roberts introduced penicillin molecule into a mRNA-display peptide library by post-translational modification and isolated peptide-drug conjugates with at least 100-fold higher activity against Staphylococcus aureus penicillin-binding protein 2a (32). This strategy, however, requires chemical reactions after in vitro translation and may produce unwanted by-products. We chose the small molecule purvalanol B (PVB), which inhibits cyclin-dependent kinase 2 (CDK2) with nanomolar IC50 values, as a component of peptide-based inhibitor because the interaction between PVB and CDK2 has been well studied, and the carboxylic acid of the 6-anilino substituent of PVB can be modified without affecting the inhibitor-kinase interaction (12). 176

Therefore, we coupled the carboxylic acid of PVB with aminophenylalanine and charged the product to suppressor tRNA for incorporation of PVB into a random-sequence peptide library during ribosome display selection. We aimed to isolate a peptide–drug conjugation with higher inhibitory activity due to increase of the interaction sites (Figure 8).

Figure 7. Binding experiments on selected photoresponsive peptides binding to glutathione beads. (A) Binding of peptides containing B09 and B69 sequences and fluorescein at N-terminus. The amount of bound peptide to the microbeads was estimated by measuring the fluorescent intensity with a microplate reader. (B) Photoresponsive binding of peptides containing either the B09 or B69 sequence. Reproduced with permission from reference (11). Copyright 2014 Elsevier.

Figure 8. Principle of enhancement of the inhibitory activity of a small molecule by a peptide. (A) Small molecule interacts with a protein. (B) Peptide–small molecule conjugate interacts with the protein. 177

After six iterative rounds of selection process, a peptide sequence A5 (SKLXRFTGCSC, where X is PVB-aminophenylalanine) was found as an effective inhibitor peptide. We examined the inhibitory activity of A5 and PVB as 15 well as control peptide, whose PVB residue was replaced to phenylalanine, using a fluorescence resonance energy transfer-based Z0-LYTE assay kit. A5 inhibited CDK2/cyclin A with an IC50 of 34 nM, while single PVB molecule was a less potent inhibitor of CDK2/cyclin A, with an IC50 value of 263 nM. As we expected, control peptide did not exhibit inhibitory activity (Figure 9). Because PVB-aminophenylalanine exhibited nearly the same inhibitory activity as PVB, the higher inhibitory activity of A5 compared with PVB arises from the peptide. Thus, we also succeeded in developing novel inhibitor peptide with higher inhibitory effect by using unnatural amino acid-modified ribosome display system.

Figure 9. Inhibitory activities of A5 peptide (circles), PVB (squares), and control peptide (triangles) against CDK2/cyclin A. The results indicate that the peptide enhances the inhibitory effect of PVB against CDK2/cyclin A. Reproduced with permission from reference (12). Copyright 2014 The Royal Society of Chemistry.

Concluding Remarks and Future Outlook In this article, we described our recent studies involving the selection of functional peptide aptamers carrying unnatural residues. These results indicated that the function of peptide aptamers could be expanded by introducing artificial, small molecules, such as fluorophores, photoresponsive module, and inhibitors into the random peptide library during in vitro selection. Our functional peptides were selected without any negative selection process; therefore their functions, especially binding affinity, would be improved even more by additional negative selection. Thus, the dual use of bioorthogonal incorporation of unnatural amino acids and evolutionary molecular engineering is a highly promising technology to produce various functional peptides and green polymers. 178

References 1. 2. 3. 4. 5. 6. 7. 8. 9.

10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21.

22. 23. 24. 25.

Uzawa, T.; Tada, S.; Wang, W.; Ito, Y. Chem. Commun. 2013, 49, 1786–1795. Smith, G. P. Science 1985, 228, 1315–1317. Mattheakis, L. C.; Bhatt, R. R.; Dower, W. J. Proc. Natl. Acad. Sci. U.S.A. 1994, 91, 9022–9026. Nemoto, N.; Miyamoto-Sato, E.; Husimi, Y.; Yanagawa, H. FEBS Lett. 1997, 414, 405–408. Roberts, R. W.; Szostak, J. W. Proc. Natl. Acad. Sci. U.S.A. 1997, 94, 12297–12302. Tawfik, D. S.; Griffiths, A. D. Nat. Biotechnol. 1998, 16, 652–656. Noren, C. J.; Anthony-Cahill, S. J.; Griffith, M. C.; Schultz, P. G. Science 1989, 244, 182–188. Bain, J. D.; Glabe, C. G.; Dix, T. A.; Chamberlin, A. R.; Diala, E. S. J. Am. Chem. Soc. 1989, 111, 8013–8014. Wang, W.; Uzawa, T.; Tochio, N.; Hamatsu, J.; Hirano, Y.; Tada, S.; Saneyoshi, H.; Kigawa, T.; Hayashi, N.; Ito, Y.; Taiji, M.; Aigaki, T.; Ito, Y. Chem. Commun. 2014, 50, 2962–2964. Liu, M.; Tada, S.; Ito, M.; Abe, H.; Ito, Y. Chem. Commun. 2012, 48, 11871–11873. Tada, S.; Zang, Q.; Wang, W.; Kawamoto, M.; Liu, M.; Iwashita, M.; Uzawa, T.; Kiga, D.; Yamamura, M.; Ito, Y. J. Biosci. Bioeng. 2015 in press. Wang, W.; Hirano, Y.; Uzawa, T.; Liu, M.; Taiji, M.; Ito, Y. Med. Chem. Commun. 2014, 5, 1400–1403. Li, S.; Millward, S.; Roberts, R. J. Am. Chem. Soc. 2002, 124, 9972–9973. Perez, J. M.; Josephson, L.; O’Loughlin, T.; Hogemann, D.; Weissleder, R. Nat. Biotechnol. 1999, 17, 375–378. Wosnick, J. H.; Mello, C. M.; Swager, T. M. J. Am. Chem. Soc. 2005, 127, 3400–3405. Zhang, J.; Shibata, A.; Ito, M.; Shuto, S.; Ito, Y.; Mannervik, B.; Abe, H.; Morgenstern, R. J. Am. Chem. Soc. 2011, 133, 14109–14119. Meador, W. E.; Means, A. R.; Quiocho, F. A. Science 1992, 257, 1251–1255. Ikura, M.; Clore, G. M.; Gronenborn, A. M.; Zhu, G; Klee, C. B.; Bax, A. Science 1992, 256, 632–638. Lehn, J. M. Chem. Soc. Rev. 2007, 36, 151–160. Bellotto, S.; Chen, S.; Rebollo, I. R.; Wegner, H. A.; Heinis, C. J. Am. Chem. Soc. 2014, 136, 5880–5883. Jafari, M. R.; Deng, L.; Kitov, P. I.; Ng, S.; Matochko, W. L.; Tjhung, K. F.; Zeberoff, A.; Elias, A.; Klassen, J. S.; Derda, R. ACS Chem. Biol. 2014, 9, 443–450. Willner, I. Acc. Chem. Res. 1997, 30, 347–356. Champin, B.; Mobian, P.; Sauvage, J. P. Chem. Soc. Rev. 2007, 36, 358–366. Shinkai, S.; Ogawa, T.; Nakaji, T.; Kusano, Y.; Nanabe, O. Tetrahedron Lett. 1979, 20, 4569–4572. Shinkai, S.; Nakaji, T.; Ogawa, T.; Shigematsu, K.; Manabe, O. J. Am. Chem. Soc. 1981, 103, 111–115. 179

26. Ballardini, R.; Balzani, V.; Gandolfi, M. T.; Prodi, L.; Venturi, M.; Phillip, D.; Ricketts, H. G.; Stoddart, J. F. Angew. Chem., Int. Ed. 1993, 32, 1301–1303. 27. Muraoka, T.; Kinbara, K.; Aida, T. Nature 2006, 440, 512–515. 28. Noble, M. E. M.; Endicott, J. A.; Johnson, L. N. Science 2004, 303, 1800–1805. 29. Zhang, J. M.; Yang, P. L.; Gray, N. S. Nat. Rev. Cancer 2009, 9, 28–39. 30. Colas, P.; Cohen, B.; Jessen, T.; Grishina, I.; McCoy, J.; Brent, R. Nature 1996, 380, 548–550. 31. Ueno, S.; Yoshida, S.; Mondal, A.; Nishina, K.; Koyama, M.; Sakata, I.; Miura, K.; Hayashi, Y.; Nemoto, N.; Nishigaki, K.; Sakai, T. Proc. Natl. Acad. Sci. U.S.A. 2012, 109, 11121–11126. 32. Li, S.; Roberts, R. W. Chem. Biol. 2003, 10, 233–239.

180