CHEMBIOCHEM COMMUNICATIONS DOI: 10.1002/cbic.201402129
Pyrrolysine-Inspired Protein Cyclization Marianne M. Lee,*[a] Tomasz Fekner,[b] Jia Lu,[a] Bradley S. Heater,[a] Edward J. Behrman,[b] Liwen Zhang,[c] Pang-Hung Hsu,[d] and Michael K. Chan*[a] The pyrrolysine translational machinery has been extensively explored for the production of recombinant proteins containing a variety of “site-specific” non-canonical amino acids for both in vitro and in vivo biochemical studies. In this study, we report the first use of this technology for the production of branched cyclic proteins with a tadpole-like topology. As a proof of concept, we fused the well-studied RGD peptide to the C terminus of an mCherry reporter protein. Previous studies have shown that cyclization of the RGD peptide enhances its internalization into cells compared to its linear counterpart. The cellular uptake efficiencies of mCherry-cyclo(RGD), mCherry-linear(RGD), and wild-type mCherry determined by flow cytometry follow the trends expected, thereby confirming the feasibility and potential utility of this cyclization approach.
In the past decade, pyrrolysine (1, Pyl; Scheme 1), the 22nd genetically encoded amino acid, and its translational machinery, comprised of a pyrrolysyl-tRNA synthetase (PylRS) and its cognate tRNA (tRNAPyl), have been established as a remarkably
Scheme 1. Pyrrolysine (1) and its translational surrogates (2–3).
[a] Prof. M. M. Lee, J. Lu, B. S. Heater, Prof. M. K. Chan The School of Life Sciences, Centre of Novel Biomaterials The Chinese University of Hong Kong Shatin, New Territories, Hong Kong SAR (China) E-mail: [email protected] [email protected]
[b] Dr. T. Fekner, Prof. E. J. Behrman Department of Chemistry and Biochemistry, The Ohio State University Columbus, OH 43210 (USA) [c] Dr. L. Zhang The Proteomics Shared Resource, The Ohio State University Columbus, OH 43210 (USA) [d] Prof. P.-H. Hsu Institute of Bioscience and Biotechnology National Taiwan Ocean University Taipei 202 (Taiwan) Supporting information for this article is available on the WWW under http://dx.doi.org/10.1002/cbic.201402129.
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
flexible biochemical tool for the synthesis of recombinant proteins incorporating a wide array of non-canonical amino acid (ncAA) residues. In particular, our research group and others have developed functionalized pyrrolysine analogues for sitespecific protein labeling and biochemical studies. Recently, we reported the translational incorporation of d-Cys-e-Lys (2)[2g, 3] and showed that the resulting protein could be specifically ubiquitylated at the site of the ncAA via expressed protein ligation (EPL). We surmised that this technology could also be used to engineer other protein frameworks, including those bearing cyclic subunits. The cyclization of a polypeptide rigidifies its structure, resulting in a lower entropic cost of binding and, as a consequence, higher binding affinity and specificity. These properties, together with the greater stability of cyclic peptides against proteolysis by exopeptidases, have made them attractive leads in the development of new therapeutics.[5a, 6] One way to generate such compounds is to employ a semisynthetic approach that relies on the split intein-mediated circular ligation of peptides and proteins (SICLOPPS). In this method, an intein-derived a-thioester at the C terminus is allowed to ligate via native chemical ligation (NCL) to the amino group of the N-terminal cysteine residue. Formation of the amide bond between the two ends of the protein leads to its unbranched cyclized analogue. SICLOPPS has shown to be quite effective in the synthesis of peptide-based inhibitors, and libraries of such cyclic peptides have yielded several promising drug leads.[7b, 8] Another approach to prepare cyclic peptides is to use two cysteine residues, each flanking the region to be rendered circular to form a disulfide linkage. One advantage of this approach over the SICLOPPS method is its ability to produce cyclic peptides with protruding branches, allowing for the generation of an agent with two distinct functionalities. A classical application of this method is the fusion of a cyclized integrinbinding RGD sequence to the C terminus of a cargo protein to enhance cellular uptake. The RGD sequence is arguably one of the most popular ligands used in cellular binding studies, due in part to the fact that its molecular target, the integrin receptor, plays a critical role in many diseases, including thrombosis, atherosclerosis, and solid tumor progression and metastasis. Studies have shown that although linear RGD peptides often have low binding affinity and selectivity, cyclization of the RGD sequence via the disulfide linkage enhances its effectiveness in promoting uptake of the cargo protein into cells. Some limitations of this general cyclization strategy, however, are the lack of specificity of the thiol chemistry, particularly when the protein contains multiple cysteine residues, and the susceptibility of disulfide bonds to reduction to the ChemBioChem 2014, 15, 1769 – 1772
CHEMBIOCHEM COMMUNICATIONS linear form in the reducing cytosolic environment of the cell, thereby limiting the utility of such cyclic peptides to extracellular interactions. Herein, we report a distinct approach to produce branched cyclic peptides/proteins that overcomes the limitations of the disulfide approach. This method, which has some similarities to SICLOPPS, is based on in vivo incorporation of pyrrolysine analogue 2 into peptides/proteins followed by NCL-based macrolactamization involving the cysteine fragment of 2 and an intein-derived C-terminal a-thioester (Scheme 2). Notably, the cyclic unit generated by the NCL is formed via an isopeptide bond that is stable in the reducing environment of the cytosol. This makes it markedly different from the cyclic peptides formed via the disulfide linkage or thioether that become linear once inside the cells. To test our proposed approach, we targeted the synthesis of a cyclic RGD peptide fused to an mCherry reporter protein. As mentioned previously, several studies have found that cyclized RGD peptides generated by the cysteine-based disulfide formation exhibit higher binding affinity and selectivity than their linear counterparts. We thus sought to investigate whether a cyclized RGD peptide lasso generated by C-terminal NCL to a pyrrolysine analogue such as 2 could provide a similar enhancement in cellular uptake of the mCherry reporter protein. To prepare the mCherry-cyclic RGD protein (Scheme 2), an expression plasmid containing the genes encoding, in series, the mCherry protein, the ncAA (encoded by the UAG codon), the RGD peptide, an intein used to generate the C-terminal athioester, and a C-terminal chitin binding domain and His6-tag was created. The resulting mCherry-2-RGD-intein-CBD construct was produced by overexpression in the presence of exogenous 2. For the linear version of the mCherry-RGD in which the RGD is not amenable to cyclization, the pyrrolysine analogue 3 was used instead to provide mCherry-3-RGD-inteinCBD. Generation and subsequent cyclization of the mCherry-2RGD-a-thioester was initiated by the addition of 100 mm sodium 2-sulfanylethanesulfonate (MESNA) to the mCherry-2RGD-intein-CBD protein (1.5 mg) to induce its thiolysis. The cysteine component of 2 was then allowed to react intramolecularly with the a-thioester functionality to yield the mCherry-cyclo(2-RGD) protein. To monitor the progress of the cyclization reaction, we took advantage of the fact that only the cysteine side chain of 2 in the desired mCherry-cyclo(2-RGD) product is not susceptible to NCL. Thus, aliquots of the reaction mixture at different time points were treated with a fluorescein thioester and then assayed by SDS-PAGE and fluorimetry to monitor the reaction progress. The assay revealed that the NCL cyclization step is relatively slow. Specifically, even after 80 h, the NCL cyclization did not appear to go to completion (Figure 1). We surmise that several factors, including the ring size of the RGD peptide and the inherent hydrolysis of the a-thioester, could contribute to the slow and incomplete cyclization. Formation of the mCherry-cyclo(2-RGD) was confirmed by MALDI-TOF mass spectrometry, based on the presence of additional residues attached to the cysteine side chain of 2 (Supporting Information). Mass spectrometry analy 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chembiochem.org sis suggests that approximately 58 % of the mCherry-2-RGD is in its cyclic form, assuming a similar detection propensity for the relevant peptides. As we were able to establish that the crude product contained a mixture of cyclic and linear mCherry-2-RGD proteins, we sought an expedient method to isolate the desired cyclic product. Once again, we took advantage of the specific NCL reactivity of the cysteine residue in the linear product. By treating the reaction mixture with biotin thioester, the protein containing the linear RGD peptide could be specifically labeled with excess biotin thioester and then removed by passing the solution through the streptavidin beads (Pierce). Mass spectrometry confirmed that the resulting solution contained only mCherry-cyclo(2-RGD), based on the absence of the mass signature of the linear side product (Supporting Information). To test whether mCherry-cyclo(2-RGD) does indeed exhibit improved uptake over both its linear analogue, mCherry-3RGD, and the wild-type mCherry, we measured each of their cellular uptake efficiencies. Briefly, the proteins were incubated with MCF-7 (human breast adenocarcinoma) cells for up to 8 h. The MCF-7 cells were then washed with ice-cold phosphatebuffered saline (PBS) to facilitate their detachment from the Upcell (Nunc) plate, followed by flow cytometric analysis. At the same concentration (2 mm) of the delivered protein, the efficiency of uptake into MCF-7 cells was much higher for mCherry-cyclo(2-RGD) (5.5 %) compared to mCherry-3-RGD (1.0 %) and wild-type mCherry (0.6 %). These results verify the ability of the cyclic RGD peptides generated by our pyrrolysine
Figure 1. Fluorescein thioester labeling of mCherry-2-RGD to monitor the formation of mCherry-cyclo(2-RGD) protein. Top: SDS-PAGE gel showing the fluorescence of mCherry-2-RGD following reaction with fluorescein thioester. Bottom: Fluorescence intensity plotted as a function of time. The decrease in the fluorescence intensity of linear mCherry-2-RGD over time results from formation of mCherry-cyclo(2-RGD) that cannot react with the fluorescein thioester.
ChemBioChem 2014, 15, 1769 – 1772
Scheme 2. Preparation of mCherry-cyclo(2-RGD).
technology to exhibit similar uptake enhancement, relative to their disulfide counterparts, without concerns associated with instability of the latter compounds in the cytosol. 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
In summary, we have developed a novel application for our pyrrolysine analogue 2 for the generation of branched cyclic peptides via intramolecular NCL. We also report that such a cycChemBioChem 2014, 15, 1769 – 1772
CHEMBIOCHEM COMMUNICATIONS lization can be used to effect changes in the binding properties of proteins. Significantly, the flexibility of inserting 2 at virtually any position within a recombinant protein makes it possible to adjust both the size of the macrocycle and the length of the sidearm. Given the important role of cyclic peptides in drug development, our work should have significant implications in that it provides a new methodology to prepare cyclic peptide-containing therapeutics. In this study, we demonstrate the generation and subsequent application of a cyclic peptide to enhance the delivery of a cargo protein into cells. As the isopeptide bond generated in our method by NCL-mediated cyclization can tolerate the reducing environment of the cytosol, which ensures the maintenance of the peptide cyclic structure, one could potentially engineer a therapeutic agent comprised of a cyclic peptide inhibitor fused to an N-terminal peptide targeting the site of action. Beneficially, as 2 is composed of Cys and Lys, any therapeutic produced would be comprised solely of natural amino acids, thus reducing potential toxicity issues associated with the compound composition.
Acknowledgements This research was supported by grants from the National Institutes of Health (GM061796 to M.K.C.) and The Chinese University of Hong Kong (4053036 to M.M.L.), as well as the Dream Programme of the CUHK School of Life Sciences to J.L. The authors thank the Flow Cytometry Core Facility of the Case Comprehensive Cancer Center (P30A043703) for their technical support. Keywords: cyclic peptides · integrin · protein engineering · pyrrolysine analogues · RGD  a) B. Hao, W. Gong, T. K. Ferguson, C. M. James, J. A. Krzycki, M. K. Chan, Science 2002, 296, 1462 – 1466; b) G. Srinivasan, C. M. James, J. A. Krzycki, Science 2002, 296, 1459 – 1462; c) S. K. Blight, R. C. Larue, A. Mahapatra, D. G. Longstaff, E. Chang, G. Zhao, P. T. Kang, K. B. GreenChurch, M. K. Chan, J. A. Krzycki, Nature 2004, 431, 333 – 335.  a) C. R. Polycarpo, S. Herring, A. Berube, J. L. Wood, D. Soll, A. Ambrogelly, FEBS Lett. 2006, 580, 6695 – 6700; b) T. Mukai, T. Kobayashi, N. Hino, T. Yanagisawa, K. Sakamoto, S. Yokoyama, Biochem. Biophys. Res. Commun. 2008, 371, 818 – 822; c) T. Yanagisawa, R. Ishii, R. Fukunaga, T. Kobayashi,
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
K. Sakamoto, S. Yokoyama, Chem. Biol. 2008, 15, 1187 – 1197; d) T. Kobayashi, T. Yanagisawa, K. Sakamoto, S. Yokoyama, J Mol Biol 2009, 385, 1352 – 1360; e) T. Fekner, X. Li, M. M. Lee, M. K. Chan, Angew. Chem. 2009, 121, 1661 – 1663; Angew. Chem. Int. Ed. 2009, 48, 1633 – 1635; f) C. C. Liu, P. G. Schultz, Annu. Rev. Biochem. 2010, 79, 413 – 444; g) T. Fekner, X. Li, M. K. Chan, Eur. J. Org. Chem. 2010, 4159 – 4162; h) X. Li, T. Fekner, M. K. Chan, Chem. Asian J. 2010, 5, 1765 – 1769; i) T. Fekner, M. K. Chan, Curr. Opin. Chem. Biol. 2011, 15, 387 – 391; j) M. M. Lee, T. Fekner, T. H. Tang, L. Wang, A. H. Chan, P. H. Hsu, S. W. Au, M. K. Chan, ChemBioChem 2013, 14, 805 – 808. a) T. Fekner, X. Li, M. K. Chan, ChemBioChem 2011, 12, 21 – 33; b) X. Li, T. Fekner, J. J. Ottesen, M. K. Chan, Angew. Chem. 2009, 121, 9348 – 9351; Angew. Chem. Int. Ed. 2009, 48, 9184 – 9187. a) P. E. Dawson, T. W. Muir, I. Clark-Lewis, S. B. Kent, Science 1994, 266, 776 – 779; b) Y. Shao, S. B. Kent, Chem. Biol. 1997, 4, 187 – 194; c) T. W. Muir, Annu. Rev. Biochem. 2003, 72, 249 – 289; d) V. Muralidharan, T. W. Muir, Nat. Methods 2006, 3, 429 – 438. a) J. Clardy, C. Walsh, Nature 2004, 432, 829 – 837; b) B. N. Tse, T. M. Snyder, Y. Shen, D. R. Liu, J. Am. Chem. Soc. 2008, 130, 15611 – 15626. a) J. Matsoukas, V. Apostolopoulos, E. Lazoura, G. Deraos, M. T. Matsoukas, M. Katsara, T. Tselios, S. Deraos, Curr. Med. Chem. 2006, 13, 2221 – 2232; b) R. J. Clark, G. C. Preza, C. C. Tan, J. W. van Dijk, E. Fung, E. Nemeth, T. Ganz, D. J. Craik, Biopolymers 2013, 100, 519 – 526. a) C. P. Scott, E. Abel-Santos, M. Wall, D. C. Wahnon, S. J. Benkovic, Proc. Natl. Acad. Sci. USA 1999, 96, 13638 – 13643; b) A. Tavassoli, S. J. Benkovic, Angew. Chem. 2005, 117, 2820 – 2823; Angew. Chem. Int. Ed. 2005, 44, 2760 – 2763; c) A. Tavassoli, S. J. Benkovic, Nat. Protoc. 2007, 2, 1126 – 1133. a) L. O. Nilsson, M. Louassini, E. Abel-Santos, Protein Pept. Lett. 2005, 12, 795 – 799; b) T. A. Naumann, A. Tavassoli, S. J. Benkovic, ChemBioChem 2008, 9, 194 – 197; c) S. H. Joo, Biomol. Ther. (Seoul) 2012, 20, 19 – 26. a) M. Oba, S. Fukushima, N. Kanayama, K. Aoyagi, N. Nishiyama, H. Koyama, K. Kataoka, Bioconjugate Chem. 2007, 18, 1415 – 1423; b) Q. H. Zhou, Y. Z. You, C. Wu, Y. Huang, D. Oupicky, J. Drug Targeting 2009, 17, 364 – 373; c) Z. Chen, J. Deng, Y. Zhao, T. Tao, Int. J. Nanomed. 2012, 7, 3803 – 3811. S. Liu, Mol. Pharm. 2006, 3, 472 – 487. a) M. Fani, D. Psimadas, C. Zikos, S. Xanthopoulos, G. K. Loudos, P. Bouziotis, A. D. Varvarigou, Anticancer Res. 2006, 26, 431 – 434; b) S. Liu, Bioconjugate Chem. 2009, 20, 2199 – 2213; c) J. Shi, Y. S. Kim, S. Zhai, Z. Liu, X. Chen, S. Liu, Bioconjugate Chem. 2009, 20, 750 – 759. D. S. Yeo, R. Srinivasan, M. Uttamchandani, G. Y. Chen, Q. Zhu, S. Q. Yao, Chem. Commun. 2003, 2870 – 2871. Y. Ye, B. Xu, G. V. Nikiforovich, S. Bloch, S. Achilefu, Bioorg. Med. Chem. Lett. 2011, 21, 2116 – 2120.
Received: March 27, 2014 Published online on July 8, 2014
ChemBioChem 2014, 15, 1769 – 1772