Engineering a Catalytically Efficient Recombinant Protein Ligase

Feb 15, 2017 - Breaking and forming peptidyl bonds are fundamental biochemical reactions in protein chemistry. Unlike proteases that are abundantly ...
0 downloads 0 Views 1MB Size
Subscriber access provided by University of Newcastle, Australia

Article

Engineering a Catalytically Efficient Recombinant Protein Ligase. Renliang Yang, Yee Hwa Wong, Giang K. T. Nguyen, James P. Tam, Julien Lescar, and Bin Wu J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/jacs.6b12637 • Publication Date (Web): 15 Feb 2017 Downloaded from http://pubs.acs.org on February 16, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Journal of the American Chemical Society is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 10

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 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of the American Chemical Society

Engineering a Catalytically Efficient Recombinant Protein Ligase. Renliang Yang1, 2, Yee Hwa Wong1,2, Giang K T Nguyen1, James P Tam1, Julien Lescar1,2*, Bin Wu1,2* 1

School of Biological Sciences, Nanyang Technological University, 60 Nanyang Drive, Singapore 636921.

2

NTU Institute of Structural Biology, Nanyang Technological University, EMB 06-01, 59 Nanyang Drive, Singapore 636921.

ABSTRACT: Breaking and forming peptidyl bonds are fundamental biochemical reactions in protein chemistry. Unlike proteases that are abundantly available, fast-acting ligases are rare. OaAEP1 is an enzyme isolated from the cyclotideproducing plant oldenlandia affinis that displayed weak peptide cyclase activity, despite having a similar structural fold with other asparaginyl endopeptidases (AEP). Here we report the first atomic structure of OaAEP1, at a resolution of 2.56 Å, in its pre-activation form. Our structure and biochemical analysis of this enzyme reveals its activation mechanism as well as structural features important for its ligation activity. Importantly, through structure-based mutagenesis of OaAEP1, we obtained an ultra-fast variant having hundreds of times faster catalytic kinetics, capable of ligating wellfolded protein substrates using only sub-micro molar concentration of enzyme. In contrast, the protein-protein ligation activity in the original wild-type OaAEP1 enzyme described previously is extremely weak. Thus, the structure-based engineering of OaAEP1 described here provides a unique and novel recombinant tool that can now be used to conduct various protein labeling and modifications that were extremely challenging before.

INTRODUCTION Protein modifications have important applications in biology and medicine, and methods development in this area has been an active field of research. Since 1990s, chemically modified proteins have led to many advances in different fields, including in vivo imaging4,5, the generation of drug derivatives6, and direct modification of protein functions for screening and therapeutic purposes7,8. Nowadays, emergent biomedical applications of modified proteins require new strategies to label the target proteins in a very specific manner, which requires additional chemical treatment of the target proteins9-15. Enzymatic catalyzed protein ligation emerges as a more promising option, compatible with native intrinsic protein substrates. Pioneering studies by several groups have provided innovative solutions to the peptide ligation problem by converting proteolytic enzymes into peptide cyclases and ligases1,16-18 or by repurposing sortase2,3, which anchors proteins to the bacterial cell wall. The first asparagine/aspartate (Asn/Asp, or Asx) peptide ligase characterized was butelase 1, purified from the cyclotide-producing plant Clitoria ternatea19. Compared to bacterial sortase which is commonly used to perform transpeptidation reactions in vitro, butelase 1 shows exceptionally high efficiency in catalyzing both peptide and protein ligation reactions, opening a wide range of new applications in

biotechnology, protein engineering, chemo-enzymatic synthesis and protein labeling20-23. Recently an asparaginyl endopeptidase (AEP) evolutionarily related to butelase 1 named OaAEP1, with also the ability to join the N- and C-termini of peptidyl substrates, was isolated from the plant Oldenlandia affinis and expressed using E. coli in an active form, following activation at acidic pH24. Both butelase 1 and OaAEP1 recognize Asx residues at the ligation site and have 66% amino-acid sequence identity. However, the reported catalytic efficiency of OaAEP1 is markedly lower than butelase 124. Surprisingly these Asx ligases share a highly conserved protein architecture, which is also found in endopeptidases such as human legumain endopeptidase (hlegum)18,25-27. When expressed in plants, Asx ligases exist in a zymogenic state comprising an N-terminal signal sequence of ∼20-30 amino acids that direct them to the vacuolar compartment, followed by an enzymatic core domain. At the C-terminus of the core domain, a ∼130residues “cap domain” or pro-domain entirely covers their active site, keeping the immature protein inactive. The core and cap domains are connected by a few dozen of residues (326-347 in OaAEP1), which we define as “linker” (Fig. 1A). Activation of Asx ligases requires proteolytic maturation at an acidic pH of about 3.7, at which the cap domain is cleaved either in cis by the proteolytic activity of the core domain itself or potentially in trans by another

ACS Paragon Plus Environment

Journal of the American Chemical Society

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 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

protease present in the milieu, leading to the release of the cap from the core domain28,29. However, the precise cleavage site(s) within the linker and pro-domain, leading to the formation of a mature active Asx ligase remain poorly characterized30. In the case of butelase 1, Asn383 was hypothesized to constitute the site of cap domain cleavage19, while several potential cleavage sites spanning region 328-351 of the protein were proposed for OaAEP1 (Fig. 1A, fig. S1A and S1B)24. Amino-acid sequence alignments of Asx ligases with AEPs such as hlegum also reveal the presence of several evolutionary conserved residues in the catalytic pocket (Fig. S1A)19,24,31-33. Therefore, what determines how these closely homologous enzymes preferentially function either as ligase or protease remains elusive. Acquiring deeper understandings of how these ligases work would render many challenging biochemical modifications possible. RESULTS AND DISCUSSION Protein Expression, Purification and Activation of OaAEP1. OaAEP1 was cloned and expressed in E. coli as an ubiquitin fusion protein as reported earlier24 (Fig. 1A). Using ion exchange and size exclusion chromatography (SEC), we obtained sufficient amounts of pure protein suitable for enzymatic and structural analysis (Fig. 1B). At the neutral pH of purification (pH=7.4), OaAEP1 was mainly present as a heterodimer in solution, where both monomers are in their zymogenic form, but only one monomer had retained in addition the ubiquitin and Nterminal His-tag, while the fusion moiety was cleaved in the other monomer (labeled as “His-Ub-AEP-FL” and ”AEP-FL” in Fig. 1B and Fig. 1C). This heterodimer was subjected to crystallization trials for structural studies (see below). In agreement with reference24, activation of OaAEP1 was optimum for pH values ranging from 3.4 to 4.0 (Fig. 1D). Following activation, OaAEP1 converts to a monomer in solution according to SEC analysis (Fig. 1C). To test the ability of the protein to be activated in trans, we mutated the catalytic Cys217 into Ser217. As anticipated, the Cys217Ser mutant only displayed background level activation and ligase activity (Fig. 1E). However, activation of the inactive mutant could be rescued in trans by addition of only 10% of wild type OaAEP1 to the inactive enzyme (Fig. 1E and fig. S1C). Thus, trans-cleavage processing events could also be at play during OaAEP1 activation and Cys217 is the active site residue responsible for both cis and trans-activation. Overall Crystal Structure of OaAEP1. Crystals of OaAEP1 diffracting to a resolution of 2.56 Å at a 3rd generation synchrotron source were obtained (table S1 and Fig. 2). Overall, clear electron density allowed the building of an unambiguous atomic model (fig. S2A, S2B and S2C). Although the OaAEP1 heterodimer was used as starting material, the crystal structure revealed a homodimer where the N-terminal (His)6 tag, the ubiquitin fusion protein as well as signal peptide sequence have been released from the two monomers present in the asymmetric unit, suggesting N-terminal proteolysis during crystal-

Page 2 of 10

lization (Fig. 1D and Fig. 2A). The OaAEP1 dimer is stabilized by intermolecular interactions between the two cap domains (Fig. 2C). This is in agreement with SEC analysis (Fig. 1C) showing that the pre-activated form of the enzyme containing the cap region is dimeric at neutral pH (with an intermolecular interface area of 1157 Å2), while the activated form, which is devoid of cap, is monomeric. Accordingly, SDS PAGE of dissolved OaAEP1 crystals (Fig. 1B) revealed the presence of both the enzymatic core and the C-terminal pro-domain with a total mass of approximately 49.7 kDa, supporting the hypothesis that the crystallized enzyme is in an immature form. The OaAEP1 core is formed by a 6-stranded β-sheet surrounded by six αhelices at its periphery 31,34,35. An overall view of the structure shown in Fig. 2A reveals that the -helical cap domain (spanning residues 347-474) entirely covers the catalytic cleft and suggests that release of the cap is required to grant access to peptide and protein substrates. The OaAEP1 catalytic site contains several residues present in other AEPs including the Cys217-His175 catalytic pair (Cys189-His148 in hlegum) and His73-Asp174 that superimpose with His45-Asp147 of hlegum (Fig. 2F). A clear difference captured here between the zymogenic forms of hlegum (PDB code: 4NOK) and OaAEP1, resides within the linker region: in the case of the human prolegumain structure, the linker is short and well ordered, whilst it is long and flexible in the present OaAEP1 structure (Fig. 2A): no electron density was visible for the 19 residues between 326 to 343, although this segment is present in the crystallized protein (Fig. 1B). OaAEP1 Linker Structurally Mimics Substrates. While most residues (326-343) of the linker region are not visible in the electron density map due to their flexibility, the last four residues of the linker (Val344-Val345Asn346-Gln347) are trapped at the interface between the cap and the core domain (Fig. 2D, 2F and 2G). Remarkably, their orientation closely overlaps with the Ac-TyrVal-Ala-Asp-chloromethylketone inhibitor that was covalently linked to the catalytic Cys189 of hlegum (Fig. 2E, PDB code: 4AWA, fig. S2D)31,34,35. Here, Gln347 penetrates deeply into the S1 pocket, establishing several interactions with surrounding active site residues, such that its amide side-chain occupies the position of the carboxylic group of the Asp residue in the Ac-Tyr-Val-Ala-Asp-CMK protease inhibitor covalently bound to hlegum and also with the amide side-chain of Asn39 from cystatin E inhibitor bound to hlegum (PDB code 4N6O)19. Taken together, the extensive set of interactions established by Gln347 in the catalytic pocket and the common orientation between residues 344-347 with hlegum inhibitors indicate that the present crystal structure can be used as a model to picture a pre-ligation conformation, with linker residues 344-347 mimicking an N-terminal substrate acceptor bound to the active site (fig. S2D). Although in other AEPs, Asx is the most favored residue for processing at the P1 position, the carbonyl group of Gln347 lies at a distance of 5.2 Å from the attacking Cys217 sulfur center (Fig. 2F and 2G), suggesting that the segment 346-351 of the pro-protein might contain auto-

ACS Paragon Plus Environment

Page 3 of 10

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 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of the American Chemical Society

cleavage sites that are proteolyzed during maturation. To determine whether Gln347 belongs to the segment recognized and cleaved during self-activation, we targeted region 346-351 of the OaAEP1 protein for mutagenesis (Fig. 2H). We particularly assessed the role of Gln347 in OaAEP1 auto-activation by mutating it to Ala, and examined whether enzyme activation was impaired. Only little impact was found on the mutant enzyme activation (Fig. 2H). We then targeted residues Asn346, Asp349 and Asp351 that are next to Gln347. While the two OaAEP1 double-mutants Asn346Ala-Gln347Ala and Asp349Ala Asp351Ala and the Asn346Ala-Asp349Ala-Asp351Ala triple-mutant can still be activated, the Asn346AlaGln347Ala-Asp349Ala-Asp351Ala mutant was most significantly impaired for auto-activation (lane “A4” in Fig. 2H, and 2I). A control experiment targeting the five Asx residues located in the flexible region of the linker (residues 325-346) showed that the “A5” mutant could still undergo activation (fig. S2F). Importantly, activation of the “A4” mutant could not be rescued in trans when wild type enzyme was added to the reaction mixture. Thus, transactivation requires the presence of at least one of the four Asx/Gln residues found in this section of the linker, with residues Asp349 and Asp351 (that had been proposed earlier to constitute possible cleavage sites36 likely to play a major role. Finally MALDI-TOF analysis of the cleaved– off cap domain confirmed that Asp351 is the primary site being targeted for cleavage during activation (fig. S2G), while secondary cleavages might take place a few residues upstream in the sequence. The peptide cyclization activity of mature OaAEP1 was then analyzed using MS with the peptide substrate “GLPVSTKPVATRNGL” (Fig. 3A). The turnover rate (kcat) for the present recombinant wild-type OaAEP1 protein is 0.052 s-1, a value in good agreement with the previous report24. Likewise, the cyclization activity of butelase 1 extracted from plant was measured using the 15 residues peptide “GLPVSTKPVATRNHV” (Fig. 3A and 3B) leading to a kcat about 90 times faster than OaAEP1. Structural Differences Between Asx Proteases and Ligases. We then analyzed the OaAEP1 3D structure to identify features that could differentiate an Asx protease from an Asx ligase (Fig. 2D and 2E). A side-by-side comparison of our structure with a previously reported proteolytic AEP-inhibitor complex (PDB code: 4AWA) revealed interesting structural features unique to Asx ligases. In contrast to Asx proteases, the molecular surface of the core region of OaAEP1 next to the catalytic Cys217 appears to be more open. Three cysteine residues (Cys247, Cys250, Cys264) are aligned along a shallow cleft at the surface of the enzyme (Fig. 2D and 2E). Unlike the surface of Asx proteases that contains several bulges, both butelase 1 (using a homology model derived from the present OaAEP1 structure) and OaAEP1 have residues with smaller side-chains along this patch. A disulfide bond is formed between Cys250 and Cys264, which are conserved in Asx ligases butelase 1 and OaAEP1 (fig. S1). This covalent bond is likely to stabilize the local structure and provide an adequate surface allowing the approach of an N-

terminal amine group. Conversely, the corresponding surface patch of AEP proteases (PDB: 4AWA) is lined with residues having bulky side-chains blocking this channel (Fig. 2E). We found that mutations around the catalytic pocket generally disrupt the ligation activity, using a peptide cyclization assay (fig. S2E). OaAEP1 Cys247Ala is an Ultra-fast Protein Ligase. The location of Cys247 attracted our attention. We hypothesized that residue Cys247 which is located at the extremity of the substrate channel could function as a final nucleophile filter (or “gate-keeper”), allowing attack from the N-terminal amine substrate to complete the ligation reaction. To test the impact of various side chains at the “gate keeper” position on maturation and ligase behavior, we performed point mutagenesis where Cys247 was mutated into Gly, Ala, Val, Ser, Thr, Met, Leu and Ile (Fig. 3C and fig. S3). While maturation of the mutants appears essentially not much affected (Fig. 3C and fig. S1D) the nature of the side chain indeed plays a critical role on ligase activity (Fig. 3D). Remarkably, the Cys247Ala mutant displayed significantly improved enzymatic properties compared to WT OaAEP1 (Fig. 3D). We attribute this phenotype to the presence of a smaller side chain that is more appropriate to accommodate an incoming amine group, while larger more hydrophobic side-chains like Met, Val, Leu or Ile abolish ligase activity and Ser and Cys have only moderate ligase activity (Fig. 3D). Interestingly, the Cys247Gly mutation appears to negatively affect ligase activity possibly through destabilizing the local protein conformation. Further investigations showed that Cys247Gly catalyzes peptide hydrolysis as favored alternative reaction, if the available N-terminal amine group is acetylated (fig. S4A). The enhanced activity of the Cys247Ala mutant might be rationalized by the more hydrophobic character of Ala, which disfavors water entering the catalytic patch, causing faster hydrolysis of the intermediate thio-ester bond, when ligation is not possible. The much-improved enzymatic properties displayed by the Cys247Ala mutant over the parent wild-type OaAEP1 protein was quantified by measuring its catalytic parameters: its kcat approximately 160 times higher than WT OaAEP1, suggested that it could form an attractive tool for a variety of challenging protein engineering and labeling applications. We first tested the ability of Cys247Ala to ligate ubiquitin with a peptide. Three residues “AsnGly-Leu” were added to the C-terminus of ubiquitin to make it recognizable by Cys247Ala. Another peptide having N-terminal residues “Gly-Leu” was used as the add-on component. At a protein: peptide molar ratio of 1:5, Cys247Ala catalyzed the ligation of more than 90% ubiquitin with the peptide within 10 min (Fig. 4A). An amount of 30 nM of Cys247Ala was enough for the reaction compared to 10-30 µM of sortase A typically required to perform productive ligations. To further demonstrate the potential of Cys247Ala, ligations between two well-folded proteins were also tested. Ubiquitin having the same “Asn-Gly-Leu” C-terminal modification and a N-terminal modified SNAP tag protein were used as substrates. Using

ACS Paragon Plus Environment

Journal of the American Chemical Society

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 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

5 µM of SNAP tag protein and 8 µM ubiquitin, 100 nM Cys247Ala was able to ligate more than 60% of the individual substrate proteins into a ligated single polypeptide product, indicating that the catalyzed protein ligation is highly efficient and irreversible (Fig. 4B). Protein ligation can be performed within 30 minutes at neutral pH and at room temperature. Remarkably, the efficiency of Cys247Ala to catalyze the ligation of two folded proteins is comparable to its ability to ligate one protein with a peptide, suggesting that its open binding site can accommodate two bulky protein substrates without hindering the ligation process. This activity is slightly better than Butelase 1, directly extracted from plant (fig. S4B). In order to further demonstrate the robust ligase activity of Cys247Ala, we performed the ligation reaction with an ubiquitin construct having recognition amino acid sequences at both its N- and C-termini. As a result, a molecular ladder up to 200 kDa could be visualized thanks to the SNAP tag conjugated at the C-terminal end of the ligated poly-ubiquitin (Fig. 4C). Control ligation experiments showed that protein ligation activity could only be observed in the presence of the Cys247Ala enzyme, but not with the OaAEP1 wild type construct (Fig. 4A and 4B). Furthermore, the catalyzed ligation reaction did not seem to be reversible and experience severe hydrolysis issue (fig. S4C), and could be completely quenched by a peptide based inhibitor (fig. S4D).

Page 4 of 10

range of biotechnological applications both for protein/peptide semisynthesis and specific protein labeling22,23,38-42.

ASSOCIATED CONTENT Supporting Information. Supplement Figures (S1-S4), Method and Material, Crystallographic Statistics Supplement Table 1. The coordinates of the WT OaAEP1 have been deposited with the PDB (www.rcsb.org) with accession number 5H0I.

AUTHOR INFORMATION Corresponding Author * [email protected]; [email protected] R.Y. and B.W. planned the experiments. R.Y. performed all biochemical experiments, and obtained the initial OaAEP1 crystals. Y.H.W. and R.Y. optimized the crystals and collected the diffraction data. J.L. and B.W. solved the OaAEP1 structure and conducted the structural analysis. G.K.T.N. and J.P.T. made critical comments on the manuscripts and conducted the kinetic measurements of Butelase 1 and OaAEP1 constructs. J.L. and B.W. wrote the manuscript.

ACKNOWLEDGMENT R.Y. acknowledges the generous support from NISB fellowship. This research was supported by Nanyang Technological University NAP startup grant to the B.W. and an AcRF Tier 1 grant RGC2/14 to the J.L..

CONCLUSION Sortase A and butelase 1 have been used to perform peptide ligations. However, both enzymes suffer from several disadvantages: Sortase A requires Ca2+, is slow and recognizes a longer -more flexible- C-terminal sequence (LPXTG). Although through extensive mutagenesis studies, several mutants were identified improving its kinetics, the catalysis was still quite inefficient in ligating two wellfolded proteins36,37. In contrast, butelase 1 extracted from plant, is extremely efficient and has a shorter recognition amino acids sequence (NHV)19. However, optimization and engineering of the recombinant construct of Butelase 1 still remains challenging because of the difficulty in producing an active recombinant enzyme. Here, we reported, to our knowledge, the first X-ray crystallographic structure of an Asx ligase, OaAEP1. The present structure can serve as a template to understand the family of Asx ligases and will facilitate the design of faster protein ligases with alternative substrate specificities. Our structure reveals the mode through which self-cleavage activation is achieved in Asx ligases and also points to key residues and structural features accounting for the functional divergence among Asx endoprotease and ligases: Asx proteases lack the flat surface near the catalytic pocket conducive to ligation, while Asx ligases have a wide and open surface able to accommodate the incoming amine group. These structural observations led us to engineer a modified ligase Cys247Ala with improved biochemical properties, which can efficiently ligate proteins. Given its fast ligation catalytic kinetics, we believe that the present Cys247Ala recombinant protein ligase can be employed for a wide

REFERENCES (1) Jackson, D. Y.; Burnier, J.; Quan, C.; Stanley, M.; Tom, J.; Wells, J. A. Science 1994, 266, 243. (2) Mao, H.; Hart, S. A.; Schink, A.; Pollok, B. A. Journal of the American Chemical Society 2004, 126, 2670. (3) Popp, M. W.; Antos, J. M.; Ploegh, H. L. Current protocols in protein science / editorial board, John E. Coligan ... [et al.] 2009, Chapter 15, Unit 15 3. (4) Rao, J.; Dragulescu-Andrasi, A.; Yao, H. Current opinion in biotechnology 2007, 18, 17. (5) Griffin, B. A.; Adams, S. R.; Tsien, R. Y. Science 1998, 281, 269. (6) Zalipsky, S. Bioconjugate chemistry 1995, 6, 150. (7) MacDonald, J. I.; Munch, H. K.; Moore, T.; Francis, M. B. Nature chemical biology 2015, 11, 326. (8) ElSohly, A. M.; Francis, M. B. Accounts of chemical research 2015, 48, 1971. (9) Yan, L. Z.; Dawson, P. E. Journal of the American Chemical Society 2001, 123, 526. (10) Dawson, P. E.; Muir, T. W.; Clark-Lewis, I.; Kent, S. B. Science 1994, 266, 776. (11) Muir, T. W. Annual review of biochemistry 2003, 72, 249. (12) Wu, P.; Shui, W.; Carlson, B. L.; Hu, N.; Rabuka, D.; Lee, J.; Bertozzi, C. R. Proceedings of the National Academy of Sciences of the United States of America 2009, 106, 3000. (13) Chan, A. O.; Ho, C. M.; Chong, H. C.; Leung, Y. C.; Huang, J. S.; Wong, M. K.; Che, C. M. Journal of the American Chemical Society 2012, 134, 2589. (14) Gilmore, J. M.; Scheck, R. A.; Esser-Kahn, A. P.; Joshi, N. S.; Francis, M. B. Angewandte Chemie 2006, 45, 5307. (15) Witus, L. S.; Netirojjanakul, C.; Palla, K. S.; Muehl, E. M.; Weng, C. H.; Iavarone, A. T.; Francis, M. B. Journal of the American Chemical Society 2013, 135, 17223. (16) Bernath-Levin, K.; Nelson, C.; Elliott, A. G.; Jayasena, A. S.; Millar, A. H.; Craik, D. J.; Mylne, J. S. Chemistry & biology 2015, 22, 571.

ACS Paragon Plus Environment

Page 5 of 10

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 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of the American Chemical Society

(17) Luo, H.; Hong, S. Y.; Sgambelluri, R. M.; Angelos, E.; Li, X.; Walton, J. D. Chemistry & biology 2014, 21, 1610. (18) Lee, J.; McIntosh, J.; Hathaway, B. J.; Schmidt, E. W. Journal of the American Chemical Society 2009, 131, 2122. (19) Nguyen, G. K.; Wang, S.; Qiu, Y.; Hemu, X.; Lian, Y.; Tam, J. P. Nature chemical biology 2014, 10, 732. (20) Nguyen, G. K.; Hemu, X.; Quek, J. P.; Tam, J. P. Angewandte Chemie 2016, 55, 12802. (21) Hemu, X.; Qiu, Y.; Nguyen, G. K.; Tam, J. P. Journal of the American Chemical Society 2016, 138, 6968. (22) Nguyen, G. K.; Kam, A.; Loo, S.; Jansson, A. E.; Pan, L. X.; Tam, J. P. Journal of the American Chemical Society 2015, 137, 15398. (23) Nguyen, G. K.; Cao, Y.; Wang, W.; Liu, C. F.; Tam, J. P. Angewandte Chemie 2015, 54, 15694. (24) Harris, K. S.; Durek, T.; Kaas, Q.; Poth, A. G.; Gilding, E. K.; Conlan, B. F.; Saska, I.; Daly, N. L.; van der Weerden, N. L.; Craik, D. J.; Anderson, M. A. Nature communications 2015, 6, 10199. (25) Ishii, S. Methods in enzymology 1994, 244, 604. (26) Chen, J. M.; Dando, P. M.; Rawlings, N. D.; Brown, M. A.; Young, N. E.; Stevens, R. A.; Hewitt, E.; Watts, C.; Barrett, A. J. The Journal of biological chemistry 1997, 272, 8090. (27) Brix, K.; McInnes, J.; Al-Hashimi, A.; Rehders, M.; Tamhane, T.; Haugen, M. H. Protoplasma 2015, 252, 755. (28) Hiraiwa, N.; Nishimura, M.; Hara-Nishimura, I. FEBS letters 1999, 447, 213. (29) Yamada, K.; Shimada, T.; Kondo, M.; Nishimura, M.; Hara-Nishimura, I. The Journal of biological chemistry 1999, 274, 2563. (30) Li, D. N.; Matthews, S. P.; Antoniou, A. N.; Mazzeo, D.; Watts, C. The Journal of biological chemistry 2003, 278, 38980. (31) Dall, E.; Fegg, J. C.; Briza, P.; Brandstetter, H. Angewandte Chemie 2015, 54, 2917. (32) Dall, E.; Brandstetter, H. Biochimie 2016, 122, 126. (33) Saska, I.; Gillon, A. D.; Hatsugai, N.; Dietzgen, R. G.; Hara-Nishimura, I.; Anderson, M. A.; Craik, D. J. The Journal of biological chemistry 2007, 282, 29721. (34) Dall, E.; Brandstetter, H. Proceedings of the National Academy of Sciences of the United States of America 2013, 110, 10940. (35) Zhao, L.; Hua, T.; Crowley, C.; Ru, H.; Ni, X.; Shaw, N.; Jiao, L.; Ding, W.; Qu, L.; Hung, L. W.; Huang, W.; Liu, L.; Ye, K.; Ouyang, S.; Cheng, G.; Liu, Z. J. Cell research 2014, 24, 344. (36) Chen, I.; Dorr, B. M.; Liu, D. R. Proceedings of the National Academy of Sciences of the United States of America 2011, 108, 11399. (37) Popp, M. W.; Ploegh, H. L. Angewandte Chemie 2011, 50, 5024. (38) Guimaraes, C. P.; Witte, M. D.; Theile, C. S.; Bozkurt, G.; Kundrat, L.; Blom, A. E.; Ploegh, H. L. Nature protocols 2013, 8, 1787. (39) Swee, L. K.; Guimaraes, C. P.; Sehrawat, S.; Spooner, E.; Barrasa, M. I.; Ploegh, H. L. Proceedings of the National Academy of Sciences of the United States of America 2013, 110, 1428. (40) Theile, C. S.; Witte, M. D.; Blom, A. E.; Kundrat, L.; Ploegh, H. L.; Guimaraes, C. P. Nature protocols 2013, 8, 1800. (41) Witte, M. D.; Theile, C. S.; Wu, T.; Guimaraes, C. P.; Blom, A. E.; Ploegh, H. L. Nature protocols 2013, 8, 1808. (42) Wagner, K.; Kwakkenbos, M. J.; Claassen, Y. B.; Maijoor, K.; Bohne, M.; van der Sluijs, K. F.; Witte, M. D.; van Zoelen, D. J.; Cornelissen, L. A.; Beaumont, T.; Bakker, A. Q.; Ploegh, H. L.; Spits, H. Proceedings of the National Academy of Sciences of the United States of America 2014, 111, 16820.

ACS Paragon Plus Environment

Journal of the American Chemical Society

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 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 10

Figure 1. Purification and activation of OaAEP1. A. Schematic representation of the His-Ub-OaAEP1 construct used in this study, with the boundaries of the three regions of the protein indicated and the deduced activation segment. A gene encoding a 6-His tag and Ubiquitin tag followed by OaAEP1 (residues 24-474) was synthesized and inserted into pET28b vector. After NiNTA affinity purification, about half of Ub-AEP-FL lost the N-terminal fusion tag and was purified in AEP-FL form (53-474), which was eventually crystallized. The enzymatic active form of OaAEP1, AEP-holo (53-347/351) is the acid-induced self-cleavage product of AEP-FL, which has lost the C-terminal auto-inhibitory cap domain. B. SDS PAGE analysis of fractions obtained over the course of chromatographic purification and of dissolved crystals. C. Size-exclusion chromatography profiles of OaAEP1 before and after activation showing Ub-AEP-FL and AEP-FL heterodimer and the AEP-holo monomer. The elution volumes of standard proteins are indicated for comparison. D. SDS PAGE analysis of OaAEP1 pH-dependent activation. The activation was performed for 3 hours, and maximum activation occurs between pH 3.4 to 4. E. Trans activation/maturation of the inactive Cys217Ser active site mutant by the active recombinant mature OaAEP1 enzyme. The active site Cys217Ala mutant is devoid of self-activation activity. Addition of 5% AEP-FL wt enables trans-cleavage and activation of most of the AEP-FL Cys217Ala after 18 hours incubation at pH 3.7.

ACS Paragon Plus Environment

6

Page 7 of 10

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 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of the American Chemical Society

Figure 2. Crystal structure of OaAEP1 in its zymogenic form. A. Overall “front” view of the OaAEP1 structure represented as “cartoon” (with α-helices as ribbons and β-strands as arrows) with the catalytic core domain colored in green, the pro-domain in purple. The flexible linker spanning residues 325-343 (not visible in the electron density) is depicted as a dashed line and residues 344-347 (at the end of the linker) that are trapped in the OaAEP1 active site are represented as sticks. The molecular surface views in panels B. and D. were obtained by separating the core domain from the cap domain and rotating by an angle of 90° in opposite directions respectively (indicated by arrows). C. The OaAEP1 inactive dimer is stabilized via intermolecular interactions (including 13 hydrogen bonds and 6 salt-bridges) provided by both cap regions. Residues 344-347 that are bound in the active site are depicted as sticks and labeled. Compared to hlegum, these linker residues are bound at the catalytic center in an orientation similar to those adopted by peptide-based legumain inhibitors. From this comparative view, the putative channel that accommodates the incoming amine could be identified on the core domain surface. The orientation of an incoming substrate, which is poised to undergo ligation, is indicated by an arrow pointing in the direction of the channel (see text). E. The hlegum molecular surface is shown in the same orientation as OaAEP1 with the peptide Ac-Tyr-Val-Ala-Asp-CMK (sticks) covalently bound to the active site Cys189 (from PDB code: 4AWA). Here, the incoming amine channel is blocked. F. The OaAEP1 active site and the activation segment. Close-up view of the OaAEP1 S1 pocket with active site residues discussed in the text displayed as sticks and labeled: The canonical AEP catalytic pair His175 and Cys217 and the linker residues 344-351 bound at the catalytic

ACS Paragon Plus Environment

7

Journal of the American Chemical Society

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 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 10

center are shown as yellow sticks. Also shown the disulfide bond formed by Cys250 and Cys264 (also conserved in butelase 1) and the “gate-keeper” residue Cys247 a major determinant for the ligase activity (note this residue is Val in butelase 1 and Ala in hlegum (fig S1A). Hydrogen bonds formed by P1 residue Gln347 which is bound in the S1 pocket are represented by dashed lines and interatomic distances are displayed in Å. G. Fo-Fc residual electron density map with the segment 344-349 of the polypeptide omitted from the phase calculation and contoured at 3σ level. The protein electrostatic surface is shown with positive charges in blue, negative charges in red and neutral in white. H. Residual activation of OaAEP1 mutants following site-directed mutagenesis targeting the “activation segment” 346-351. Single or double mutations of Asn346, Gln347, Asp349 and Asp351 demonstrated moderate loss of self-cleavage induced activation. Triple mutant “A3” (Asn346Ala, Asp349Ala and Asp351Ala) and quadruple mutant “A4” (Asn346, Gln347, Asp349 and Asp351) are more severely impaired, indicating the activation site could be any residue with this segment of the linker. Asp349 and Asp351 are the most likely cleavage sites, while Asn346 and Gln347 are potential cleavage sites. I. Addition of wild type OaAEP1 was not able to rescue and activate “A4” mutant further.

A B

kcat (s-1) 0.052 ± 0.008 13.9 ± 1.1 4.83 ± 0.62

Ligase WT AEP AEP (Cys247Ala) Butelase 1

L G

NH2

N

R T A V

N G L

G L P

R T A

oaAEP

V P K T

S

V P K

G L

Precursor Substrate

100 90

P V S T

Cyclized Peptide

100 4.6E+4 90

80

80

70

70

60

60

50

50

40

40

30

30

20

20

2005.1

Precursor Substrate

10

10 0 999

kcat/Km (M-1 s-1) 215 34,209 99,793

Km (µM) 245 ± 60 407 ± 49 48.4 ± 14

1201

1403

1605

1807

0 999

2009

1201

1403

1605

C

Cys247 Activation

-

+

1807

2009

Mass (m/z)

Mas s (m/z)

Gly -

+

Ala -

+

Ser -

+

Thr -

+

Met -

+

Val -

+

Leu -

+

Ile -

+

(kDa) 75 His-Ub-AEP-FL AEP-FL

50 37

functional AEP-holo

25 20

Cleaved C-terminal Cap domain

15

D

Gate-Keeper mutation ligation efficiency at 10 min and 1 hour

Figure 3. Comparison of peptide cyclization activity of WT OaAEP1, Cys247Ala and butelase 1 and the influence of the “gate-keeper” residue Cys247 on ligase activity. A. Kinetic parameters for the cyclization of “GLPVSTKPVATRNGL” by WT OaAEP1, Cys247Ala and butelase 1 freshly extracted from plant. Cys247Ala has a kcat/Km value 160 times faster than WT OaAEP1. MS analysis of the reaction is given in panel B. Schematic representation of the peptide used to determine the ligation efficiency, together with MS results demonstrating the backbone ligation achieved by wild type OaAEP1 and the Cys247Ala mutant. C. Role of the “gate-keeper” residue in OaAEP1 activation. Self-activation appears not much affected by mutations of the gate-keeper residue. D. Comparison of cyclization efficiency for various residues at the “gate-keeper” Cys247 position. The ligase catalytic activity appears to be inversely correlated with the size of the side chain at the Cys247 position.

ACS Paragon Plus Environment

8

Page 9 of 10

A

Ub

NH2

NGL

+

COOH

GL PVK AR

NH2

CONH2

Biotin

GL

Ub

NH2

NGL PVK AR

CONH2

Biotin

AEP wt or AEP* (nM) (kDa) 37

0

3000

1000

300

100

30 100

AEP-holo

25 20

666.7

Ub

3 µM AEP-holo ligation

90 80 70 60

Commassie-Blue stained

50

15

His-(3C)-Ub contaminant

10

Ub+peptide

40

Ub/2

30 20

Ub + peptide

10

Ub

0 3996.0

(kDa) 37

5610.6

7225.2

8839.8

10454.4

12069.0

Mass (m/z)

100

287.8

90

AEP*-holo 25

70

20

Ub+peptide

3 µM AEP*-holo ligation

80

60 50

His-(3C)-Ub contaminant

15

40 30

(Ub+peptide)/2

20

Ub + peptide

10

Ub

10

Ub

0 3996.0

5610.6

7225.2

8839.8

10454.4

12069.0

Mass (m/z)

B

C BG-Alexa647 labelled SNAP tag NH2

Ub

NGL

COOH

+

NH2

GL

SNAP

NH2

GL

Ub

NGL

+

COOH

NH 2

GL

Ub SNAP

NGL

SNAP

1 + + -

2 + + -

4 + + -

8 16 + + + + -

50 100 + + + + -

50 100 + + + +

Ub (100 µM) + SNAP (µM) 3 AEP* (µM) 0 AEP-holo SNAP + Ub

25

SNAP

20 15 10

COOH

GL

Ub SNAP

Ub

COOH

Ub Ub Ub Ub Ub SNAP

Ub

(kDa)

BG-Alexa647 signal

Commassie-Blue stained

37

0 + -

SNAP

COOH

BG-Alexa647 labelled Ub-SNAP

Ub (µM) SNAP (5 µM) AEP* (0.1 µM) AEP (4 µM)

GL

COOH NH2

Ub

NH2

GL

COOH

GL NH2

BG-Alexa647 signal

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 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of the American Chemical Society

(kDa) 100 75 50 37

+ + + 1 3 10 1 1 1 PolyUb-SNAP 3Ub-SNAP 2Ub-SNAP Ub-SNAP

25 20 15

SNAP

SNAP fragment

SNAP + Ub SNAP

10

Figure 4. Ligase activity of Cys247Ala using well-folded proteins as substrates. A. Lower panel: As little as 30 nM Cys247Ala is able to catalyze ligation of 100 µM Ubiquitin and 500 µM biotin-labeled peptide efficiently in 30 min, at room temperature, in PBS buffer. By contrast the upper panel shows that wild type OaAEP1 is essentially unable to perform the reaction: only minute amount of ligated product is detectable at a concentration of 3000 nM (The faint upper band labeled His-3C-Ub corresponds to an Ub contaminant). The engineered Ub construct (three left over residues after 3C cleavage at the N-terminal, and extra NGL residues at the C-terminal) has the molecular weight of 9095 Da, while the (Ub + peptide) has the molecular weight of 9865 Da. B. Fluorescence detection using O6 benzylguanosine-Alexa647 of the SNAP protein and SDS PAGE analysis with Coomassie blue staining detection of the protein substrate and ligated products. The reaction is schematically depicted as inset. 100 nM of Cys247Ala, but not 4000 nM of wild type OaAEP1, efficiently ligates 5 µM of C-terminally modified SNAP tag protein with a N-terminally modified ubiquitin. C. Cys247Ala efficiently ligated multiple ubiquitin molecules bearing recognition amino acid sequences at their N- and C- termini. By mixing N-terminal compatible SNAP tag protein with bi-functional ubiquitin molecules (both N-terminal and C-terminal are available for ligation), we could obtain a well-defined ladder pattern on SDS-PAGE (SNAP + nUb) using BG-Alexa647 fluorescent stain for detection.

ACS Paragon Plus Environment

9

Journal of the American Chemical Society

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 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 10

Table of Contents artwork

ACS Paragon Plus Environment

10